Vous êtes sur la page 1sur 170

Journal of Cereal Science

Volume 50, Issue 2, Pages 139-304 (September 2009)

1.

IFC - Editorial Board Page IFC

Research Papers 2. Impact of amylose content on starch retrogradation and texture of cooked milled rice during storage Pages 139-144 Shifeng Yu, Ying Ma, Da-Wen Sun QTL mapping of grain quality traits in rice Pages 145-151 Jue Lou, Liang Chen, Gaohong Yue, Qiaojun Lou, Hanwei Mei, Liang Xiong, Lijun Luo Optimization of air classification for the production of -glucan-enriched barley flours Pages 152-158 B. Ferrari, F. Finocchiaro, A.M. Stanca, A. Gianinetti Study of the physical properties of kafirin during the fabrication of tablets for pharmaceutical applications Pages 159-165 Abd Elmoneim O. Elkhalifa, Dominique M.R. Georget, Susan A. Barker, Peter S. Belton Formation of grain chalkiness and changes in water distribution in developing rice caryopses grown under high-temperature stress Pages 166-174 Tsutomu Ishimaru, Akemi K. Horigane, Masashi Ida, Norio Iwasawa, Yumiko A. San-oh, Mikio Nakazono, Naoko K. Nishizawa, Takehiro Masumura, Motohiko Kondo, Mitsuru Yoshida Simulation of the factors affecting -glucan levels during the cultivation of oats Pages 175-183 Uma Tiwari, Enda Cummins Differences in functional properties and biochemical characteristics of congenetic rice proteins Pages 184-189 Xiaohong Cao, Huanbin Wen, Cuijuan Li, Zhenxin Gu Role of oxidative cleavage and acid hydrolysis of oat beta-glucan in modelled beverage conditions Pages 190-197 R. Kivel, L. Nystrm, H. Salovaara, T. Sontag-Strohm

3.

4.

5.

6.

7.

8.

9.

10.

Effect of particle size on kinetics of starch digestion in milled barley and sorghum grains by porcine alpha-amylase Pages 198-204 Ghaid J.S. Al-Rabadi, Robert G. Gilbert, Michael J. Gidley Characteristics of enzymatic hydrolysis of thermal-treated wheat gluten Pages 205-209 Jin-shui Wang, Zhi-yan Wei, Lu Li, Ke Bian, Mou-ming Zhao Genetic variability in yellow pigment components in cultivated and wild tetraploid wheats Pages 210-218 A.M. Diges, C. Platani, L. Cattivelli, G. Mangini, A. Blanco The gluten protein and interactions between components determine mixograph properties in an F6 recombinant inbred linepopulation in bread wheat Pages 219-226 Yong Zhang, Jianwei Tang, Jun Yan, Yelun Zhang, Yan Zhang, Xianchun Xia, Zhonghu He Importance of extensional rheological properties on fiber-enriched corn extrudates Pages 227-234 Dhananjay A. Pai, Orane A. Blake, Bruce R. Hamaker, Osvaldo H. Campanella Impact of the baking kinetics on staling rate and mechanical properties of bread crumb and degassed bread crumb Pages 235-240 Alain Le-Bail, Kahina Boumali, Vanessa Jury, Fadhel Ben-Aissa, Ruben Zuniga Morphologies and microstructures of cornstarches with different amyloseamylopectin ratios studied by confocal laser scanning microscope Pages 241-247 Pei Chen, Long Yu, George Simon, Eustathios Petinakis, Katherine Dean, Ling Chen Atmospheric CO2 enrichment changes the wheat grain proteome Pages 248-254 P. Hgy, C. Zrb, G. Langenkmper, T. Betsche, A. Fangmeier Rheological characterization of refrigerated and frozen non-fermented gluten-free dough: Effect of hydrocolloids and lipid phase Pages 255-261 Gabriel Lorenzo, Noem E. Zaritzky, Alicia N. Califano Comparison of color techniques to measure the color of parboiled rice Pages 262-265 Bin Lv, Bin Li, Sha Chen, Jian Chen, Bo Zhu Optimization of ultrasound-assisted extraction of defatted wheat germ proteins by reverse micelles Pages 266-271 Ke-Xue Zhu, Xiao-Hong Sun, Hui-Ming Zhou Wholemeal wheat bread: A comparison of different breadmaking processes and fungal phytase addition Pages 272-277 Cristina M. Rosell, Eva Santos, Juan M. Sanz Penella, Mnica Haros

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

Yield dilution of grain Zn in wheat grown in open-top chamber experiments with elevated CO2 and O3 exposure Pages 278-282 Hkan Pleijel, Helena Danielsson Association analysis reveals effects of wheat glutenin alleles and rye translocations on dough-mixing properties Pages 283-290 Shusong Zheng, Patrick F. Byrne, Guihua Bai, Xueyan Shan, Scott D. Reid, Scott D. Haley, Bradford W. Seabourn

23.

Research notes 24. Contribution of cereals to dietary fibre and antioxidant intakes: Toward more reliable methodology Pages 291-294 F. Saura-Calixto, J. Prez-Jimnez, I. Goi A MALDI-TOF based analysis of high molecular weight glutenin subunits for wheat breeding Pages 295-301 Li Liu, Aili Wang, Rudi Appels, Junhong Ma, Xianchun Xia, Ping Lan, Zhonghu He, Frank Bekes, Yueming Yan, Wujun Ma

25.

Letter to Editor 26. Ustilago and the accidental domestication of maize Pages 302-303 Juan Pablo Ricardo Martnez-Soriano, Katia Avia-Padilla

Book Review 27. K. Khan and P.R. Shewry, Editors, Wheat Chemistry and Technology, fourth ed., AACC International (2009) ISBN 978-1-891127-55-7 480 pp., 196 figures, 69 tables, 27 color images. Page 304 Robert Graybosch

Journal of Cereal Science


Aims and Scope
The Journal of Cereal Science was established in 1983 to provide an international forum for the publication of original research papers of high standing covering all aspects of cereal science related to the functional and nutritional quality of cereal grains and their products. The journal also publishes concise and critical review articles appraising the status and future directions of specific areas of cereal science and short rapid communications that present news of important advances in research. The journal aims at topicality and at providing comprehensive coverage of progress in the field. Research areas include: Composition and analysis of cereal grains in relation to quality in end use Morphology, biochemistry, and biophysics of cereal grains relevant to functional and nutritional characteristics Structure and physicochemical properties of functionally and nutritionally important components of cereal grains such as polysaccharides, proteins, oils, enzymes, vitamins, and minerals Storage of cereal grains and derivatives and effects on nutritional and functional quality Genetics, agronomy, and pathology of cereal crops if there is a substantive relationship to end-use properties of cereal grains Functional and nutritional aspects of cereal-based foods and beverages, whether baked, fermented, or extruded Industrial products (e.g. starch derivatives, syrups, protein concentrates, and isolates) from cereal grains, and their technology

Editor-in-Chief
F. MacRitchie
Grain Science and Industry Department Kansas State University Manhattan, Kansas, USA

Editors
G. B. Fincher
Waite Agricultural Research Institute, University of Adelaide, Australia

R. A. Graybosch
USDA-ARS University of Nebraska, Lincoln, Nebraska, USA

R. J. Hamer
Wageningen Centre for Food Sciences Wageningen, The Netherlands

D. Lafiandra
Dipartimento di Agrobiologia e Agrochimica University of Tuscia Viterbo, Italy

J. R. N. Taylor
University of Pretoria, South Africa

Reviews Editor
P. R. Shewry
Rothamsted Research Harpenden, UK

Editorial Board
S. Bean USDA-ARS GMPRC, Manhatten, USA A. Blechl US Department of Agriculture (USDA), Albany, CA, USA P. Colonna INRA, Laboratoire de Biochimie et Technologies des Glucides Nantes, France K. Denyer John Innes Centre, Norwich, UK J. Dexter Canadian Grain Commission, Winnipeg, Manitoba, Canada R. D'Ovidio Universita' degli Studi della Tuscia, Viterbo, Italy P. J. Frazier Dalgety Food Technology Centre, Cambridge, UK Zhonghu He International Maize and Wheat Improvement Center, Chinese Academy of Agricultural Sciences, Beijing, China B. O. Juliano Philippines Rice Research Institute, Laguna, Philippines P Koehler . German Research Centre for Food Chemistry, Garching, Germany J. Kokini Cook College, Rutgers University, New Brunswick, New Jersey, USA O. R. Larroque CSIRO Plant Industry, Canberra, Australia A. W. MacGregor Livingston, Scotland, UK M.-H. Morel iNRA, Joint Research Unit "Agropolymers Engineering and Emerging Technologies", Montpellier, France Y. Popineau INRA, Unite de Recherche sur les Proteines Vegetales et leurs Interactions, Nantes, France K. Poutanen VTT Biotechnology, Espoo, Finland C. M. Rosell Food Science Department, (IATA-CSIC), Valencia, Spain S. O. Serna Saldivar ITESM, Monterrey, Mexico B. Svensson BioCentrum-DTU, The Technical University of Denmark, Kgs. Lyngby, Denmark

Founding Editors
T. Galliard and J. D. Schofield Author enquiries
For enquiries relating to the submission of articles (including electronic submission where available) please visit this journals homepage at http://www.elsevier.com/locate/jcs. You can track accepted articles at http://www.elsevier.com/trackarticle and set up e-mail alerts to inform you of when an articles status has changed. Also accessible from here is information on copyright, frequently asked questions and more. Contact details for questions arising after acceptance of an article, especially those relating to proofs, will be provided by the publisher. For a full and complete Guide for Authors, please go to: http://www.elsevier.com/wps/find/journaldescription.cws_home/622859/authorinstructions

Journal of Cereal Science 50 (2009) 139144

Contents lists available at ScienceDirect

Journal of Cereal Science


journal homepage: www.elsevier.com/locate/jcs

Impact of amylose content on starch retrogradation and texture of cooked milled rice during storage
Shifeng Yu a, Ying Ma a, *, Da-Wen Sun a, b
a b

School of Food Science and Engineering, Harbin Institute of Technology, 202 Haihe Road, Harbin 150090, PR China Food Refrigeration and Computerised Food Technology, University College Dublin, National University of Ireland, Earlsfort Terrace, Dublin 2, Ireland

a r t i c l e i n f o
Article history: Received 3 October 2008 Received in revised form 30 March 2009 Accepted 3 April 2009 Keywords: Amylose Cooked rice Retrogradation Storage

a b s t r a c t
Milled rice from 11 varieties, with amylose levels from 1.2 to 35.6% dry base, were collected to study the impacts of amylose content on starch retrogradation and textural properties of cooked rice during storage. The relationship between amylose content and different properties was determined using Pearson correlation. Starch retrogradation enthalpy (DHr) of cooked rice was determined by differential scanning calorimetry. DHr values were found to be positively correlated with amylose content (0.603 r 0.822, P < 0.01) during storage. Textural properties were determined by a Texture Analyser. The hardness of cooked rice showed a positive correlation with amylose content (0.706 r 0.866, P < 0.01) and a positive correlation with DHr of cooked rice (r 0.650, P < 0.01) during storage. The adhesiveness showed a negative correlation with amylose content (0.929 r 0.678, P < 0.01) and a negative correlation with DHr of cooked rice (r 0.833, P < 0.01) during storage. Hardness showed a negative correlation with adhesiveness (r 0.820, P < 0.01). These results indicated that amylose content has signicant effects on starch retrogradation and textural properties of cooked rice. The cooked rice with high amylose content is easy to retrograde, the cooked rice with low amylose content retrograded slowly. Sarch retrogradation contributes to the changes of textural properties of cooked rice during storage. 2009 Elsevier Ltd. All rights reserved.

1. Introduction In recent years, there has been an increased demand for ready to eat meals and many products with cooked rice have been developed. Cooked rice is one of the most important ready to eat meals in South East Asia including China. These products usually are packaged in polyethylene bags after cooking and refrigerated storage to prolong their shelf life. During refrigerated storage, desired quality of cooked rice, e.g. starch retrogradation, loss of avor/aroma and colour, increased hardness and decreased adhesiveness can change. Starch retrogradation and texture are two of the most important quality parameters of cooked rice. However, many factors inuence starch retrogradation and texture of cooked rice during storage, such as rice variety, storage conditions, amylose content, starch type, degree of milling, water to rice ratio, cooking methods, cooling methods and so on. Recently, some studies have focused on the effects of rice cultivars (Singh et al., 2005), rice starch structure (Ong and

* Corresponding author. Tel.: 86 0451 86682903; fax: 86 0451 86682906. E-mail addresses: shifengyu@hit.edu.cn (S.F. Yu), maying@hit.edu.cn (Y. Ma). 0733-5210/$ see front matter 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2009.04.003

Blanshard, 1995; Ramesh et al., 1999) and proteins on the textural properties of cooked rice (Martin and Fitzgerald, 2002), and the storage temperatures also affected the starch retrogradation and textural properties of cooked milled rice (Perdon et al., 1999). Other studies reported that amylose inuences the textural properties of cooked rice (Ong and Blanshard, 1995; Singh et al., 2005), and the amylose content was correlated with the retrogradation behaviour of rice our (Varavinit et al., 2003). The proportion of amylose and amylopectin affected the hardness of rice starch gel (Hibi and Hikone, 1998), and the ratio of amylose and amylopectin inuenced starch gel retrogradation rate after storage (Mariotti et al., 2009). Varavinit et al. (2003) found gelatinization temperatures of Thai rice starch were highly positively correlated with amylose levels and low amylose starch showed a low degree of retrogradation. Lu et al. (2009) reported that amylose content inuences the dynamic viscoelasticity of rice starch gel. High amylose starch showed higher moduli, lower loss tangent values and higher retrogradation rate, and Japonica rice starch showed slow retrogradation rate although it contains a similar amount of true amylose. Based on the literature cited above, amylose levels could signicantly inuence the starch or our gel properties and retrogradation properties. However, there have been few investigations done on the impacts of amylose

140

S.F. Yu et al. / Journal of Cereal Science 50 (2009) 139144

content on starch retrogradation and textural properties of cooked milled rice during storage. A relationship between amylose content and starch retrogradation and textural properties of cooked rice during storage are still not very clear. Therefore, the objective of this study was to correlate amylose contents of milled rice from various rice cultivars, in order to further understand the role of amylose on starch retrogradation and textural properties of cooked rice during storage. 2. Materials and methods

with 1300 ml of tap water. After the rice cooked for 25 min, the thermostat coupled with micro-switch automatically switched off the automatic rice cooker. The cooked rice samples were held in the rice cooker for an additional 15 min. Finally, about 2000-g cooked rice was removed from the rice cooker to a stainless tray for precooling in a 18  C refrigerator, to room temperature. After cooling, the cooked rice in the centre layer of the tray was taken out, packaged in polyethylene bags and sealed as quickly as possible (0 h of storage). 2.4. Cooked rice storage

2.1. Materials Eleven different cultivars of rice were used in the study (Table 1). The cultivars of YYou1 (YY1), Yixiang3003 (YX3003) and Keyou21 (KY21) were collected from Hunan Rice Research Institute, Hunan province, China; The Songnuo2 (SN2), Suijing4 (SJ4), WuYou1 (WY1) and Daohuaxiang2 (DH2), were collected from Heilongjiang Rice Research Institute, Heilongjiang province, China; NingJing42 (NJ42), was collected from the farm of NingxiaLianhu, Ningxia province, China; TYou15 (TY15) was collected from Hanzhong city, Shanxi province, China; The ZengyeSimiao6 (SM6) was purchased in the city of Zengcheng, Guang dong province, China. Thai Jasmine rice (TJ) (Produced by Thaihefa Co. LTD, Thailand) was purchased in Wal-Mart super-market. The moisture content of all milled rice was 1113%. All the samples were stored at 4  C in a refrigerator before experiments. 2.2. Chemical analysis Total carbohydrates content (TCC) was determined by the phenolsulphuric acid method (Adebooye and Singh, 2008). Apparent amylose contents (AAC) (absolute content of amylose in rice) were determined by iodine colorimeter at 620 nm using a potato starch standard mixture (Juliano et al., 1981). Protein contents (PC) (N*5.95) were determined by Micro-Kjeldahl method 920.87 of AOAC (2000) and lipid contents of samples were determined by ofcial method 923.05 of AOAC (2000). The results were reported on a dry base. All the experiments were performed at least in triplicate and results are presented as mean values. 2.3. Rice cooking Experiments were conducted with an automatic rice cooker (CFXB4003-A1, 4.0L, 700 W, 220 V, 50 Hz, Guangzhou domestic appliance Ltd., China). 1000-g rice was soaked in a pot for 30 min Packed cooked rice samples were stored in a refrigerator at 4 1  C for 0, 1, 4, 7, 11 and 14 days. The sealed bags of cooked rice were taken out at different storage times, and allowed to equilibrate for 1.5w2.0 h in an incubator (22 0.2  C) before texture determination. All experiments were performed in triplicate. 2.5. Textural prole analysis Textural prole analysis (TPA) of the cooked rice was performed using a Texture Analyser (TA.XT.plus, Texture Technologies Corp., UK) with a 50 kg load cell with a two-cycle compression. The analyser was linked to a computer that recorded the data via a software program called Texture Expert Excede Version 1.0 (Stable Micro Systems Software). A two-cycle compression force versus time program was used to compress the samples till 90% of the original cooked grain thickness, returned to the original position and again compressed. A 6-mm diameter ebonite probe was used to compress 3 grains, with pre-test speed of 1.0 mm/s, test speed and post-test speed of 0.5 mm/s. Parameters recorded from the test curves were hardness and adhesiveness. All textural analyses were replicated ten times per sample and results are presented as mean values. 2.6. Differential scanning calorimetry Starch retrogradation properties of cooked rice was analyzed by a Perkin Elmer pyris 6 differential scanning calorimeter (DSC) (Perkin Elmer, USA). The DSC was calibrated with indium (melting point 156.6  C, DHf 28.6 J/g) and an empty pan as a reference. The cooked rice samples were prepared using the method of Kim et al. (1997) after texture determination, cooked rice was mixed with 99% ethanol (1:4, v/v) and dehydrated for 12 h, then passed through a Buchner funnel, dried at 37  C in an air drier for 24 h and passed through a 100-mesh sieve after milling. A total weight of 4.0 mg cooked rice samples and distilled water (1:2, w/w) was placed in pre-weighed aluminium sample pans (PE0219-0062). The pans were sealed hermetically to prevent moisture loss and kept overnight. For all DSC runs, a sealed empty aluminium pan was used as reference. The sample was held isothermally at 20  C for 1 min before heating from 20 to 140  C at 10  C/min. The peak temperature and the enthalpy (DHr, J/g) associated with the retrograded starch melting peak appearing between 40 and 70  C were calculated. DHr was used to indicate the degree of starch retrogradation. The DSC measurements were performed in triplicate, and results are presented as mean values. 2.7. Statistical analysis All tests were performed at least in triplicate. Pearsons correlation coefcient analyses and analysis of variance (ANOVA) using Duncans multiple range test were performed at P < 0.05 or P < 0.01 using SPSS 12.0 software (SPSS Inc., USA).

Table 1 Chemical composition of the different rice varieties (%, dry base). Type Waxy Very low amylose Low amylose Variety SN2 TJ SJ4 DH2 NJ42 WY1 YY1 YX3003 SM6 TY15 KY21 AAC 1.2a 0.2 5.96b 0.56 11.67c 0.68 12.78c 0.80 13.18d 0.63 13.50d 0.50 14.85e 0.53 15.20e 0.50 23.94f 0.21 27.59g 0.87 35.73h 0.68 TCC 80.75a 1.47 80.51a 1.91 80.75a 1.97 80.17a 1.28 79.74a 1.25 81.03a 1.65 79.74a 1.25 79.71a 1.28 81.33a 2.02 79.75a 1.02 80.33a 1.22 PC 6.93a 0.24 7.94b 0.13 8.36c 0.19 6.93a 0.24 7.40d 0.23 7.66d 0.29 8.16c 0.02 8.14c 0.03 8.63e 0.39 8.33e 0.21 6.55f 0.02 LC 1.11a 0.07 0.87b 0.19 1.20a 0.12 1.44c 0.10 0.78b 0.12 1.04a 0.23 0.76b 0.12 0.78b 0.22 0.74b 0.04 0.68b 0.2 0.78b 0.14

Intermediate amylose High amylose

Data are means standard deviation, n 3. ahMeans in the same column followed by the same lowercase superscript letters are not different (P > 0.05). AAC: apparent amylose content (absolute content of amylose in rice); TCC: total carbohydrates content; PC: protein content; LC: lipid content.

S.F. Yu et al. / Journal of Cereal Science 50 (2009) 139144

141

3. Results and discussion 3.1. Chemical composition of milled rice Table 1 shows the apparent amylose (AA), total carbohydrates (TCs), protein and lipid contents in milled rices. In the classication by Juliano (2005), SN2 is a waxy rice (1.2%), TJ is a very low amylose content rice (5.96%), SJ4, DH2, NJ42, WY1, YY1 and YX3003 are low amylose content rices (11.67%, 12.78%, 13.18%, 13.50%, 14.85% and 15.20%, respectively), SM6 and TY15 are intermediate amylose content rices (23.94% and 27.59%, respectively), and KY21 is a high amylose content rice (35.73%). TCC of milled rice ranged from 79% to 81% and were not signicantly different among rice cultivars (P > 0.05). Protein levels of milled rice ranged from 6.55% to 8.60%. KY21 had the lowest protein content (6.55%), and SM6 showed the highest protein content (8.60%). The lipid contents ranged from 0.68% to 1.44%. Among the different rice cultivars, lipid content of DH2 was observed to be highest (1.44%), whereas TY15 had the lowest (0.68%). 3.2. Retrogradation properties Retrogradation properties were studied by analyzing the melting endotherm of recrystallized amylopectin by DSC and results are shown in Table 2. There were no signicant differences in onset, peak or conclusion temperatures for all rice varieties with storage time (P > 0.05). The mean melting onset, peak and conclusion temperature were no signicantly different among waxy rice (SN2), very low amylose content rice (TJ), low amylose content rice (SJ4, DH2, NJ42, WY1 and YY1) (P > 0.05), and there were no signicant differences in intermediate amylose content rices (SM6 and TY15) (P > 0.05).

Table 2 (continued) Variety Storage Amylopectin time (day) To( C) 0 1 3 7 11 14 0 1 3 7 11 14 0 1 3 7 11 14 0 1 3 7 11 14 0 1 3 7 11 14 0 1 3 7 11 14 45.43 0.85 45.53a 0.65 45.37a 0.66 44.84a 0.55 45.24a 0.45 44.44a 0.66 44.51a 0.89 44.09a 0.80 45.19a 0.82 44.12a 0.89 45.29a 0.91 44.54a 0.88 46.54b 0.77 46.30b 0.60 46.02b 0.58 45.66b 0.46 45.31b 0.65 46.26b 0.67 45.42b 0.77 44.24b 0.77 44.73b 0.68 44.70b 0.37 44.84b 0.60 44.84b 0.24
a

Tp( C) 52.70 0.73 52.20a 0.65 52.20a 0.55 51.86a 0.72 52.38a 0.47 51.71a 0.89 52.05a 0.42 52.19a 0.52 52.19a 0.44 51.85a 0.65 52.37a 0.42 52.21a 0.41 54.55b 0.78 54.37b 0.58 53.53b 0.44 53.54b 0.35 53.53b 0.52 53.71b 0.34 53.67b 0.97 54.03b 0.18 53.95b 0.18 53.90b 0.51 53.74b 0.11 53.32b 0.12
a

Tc( C)
a

DHr(J/g)

WY1

59.53 0.65 1.09e 0.15 59.68a 0.95 5.11n 0.16 59.41a 0.69 6.34j 0.12 59.79a 0.42 6.93k,m 0.11 59.58a 0.65 7.18m 0.12 59.12a 0.70 7.61n 0.13 60.81a 0.92 60.68a 0.42 60.85a 0.44 60.29a 0.42 60.37a 0.74 60.17a 0.85 63.04b 0.58 62.53b 0.69 62.75b 0.75 62.35b 0.34 62.88b 0.25 62.95b 0.65 62.60b 0.87 62.87b 0.42 63.38b 0.21 63.01b 0.84 63.49b 0.06 63.18b 0.07 63.67b 0.22 63.15b 0.84 62.12b 0.45 62.46b 0.54 62.15b 0.65 62.39b 0.10 64.35c 0.95 64.62c 0.84 63.11c 0.74 63.28c 0.54 63.28c 0.64 63.87c 0.44 1.21a 0.12 5.90i 0.15 7.36n 0.22 8.50c,d 0.12 8.67d 0.23 8.86d,t 0.19 2.36 0.05 8.52c,t 0.12 9.86p 0.14 10.02q 0.13 10.51r 0.23 10.42r 0.13 1.79s 0.08 7.46n 0.18 8.89t 0.11 9.52u 0.17 9.79p 0.19 10.39r 0.15 3.11v 0.02 7.87n 0.08 9.44u 0.07 9.76p 0.12 10.19q 0.03 10.63r 0.20 3.41q 0.05 9.71p 0.13 9.91p 0.18 10.62r 0.14 11.02w 0.15 11.13w 0.08

YY1

YX3003

SM6

TY15

45.23b 0.54 53.66b 0.34 45.13b 0.38 53.54b 0.50 45.48b 0.65 53.04b 0.48 44.52b 0.76 53.17b 0.53 44.90b 0.41 53.21b,c 0.24 44.90b 0.11 53.25b,c 0.12 46.27c 0.92 45.67c 0.81 46.00c 0.54 46.47c 0.94 45.54c 0.90 45.66c 0.58 55.38c 0.84 54.70c 0.70 54.54c 0.64 54.20c 0.90 54.37c 0.50 54.53c 0.57

KY21 Table 2 Retrogradation characteristics of cooked rice from different rice varieties during storage for 0, 1, 3, 7, 11 and 14 days. Variety Storage Amylopectin time (day) To( C) 0 1 3 7 11 14 0 1 3 7 11 14 0 1 3 7 11 14 0 1 3 7 11 14 0 1 3 7 11 14 44.24a 0.91 44.09a 0.52 44.19a 0.43 44.34a 0.86 44.12a 0.61 44.43a 0.38 44.03a 0.60 44.88a 0.50 44.55a 0.30 44.85a 0.24 44.35a 0.16 44.98a 0.56 44.18a 0.16 45.08a 0.21 44.98a 0.30 44.76a 0.46 44.70a 0.50 44.75a 0.36 44.19a 0.58 44.08a 0.76 44.26a 0.58 44.17a 0.63 44.99a 0.55 44.61a 0.35 44.43a 0.67 44.54a 0.55 45.00a 0.72 44.32a 0.37 45.23a 0.87 44.83a 0.60

Tp( C) 52.23a 0.92 52.22a 0.93 51.54a 0.85 51.44a 0.87 51.38a 0.65 51.72a 0.56 51.65a 0.90 51.61a 0.30 51.61a 0.06 51.91a 0.12 51.99a 0.36 51.77a 0.51 52.24a 0.22 52.04a 0.16 52.04a 0.18 52.20a 0.20 52.03a 0.30 52.52a 0.56 51.38a 0.86 51.86a 0.53 51.53a 0.26 52.22a 0.58 52.04a 0.48 51.70a 0.77 52.37a 0.37 52.86a 0.87 52.54a 0.67 51.70a 0.74 52.37a 0.47 51.54a 0.65

Tc( C) 60.52a 0.88 60.14a 0.72 59.02a 0.92 59.34a 0.72 59.03a 0.93 60.46a 0.62 60.71a 0.87 59.93a 0.92 60.29a 0.07 60.19a 0.40 60.44a 0.50 60.58a 0.30 60.00a 0.46 59.60a 0.65 59.39a 0.50 59.62a 0.44 59.35a 0.56 59.68a 0.78 60.03a 0.57 59.81a 0.87 59.76a 0.92 59.70a 0.37 60.04a 0.57 59.60a 0.60 60.44a 0.80 60.21a 0.60 59.82a 0.66 59.43a 0.62 59.97a 0.23 59.80a 0.98

DHr(J/g)
1.49a 0.11 5.72b 0.12 8.43c 0.06 8.67d,t 0.15 8.94d.t 0.20 8.64d,t 0.48 1.20e 0.06 4.73f 0.13 8.16c 0.12 8.65d 0.21 8.73d,t 0.11 8.61c,d 0.12 0.24g 0.10 3.74h 0.11 5.56b 0.13 6.09i 0.16 6.25j 0.10 6.59j,k 0.22 0.36g 0.05 4.37l 0.12 5.94i 0.37 6.38j 0.14 6.77k 0.23 7.12m 0.19 1.04e 0.09 4.37l 0.08 6.26j 0.23 6.83k 0.09 6.88k 0.18 7.15m 0.14

SN2

Data are means standard deviation, n 3. awMeans in the same column followed by the same lowercase superscript letters are not different (P > 0.05). To onset temperature; Tp peak temperature; Tc conclusion temperature; DHr retrogradation enthalpy.

TJ

SJ4

DH2

NJ42

KY21 had the highest mean melting onset, peak and conclusion temperature (45.94 0.38  C, 54.62 0.41 and 63.75 0.62  C, respectively). The mean onset, peak and conclusion temperature of high amylose rice (KY21) > intermediate amylose content rice (TY15 and SM6) > low amylose content rice (SJ4, DH2, NJ42, WY1 and YY1), very low amylose content rice (TJ) and waxy rice (SN2). From these results, the onset, peak and conclusion temperatures of cooked rice are not affected by storage time. This is in accordance with the studies of Karlsson and Eliasson (2003) who reported no changes in the onset, peak and conclusion temperature of potato retrogradation with storage time. But the onset, peak and conclusion temperature of cooked rices are signicantly affected by the amylose content of the rice variety. The cooked rice with high amylose content has higher melting onset, peak and conclusion temperature. For YY1 and YX3003, the amylose content, protein content and lipid content were not signicantly different (Table 1), but the onset, peak and conclusion temperature are signicantly different (P < 0.05). This indicated that the starch structure properties may affect the onset, peak and conclusion temperature. Amylopectin retrogradation enthalpy (DHr) of cooked milled rice is shown in Table 2. Most recrystallization within cooked rices occurred during the rst 3 days of storage. The retrogradation

142

S.F. Yu et al. / Journal of Cereal Science 50 (2009) 139144

Table 3 Changes of hardness (N) of cooked rice during storage at 4  C for 0, 1, 3, 7, 11, and 14 days. Rice varieties Storage time (day) 0 SN2 TJ SJ4 DH2 NJ42 WY1 YY1 YX3003 SM6 TY15 KY21 0.37a,A 46.66b,A 49.37c,A 49.37c.A 48.54c,A 50.31c,A 53.64c,A 59.16d,A 65.29e,A 76.58f,A 82.43g,A (0.01) (1.54) (5.47) (3.98) (3.77) (4.62) (4.87) (2.61) (3.39) (6.16) (5.59) 1 0.16a,A 67.99b,B 69.92b,B 73.79b,B 70.87b,B 67.12b,B 71.40b,B 83.42c,B 79.76c,B 97.82d,B 100.01d,B (0.02) (4.58) (4.17) (5.70) (2.90) (5.63) (2.83) (3.04) (1.76) (4.18) (3.14) 3 0.02a,A 74.03b,C 81.30c,C 92.09d,C 93.54d,C 80.41c,C 84.74c,C 93.22d,C 89.91d,C 100.17e,B 103.64e,B (0.01) (3.23) (4.48) (6.39) (3.80) (5.79) (3.49) (2.35) (7.42) (5.13) (1.38) 7 65.41a,B 84.63b,D 94.15c,D 94.54c.C 94.47c,C 94.52c,D 101.60d,D 104.22d,D 102.23d,D 111.39d,C 110.27d,C (4.30) (4.65) (5.35) (5.97) (6.84) (6.15) (2.27) (6.15) (7.37) (3.37) (2.19) 11 68.46a,B 88.23b,D 93.89b,D 98.33c,C 95.62c,C 95.06c,D 105.88d,D 109.54d,D 103.24d,D 112.09e,C 115.37e,C (1.86) (5.12) (4.64) (4.99) (6.62) (5.20) (3.12) (3.54) (7.37) (2.29) (1.93) 14 72.04a,C 87.99b,D 92.47b,D 99.52c,C 94.37b,C 96.56b,D 102.35c,D 101.46c,D 101.34c,D 106.65c,C 108.76d,C (1.54) (5.77) (4.45) (5.75) (4.62) (5.36) (2.19) (2.46) (6.05) (3.32) (1.12)

Data are the mean standard deviation, n ! 3. ADNumbers followed by the same uppercase superscript letters in the same row are not different (P > 0.05). aeMeans in the same column followed by the same lowercase superscript letters are not different (P > 0.05).

process continued after this, but was then slower, and the retrogradation process nished in 14 days. Comparing the retrogradation enthalpy (DHr) of cooked milled rice from different cultivars varied from 0.24 to 3.41 J/g after pre-cooling (storage 0 day); KY21 had the highest DHr (3.41 J/g), followed by TY15 (3.11 J/g) and SM6 (2.79 J/g) while SJ4 showed the lowest DHr of 0.24 J/g. From these results, retrogradation of cooked rice occurred with the cooling process. This result was in agreement with research on rice starch retrogradation where rice starch gel began to retrograde at the temperature of 80w95  C during the cooling process (Ding et al., 2003). Increasing storage time from 0 to 14 days increased the starch retrogradation enthalpy for each variety of cooked rice. The amylose content showed a positive correlation with DHr during storage, as shown in Table 5. KY21 showed the highest retrogradation enthalpy from 3.41 to 11.13 J/g, followed by TY15, SM6 and YX3003, SN2, TJ. SJ4 showed the lowest retrogradation enthalpy from 0.24 J/g to 6.59 J/g, followed by DH2, NJ42, WY1 and YY1. It is interesting that the waxy rice, very low amylose content rice, intermediate amylose content and high amylose content rice retrograded quickly, the low amylose content rice (except YX3003) retrograded slowly. This suggests that retrogradation of cooked rice could be suppressed by a suitable amylose content of rice, the cooked rice with high amylose content is easy to retrograde, the cooked rice with low amylose content retrograded slowly. Starch retrogradation consists of two different processes: gelation of amylose solubilized during gelatinization and amylopectin recrystallization within the gelatinized granules which takes place on cooling and during storage (Ji et al., 2007; Perdon et al., 1999). Indeed, crystallisation consists of (1) nucleation, i.e. crystal nuclei formation, (2) propagation, i.e. nuclei crystal growth, and (3)

maturation, i.e. further crystal growth and/or perfection (Vandeputte et al., 2003). The amylose retrograded within <1 day (Biliaderis, 1992), and the retrograded amylose prepared for the crystal nuclei, which can increase the nuclei crystal growth, so the higher amylose content, the faster starch retrograded and the higher the DHr value. For waxy rice (SN2) and very low amylose content rice (TJ), the amylopectin content is very high, so amylopectin recrystallization is very easy, and showed high retrogradation enthalpy. On the other hand, the other components and rice starch structure, i.e. protein and lipid may also affect the retrogradation properties of cooked rice, but the relationship between protein, lipid and retrogradation are very complex. For YY1 and YX3003 rice varieties, the amylose content, protein and lipid were not signicantly different, but the onset, peak, conclusion temperatures and enthalpy are very different. This indicated that the starch molecular characteristics may affect the retrogradation properties (onset, peak, conclusion temperature and enthalpy). This is conrmed by the research of Vandeputte et al. (2003) who reported amylopectin chains with degree of polymerisation (DP) 6 9 and DP > 25 inhibited retrogradation, whereas relative amounts of DP 1222 were positively correlated with retrograded amylopectin conclusion temperatures and enthalpies. 3.3. Texture properties Cooked milled rice stored for 14 days was evaluated at intervals for textural properties using a texture analyser. Results of two major textural parameters (hardness and adhesiveness) are shown in Tables 3 and 4. The hardness of cooked rice increased as a consequence of storage. Fresh cooked rice had the lowest

Table 4 Changes of adhesiveness (N.s) of cooked rice during storage at 4  C for 0, 1, 3, 7, 11, and 14 days. Rice varieties Storage time (day) 0 SN2 TJ SJ4 DH2 NJ42 WY1 YY1 YX3003 SM6 TY1 KY21 7.41a,A 6.38b,A 6.96b,A 6.68b,A 6.60b,A 6.63b,A 5.24c,A 4.36d,A 2.29e,A 1.41f,A 1.38f,A (0.15) (0.22) (0.64) (0.80) (0.58) (0.67) (0.68) (0.41) (0.01) (0.26) (0.30) 1 5.24a,B 1.16b,B 3.48c,B 2.17d,B 2.18d,B 2.17d,B 1.18b,B 0.37e,B 0.40e,B 0.31e,B 0.14f,B (0.30) (0.36) (0.50) (0.33) (0.49) (0.21) (0.26) (0.04) (0.11) (0.04) (0.01) 3 3.26a,C 0.35b,C 1.03c,C 0.89c,d,C 0.68d,C 1.06c,C 0.34b,C 0.26b,C 0.21b,C 0.18b,d,C 0.07e,C (0.37) (0.09) (0.24) (0.10) (0.21) (0.31) (0.06) (0.03) (0.05) (0.01) (0.01) 7 1.29a,D 0.24b,D 0.76c,D 0.39b,D 0.25b,D 0.30b,D 0.24b,D 0.21b,C 0.14d,C 0.13d,C 0.06e,C (0.35) (0.06) (0.02) (0.10) (0.09) (0.14) (0.01) (0.02) (0.03) (0.03) (0.01) 11 0.69a,E 0.22b,D 0.61a,E 0.36c,D 0.19b,D 0.24b,D 0.20b,D 0.19b,C 0.12b,C 0.14b,C 0.05d,C (0.11) (0.06) (0.11) (0.11) (0.08) (0.08) (0.02) (0.02) (0.03) (0.01) (0.01) 14 0.24a,F 0.21a,D 0.35a,F 0.30a,D 0.19b,D 0.17b,D 0.14b,D 0.14b,C 0.08c,C 0.06c,C 0.04c,C (0.04) (0.06) (0.11) (0.08) (0.06) (0.04) (0.03) (0.02) (0.02) (0.02) (0.01)

Data are the means standard deviation, n ! 3. AFNumbers followed by the same uppercase superscript letters in the same row are not different (P > 0.05). afMeans in the same column followed by the same lowercase superscript letters are not different (P > 0.05). N.s: the unit of adhesiveness, N is the unit of force, s is the time sec.

S.F. Yu et al. / Journal of Cereal Science 50 (2009) 139144

143

hardness values and these values increased signicantly after 7 days of storage, slightly increased from 7 to 14 days. Among the cooked rice of different varieties, hardness of high amylose content rice (KY21) was observed to be highest, followed by intermediate amylose content rice (TY15 and SM6), low amylose content rice (YY1, WY1, NJ42, DH2 and SJ4) and very low amylose content rice (TJ), whereas waxy rice (SN2) had the lowest hardness during storage. Most of the loss of texture occurred during the rst 7 days of storage at 4  C. The Pearson correlation coefcients for the relationship between hardness and amylose content during storage are shown in Table 5. The hardness of cooked rice showed a positive correlation with amylose content (0.706 r 0.866, P < 0.01). From these results, hardness of cooked rice is affected by the amylose content of rice, the cooked rice with high amylose content is easy to harden, the cooked rice with lower amylose content hardens slowly. This suggests that hardness of cooked rice was mainly inuenced by amylose content. Amylose is known to leach out during cooking (Leelayuthsoontorn and Thipayarat, 2006). Rice with higher amylose content is liable to leach more into the cooking water, and generate a coated lm on rice grains (Leelayuthsoontorn and Thipayarat, 2006). Therefore, higher amylose leakage may form a thicker lm on the rice grains. Amylose retrograded rapidly and increased the hardness of cooked rice in a short time. On the other hand, the hardness is positively correlated with amylopectin retrogradation (r 0.650, P < 0.01) (Table 6). It is in agreement with the report of Perdon et al. (1999). The hardness of cooked rice increased continually with amylopectin retrogradation during storage time. This indicated that amylopectin retrogradation contributes to the hardness increase during storage, in agreement with the research of Keetels et al. (1996) who reported that amylopectin recrystallization resulted in an increased hardness of starch gels. From the results above, this suggested that both the amylose and amylopectin retrogradation contribute to the hardness of cooked rice. The hardness of fresh cooked rice (storage 0 day) may be mainly caused by amylose retrograded in a short time, and amylopectin retrogradation may contribute to the hardness increase during storage. However, an alternative explanation may be that amylose and amylopectin reaction occurs contributing to the hardness increase of cooked rice. The long amylopectin chains may crystallize with an amylose molecule, which might extend through several adjacent clusters, thereby contributing to double helices in several crystallites and which could result in a lower degree of swelling, a reduction in the leaching of material and ultimately giving rise to a harder cooking rice grain (Ong and Blanshard, 1995). However, it is difcult to explain the hardness increasing during storage. As shown in Table 4, the adhesiveness values with storage were observed to range from 1.38 to 7.41 N.s for cooked rice from different varieties. Fresh cooked rice grains (storage 0 day) from a waxy variety (SN2) was found to be more adhesive (7.41 N.s) than the fresh cooked rice from other varieties. High amylose content cooked rice (KY21) showed the lowest value for adhesiveness (1.38 N.s). Most adhesiveness decrease occurred during the rst 3 days storage. The adhesiveness decrease continued after this, but
Table 5 Pearson correlation coefcients for the relationship between amylose content, hardness, adhesiveness and DHr during storage. Hardness Amylose content 0.833** 0.802** 0.706** 0.847** 0.847** 0.866** (0 day) (1 day) (3 day) (7 day) (11 day) (14 day) Adhesiveness 0.929** 0.761** 0.678** 0.728** 0.721** 0.751** (0 day) (1 day) (3 day) (7 day) (11 day) (14 day)

Table 6 Pearson correlation coefcients for the relationship between DHr, hardness and adhesiveness of cooked rice.

DHr
Hardness Adhesiveness **P 0.01, *P 0.05. 0.650* 0.833**

Hardness 0.802**

was then slower from 7 to 14 days. As shown in Table 5, the adhesiveness value was negatively correlated with amylose content (0.929 r 0.678, P < 0.01). The rice cultivars with high amylose content showed lower adhesiveness. On the other hand, the adhesiveness of cooked rice showed a negative correlation with the amylopectin retrogradation enthalpy (r 0.833, P < 0.01) (Table 6). These results indicated that the adhesiveness was related to starch retrogradation. During cooking, the amylose and amylopectin leached out from the inside of rice grains and a coated lm of cooking liquid may form three dimensional networks on the surface of rice grains (Leelayuthsoontorn and Thipayarat, 2006). Then the amylose retrograded rapidly within less than 1 day (Biliaderis, 1992; Ding et al., 2003), The adhesiveness decreased rapidly on the rst day, which may have been caused by amylose retrogradation, then decreased slowly, possibly caused by amylopectin retrogradation. In summary, the texture properties (hardness and adhesiveness) of cooked rice correlated with amylose content and starch retrogradation. Starch retrogradation contributes to hardness increase and adhesiveness decrease of cooked rice, and the rice varieties with high amylose content were observed to have a higher hardness and lower adhesiveness. However, the other components (e.g. proteins and lipids) and rice starch structure also inuence the texture properties of cooked rice. It is very complex and the relationship needs to be further studied. 4. Conclusions The amylose levels affect starch retrogradation enthalpy and texture properties of cooked milled rice with different cultivars signicantly. The starch retrogradation enthalpy (DHr) was correlated with amylose content, the rice with high amylose content having a higher retrogradation enthalpy, the rice with a suitable amylose content retrograded slowly. The hardness of cooked milled rice positively correlated with amylose content (0.706 r 0.866, P < 0.01) and with DHr of cooked rice (r 0.650, P < 0.01) during storage. The adhesiveness had a negative correlation with amylose content (0.929 r 0.678, P < 0.01) and a negative correlation with DHr of cooked rice (r 0.833, P < 0.01) during storage. The starch retrogradation contributes to the changes of texture properties (hardness increase and adhesiveness decrease) of cooked rice during storage. Starch retrogradation of cooked rice occurred during the pre-cooling process. These results may help guide food processors to produce a good quality of cooked rice by controlling the amylose content and pre-cooling process. Acknowledgements This study was supported by the Science & Technology Council of Heilongjiang Province, China, under the contact No.C200804. Appendix. Supplementary information Supplementary information associated with this article can be found, in the online version, at doi: 10.1016/j.jcs.2009.04.003.

DHr
0.603** 0.822** 0.612** 0.625** 0.657** 0.708** (0 day) (1 day) (3 day) (7 day) (11 day) (14 day)

**P

0.01.

144

S.F. Yu et al. / Journal of Cereal Science 50 (2009) 139144 Kim, J.-O., Kim, W.-S., Shin, M.-S., Kwangiu, 1997. A comparative study on retrogradation of rice starch gels by DSC, X-ray and a-amylase methods. Starch/ Starke 49, 7175. Leelayuthsoontorn, P., Thipayarat, A., 2006. Textural and morphological changes of Jasmine rice under various elevated cooking conditions. Food Chemistry 96, 606613. Lu, Z.-H., Sasaki, T., Li, Y.-L., Yoshihashi, T., Li, L.-T., Kohyama, K., 2009. Effect of amylose content and rice type on dynamic viscoelasticity of a composite rice starch gel. Food Hydrocolloids. doi:10.1016/j.foodhyd.2009.01.009. Mariotti, M., Sinelli, N., Catenacci, F., Pagani, M.A., Lucisano, M., 2009. Retrogradation behaviour of milled and brown rice pastes during ageing. Journal of Cereal Science 49, 171177. Martin, M., Fitzgerald, M.A., 2002. Proteins in rice g grains inuence cooking properties. Journal of Cereal Science 36, 285294. Ong, M.H., Blanshard, J.M.V., 1995. Texture determinants in cooked, parboiled rice. I: Rice starch amylose and the ne structure of amylopectin. Journal of Cereal Science 21, 251260. Perdon, A.A., Siebenmorgen, T.J., Buescher, R.W., Gbur, E.E., 1999. Starch retrogradation and texture of cooked milled rice during storage. Journal of Food Science 64 (5), 828831. Ramesh, M., Ali, Z.S., Bhattacharya, K.R., 1999. Structure of rice starch and its relation to cooked rice texture. Carbohydrate Polymers 38, 337347. Singh, N., Kaur, L., Sodhi, N.S., Sekhon, K.S., 2005. Physicochemical, cooking and textural properties of milled rice from different Indian rice cultivars. Food Chemistry 89, 253259. Vandeputte, G.E., Vermeylen, R., Geeroms, J., Delcour, J.A., 2003. Rice starches. III. Structural aspects provide insight in amylopectin retrogradation properties and gel texture. Journal of Cereal Science 38, 6168. Varavinit, S., Shobsngob, S., Varanyanond, W., Chinachoti, P., Naivikul, O., 2003. Effect of amylose content on gelatinization, retrogradation and pasting properties of ours from different cultivars of Thai rice. Starch/Starke 55, 410415.

References
Adebooye, O.C., Singh, V., 2008. Physico-chemical properties of the ours and starches of two cowpea varieties (Vigna unguiculata (L.) Walp). Innovative Food Science and Emerging Technologies 9, 92100. AOAC, 2000. Association of Ofcial Analytical Chemists Ofcial Methods of Analysis. Method 920.87, 923.05, 17th ed. The Association, Washington, DC. Biliaderis, C.G., 1992. Structures and phase transitions of starch in food systems. Food Technology 46, 98109. Ding, W.P., Wang, Y.H., Ding, X.L., 2003. Study on the mechanism of gelatinization and retrogradation of rice starch. Cereal & Feed Industry (Chinese) 3, 1116. Hibi, Y., Hikone, 1998. Roles of water-soluble and water-insoluble carbohydrates in the gelatinization and retrogradation of rice starch. Starch/Starke 50, 474478. Ji, Y., Zhu, K.X., Qian, H.F., Zhou, H.M., 2007. Staling of cake prepared from rice our and sticky rice our. Food Chemistry 104, 5358. Juliano, B.O., 2005. Overview of rice and rice-based products. In: Proceedings of the World Rice Research Conference. Rice is Life: Scientic Perspectives for the 21st Century. International Rice Research Institute and Japan International Research Center for Agricultural Sciences, Tokyo and Tsukuba, Japan, pp. 268270. Juliano, B.O., Perez, C.M., Blankeney, A.B., Castello, T., Kongseree, N., Laihnelet, B., Laihnelet, B., Webb, D.B., 1981. International cooperative testing on the amylose content of milled rice. Starch/Starke 33, 157162. Karlsson, M.E., Eliasson, A.-C., 2003. Gelatinization and retrogradation of potato (Solanum tuberosum) starch in situ as assessed by differential scanning calorimetry (DSC). Lebensmittel-Wissenschaft Und-Technologie 36, 735741. Keetels, C.J.A.M., Oostergetel, G.T., Vliet, T.V., 1996. Recrystallization of amylopectin in concentrated starch gels. Carbohydrate Polymers 30, 6164.

Journal of Cereal Science 50 (2009) 145151

Contents lists available at ScienceDirect

Journal of Cereal Science


journal homepage: www.elsevier.com/locate/jcs

QTL mapping of grain quality traits in rice


Jue Lou a, b, Liang Chen b, *, Gaohong Yue c, Qiaojun Lou b, Hanwei Mei b, Liang Xiong b, Lijun Luo a, b, *
a

College of Plant Science and Technology, Huazhong Agricultural University, No. 1 Lion Mountain Street, Hongshan District, Wuhan 430070, China Shanghai Agrobiological Gene Center, 2901 Beidi Road, Shanghai 201106, China c Wenzhou Vocational College of Science and Technology, Wenzhou 325006, China
b

a r t i c l e i n f o
Article history: Received 10 December 2008 Received in revised form 18 April 2009 Accepted 22 April 2009 Keywords: Grain quality QTL mapping Rice

a b s t r a c t
Grain quality improvement is one of the most important goals in a rice breeding program. An indica variety with small grain size was crossed to a japonica variety with large grain size to construct a set of recombinant inbred lines (RILs) which was used to identify quantitative trait loci (QTLs) controlling eight grain quality traits. Based on a linkage map of 185 SSR markers, a total of 16 QTLs were mapped on six chromosomes. A pleiotropic main effect QTL (M-QTL) anked by RM3204 and RM16 on chromosome 3 inuences the grain length (GL), length width ratio (LWR) and head rice ratio (HRR), explaining the phenotypic variation of 46.0, 36.1 and 29.7%, respectively. A total of 18 epistatic QTLs were identied for all the traits except MRR, distributed on all the chromosomes except chromosome 10. Two M-QTLs for GL and one M-QTL for GW were involved in epistatic QTL. No signicant interaction between M-QTL or epistatic QTL and environment was detected except AC having signicant M-QTL by environment interaction with minor effect. GL and LWR have a signicant negative relation with HRR which might make it difcult to develop long grain with higher HRR in the rice breeding practice. 2009 Elsevier Ltd. All rights reserved.

1. Introduction Rice is one of the major staple cereal foods, feeding more than half of the world population. Both yield potential and grain quality are the priority traits in a rice breeding program. Selection on grain quality traits based on direct observation in the eld is inefcient because the quality traits are very complex and easily affected by the environment. With the recent development of DNA markers and linkage maps of rice, it has become possible for complex polygenic traits to be dissected into single Mendelian quantitative trait loci (QTL). Many QTLs for traits of agronomic importance have been detected and used in rice improvement by marker-assisted selection (MAS) (Bernardo, 2008). The primary components of rice grain quality inuencing the commercial value include appearance quality, milling quality, cookingeating quality and nutritional quality, which are determined by their physicalchemical properties and other sociocultural factors.

Abbreviations: AC, amylose content; BRR, brown rice ratio; GL, grain length; GW, grain width; HRR, head rice ratio; LWR, length width ratio; MAS, marker-assisted selection; M-QTL, main-QTL; MRR, milled rice ratio; QTL, quantitative trait locus; PC, protein content; PCR, polymerase chain reaction; RIL, recombinant inbred line. * Corresponding authors. Shanghai Agrobiological Gene Center, 2901 Beidi Road, Shanghai 201106, China. Tel.: 86 21 52230526; fax: 86 21 62204010. E-mail addresses: cl@sagc.org.cn, lijun@sagc.org.cn (L. Chen). 0733-5210/$ see front matter 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2009.04.005

Generally, the appearance quality of rice grain is essentially composed of grain length (GL), grain width (GW), grain thickness, grain shape dened as length:width ratio (LWR), the chalkiness of the endosperm and the translucency of the endosperm. The genetic basis of rice grain size has been studied extensively in the last decade (Aluko et al., 2004; Huang et al., 1997; Li et al., 2004a; Tan et al., 2000; Wan et al., 2005, 2006;). Among them, one QTL for GL was consistently detected around the pericentromeric region of chromosome 3, usually explaining the largest phenotypic variation. Eventually, GS3 underlying the QTL was cloned by using a BC3F2 population from a cross between Minghui63 and Chuan7 (Fan et al., 2006). Recently, GW2 (Song et al., 2007) for grain width and qSW5 (Shomura et al., 2008) for seed width were cloned. Milling quality is assessed by using three principal characteristics brown rice ratio (BRR), milled rice ratio (MRR), and head rice ratio (HRR). Much research on QTL mapping for milling quality has been reported (Aluko et al., 2004; Dong et al., 2004; Kepiro et al., 2008; Li et al., 2004a,b; Mei et al., 2002; Septiningsih et al., 2003; Tan et al., 2001). Some of it has studied the milling quality and appearance quality simultaneously. The eatingcooking quality of rice is usually evaluated by three major physical and chemical characteristics of the starch as indirect indices: amylose content (AC), gel consistency, and gelatinization temperature. The AC of rice, recognized as one of the most important determinants of eatingcooking quality, has been reported to be mainly controlled by the Wx gene on chromosome 6 (Fan et al., 2005; Tan et al., 1999; Wang et al., 2007).

146

J. Lou et al. / Journal of Cereal Science 50 (2009) 145151

There are several components inuencing nutritional quality of rice such as protein content (PC), amino acid content and fat content. Among them, PC has been considered as a main component. Although various researchers have so far shown different results on the genetic basis of PC, this trait displayed typical normal distribution and was affected by many small effect QTLs (Aluko et al., 2004; Hu et al., 2004; Tan et al., 2001). QTLs for PC on chromosome 1 and 6 have been detected repeatedly in those investigations. However, the studies mentioned above mainly focused on grain appearance and AC. Less attention was paid to grain milling and PC while fewer researchers have studied the four kinds of grain quality traits simultaneously and analyzed the mutual relationships among these traits. Here we report the QTL mapping of eight quality traits by using an RIL population derived from an indica/japonica cross grown under three environments, and analyzing the correlations among the eight quality traits. 2. Materials and methods 2.1. Plant materials and eld experiments A set of 286 F8 recombinant inbred lines (RILs) of rice was developed from a cross between indica cultivars Chuan7 (1000-grain weight is 10.4 g) and japonica variety Nanyangzhan (1000-grain weight is 41.6 g). In October of 2004 in Shanghai (E1), this RIL population and parents were used for genotypic analysis and phenotypic evaluation. Then these 286 lines and parents were used for trait investigation in October of 2006 in Shanghai (E2) and May of 2007 in Hainan (E3), respectively. Each plot consisted of three rows of 21 plants at a spacing pattern of 25 cm (between rows) by 20 cm (within rows). The eld trial was arranged in a randomized block design with three replications. Field management followed normal eld production practice. At maturity, each plot was harvested in bulk. 2.2. Quality evaluation Fully lled grains of individual lines were used for grain quality evaluation. BRR, MRR and HRR were measured in E2 and E3 while other traits were evaluated in three environments. GL and GW were measured by using a video system (Jeda801). The milling quality traits were investigated according to Chinese National Standard NY 147-88. Hulls were removed from 100 g of grains in duplicate by using a huller (SDL-A; CNRRI, Hangzhou, Zhejiang) to obtain brown rice. BRR is the ratio of brown rice weight to rice grain weight. Brown rice samples (50 g 2) were processed into milled rice in a desk-top rice miller (JNMJ 6; Taizhou, Zhejiang, China) removing embryo and bran. MRR is the ratio of milled rice weight to rice grain weight. Head rice includes the whole kernels and those having 80% of the complete kernels in milled rice. HRR is the ratio of head rice weight to rice grain weight. To measure AC and PC, brown rice was placed in a dry room with a constant humidity of 12% for 1 week to balance the moisture content. Samples were scanned by Vector 22/N-I FT-NIR (Bruker Optics, Germany). AC and PC were predicted from models of amylose content and protein content developed by Zhang et al. (2005) and Wu et al. (2006), respectively. 2.3. DNA preparation and PCR amplication DNA was extracted from fresh leaves of 286 RIL individuals and their parents using the CTAB method as described by Murray and Thompson (1980). The extracted DNA was dissolved in TE

buffer and tested for quality and quantity using a DU 640 nucleic acid and protein analyzer (Beckman Coulter Co.). Then these 286 DNA samples were diluted into 25 ng/ml with sterilized double distilled water and stored at 4  C for the polymerase chain reaction (PCR). PCR was performed with an initial 5-min period at 94  , followed by 35 cycles of 30 s of denaturing at 94  C, 30 s of annealing at 55  C, and 45 s of extension at 72  C, and a nal 5-min extension at 72  C. PCR products with large difference were separated on 3% agarose gel and detected by using a UV-GIS detection system (Shanghai Tanon Science and Technology Co., Ltd.). Otherwise, PCR products were separated on 5% denatured polyacrylamide gel electrophoresis and detected by silver staining (Xu et al., 2002). 2.4. Linkage map construction and data analysis A total of 185 SSR markers covering all 12 chromosomes were analyzed for this population. The genetic linkage map was constructed by using MapMaker/Exp V3.0 (Lincoln et al., 1992). It spanned a total of 1585.6 cM of genome size, with an average interval of 8.57 cM between adjacent markers. Phenotypic correlation analysis was calculated by using S-Plus for Windows V6.1. QTL analysis was conducted by QTLMapper V1.6 on the basis of the mixed model approach (Wang et al., 1999). A threshold of P 0.005 and LOD ! 2.5 was used to declare the signicant main effect QTL (M-QTL), digenic epistatic QTLs, and QTL (M-QTL or epistatic QTL) environment interaction. Contribution rate (H2) was estimated as percentage of variance explained by each locus or epistatic pair in proportion to the total phenotypic variance. QTLs were named following the popular nomenclature but in alphabetic order for QTLs on the same chromosome (McCouch et al., 1997). 3. Results 3.1. Trait performance of the parents and RILs There were distinct differences between parents on GL, GW, LWR, HRR, and moderate differences on BRR, MRR, AC and PC (Table 1, Figs. 1 and 2). Nanyangzhan grain is twice as long as that of Chuan7. The mean LWR of Chuan7 and Nanyangzhan are 2.27 and 3.63, respectively. Chuan7 has an HRR of 67.1% in E2 and 45.1% in E3, whereas the HRR of Nanyangzhan is only 10.4% in E2 and 5.7% in E3. The distributions of the eight traits in the population were continuous, indicating quantitative inheritance of these characters. Transgressive segregation with one or both directions occurred in all of the traits except GL in E3 (Table 1, Fig. 2). GL, LWR and HRR seemed to have bimodal distribution while GW, BRR, MRR, AC and PC showed a unimodal normal distribution (Fig. 2). 3.2. Correlations among eight rice quality characters The pairwise phenotypic correlation coefcients were generally consistent in different environments (Table 2). On the whole, there were signicant correlations between appearance quality and milling quality, appearance quality and nutritional quality. GL and LWR had weaker positive relation with BRR, but had a signicant negative relation with MRR and a very strong negative relation with HRR. The correlation coefcients between GL, LWR and HRR were 0.582, 0.490 in E2 and 0.840, 0.797 in E3, respectively. GW had no signicant correlation with grain milling quality traits. Among the three appearance quality traits, GL was signicantly positively correlated with LWR and not signicantly correlated with GW. GW had a signicant negative relation with LWR. For the three

J. Lou et al. / Journal of Cereal Science 50 (2009) 145151 8.01 1.53 3.00 0.25 2.70 0.59 76.71 1.77 61.50 3.44 27.96 14.90 18.88 3.13 12.07 1.70

147

Average Population Nanyangzhan Range

12.40 0.92 3.43 0.26 3.62 0.33 79.17 0.22 61.67 0.54 5.75 0.64 19.52 1.20 14.04 0.21

5.7011.91 2.373.65 1.824.12 71.0781.22 51.1368.75 2.3654.28 8.9626.60 9.1919.25

Fig. 1. Grain shape of the two parents used to construct the RI population.

5.64 0.41 2.56 0.17 2.20 0.21 75.73 0.02 63.24 0.08 45.11 0.14 21.26 1.46 13.94 0.26

8.25 1.46 2.89 0.26 2.88 0.61 78.40 1.63 63.76 2.80 38.76 12.95 17.78 3.45 13.56 1.20

milling quality traits, signicant positive correlation was observed between BRR and MRR, MRR and HRR. BRR was negatively related with HRR at a signicance level of 5%. There was no signicant correlation between AC and other quality traits. Signicant negative correlation was observed between PC and GL, PC and LWR, but unstably under different environments. 3.3. QTL mapping A total of 16 QTLs were identied for the eight traits across the three environments, distributed on six chromosomes with LOD values varying from 4.17 to 90.10 (Table 3, Fig. S1). The phenotypic variation of QTLs ranged from 1.90 to 46.00%. No signicant QTL by environment interaction was detected except AC. Two QTLs were detected for GL, collectively accounting for 48.45% of the total phenotypic variation (Table 3, Fig. S1). The larger effect QTL, qGL-3, anked by RM6283 and RM16 on chromosome 3, explained 46% of the variation, indicating that this QTL played a decisive role for GL in the RIL population. The allele from Nanyangzhan increased grain length by 0.99 mm with a LOD value of 90.1. Another QTL mapped to RM588RM540 on chromosome 6 had a contribution of 2.45% of the phenotypic variation. The allele from Nanyangzhan had a positive additive effect for qGL-6 by 0.23 mm. Four QTLs for GW on chromosomes 2, 6 and 9 totally explained 8.99% of the phenotypic variance, including two anked by RM221RM526 (qGW-2a) and RM3874RM5651 (qGW-2b) both on chromosome 2, one mapped to RM541RM3183 (qGW-6) on chromosome 6 and the other located to RM3808RM215 (qGW-9) on chromosome 9 (Table 3, Fig. S1). These QTLs accounted for 2.112.44% of the variance with LOD values varying from 4.17 to 5.14. The Nanyangzhan allele at each locus could increase GW by 0.04 mm. Only one QTL mapped to RM6283RM16 on chromosome 3 was detected for LWR (Table 3, Fig. S1). The major QTL, qLWR-3, accounted for 36.09% of the variation with an LOD score of 43.1. The allele from Nanyangzhan had a positive effect, which could increase LWR by 0.32. Two QTLs for BRR were identied (Table 3, Fig. S1). qBRR-1 in the interval of RM572RM582 explained 1.9% of the variation with a LOD value of 4.47. Another QTL (LOD 8.04), qBRR-3, located in the interval of RM16RM6266 explained 3.19% of the phenotypic variation. The additive effect of alleles from Nanyangzhan increased BRR by 0.30 and 0.39%, respectively. A minor QTL qMRR-3 anked RM3204RM6283 on chromosome 3 was detected for MRR and accounted for 6.65% of the total variation (Table 3, Fig. S1). An allele from Chuan7 contributed the positive effect for MRR at this locus by 0.86%. A major QTL qHRR-3 was associated with HRR (Table 3, Fig. S1). This locus in the interval of RM3204RM6283 on chromosome 3

Parents

Average Population Range

Chuan7

E3a

5.7911.79 2.293.70 1.874.48 74.7783.69 53.4470.13 6.1464.54 7.9424.52 10.1817.12 GL (mm) GW (mm) LWR BRR (%) MRR (%) HRR (%) AC (%) PC (%)
a

Table 1 Performance of eight grain quality traits of the parents and their RIL population in three cropping seasons.

Traits

E1, E2 and E3 represent Shanghai in October 2004, Shanghai in October 2006 and Hainan in May 2007, respectively.

Nanyangzhan Parents Chuan7 E2a

Population

Nanyangzhan

Range

6.13 0.37 2.61 0.15 2.38 0.21

22.31 1.57 12.86 0.23

E1a

Parents

Chuan7

20.61 0.90 13.19 0.03

13.30 0.68 3.69 0.21 3.61 0.21

10.72125.40 9.1118.39

5.3711.88 2.063.78 1.654.43

18.45 2.69 13.38 1.22

Average

8.82 1.57 3.10 0.26 2.86 0.59

6.11 0.44 2.46 0.17 2.48 0.23 77.94 0.01 68.25 0.38 67.14 0.64 22.36 1.14 13.38 0.11

13.01 0.41 3.55 0.23 3.66 0.20 83.20 0.02 65.17 0.17 10.39 .63 20.70 0.92 13.40 0.12

148

J. Lou et al. / Journal of Cereal Science 50 (2009) 145151

90 80 70 60 50 40 30 20 10 0

100 80

No. of lines

No. of lines
5.5 6.5 7.5 8.5 9.5 10.5 11.5

60 40 20 0 2.2 2.4 2.6 2.8 3 3.2 3.4 3.6

GL (mm)
100 80 80 70 60 50 40 30 20 10 8 10.5 13 15.5 18 20.5 23 25.5 0

GW (mm)

No. of lines

60 40 20 0

No. of lines

9.5 10.5 11.5 12.5 13.5 14.5 15.5 16.5 17.5 18.5

AC (%)
80 70 60 50 40 30 20 10 1.8 2.2 2.6 3.0 3.4 3.8 4.2 4.6 0 72 73.5 75

PC (%)

100

No. of lines

60 40 20 0

No. of lines

80

76.5

78

79.5

81

82.5

LWR
80 70 40 35 30 25 20 15 10 5 52 55 58 61 64 67 70 73 0 5

BRR (%)

No. of lines

50 40 30 20 10 0

No. of lines

60

10 15 20 25 30 35 40 45 50 55 60

MRR (%)

HRR (%)

Fig. 2. Phenotypic distribution of eight traits in the Chuan7/Nanyangzhan RI population across three environments. Mean values of Chuan7 (open arrow) and Nanyangzhan (Closed arrow) from three environments were shown above. White, grey and black columns represent traits evaluation in E1, E2 and E3, respectively.

explained 29.74% of the variation. The Chuan7 allele had a positive additive effect of 8.39% for HRR. In total, three QTLs were detected for AC. They were designated as qAC-2, qAC-6a and qAC-6b, respectively. The total contribution was 13.22% (Table 3, Fig. S1). qAC-2 (LOD 10.7) was identied in the region of RM525RM221 on chromosome 2, which explained 2.55% of the total phenotypic variation. The allele from Nanyangzhan had an additive effect of 0.54%. The other two QTLs on chromosome 6 accounted for 2.84 and 7.83% of the variation with LOD values of 14 and 26.7. Chuan7 alleles could increase the AC value at these two loci by 0.56 and 0.94%. All of the three loci were detected with signicant QTL by environment interaction, totally explaining 7.30% of the phenotypic

variance. Interaction between individual QTL and environment accounted for from 1.83 to 3.58% of the phenotypic variation. Signicant additive effects of AEi are listed in Table 3. Compared with the total effect of M-QTLs, the effect of QTL by environment interaction was smaller, implying that AC was inuenced by some M-QTLs in this RI population. Two QTLs for PC were detected, collectively explaining 7.19% of the variance (Table 3, Fig. S1). One was anked by RM588 RM540 on chromosome 6 and accounted for 4.5% of the variation. The Chuan7 allele at this locus increased PC by 0.30%. Another QTL qPC-7 was responsible for 2.69% of the total variation. The Nanyangzhan allele could increase the PC at this locus by 0.23%.

J. Lou et al. / Journal of Cereal Science 50 (2009) 145151 Table 2 Coefcients of pairwise correlation among the eight grain quality traits from a RI population derived from the cross of Chuan7 Nanyangzhan observed in E1a (upper), E2a (middle) and E3a (lower). GL GW 0.113 0.086 0.094 0.928**b 0.902** 0.914** 0.470** 0.211** MRR 0.280** 0.247** HRR 0.582** 0.840** 0.061 0.01 0.035 0.047 0.217** 0.339** 0.022 0.144* 0.165** 0.261** 0.097 0.134* 0.158* 0.004 0.490** 0.797** 0.122* 0.106 0.065 0.095 0.248** 0.302** 0.154* 0.141* 0.012 0.099 0.049 0.219** 0.477** 0.487** 0.043 0.029 0.115 0.023 0.064 0.035 0.231** 0.363** 0.109 0.01 0.029 0.155* 0.194** 0.190* 0.138* 0.376** 0.492** GW LWR BRR MRR HRR AC

149

LWR

0.437** 0.498** 0.481** 0.092 0.089 0.367** 0.148*b

BRR

AC

PC

a E1, E2 and E3 represent Shanghai in October 2004, Shanghai in October 2006 and Hainan in May 2007, respectively. b * and ** are signicant at 0.05 and 0.01 levels, respectively.

3.4. Digenic interaction Signicant epistatic QTLs are summarized in Table 4. No signicant epistatic QTL by environment interaction was detected. In total, 18 QTL pairs had signicant epistatic effect on all traits except MRR, covering all chromosomes except chromosome 10. The contribution rate of the epistatic QTLs ranged from 0.71 to 3.42%. Among these epistatic QTLs, two pairs of loci were for GL, ve for GW, two for LWR, two for BRR, one for HRR, four for AC and two for PC. But only three pairs were involved in M-QTL, two for GL and one for GW, accounting for 1.10, 1.02 and 0.93% of the phenotypic variation, respectively. 4. Discussion 4.1. QTL mapping of eight rice grain quality traits A total of 16 M-QTLs for the eight rice grain quality traits were identied in the present study (Table 3, Fig. S1). Three QTLs, qGW-9,

qBRR-1 and qPC-7, were novel compared with previous research (Huang et al., 1997; Li et al., 2004a,b; Redona and Mackill, 1998; Tan et al., 2000; Yu et al., 1997). Three QTL clusters were observed on chromosomes 3, 2 and 6. The rst QTL cluster anked by RM3204 and RM6266 covering the centromeric region of chromosome 3 contained ve QTLs for GL, LWR, BRR, MRR and HRR. The allele from Nanyangzhan at these loci increased GL, LWR and BRR, but decreased MRR and HRR. qGL-3 explained the largest phenotypic variation of 46% for GL. GS3 was initially mapped to the interval as qGL-3 by using a population derived from Minghui63 and Chuan7 which was also used in the present study (Fan et al., 2005, 2006). Therefore, GS3 and qGL-3 should be the same locus. The second QTL cluster was anked by RM525 and RM526 on chromosome 2 including qGW-2a, qAC-2 and qGW-2b. The QTLs in this cluster explained 4.66% of phenotypic variation for GW and 2.55% for AC. The Nanyangzhan allele of these QTLs could increase GW and AC. The third QTL cluster located to the interval of RM584RM540 on chromosome 6 spanned the Wx locus. It harbored QTLs for AC, GL and PC with small effect. The Chuan7 allele had a positive effect on AC and PC, but a negative effect on GL. Main effect QTLs for AC were not identied in this research. It may be due to the fact that Chuan 7 and Nanyangzhan contain the same indica Wx allele and have a similar amylose content. This result was also found by Bao et al. (2002) and Wan et al. (2004). QTLs with major effect for PC were not detected in the present research as others have reported (Aluko et al., 2004; Hu et al., 2004; Tan et al., 2001; Wang et al., 2007). Epistasis has been demonstrated as an important factor in the genetic basis for rice owering time and heterosis (Yamamoto et al., 2000; Yu et al., 1997). In this research, epistatic interaction also played an important role in determining rice grain quality (Tables 3 and 4). Among the eight traits surveyed, only MRR was not inuenced by epistatic interaction. Epistatic QTLs explained a larger proportion of the phenotypic variation than M-QTLs for GW. As to QTL environment interaction, only AC was affected signicantly by M-QTL environment interaction which explained 7.3% of the variation. 4.2. Quality improvement through marker-assisted selection based on QTL mapping QTL mapping results and correlation analysis showed that grain appearance quality had a close relation with milling quality in the present study. A QTL anked by RM6283 and RM16 had a major

Table 3 M-QTLs and Q E interaction inuencing the eight grain quality traits in an RILs population derived from the cross of Chuan7 Nanyangzhan. Trait GL GW QTL qGL-3 qGL-6 qGW-2a qGW-2b qGW-6 qGW-9 qLWR-3 qBRR-1 qBRR-3 qMRR-3 qHRR-3 qAC-2 qAC-6a qAC-6b qPC-6 qPC-7 Chr. 3 6 2 2 6 9 3 1 3 3 3 2 6 6 6 7 Interval RM6283RM16 RM588RM540 RM221RM526 RM3874RM5651 RM541RM3183 RM3808RM215 RM6283RM16 RM572RM582 RM16RM6266 RM3204RM6283 RM3204RM6283 RM525RM221 RM584RM585 RM588RM540 RM588RM540 RM5436RM6776 LOD 90.13 7.79 4.63 4.30 4.17 5.14 43.13 4.47 8.04 6.69 35.90 10.71 13.97 26.73 9.03 5.96 Aia 0.99 0.23 0.04 0.04 0.04 0.04 0.32 0.30 0.39 0.86 8.39 0.54 0.56 0.94 0.30 0.23 AEi1b AEi2b AEi3b H2(Ai)c 46.00 2.45 2.44 2.22 2.11 2.22 36.09 1.90 3.19 6.65 29.74 2.55 2.84 7.83 4.50 2.69 H2(AEi)c H2(A)d 48.45 8.99 H2(AE)d

LWR BRR MRR HRR AC

36.09 5.09 6.65 29.74 13.22

0.53 0.51 0.71 0.52

1.89 1.83 3.58

7.30

PC
a b c d

7.19

Ai is the additive effects of QTL. Positive values of additive effects indicate that the Chuan7 genotype have a positive effect on that trait. AEil, AEi2 and AEi3 are the additive effects of the environmental interaction from Ai and E1, Ai and E2, Ai and E3, respectively. H2(Ai) and H2(AEi) are the percentage of the phenotypic variation explained by Ai and AEi. H2(A) and H2(AE) are the collective percentages of the phenotypic variation explained by Ai and AEi for the trait.

150

J. Lou et al. / Journal of Cereal Science 50 (2009) 145151

Table 4 Digenic epistasis involved in rice grain quality traits in an RILs population derived from the cross of Chuan7 Nanyangzhan across different environments. Trait GL GW Ch-In i 35 311 16 25 27 315 57 11 54 310 47 25 213 315 316 49 218 218 Interval i RM282RM6080 RM6283RM16 RM582RM579 RM3874RM5651 RM6023RM475 RM168RM520 RM164RM430 RM24RM562 RM5770RM421 RM3204RM6283 RM252RM1223 RM3874RM5651 RM29RM6639 RM168RM520 RM520RM422 RM6089RM1153 RM5897RM6247 RM5897RM6247 QTL qGL-3 qGW-2b Ch-In j 311 126 74 813 91 113 816 81 612 1117 53 91 47 615 83 93 35 124 Interval j RM6283RM16 RM1103RM2854 RM1377RM542 RM547RM5556 RM316RM219 RM1812RM7557 RM1376RM152 RM3120RM3155 RM587RM584 RM2064RM224 RM6545RM5770 RM316RM219 RM252RM1223 RM586RM588 RM6948RM5493 RM6771RM7039 RM282RM6080 RM277RM4589 QTL qGL-3 LOD 92.83 89.02 8.42 6.39 7.21 7.64 6.63 6.49 8.21 7.29 6.37 8.48 7.55 17.91 9.91 8.33 6.69 7.55 Aia 0.11 0.98 0.04 0.03 0.04 Aja 0.972 AAijb 0.22 0.19 0.05 0.03 0.04 0.05 0.03 0.09 0.10 0.25 0.27 2.77 0.42 0.37 0.31 0.28 0.19 0.22 H2(Ai)c 0.32 27.6 2.19 1.07 2.0 H2(Aj)c 27.04 H2(AAij)c 1.1 1.02 2.77 0.93 1.77 3.42 1.23 2.44 3.41 1.25 1.47 2.88 1.66 1.26 0.87 0.71 1.51 2.06 H2(AA)d 2.12 10.12

LWR BRR HRR AC

5.85 2.72 2.88 4.50

0.26 2.32 0.734 0.32 0.393

1.32 2.02 4.98 0.96 1.43

PC
a

3.57

Ai and Aj are the additive effects of the test points i and j, respectively. Positive values of Ai and Aj imply that the Chuan7 genotype has a positive effect on that trait. AAij is the effect of additive-by-additive interaction between points i and j; a positive value indicates that the parental two-locus genotypes have a positive effect and that the recombinants had a negative effect. c H2(Ai), H2(Aj) and H2(AAij) are the percentages of the phenotypic variations explained by Ai, Aj and AAij, respectively. d H2(AA) is the collective percentage of the phenotypic variation explained by AAij for the trait.
b

effect on GL, LWR and HRR and a minor effect on BRR and MRR. But this QTL functions reversely on GL, LWR, BRR and MRR, HRR (Tables 2 and 3, Fig. 2). The coefcient between GL and HRR, LWR and HRR reached 0.840 and 0.797 in E3, respectively. However, several other researchers have shown different results from this. Aluko et al. (2004) conducted similar research to the present study and detected a QTL around the Wx on chromosome 6 inuencing AC, GC and HRR with major effect, PC, GL and LWR with minor effect, but the QTL functioned in the same way for all six traits. However, the results didnt coincide with the general understanding that long grain cultivars usually have lower head rice percentage. It therefore needs more studies to elucidate the genetic basis of grain quality traits and care should be taken for the application of qGL-3 to improve grain appearance quality by marker-assisted selection (MAS) in rice breeding programs while running the risk of decreasing HHR. No signicant correlation was observed between AC and other quality traits in the present research. AC is usually considered to be mainly controlled by the Wx gene with modications by other minor QTLs (Aluko et al., 2004; Septiningsih et al., 2003;Tan et al., 1999). Therefore, AC is suitable for MAS in a rice breeding program (Zhou et al., 2003). The present and other investigations (Aluko et al., 2004; Septiningsih et al., 2003; Tan et al., 2001) showed that PC had no or weak correlation with other quality traits and was governed by multiple minor QTLs. It means that PC is also suitable for MAS but it is relatively hard to get an obvious improvement by introgressing a single QTL.

References
Aluko, G., Martinez, C., Tohme, J., Castano, C., Bergman, C., Oard, J.H., 2004. QTL mapping of grain quality traits from the interspecic cross Oryza sativa O. glaberrima. Theoretical and Applied Genetics 109, 630639. Bao, J.S., Wu, Y.R., Hu, B., Wu, P., Cui, H.R., Shu, Q.Y., 2002. QTL for rice grain quality based on a DH population derived from parents with similar amylose content. Euphytica 128, 317324. Bernardo, R., 2008. Molecular markers and selection for complex traits in plant: learning from the last 20 years. Crop Science 48, 16491664. Dong, Y.J., Tsuzukia, E., Lin, D.Z., Kamiuntena, H., Teraoa, H., Matsuoa, M., Cheng, S.H., 2004. Molecular genetic mapping of quantitative trait loci for milling quality in rice (Oryza sativa L.). Journal of Cereal Science 40, 109114. Fan, C.C., Yu, X.Q., Xing, Y.Z., Xu, C.G., Luo, L.J., Zhang, Q.F., 2005. The main effects, epistatic effects and environmental interactions of QTLs on the cooking and eating quality of rice in a doubled-haploid line population. Theoretical and Applied Genetics 110, 14451452. Fan, C.C., Xing, Y.Z., Mao, H.L., Lu, T.T., Han, B., Xu, C.G., Li, X.H., Zhang, Q.F., 2006. GS3, a major QTL for grain length and weight and minor QTL for grain width and thickness in rice, encodes a putative transmembrane protein. Theoretical and Applied Genetics 112, 11641171. Hu, Z.L., Li, P., Zhou, M.Q., Zhang, Z.H., Wang, L.X., Zhu, L.H., Zhu, Y.G., 2004. Mapping of quantitative trait loci (QTLs) for rice protein and fat content using doubled haploid lines. Euphytica 135, 4754. Huang, N., Parco, A., Mew, T., Magpantay, G., McCouch, S.R., Guiderdoni, E., Xu, J.C., Subudhi, P., Angeles, E.R., Khush, G.S., 1997. RFLP mapping of isozymes, RAPD, and QTLs for grain shape, brown planthopper resistance in a doubled-haploid rice population. Molecular Breeding 3, 105113. Li, J.M., Xiao, J.H., Grandillo, S., Jiang, L.Y., Wan, Y.Z., Deng, Q.Y., Yuan, L.P., McCouch, S.R., 2004a. QTL detection for rice grain quality traits using an interspecic backcross population derived from cultivated Asian (O. sativa L.) and African (O. glaberrima S.) rice. Genome 47, 697704. Li, Z.F., Wan, J.M., Xia, J.F., Zhai, H.Q., Ikehashi, H., 2004b. Identication of quantitative trait loci underlying milling quality of rice (Oryza sativa) grains. Plant Breeding 123, 229234. Lincoln, S.E., Daly, M.J., Lander, E., 1992. Constructing genetic maps with MAPMAKER/EXP 3.0. Whitehead Institute Technical report, third ed. Whitehead Institute, Cambridge. Kepiro, J.L., McClung, A.M., Chen, M.H., Yeater, K.M., Fjellstrom, R.G., 2008. Mapping QTLs for milling yield and grain characteristics in a tropical japonica long grain cross. Journal of Cereal Science 48, 477485. McCouch, S.R., Cho, Y.G., Yano, M., Paul, E., Blinstrub, M., Morishima, H., Kinoshita, T., 1997. Suggestion for QTL nomenclature. Rice Genetics Newsletter 14, 1113. Mei, H.W., Luo, L.J., Guo, L.B., Wang, Y.P., Yu, X.Q., Ying, C.S., Li, Z.K., 2002. Molecular mapping of QTLs for rice milling yield traits. Acta Genetica Sinica 29, 791797. Murray, M.G., Thompson, W.F., 1980. Rapid isolation of high molecular weight plant DNA. Nucleic Acids Research 8, 43214325. Redona, E., Mackill, D.J., 1998. Quantitative trait locus analysis for rice panicle and grain characteristics. Theoretical and Applied Genetics 96, 957963. Septiningsih, E.M., Trijatmiko, K.R., Moeljopawiro, S., McCouch, S.R., 2003. Identication of quantitative trait loci for grain quality in an advanced backcross

Acknowledgements This research was supported by grants from State Key Development Program for Basic Research of China (2004BC117204) and from Shanghai Science and Technology Development Funds (06JC14062).

Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.jcs.2009.04.005.

J. Lou et al. / Journal of Cereal Science 50 (2009) 145151 population derived from the Oryza sativa variety IR64 and the wild relative O. rupogon. Theoretical and Applied Genetics 107, 14331441. Shomura, A., Izawa, T., Ebana, K., Ebitani, T., Kanegae, H., Konishi, S., Yano, M., 2008. Deletion in a gene associated with grain size increased yields during rice domestication. Nature Genetics 40, 10231028. Song, X.J., Wang, H., Shi, M., Zhu, M.Z., Lin, H.X., 2007. A QTL for rice grain width and weight encodes a previously unknown RING-type E3 ubiquitin ligase. Nature Genetics 39, 623630. Tan, Y.F., Li, J.X., Yu, S.B., Xing, Y.Z., Xu, C.G., Zhang, Q.F., 1999. The three important traits for cooking and eating qualities of rice grain are controlled by a single locus. Theoretical and Applied Genetics 99, 642648. Tan, Y.F., Xing, Y.Z., Li, J.X., Yu, S.B., Xu, C.G., Zhang, Q.F., 2000. Genetic bases of appearance quality of rice grains in Shanyou 63, an elite rice hybrid. Theoretical and Applied Genetics 101, 823829. Tan, Y.F., Sun, M., Xing, Y.Z., Hua, J.P., Sun, X.L., Zhang, Q.F., Corke, H., 2001. Mapping quantitative trait loci for milling quality, protein content and color characteristics of rice using a recombinant inbred line population derived from an elite rice hybrid. Theoretical and Applied Genetics 103, 10371045. Wan, X.Y., Wan, J.M., Su, C.C., Wang, C.M., Shen, W.B., Li, J.M., Wang, H.L., Jiang, L., Liu, S.J., Chen, L.M., Yasui, H., Yoshinura, A., 2004. QTL detection for eating quality of cooked rice in a population of chromosome segment substitution lines. Theoretical and Applied Genetics 110, 7179. Wan, X.Y., Wan, J.M., Weng, J.F., Jiang, L., Bi, J.C., Wang, C.M., Zhai, H.Q., 2005. Stability of QTLs for rice grain dimension and endosperm chalkiness characteristics across eight environments. Theoretical and Applied Genetics 110, 13341346.

151

Wan, X.Y., Wan, J.M., Jiang, L., Wang, J.K., Zhai, H.Q., Weng, J.F., Wang, H.L., Lei, C.L., Wang, J.L., Zhang, X., Cheng, Z.J., Guo, X.P., 2006. QTL analysis for rice grain length and ne mapping of an identied QTL with stable and major effects. Theoretical and Applied Genetics 112, 12581270. Wang, D.L., Zhu, J., Li, Z.K., Paterson, A.H., 1999. Mapping QTLs with epistatic effects and QTL environment interactions by mixed linear model approaches. Theoretical and Applied Genetics 99 551264. Wang, L.Q., Liu, W.J., Xu, Y., He, Y.Q., Luo, L.J., Xing, Y.Z., Xu, C.G., Zhang, Q.F., 2007. Genetic basis of 17 traits and viscosity parameters characterizing the eating and cooking quality of rice grain. Theoretical and Applied Genetics 115, 463476. Wu, J.H., Zhang, H.J., Mei, H.W., Li, Y., Yang, H., Wang, X.S., Lin, R.H., Luo, L.J., 2006. Study on adaptability of NIR models of protein content in rice. Scientia Agricultura Sinica 39, 24352440. Xu, S.B., Tao, Y.F., Yang, Z.Q., Chu, J.Y., 2002. A simple and rapid methods used for silver staining and gel preservation. Hereditas 24, 335336. Yamamoto, T., Lin, H.X., Sasaki, T., Yano, M., 2000. Identication of heading date quantitative trait loci Hd-6 and characterization of its epistatic interaction with Hd-2 in rice using advanced backcross progeny. Genetics 154, 885891. Yu, S.B., Li, J.X., Xu, C.G., Tan, Y.F., Gao, Y.J., Li, X.H., Zhang, Q.F., Saghai Maroof, M.A., 1997. Importance of epistasis as the genetic basis of heterosis in an elite rice hybrid. Proceedings of the National Academy of Sciences USA 94, 92269231. Zhang, H.J., Wu, J.H., Mei, H.W., Huang, F., Luo, L.J., 2005. Application of FT-NIRS in analysis of amylose content in rice. Plant Genetic Resources 6, 9195. Zhou, P.H., Tan, Y.F., He, Y.Q., Xu, C.G., Zhang, Q.F., 2003. Simultaneous improvement for four quality traits of Zhenshan 97, an elite parent of hybrid rice, by molecular marker-assisted selection. Theoretical and Applied Genetics 106, 326331.

Journal of Cereal Science 50 (2009) 152158

Contents lists available at ScienceDirect

Journal of Cereal Science


journal homepage: www.elsevier.com/locate/jcs

Optimization of air classication for the production of b-glucan-enriched barley ours


B. Ferrari, F. Finocchiaro*, A.M. Stanca, A. Gianinetti
CRA GPG, Genomic & Postgenomic Research Centre, Via S. Protaso 302, 29017 Fiorenzuola DArda (PC), Italy

a r t i c l e i n f o
Article history: Received 29 October 2008 Received in revised form 2 April 2009 Accepted 6 April 2009 Keywords: Barley b-Glucan enrichment Air classication

a b s t r a c t
Barley grains contain a relevant amount of soluble bre (i.e. b-glucans) and can be used for production of foods with healthy functions. To produce barley ours enriched in b-glucans, grain micronization and air classication of the our was used, and a method to predict the relationship between the yield of the enriched our and its b-glucan content was developed. This method is based on a series of sortings of micronized our, with a progressive increase in the yield of a selected fraction; a curve for b-glucan enrichment vs. yield can then be calculated. Thus, the most suitable combination of yield and b-glucan content can be chosen. We tested this approach on two hull-less barleys with different starch type, and obtained barley our fractions with twice the b-glucan concentration that the grain had. Enriched our fractions with 11.2 and 15.6% b-glucans were obtained from cultivars Priora (hull-less) and CDC Alamo (waxy, hull-less), respectively, with a good our yield (about 30%). We suggest that our method can be adopted for nely tuning the enrichment process to meet the industrys needs. 2009 Elsevier Ltd. All rights reserved.

1. Introduction In developed countries people tend to consume excessive amounts of carbohydrates, proteins and lipids, which increase their risk of suffering from chronic diseases (obesity, type 2 diabetes mellitus and coronary heart disease; Liu, 2002). In particular, cereals are used to produce rened ours containing almost only endosperm (starch), and not the germ and bran that provide bre, minerals, vitamins and other phytochemicals (Cordain et al., 2005). The evidence regarding the role of diet in the occurrence of several chronic diseases prompted interest in nutraceuticals (a term obtained by the fusion of nutrition and pharmaceutical) or functional foods. These are foods that provide a health benet beyond basic nutrition if they are regularly included in the normal daily diet, in sufcient quantities. Functional foods can be obtained using cereals as the main ingredients. In this regard, barley contains (1 / 3)(1 / 4) mixed linked b-glucans (referred to as b-glucans) in its caryopsis. These are non-starchy polysaccharides that represent the major structural components of endosperm cell walls (Fincher, 1975; MacGregor and Fincher, 1993). These polysaccharides are high molecular weight polymers of b-D-glucopyranose (Brennan and Cleary, 2005;

Abbreviations: CF, coarse fraction; FF, ne fraction; UBF, unsorted barley our. * Corresponding author. Tel.: 39 0523 983758; fax: 39 0523 983750. E-mail address: n.fra@libero.it (F. Finocchiaro). 0733-5210/$ see front matter 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2009.04.007

MacGregor and Fincher, 1993), whose linkages are about 30% b(1 / 3) and 70% b-(1 / 4). The (1 / 4)-linked b-D-glucopyranosyl residues (cellulose-like portions) occur in groups of two or three, separated by single (1 / 3) linkages (Fincher, 1975; Lazaridou and Biliaderis, 2007; Parrish et al., 1960), which make the molecule more exible and soluble. In fact, barley b-glucans form highly viscous solutions (Burkus and Temelli, 2005; MacGregor and Fincher, 1993) that can trap glucose and cholesterol and reduce their absorption (Jadhav et al., 1998). Barley b-glucans are considered soluble dietary bres with hypoglycaemic and hypocholesterolemic effects (Baik and Ullrich, 2008; Brennan and Cleary, 2005; Burkus and Temelli, 2005; Lazaridou and Biliaderis, 2007). Thus, b-glucans are regarded as important functional ingredients for the cereal foods industry, and interest in the inclusion of b-glucans in both cereal- and dairybased food systems has grown in recent years (Baik and Ullrich, 2008; Brennan and Cleary, 2005). The level of b-glucans varies among genotypes (mostly from about 2 to 10%; Baik and Ullrich, 2008; Henry, 1986, 1987; McCleary et al., 2006; Stuart et al., 1988). Hull-less barleys often have high b-glucan contents (Bhatty, 1992), and are mainly used as human food thanks to the ease in processing and edibility (Taketa et al., 2004). Barley genotypes also vary in the amylose/amylopectin ratio; both waxy starch (with up to 100% amylopectin) and high amylose (over 35%) varieties are known. Interestingly, these genotypes with anomalous starch composition contain higher b-glucan levels (Baik and Ullrich, 2008; Izydorczyk and Dexter, 2008; Wood et al., 2003; You and Izydorczyk, 2002).

B. Ferrari et al. / Journal of Cereal Science 50 (2009) 152158

153

Barley fractions enriched in b-glucans can be obtained by extraction with aqueous solvents (Burkus and Temelli, 1998). Cavallero et al. (2002) obtained an enriched fraction by water extracting the barley our coarse fraction with subsequent centrifugation and freeze-drying of the supernatant. However, these processes are highly energy-consuming and sometimes use nonedible extractants like NaOH or organic solvents, residues of which are prohibited in human foods. Another way to increase b-glucan content is milling and sieving of ours to discriminate b-glucan-rich fractions (Andersson et al., 2003; Bhatty, 1992; Izydorczyk et al., 2003; Kiryluk et al., 2000; Klamczynski and Czuchajowska, 1999; Knuckles and Chiu, 1995; Knuckles et al., 1992; Sundberg and man, 1994; Wu et al., 1994; Yoon et al., 1995). Air classication was sometimes reported to be more effective (Andersson et al., 2000; Vasanthan and Bhatty, 1995; Wu and Doehlert, 2002; Wu et al., 1994), but several steps had to be used to reach a good b-glucan increment in the our. In this work, we optimized a b-glucan enrichment system based on micronization of barley grains and subsequent air classication of ours to improve both b-glucan concentration and our yield, with a few air classication steps. 2. Experimental 2.1. Barley materials Priora (a hull-less Italian variety) and CDC Alamo (a waxy, hullless, Canadian genotype) were grown at Fiorenzuola DArda (northern Italy) in 2004. 2.2. Barley b-glucan enrichment Barley grains were micronized in a KMX-500 device (100200 kg/h; Separ Microsystem, Brescia, Italy), which consists of a steel drum containing a rotor operating at a peripheral speed of approximately 175 m/s. Particle size reduction is accomplished by mechanical impact against stator and rotor serrated surfaces and by turbulent collision between particles. The turbulence and large air volume ensure minimal temperature increase, and guarantee a continuous discharge of ne powder through a screenless drain pipe to a cyclone separator and a collector. The obtained our was fractioned by air classication with an SX-500 apparatus (Separ Microsystem, Brescia, Italy). This apparatus has a working capacity of 100200 kg/h and consists of a turboseparator (a highly modied cyclone) and a cyclone assembled in series. An aspirating pump mounted at the end of the system drives the air ow, which is modulated by means of an inlet valve. A further cyclone and a lter are placed before the pump to remove any very ne powder, which in our work never exceeded 1% of the initial our and is thus not discussed at any point. The apparatus sorted the our into two portions: a coarse fraction (CF) and a ne fraction (FF), which were collected from the turboseparator and the cyclone, respectively. The air ow inlet restriction valve was the only adjustable setting of the apparatus. The relative yield of the coarse and ne fractions could be changed in a consistent way by varying the air ow. As the set value of the air ow valve is of signicance only for the apparatus used here, the yield of the coarse fraction was assumed as a more correct reference parameter for the air classication process. The yield percentages reported for our fractions always refer to the initial un-fractionated barley our. Sorting was done in duplicate and a CV < 5% between replicates was obtained for all samples. 2.3. Chemical analyses

according to McCleary and Codd (1991). The moisture content of all samples was determined with a Precisa HA60 IR moisture analyser (Precisa Instruments, Diekinton, Germany). Data are reported on a dry basis and are the mean values of at least duplicate determinations. Total dietary bre (TDF) was determined by the Prosky method (Prosky et al., 1988). Crude protein was determined by the Kjeldhal method, using the nitrogen composition on a dry weight base (N 6.25). Starch was determined by hydrochloric acid dissolution according to ICC method no. 123/1 (ICC, 1994). Acceptable maximum difference among duplicate results was 0.2% for protein, starch and TDF and 0.4% for b-glucans. 2.4. Particle size distribution analysis The particle size distribution was established with laser diffraction using a laser granulometer Master Sizer 2000 Ver. 5.22 (model Hydro 2000S AWA, Malvern Instruments Ltd., UK) on ours dispersed in ethanol by means of ultrasound. Before the measurements, a blank (i.e. the dispersion liquid) was analysed to ensure that the system was clean. 2.5. Optical microscope observations Flours were dispersed in absolute ethanol and observed on a microscope glass slide, after addition of some drops of DMSO (Prolabo, Paris, France) and covering with a coverslip. Observations were done with a transmitted light brighteld microscope (Olympus BX51, Olympus Optical Co., Ltd., Tokyo, Japan) equipped with a Digital Camera System DP50 and images were captured with Studio Lite software (Pixera co., Osaka, Japan). 2.6. Apparent specic weight The apparent specic weight of ours was measured by progressively adding and settling (with gentle uttering of the recipient) the our into a volumetric beaker, to the 0.2 L level. The weight was registered when the apparent volume stabilized at the prexed level. 2.7. Statistical analysis All the data presented are mean values of at least two replicates. Raw values of chemical determinations and apparent specic weights were statistically analysed according to a two-way analysis of variance (ANOVA) with interaction, by using Systat 9.0 software (SPSS, Chicago, IL, USA). Genotype and our fraction were considered as xed factors. Within each variable, signicant differences among the means were assessed with the Tukeys test for pairwise comparisons (p 0.05). 3. Results 3.1. Barley b-glucan enrichment Preliminary experiments showed that two successive micronizations of the barley grains provided a sufciently ne our. Additional micronization tended to cause aggregation of the small particles into tiny pellets, with reduced efcacy of the subsequent fractioning (not shown). The double-micronized our of cultivar Priora was sorted with the air classier to obtain a series of coarse/ ne fraction pairs. Increasing yields of the coarse fraction (CF) were obtained by reducing the air ow setting (Fig. 1A). The increased yield was accompanied by an increased b-glucan level

b-Glucans were determined with a mixed-linkage b-glucan detection assay kit (Megazyme International Ltd., Wicklow, Ireland),

154

B. Ferrari et al. / Journal of Cereal Science 50 (2009) 152158

A
100% 80% yield beta-glucans

12 10

-glucans (%)

yield (%)

8 6 4 2 0

60% 40% 20% 0%

0 5% 14

valve regulation

B
-glucans (%)

18 16 14 12 10 8 6 4 2 0

6%

9%

2%

8%

4%

0%

9%

8%

6%

2%

15

0-

6-

-2

-2

-3

-4

-4

-5

-6

-7

9-

-7 72

15

22

28

34

40

49

58

66

yield coarse fraction (%)


Fig. 1. Optimization of the enrichment process by air classication. (A) Relationship between the yield (%) of coarse fraction (histogram) and inlet valve regulation. UBF is the unsorted barley our. The curve represents the measured b-glucan content (%) of the coarse fraction. (B) b-Glucan contents (%) for every our portion progressively incorporated in the coarse fraction were calculated from the corresponding data in plot A by using Eq. (2); the dashed line shows the b-glucan content of the unsorted barley our. On the x-axis, the two extremes of each yield interval indicate, for each our portion, the yield of the rst coarse fraction that includes such our portion (upper extreme) and the yield of the previous coarse fraction (lower extreme), respectively. Values are presented as mean standard error (n 2 for yield, and n 4 for b-glucan content).

up until the valve setting of 210, after which the b-glucan level peaked and then steadily decreased (Fig. 1A). Complementary results were obtained with the ne fraction (FF; not shown). The sequential increments in the yield of the coarse fraction consisted in increasing proportions of the ne fraction that were progressively included in the coarse fraction. Thus, the measured b-glucan concentrations in Fig. 1A were values averaged over the bulk of each sequentially collected coarse fraction. Therefore, the b-glucan concentrations of the coarse fractions changed in consequence of the b-glucan concentration in each of the our portions that were progressively included. Hence, the peak in the concentration of b-glucans in Fig. 1A revealed that our portions particularly rich in b-glucans had just been included. The actual b-glucan concentrations of the our portions that had been progressively incorporated in the coarse fraction were therefore calculated, based on the fact that the measured b-glucan concentration of each coarse fraction was a weighted average between the (unmeasured) b-glucan concentration of the newly added our portion and the measured b-glucan concentration of the previous coarse fraction. That is:

refers to the portion of our that is included in (CFx) over (CFx 1). Rearranging Eq. (1), we obtain:

%bGPx %bGCFx ,YCFx %bGCFx1 ,YCFx1 = YCFx YCFx1

75

-1

00

U BF

25

24

23

22

21

20

19

18

17

16

15

(2)

%bGCFx f%bGCFx1 ,YCFx1 %bGPx ,YCFx YCFx1 g=YCFx (1)

where %bG is b-glucan concentration, and Y is the yield. These two variables are denoted as follows: (CFx) refers to a given coarse fraction, (CFx 1) refers to the previous coarse fraction, and (Px)

The b-glucan concentrations of each portion of our are shown in Fig. 1B; they were calculated with Eq. (2) from the measured values of yield and b-glucan concentration of the corresponding coarse fractions in Fig. 1A. In this way it was possible to highlight the portions of our with the highest b-glucan concentrations (Fig. 1B). The graph in Fig. 1B enabled us to choose how the our should be sorted out, i.e. how close to the peak in b-glucan concentration, the left and the right cut-offs should be. Thus, the relationship between our yield and b-glucan concentration was made explicit and the best combination between these two factors could be adopted. Theoretically it should be possible to sort out the our fraction corresponding to the 2228% yield (with an actual yield of 6%), which had the maximum accumulation of b-glucans (Fig. 1B). It should be noted, however, that the presence/absence of fractions with quite different particle sizes in the sorting our may affect the working conditions inside the air classier and thus the subsequent sorting, especially when the fraction of sorted our is quite small. Based on these ndings it was decided to collect three fractions with high b-glucan contents, i.e. those corresponding to the progressive yield intervals of 22.128%, 28.134% and 34.140% (see Fig. 1B), for a total theoretical yield of about 18% and a theoretical

B. Ferrari et al. / Journal of Cereal Science 50 (2009) 152158

155

b-glucan concentration of 11.2% (as average of the b-glucan content


of the three our portions). In making this choice, it was considered that our portions to the left of the peak (coarser particles) had been subjected to a suboptimal micronization, which in an industrial process should be improved. The ow-sheet of the improved enrichment process is reported in Fig. 2. The settings identied in the previous experiment (Fig. 1) were used to remove both the coarse fraction (CF1) obtained when the air ow valve is set to 220 (corresponding to a yield of 22% in Fig. 1A), as well as the ne fraction (FF2) obtained when the air ow valve is set to 190 (this fraction represents approximately 60% of the whole our (Fig. 1A)). The coarse fraction (CF1) should be subjected to re-milling in industrial processing to increase overall yield. The coarse fraction CF2 was retained because it corresponds with the b-glucan peak of Fig. 1B, whilst the ne fraction FF2 had a low b-glucan concentration (Table 1) and could be used, for example, for monogastric feeding. There was a signicant rise in b-glucan concentration in the enriched fraction (CF2); from 7.8 to 15.6% for CDC Alamo and from 5.4 to 11.2% for Priora. 3.2. Chemical analyses The composition of our fractions, obtained with the enrichment process described above, is given in Table 1. Table 1 shows that, besides the b-glucan content, total dietary bre (TDF) also doubled in the CF2 fraction, whilst in the FF2 fraction, both b-glucan and TDF dropped. Hull-less or de-hulled barley grains contain 1120% TDF (Baik and Ullrich, 2008). In addition to b-glucans, the most notable components of barley bre are arabinoxylans (Bhatty, 1993; Izydorczyk and Dexter, 2008; Sundberg et al., 1995). The high level of TDF in CF2 is expected to contribute to the functional properties of these enriched ours (Baik and Ullrich, 2008; Brennan and Cleary, 2005). On the other hand, the starch content decreased in the coarse fractions (especially for cultivar CDC Alamo), whereas it increased in the FF2 (from 53.2 to 61.2% for CDC Alamo and from 55.3 to 66.7% for Priora). A higher level of starch in the nest fractions is also reported in studies utilizing oats (Wu and Doehlert, 2002; Wu and Stringfellow, 1995).

Table 1 Main components of barley ours. Genotype Flour fraction Yield (%) Protein (%)* Starch (%)* TDF (%)* b-Glucans (%)* 10.4 28.4 61.2 16.5 29.8 53.7 12.6e 14.0c 13.6d 12.6e 13.7d 15.6a 14.4b 12.4e 53.2d 42.8h 43.5g 61.2b 55.3c 51.1e 46.4f 66.7a 12.8d 25.5a 25.5a 3.8f 11.5e 23.4b 21.8c 3.3g 7.8c 11.1b 15.6a 2.3e 5.4d 8.1c 11.2b 3.0e

CDC Alamo Micronized CF1 CF2 FF2 Priora Micronized CF1 CF2 FF2

CF1 and CF2 are the coarse fractions from the rst and second air classication, respectively; FF2 is the ne fraction from the second air classication (see Fig. 2). ANOVA showed a signicant effect of genotype and our fraction as well as their interaction on each component (p < 0.001). * Values followed by the same letter are not signicantly different (p < 0.05; Tukey test).

3.3. Particle size distribution analysis and optical microscopic observations In Fig. 3 the particle size distribution of the double-micronized our and of the our fractions obtained after air classication (CF1, CF2 and FF2) are reported for the Priora sample. The particle size distribution of Priora and CDC Alamo micronized ours was very similar (not shown). Two main peaks were observed at 1520 mm and 200 mm and a lower peak occurred in the 24 mm range (Fig. 3A). Microscopic observations of ours (Fig. 4) showed that the largest particles (200 mm peak) are heterogeneous endosperm fragments, while the particles of about 20 mm correspond to free lenticular starch granules. Indeed, two types of starch granules are present in barley; large A-granules that represent most of the endosperm starch, and small B-granules that are present in large numbers but represent only a small percentage of total starch (Lindeboom et al., 2004). The size limit between the two types of granules is about 7 mm (Bohacenko et al., 2006). The peak at 20 mm (Fig. 3A,D) thus appears to be due to A-type starch granules, whereas the 3 mm particles are B-granules. It is also worth noting that the apparent density of the ours is inversely linked to their b-glucan content (compare Tables 1 and 2). The b-glucan-enriched fraction (CF2) is lighter than the other fractions because of the presence of partially emptied endosperm cells (Fig. 4C). Since both lighter (compare Fig. 3C and Table 2) and smaller particles (Fig. 3D) were separated from the coarse fraction CF1 (Fig. 3B), the rst air classication appears to have sorted out the our not only according to the particle size, but also to specic weight. After the second air classication, enriched fraction CF2 contained particles with a peak at 100200 mm (Fig. 3C). The nest fraction FF2 (Fig. 3D) was composed of A-type granules, and was poor in b-glucans and rich in starch (see Table 1). Thus, in the overall process, particles were separated both based on their size (the particles corresponding to the peak at 20 mm in Fig. 3A were isolated as a single peak in Fig. 3D) and within similar sizes (the peak on the right in Fig. 3A was split in two, as shown in Fig. 3B,C). Indeed, air classication separates our particles based on both their size and specic weight (Kuo and Tsai, 2001). Therefore, yield rather than particle size appears to be a more appropriate reference variable to characterize a our during the air classication process. 4. Discussion The enriched Priora our fraction (CF2), in which b-glucans are almost doubled (Table 1), had a yield of about one-third of the whole barley our (29.8%, Table 1). Although the b-glucan content was 11.2%, as theoretically expected, the yield was higher than the theoretical yield calculable from Fig. 1 summing the three our

Fig. 2. Flow-sheet of the enrichment process optimized for the cultivar Priora. The nal enriched fraction (CF2) is highlighted in grey.

156

B. Ferrari et al. / Journal of Cereal Science 50 (2009) 152158

Fig. 3. Particle size distribution (obtained by laser diffraction) of Priora ours. (A) Double-micronized whole our; (B) coarse fraction from the rst air classication (CF1); (C) coarse fraction from the second air classication (CF2); (D) ne fraction from the second air classication (FF2). In each plot, both the percentage distribution (left y-axis) and the cumulative percentage distribution (right y-axis) are shown.

Fig. 4. Microscope images of Priora ours. (A) Whole double-micronized our (both endosperm fragments and free starch granules are present); (B) coarse fraction CF1, with endosperm fragments; (C) coarse fraction CF2, with endosperm fragments having broken cell walls and less starch granules; (D) ne fraction FF2, wherein only free starch granules are visible. Black bars correspond to 100 mm.

B. Ferrari et al. / Journal of Cereal Science 50 (2009) 152158 Table 2 Apparent specic weight of the ours. Genotype Priora Flour fraction Micronized 2 CF1 CF2 FF2 Micronized 2x CF1 CF2 FF2 Apparent density (kg/m3)* 440c 491b 354d 570a 446bc 481bc 319d 566a

157

CDC Alamo

CF1, CF2 and FF2 are coarse, enriched and ne fractions, respectively. ANOVA showed a signicant effect of our fraction, but not of genotype as well as of the factor interaction, on the apparent specic weight (p 0.05). * Values followed by the same letter are not signicantly different (p < 0.05; Tukey test).

portions included in CF2, because of the tendency of the sorting to be slightly affected when the particle size distribution of the sorted our is changed (see Section 3.1.) such that a portion of coarser our was also probably included. Thus, the graph reported in Fig. 1B provides a straightforward way to approximate the results obtained by air classication and to modulate the process, allowing for optimization in the pilot plant before industrial processing. In fact, this method enables the two key cut-off points on either side of the b-glucan peak to be dened using a single series of samples, in which the yield is progressively increased. The optimization procedure is, therefore, greatly reduced with respect to a full twofactors approach in which all the possible combinations of the two yield cut-offs are tested. Several air classication systems for obtaining our fractions enriched in b-glucans are reported in the literature. Most of them require many cycles to reach a good b-glucan enrichment, in some cases with a low yield. Andersson et al. (2000) used air classication with seven different barley genotypes. In most cases, the b-glucan content was doubled but the our yield was not very high, less than 20%. A better result was obtained with the hull-less waxy cultivar Prowashonupana, which is a barley genotype that, although of poor agronomical performance, has a very high b-glucan level. However, the air classication system adopted by Andersson et al. (2000) was not very effective; the enriched our fraction had a yield of about 33%, but despite the very high initial b-glucan content (17%), it was only enriched to a content of about 22%. On the other hand, Wu et al. (1994) obtained, from Prowashonupana (17.4% b-glucans), a pooled fraction with 31% b-glucans and a yield of 62%, but defatting with hexane was required before the our could be processed. The same authors used air classication to obtain an enriched our fraction with 14.6% b-glucans and a our yield of 27.3%, using the hull-less high b-glucan (8.0%) barley CI 4362. This latter enrichment is similar to those presented in this paper, though Wu et al. (1994) used four air classication steps to obtain this result. Performing air classication on pin milled barleys with different amylose percentages, Vasanthan and Bhatty (1995) obtained nal coarse fractions with 1324% b-glucans. However, their best result (23.8% b-glucans) was obtained with the waxy cultivar SB89528 which gave a yield of only 7.6%. Similarly, a our fraction with 13.1% b-glucans and 10.4% yield was obtained with the normal-starch barley Condor. Our predictive graph (Fig. 1B) indicates that a similar enriched fraction could be obtained for Priora by combining the our portions corresponding to 22.128% and 28.134% yield intervals, with a yield of about 12%. This example demonstrates how the link between yield and enrichment can be made numerically explicit by building up a graph like the one shown in Fig. 1B. Optimization of roller milling and sieving using high b-glucan varieties accomplished yields >20% for bre-rich fractions, and the

resulting b-glucan contents were >15% (Izydorczyk et al., 2003). These results are similar to those obtained by air classication, though we set the air classier to obtain higher yields with slightly lower b-glucan contents. It therefore seems that in the continuous process of technical improvement, both sieving and air classication are tending to produce similar results. Once a our has been milled to a sufciently ne degree, both techniques appear to be able to sort out comparable b-glucan-enriched fractions. The quantitative prediction of the relationship between the enrichment and yield can then be computed by a b-glucan/yield diagram (or table) for sieving (Wu et al., 1994) as well as for air classication (Fig. 1). However, air classication has the advantage that the enrichment/yield combination can be easily adjusted by modulating the air ow. In addition, as air classication is based on both size and specic weight (Kuo and Tsai, 2001), the sorting out of the lighter b-glucan-enriched our fraction (Table 2) is straightforward and requires less successive steps. 5. Conclusion With the enrichment system reported in this study, a barley our fraction with a doubled b-glucan content and a yield of about one-third of the whole barley our was obtained from two commercial varieties, one waxy variety (CDC Alamo) and one with normal starch (Priora). We propose this method to separate our portions having different b-glucan contents. It can be adapted to different hull-less varieties and the enrichment/yield combination can be optimized in the pilot plant to meet the needs of various industrial applications. Using the b-glucan-enriched fractions as an ingredient in bakery products is the subject of further work in our laboratories. Acknowledgements This work has been supported by the following projects: Alimenti funzionali a base di cereali of the Ministry of University and Research (MIUR) and Filiera orzo per alimenti funzionali of the Regione Emilia Romagna and CRPV. We thank Dr Claudia Salati (Progeo) for providing part of the analyses on the chemical composition of the ours. References
Andersson, A.A.M., Andersson, R., man, P., 2000. Air classication of barley ours. Cereal Chemistry 77, 463467. Andersson, A.A.M., Armo, E., Grangeon, E., Fredriksson, H., Andersson, R., man, P., 2003. Milling performance of North European hull-less barleys and characterization of resultant millstreams. Cereal Chemistry 80, 667673. Baik, B.-K., Ullrich, S.E., 2008. Barley for food: characteristics, improvement, and renewed interest. Journal of Cereal Science 48, 233242. Bhatty, R.S., 1992. b-Glucan content and viscosities of barleys and their roller-milled our and bran products. Cereal Chemistry 69, 469471. Bhatty, R.S., 1993. Physicochemical properties of roller-milled barley bran and our. Cereal Chemistry 70, 397402. Bohacenko, I., Chmelk, J., Psota, V., 2006. Determination of the contents of A- and B-starches in barley using low angle laser light scattering. Czech Journal of Food Science 24, 1118. Brennan, C.S., Cleary, L.J., 2005. The potential use of cereal (1 / 3,1 / 4)-b-D-glucans as functional foods ingredients. Journal of Cereal Science 42, 113. Burkus, Z., Temelli, F., 1998. Effect of extraction conditions on yield, composition, and viscosity stability of barley b-glucan gum. Cereal Chemistry 75, 805809. Burkus, Z., Temelli, F., 2005. Rheological properties of barley b-glucan. Carbohydrate Polymers 59, 459465. Cavallero, A., Empilli, S., Brighenti, F., Stanca, A.M., 2002. High (1 / 3,1 / 4)-b-glucan barley fractions in bread making and their effects on human glycemic response. Journal of Cereal Science 36, 5966. Cordain, L., Eaton, S.B., Sebastian, A., Mann, N., Lindeberg, S., Watkins, B.A., OKeefe, J.H., Brand-Miller, J., 2005. Origins and evolution of the Western diet: health implications for the 21st century. American Journal of Clinical Nutrition 81, 341354.

158

B. Ferrari et al. / Journal of Cereal Science 50 (2009) 152158 McCleary, B.V., Codd, R., 1991. Measurement of (1 / 3),(1 / 4)-b-D-glucan in barley and oats: a streamlined enzymic procedure. Journal of the Science of Food and Agriculture 55, 303312. McCleary, B.V., Charnock, S.J., Rossiter, P.C., OShea, M.F., Power, A.M., Lloyd, R.M., 2006. Measurement of carbohydrates in grain, feed and food. Journal of the Science of Food and Agriculture 86, 16481661. Parrish, F.W., Perlin, A.S., Reese, E.T., 1960. Selective enzymolysis of poly-b-Dglucans, and the structure of the polymers. Canadian Journal of Chemistry 38, 20942104. Prosky, L., Asp, N.G., Schweizer, T.F., Devries, J.W., Furda, I., 1988. Determination of insoluble, soluble and total dietary ber in foods and food products: interlaboratory study. Journal of the Association of Ofcial Analytical Chemists 71, 10171023. Stuart, I.M., Loi, L., Fincher, G.B., 1988. Varietal and environmental variations in (1 / 3, 1 / 4)-b-Glucan Levels and (1 / 3, 1 / 4)-b-glucanase potential in barley: relationship to malting quality. Journal of Cereal Science 7, 6171. Sundberg, B., man, P., 1994. Fractionation of different types of barley by roller milling and sieving. Journal of Cereal Science 19, 179184. Sundberg, B., Tilly, A.-C., man, P., 1995. Enrichment of mixed-linked (1 / 3), (1 / 4)- b-D-glucans from a high-bre barley-milling stream by air classication and stack-sieving. Journal of Cereal Science 21, 205208. Taketa, S., Kikuchi, S., Awayama, T., Yamamoto, S., Ichii, M., Kawasaki, S., 2004. Monophyletic origin of naked barley inferred from molecular analyses of a marker closely linked to the naked caryopsis gene (nud). Theoretical and Applied Genetics 108, 12361242. Vasanthan, T., Bhatty, R.S., 1995. Starch purication after pin milling and air classication of waxy, normal and high amylose barleys. Cereal Chemistry 72, 379384. Wood, P.J., Weisz, J., Beer, M.U., Newman, C.W., Newman, R.K., 2003. Structure of (1 / 3) (1 / 4)-b-D-glucan in waxy and nonwaxy barley. Cereal Chemistry 80, 329332. Wu, Y.V., Doehlert, D.C., 2002. Enrichment of b-glucan in oat bran by ne grinding and air classication. Lebensmittel-Wissenschaft & Technologie 35, 3033. Wu, Y.V., Stringfellow, A.C., 1995. Enriched protein- and b-glucan fractions from high-protein oats by air classication. Cereal Chemistry 72, 132134. Wu, Y.V., Stringfellow, A.C., Inglett, G.E., 1994. Protein and b-glucan enriched fractions from high-protein, high b-glucan barleys by sieving and air classication. Cereal Chemistry 71, 220223. Yoon, S.H., Berglund, P.T., Faustnaught, C.E., 1995. Evaluation of selected barley cultivars and their fractions for b-glucan enrichment and viscosity. Cereal Chemistry 72, 187190. You, S., Izydorczyk, M., 2002. Molecular characteristics of barley starches with variable amylose content. Carbohydrate Polymers 49, 3342.

Fincher, G.B., 1975. Morphology and chemical composition of barley endosperm cell walls. Journal of the Institute of Brewing 81, 116122. Henry, R.J., 1986. Genetic and environment variation in the pentosan and b-glucan contents of barley, and their relation to malting quality. Journal of Cereal Science 4, 269277. Henry, R.J., 1987. Pentosan and (1-3), (1-4)-b-glucan concentrations in endosperm and whole grain of wheat, barley, oats and rye. Journal of Cereal Science 6, 253258. ICC, 1994. Standard methods of the International Association of Cereal Science and Technology. Determination of starch content by hydrochloric acid dissolution. The Association, Vienna, Austria. Izydorczyk, M.S., Dexter, J.E., 2008. Barley b-glucans and arabinoxylans: molecular structure, physicochemical properties, and uses in food products a review. Food Research International 41, 850868. Izydorczyk, M.S., Dexter, J.E., Desjardins, R.G., Rossnagel, B.G., Lagassee, S.L., Hatcher, D.W., 2003. Roller milling of Canadian hull-less barley: optimization of roller milling conditions and composition of mill streams. Cereal Chemistry 80, 637644. Jadhav, S.J., Lutz, S.E., Ghorpade, V.M., Salunkhe, D.K.,1998. Barley: chemistry and valueadded processing. Critical Reviews in Food Science and Nutrition 38, 123171. Kiryluk, J., Kawka, A., Gasiorowski, H., Chalcarz, A., Aniola, J., 2000. Milling of barley to obtain b-glucan enriched products. Nahrung/Food 44, 238241. Klamczynski, A.P., Czuchajowska, Z., 1999. Quality of ours from waxy and nonwaxy barley for production of baked products. Cereal Chemistry 76, 530535. Knuckles, B.E., Chiu, M.M., 1995. b-Glucan enrichment of barley fractions by air classication and sieving. Journal of Food Science 60, 10701074. Knuckles, B.E., Chiu, M.M., Betshart, A.A., 1992. b-Glucan-enriched fractions from laboratory-scale dry milling and sieving of barley and oats. Cereal Chemistry 69, 198202. Kuo, K., Tsai, C., 2001. On the theory of particle cutoff diameter and collection efciency of cyclones. Aerosol and Air Quality Research 1, 4756. Lazaridou, A., Biliaderis, C.G., 2007. Molecular aspects of cereal b-glucan functionality: physical properties, technological application and physiological effects. Journal of Cereal Science 46, 101118. Lindeboom, N., Chang, P.R., Tyler, R.T., 2004. Analytical, biochemical and physicochemical aspects of starch granule size, with emphasis on small granule starches: a review. Starch/Starke 56, 8999. Liu, S., 2002. Intake of rened carbohydrates and whole grain foods in relation to risk of type 2 diabetes mellitus and coronary heart disease. Journal of the American College of Nutrition 21, 298306. MacGregor, A.W., Fincher, G.B., 1993. Carbohydrates of the barley grain. In: MacGregor, A.W., Bhatty, R. (Eds.), Barley Chemistry and Technology. AACC International, St. Paul, pp. 73130.

Journal of Cereal Science 50 (2009) 159165

Contents lists available at ScienceDirect

Journal of Cereal Science


journal homepage: www.elsevier.com/locate/jcs

Study of the physical properties of karin during the fabrication of tablets for pharmaceutical applications
Abd Elmoneim O. Elkhalifa a, Dominique M.R. Georget b, *, Susan A. Barker b, Peter S. Belton b
a b

School of Pharmacy, Ahfad University for Women, Omdurman, Sudan School of Chemical Sciences and Pharmacy, University of East Anglia, Norwich, Norfolk NR4 7TJ, UK

a r t i c l e i n f o
Article history: Received 16 January 2009 Received in revised form 25 March 2009 Accepted 28 March 2009 Keywords: Sorghum Karin Protein body Tablets Mechanical properties Drug release

a b s t r a c t
Karin and protein bodies were extracted from a condensed tannin-free white Sudanese cultivar of sorghum (Dabar). The extracted materials were characterized by SDS-PAGE. The potential of karin as a tablet matrix for pharmaceutical applications was studied. Tablets composed of karin, calcium hydrogen orthophosphate, caffeine and magnesium stearate were prepared by direct compression. The tablets showed appropriate levels of hardness and friability. Drug release studies showed that caffeine dissolution was greater in 0.1 M HCl than in either phosphate buffer (pH 6.8) or distilled water. Deamidation of the protein in acid conditions might explain this observation. FTIR analysis showed that the secondary structure of karin was found to be mainly governed by a helices with some b sheets. Upon tabletting, there was a change in protein conformation, which was recovered upon dissolution irrespective of the dissolution media. This might be explained by the loss of protein coil to coil interaction during tabletting (possibly due to the diluting effect of calcium hydrogen orthophosphate). This was later recovered when tablets were dissolved due to the hydrophobic interactions between the karin proteins. In summary, this work has shown that karin has a potential for use as a tablet for drug delivery. 2009 Elsevier Ltd. All rights reserved.

1. Introduction There is a continuing interest in the production of biodegradable materials based on proteins (Gennadios, 2002; Taylor et al., 2006), which includes a wide range of plant-derived proteins such as soy protein, wheat gluten, and maize zein. The qualities of renewability, degradability, compostability and edibility make those biopolymers particularly appealing for both food and non-food applications (Taylor and Belton, 2002). Moreover, wide commercialisation of these macromolecules for non-traditional use will provide a valueadded innovative use for the traditional agricultural commodities that cultivate these plants. Although it is well adapted to the temperate European climate (Berenji and Dahlberg, 2004), sorghum (Sorghum bicolor (L) Moench) is a major crop in sub-Saharan Africa and is well suited to the climatic and soil conditions there. In Sudan, it accounts for 25% of the total production in Africa, the latter being the major world producer with an annual tonnage of 23 million tonnes (http:// www.fao.org). Karin, the prolamin protein fraction of sorghum,

* Corresponding author. Tel.: 44 01603 593 145; fax: 44 01603 592 003. E-mail address: d.georget@uea.ac.uk (D.M.R. Georget). 0733-5210/$ see front matter 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2009.03.010

which constitutes 6575% of the total protein in the whole grain (Hamaker et al., 1995), shares similarities with maize storage proteins in molecular weight, solubility, amino acid composition and structure (Belton et al., 2006; Shull et al., 1991). However, based on its hydration free energy (Duodu et al., 2003), it is considerably more hydrophobic than gluten. It may thus form an effective water barrier. Four classes of karin protein (Belton et al., 2006) have been identied. The a, b and g have been classied based on their solubility, electrophoretic mobility, amino acid composition and sequences, molecular masses and immunochemical cross-reactions. A fourth group, d karin, similar to d zein has been identied based on sequences of cloned DNA (Izquierdo and Godwin, 2005). This protein has not been characterized yet at the protein level. Karin is used interchangeably in the literature with karin proteins to denote the unfractionated material and will be used in this sense here. Like zein (Georget et al., 2008), karin may be considered for use as a novel excipient for pharmaceutical applications. Being a food commodity, it is generally recognised as safe (http://www.fda.gov) and it has previously been used as a substitute for gluten proteins to prepare products suitable for ingestion by people with coeliac disease (Taylor et al., 2006). Thus, karin would appear to be a suitable choice for investigation as a novel pharmaceutical excipient. Additionally, as pharmaceutical products

160

A.E.O. Elkhalifa et al. / Journal of Cereal Science 50 (2009) 159165

are typically value-added compared to food products, if karin were to be used widely in this context, it would result in increased income for the growers of sorghum. Solid oral dosage forms, such as tablets, are regarded as being the dosage forms most acceptable to patients, and may be designed to be either immediate-release or controlled-release. Controlledrelease tablet formulations are generally based on swellable or erodible polymer matrices, with hypromellose being a common polymer (Li et al., 2005). As the behaviour of a matrix tablet in a biological system will be governed by the chemical and physical nature of its excipients, it is important to study potential new matrix excipients to widen the formulation choice and the ability to design the drug release prole required for a particular application. There are few reports in the literature of plant proteins being used as primary tablet excipients. Inclusion of zein proteins in tablets was investigated by both Katayama and Kanke (1992) and Georget et al. (2008). The latter study showed that wet granulated tablets could successfully be produced with zein proteins comprising circa 60%w/w of the formulation. These tablets showed a sustained drug release prole controlled by polymer relaxation, which could be tted using the PeppasSahlin model. In the current research, two protein extracts, namely protein bodies and unfractionated karin, were produced from sorghum our and their biophysical properties were assessed. Protein bodies are storage organelles found in the endosperm of sorghum grains in which karin lay in a native state. The secondary structure of karin was studied using FTIR. Furthermore, the potential use of sorghum karin as a primary excipient for sustained-release tablets was explored. Direct compression tablets comprising karin, calcium hydrogen orthophosphate and caffeine (as a model drug) were produced and their physical and drug release proles studied. 2. Materials and methods 2.1. Materials Sorghum variety Dabar was grown in Sudan and harvested in the year 2006. Calcium hydrogen orthophosphate CaHPO4 (BDH Laboratory Supplies, Poole, UK), magnesium stearate (Sigma Aldrich Company Ltd, Gillingham, UK), sodium metabisulphite (BDH Laboratory Supplies, Poole, UK), calcium chloride (Sigma Aldrich Company Ltd, Gillingham, UK), a-amylase (Bacillus species, EC 3.2.1.1, 15003000 U/mg; SigmaAldrich, St. Louis, MO, USA), and caffeine (Alfa Aesar, Heysham, UK) were used as received. All other reagents (e.g. buffer components) used in this study were analytical grade and were used without further purication. 2.2. Karin extraction Sorghum karin was extracted from a condensed tannin-free white Sudanese cultivar (Dabar). Milled sorghum was extracted using a modied method (Carter and Reck, 1970) with a solution of 70% (v/v) absolute ethanol in distilled water and 0.5% (w/v) sodium metabisulphite at a ratio of 1:5 (w/w) our to extractant with vigorous stirring at 70 0.1  C for 1 h. Insoluble matter was removed by centrifugation at 1000 g for 5 min at ambient temperature. The clear supernatant was decanted into large glass Petri dishes, and the ethanol was removed from the extract by evaporation in a fume cabinet at ambient temperature until a viscous liquid sediment was formed. Chilled distilled water (10  C) of approximately equal volume was used to dilute the sediment. The pH was then reduced to pH 5 with 1 M HCl to precipitate all the proteins from the suspension. The proteins were then recovered by vacuum ltration through a Buchner funnel using two layers of Whatman No. 4 lter paper. The resulting wet

protein fraction was frozen at 20  C and then freeze-dried. The freeze-dried karin preparation was weighed and ground using a mortar and pestle into a ne powder, and then defatted three times for 1 h each with hexane at ambient temperature. The defatted protein powder was then air-dried in a fume cabinet and stored in an airtight bottle at 10  C. 2.3. Preparation of protein body-enriched our Protein bodies from condensed tannin-free white Sudanese sorghum cultivar (Dabar) were prepared as follows. The whole grain our was cooked in distilled water (1:5 our:water ratio) with constant stirring for 10 min at approximately 95  C to gelatinize the starch. The cooked sample was incubated with a-amylase, to solubilise gelatinized starch. To 240 g cooked sorghum porridge, 15.1 mL a-amylase solution (100 mg enzyme and 5 mg CaCl2 dissolved in distilled water and made up to 25 mL) was added and the mixture stirred at ambient temperature overnight. The sample was then centrifuged at 3500 g for 10 min, the supernatant discarded and the residue was washed several times with distilled water and centrifuged after each washing at 3500 g for 10 min. The nal residue was then freeze-dried. 2.4. SDS-PAGE The protein prole of karin and protein bodies was analysed by SDS-PAGE under reducing conditions using 12 cm long, 1 mm thick gels on a Hoeffer/Pharmacia Biotech vertical electrophoresis system (SE600) (Uppsala, Sweden) as earlier described (Duodu et al., 2002). 2.5. Scanning electron microscopy (SEM) Samples were mounted on an aluminum pin stub using conductive self-adhesive carbon labels. The powder specimens were sputter coated with a layer of gold approximately 50 nm thick in a sputter coater S150B (Edwards, UK). Fractured samples from tablets were coated with a layer of carbon approximately 50 nm thick in a Polaron CC7650 (Quorum Technologies Ltd, UK). All samples were examined in a JEOL 5900LV scanning electron microscope (JEOL Ltd, UK) at an accelerating voltage of 20 kV. 2.6. Fourier transform infrared (FTIR) spectroscopy FTIR was performed to characterize the extracted karin and to investigate any potential changes in the secondary structure of proteins on processing. Spectra were recorded on a Bruker FTIR spectrometer (Bruker Optics Limited, Coventry, UK, model IFS 66/S). Powder samples were placed on a single-reection diamond ATR (attenuated total reectance) accessory (SPECAC, Orpington, UK) and carefully pressed down to ensure a good contact with the ATR crystal. For each sample, 200 spectra at 2 cm1 resolution were averaged. The empty ATR crystal served as a reference. For further analysis, spectra for water vapour were measured and subtracted from the sample spectra using Omnic v6.1A software (Thermo Nicolet Cooperation, Madison, USA). Each spectrum was zeroed at 1800 cm1 where the baseline was relatively at. Fourier self deconvolution (FSD) was also carried out with an enhancement factor of 1.3 and a bandwidth of 30 (Georget and Belton, 2006). The positions of the absorbance peaks located in the amide I region were determined using the second derivative. The relative intensities at different wavenumbers (1670, 1650 and 1620 cm1) were calculated as the percentage of their total sum (Wellner et al., 2005). Measurements were triplicated and bars representing standard error of the mean are reported.

A.E.O. Elkhalifa et al. / Journal of Cereal Science 50 (2009) 159165

161

2.7. Tablet formulation and production Since it was not possible to generate tablets from protein bodies due to the signicant elastic nature of the material, what follows in this study describes the work undertaken on karin only. A Manesty E-2 single station tablet press (Manesty Machines Ltd, Liverpool, UK) tted with 12.7 mm normal concave punches, was used to manufacture the compressed material samples. A direct compression tablet formulation was used. Calcium hydrogen orthophosphate, which shows brittle fracture behaviour, was added to the formulation to compensate for the plastic deformation shown by the karin. Caffeine was used as a model drug and magnesium stearate was used as a lubricant. The basic formulation was as follows: karin 65.1% (w/w), calcium hydrogen orthophosphate 27.9% (w/w), caffeine 6% (w/w) and magnesium stearate 1% (w/w). Based on preliminary experiments, this level of karin was found to be optimum for processing purposes. The calcium hydrogen orthophosphate, karin proteins and caffeine were mixed using a glass mortar and pestle, and the magnesium stearate added just prior to compression. Sufcient material was prepared to load the feeding shoe. The target tablet weight was 500 mg. 2.8. Tablet physical properties

2.10. Statistical analysis Statistical analysis of FTIR and dissolution data was based on ANOVA using the software package SPSS. Signicant difference was statistically considered at the level of p < 0.05 using a Scheffe test. 3. Results and discussion 3.1. SDS-PAGE The extracted karin and protein bodies were evaluated by SDSPAGE to determine the general characteristics of the extracted proteins. Results obtained, shown in Fig. 1, demonstrated the presence of g-, a1-, a2-, and b-karin in the two samples, with a1-karin (Mr 2425 kDa) predominating in the protein bodies. The band patterns were in agreement with previous studies (Byaruhanga et al., 2007; El Nour et al., 1998; Mazhar and Chandrashekar, 1995; Oria et al., 2000). The lack of a2 karin in the protein bodies is accompanied by very weak bands appearing between 37 and 50 kDa and might well be explained by the formation of a-karin dimers during extraction as earlier reported (Duodu et al., 2002). 3.2. Microstructure

Tablet hardness was measured using an Erweka TBH 28 hardness tester (Erweka GmbH, Germany) (n 3). Friability was determined using Erweka TAR friabilator (Erweka GmbH, Germany). Ten tablets were preweighed before being placed in a perspex drum which was positioned vertically through a power controlled axis. After completion of 100 rotations, the dusted tablets were weighed. The friability was dened as follows:

wi wf 100 wf

(1)

where wi is the weight of tablets prior to testing and wf is the weight of tablets after test completion. For pharmaceutical applications, friability values should not exceed 1%. The disintegration time was measured in puried water, 0.1 M HCl and phosphate buffer pH 6.8 at 37 0.5  C in a Copley DTG 2000 apparatus using disks (n 3) (Copley Scientic Ltd, Nottingham, UK). One tablet was placed per glass tube of a basket rack. The latter was moved up and down at a frequency of 0.5 Hz in a beaker containing the uid. The time was recorded when no tablet residues were detected in the glass tubes. The standard error of the mean was calculated for all replicates. 2.9. Dissolution prole The tablet dissolution study was carried out using a British Pharmacopoeia apparatus II dissolution bath (Dis 8000 combined with a heating unit FH16-D, Copley Scientic Ltd, Nottingham, UK). The media used were distilled water, phosphate buffer (pH 6.8) and simulated gastric medium (0.1 M HCl pH 1.3) (British Pharmacopoeia, 2007). The volume of media used was 900 mL, the stirring speed was 50 rpm and the temperature was maintained at 37.0 0.5  C. At appropriate time intervals, aliquots of 10 mL were withdrawn and ltered with a 0.22 mm syringe-driven lter unit (Millipore, Cork, Ireland) and measured spectrophotometrically (S-22 Boeco UV/Vis spectrophoto-meter, Boeckel and Co, Hamburg, Germany) at lmax 270 nm corresponding to the lmax of caffeine. The experiments were carried out in triplicate. Hence the mean values with the standard error of the mean represented as bars are reported. Fitting of the dissolution proles was carried out using the software package SPSS (SPSS UK Ltd, Woking, UK).

SEM micrographs are shown in Fig. 2 for sorghum our, dry karin extract and the fracture surface of a karin tablet containing caffeine. Sorghum our (Fig. 2A) comprises mainly compacted starch granules with size greater than 10 mm, mixed with protein bodies with diameter less than 5 mm. These micrographs are very similar to previously published work (Elkhalifa et al., 2006; Glennie et al., 1983; Rom et al., 1992). The extract containing karin (Fig. 2B) comprises mainly spherical particles with diameters ranging from 1 to 5 mm. It is likely that the karin proteins underwent aggregation giving rise to particles with diameters greater than those of the protein bodies. Treatment in 70% ethanol led to solubilization of karin and consequent disruption of the organization of the protein bodies might have taken place. Upon precipitation, karin proteins aggregated via weak interactions or possibly covalent bonds as alluded to by Duodu et al. (2003), which resulted in the formation of particles with size greater than that of protein bodies. The micrograph of the fracture surface of the tablet (Fig. 2C) shows a continuous matrix with some spherical particles with diameters

Fig. 1. SDS-PAGE of karin extracted from sorghum and protein-body-enriched samples extracted from whole grain under reducing conditions. 1 molecular weight standards; 2 karin proteins; 3 protein-body-enriched our.

162

A.E.O. Elkhalifa et al. / Journal of Cereal Science 50 (2009) 159165

tablets with suitable hardness value (50.9 N 0.57), indicating that the tablets would be able to resist the attrition possible in future handling operations such as lm-coating, packaging and transport. Given the high content of proteins in the tablet, it is not surprising to observe a low value of hardness. This is in line with work previously published (Georget et al., 2008).

3.4. Secondary structure of karin FTIR was used to investigate the secondary structure of the sorghum our, extracted karin and karin tablets containing caffeine. Since caffeine absorbs in the same region as the karin proteins (Nolasco et al., 2006), the spectrum of caffeine was subtracted from spectra obtained on tablets. Absorbance peaks at 745 and 759 cm1 from caffeine were used as markers to assess their absence in the subtracted spectra. The spectra in the amide I and amide II regions are shown in Fig. 3. The amide I region is assigned to a vibrational mode that originates from the CO stretching of the amide group, coupled to the in-phase bending of the NH bond and the stretching of the CN bond, and gives rise to infrared bands in the region between approximately 1600 and 1700 cm1 (Surewicz et al., 1993). The amide II appears between 1480 and 1575 cm1 and is attributed to NH bending combined with CN stretching. In the spectrum for sorghum our, an absorbance peak occurred at 1748 cm1 and is assigned to C]O vibration from the lipids. Upon defatting during extraction, this peak disappeared. The assignment of the valleys appearing in the second derivative followed that previously described (Georget et al., 2008). Two valleys predominate, one at 1651 cm1 attributed to random coils and a helices (Duodu et al., 2001; Gao et al., 2006) and the other at 1620 cm1 assigned to intermolecular b sheets (data not shown) (Byaruhanga et al., 2006; Gao et al., 2005; Rahmelow and Hubner, 1996; Taylor and Belton, 2002). The percentages of b turns, a helices and intermolecular b sheets were calculated to give a fuller picture of the changes in the secondary structure of the karin proteins upon extraction and processing. In Fig. 4, the percentages of the different conformations for the various samples are represented. Within the sample of untreated sorghum our, there was a predominance of random coils and a helices at circa 40%, with essentially equal contributions from b turns and intermolecular b sheets, at circa 30% each. After extraction of the karin from the sorghum our, the levels of both random coils and a helices and of b turns decreased, whereas there was an increase in the contribution of intermolecular b sheets, all effects being signicant. The occurrence of b turns is strongly associated with polyglutamine segments found between
Fig. 2. (A) Micrograph of sorghum our, s: starch granule, pb: protein bodies; (B) micrograph of dry karin extract; and (C) micrograph of fracture surface of karin tablet containing caffeine. Scale bar: 10 mm.

0.18

Absorbance (Arbitrary units)

0.16 0.14 0.12 0.1 0.08 0.06 0.04 0.02 0 1850 1800 1750 1700 1650 1600 1550 Sorghum flour Kafirin Kafirin tablet 1500 1450

in the region of 15 mm. These are most likely to be the protein aggregates and their spherical nature in this micrograph indicates that during compression they deformed elastically, rather than fracturing. In fact, micrographs of the current karin tablet formulation are very similar to those previously observed (Georget et al., 2008) for tablets composed of zein and theophylline, suggesting that zein and karin show similar compaction properties. 3.3. Tablet physical properties The prepared tablets complied with the general British Pharmacopeia (BP) limits for weight uniformity (British Pharmacopoeia, 2007) (498.86 6.29 mg) and thickness (4.52 0.05 mm). Moreover, an acceptable value of friability (0.21%) was obtained for the

Wavenumber (cm-1)
Fig. 3. FTIR spectra of sorghum our, karin and karin tablets in the amide I (1650 cm1) and amide II (between 1550 and 1500 cm1) regions.

A.E.O. Elkhalifa et al. / Journal of Cereal Science 50 (2009) 159165

163

45 a 40 b a b c d c,e , d,e b b,c c d b c,d b,c d b

Percentage of conformation

35 a 30 25 20 15 10 5 0
Sorghum flour Kafirin extract Kafirin tablets Kafirin tablets Kafirin tablets Kafirin tablets dissolved in water dissolved in pH 6.8 dissolved in pH 1

Various samples
beta turns random coils and alpha helices intermolecular beta sheets

Fig. 4. Changes in the secondary structure of proteins from sorghum our, karin extract, karin proteins from tablets and karin proteins from tablets dissolved in water, pH 6.8 and pH 1.3. For a given conformation, means with the same letter are not signicantly (p < 0.05) different.

two consecutive a helices in maize (Momany et al., 2006) and millet prolamins (Bugs et al., 2004). In the case of karin, sequences of amino acids containing glutamine are also present (Belton et al., 2006) and it is plausible that unwinding of the turns/loops might occur upon extraction. It is noteworthy that b turns could be associated with coil to coil interactions. In sorghum our, protein bodies remain intact and karin proteins are in their native state as conrmed by the SEM (Fig. 2A). The coil to coil interactions will be favored due to the tight packing of the karin proteins in the protein bodies and hence more b turns. After tabletting, the percentage of intermolecular b sheets decreased slightly with a commensurate increase in b turns. This may possibly be ascribed to a reduced interaction between protein coils, consequent upon the diluting effect of the remainder of the tablet components. At the end of the dissolution study, the original karin conformation was recovered, irrespective of which dissolution medium was used. Recovery of the hydrophobic protein coil to coil interactions would be a reasonable explanation for this. It is interesting to note that the valleys ascribed to b turns (1662 cm1) and intramolecular b sheets (1637 cm1) in zein (Georget et al., 2008) were detected to a lesser extent in the karin extract (Fig. 4). The valley ascribed to intramolecular b sheets was only present as a shoulder in karin, perhaps indicating a difference in the secondary structure between zein and karin. 3.5. Disintegration and dissolution studies

stomach in vivo. Taken together, the disintegration data suggest that karin will provide a sustained-release matrix for oral tablets only if it could be protected against interaction with HCl in the stomach, by, for example, an enteric lm coat. However, as the dissolution data show (discussed below), the release of caffeine from the karin matrix was slow enough, even at pH 1.3, to suggest that the tablet was not disintegrating rapidly in the dissolution test. The BP disintegration test is quite vigorous and imparts more shear to the tablets than the BP dissolution test, hence it would be expected that tablets would show increased fragmentation in the disintegration test than the dissolution test. The potential for the use of karin as a sustained-release tablet matrix is based on its relative lack of aqueous solubility. However, to date there are no reports in the literature examining this potential. In this study, the release of a model drug, caffeine, from the karin tablet matrix was tested in three different media: distilled water, 0.1 M HCl pH 1.3 to simulate the stomach conditions, and phosphate buffer pH 6.8 to simulate the small intestine conditions. In all three media, the dissolution data were extremely reproducible and a sustained-release effect was observed, as shown in
90 80 70 60 a d

% release

The disintegration time indicates how long the tablets take to fragment in various media and, together with the results from the dissolution test, provide an indication of whether the karin will act as an immediate-release or a controlled-release oral tablet matrix. In distilled water the tablets were still intact after 4 h, which is ascribed to the fact that karin is highly hydrophobic. Tablet disintegration took circa 76 0.50 minin pH 6.8 buffer, indicating some interaction of the protein with the medium. However, tablet disintegration in 0.1 M HCl was much faster at circa 11 0.57 min. It is suspected that under acidic conditions, deamidation of the karin proteins might take place and consequently, this will enhance tablet disintegration. The measured disintegration time in 0.1 M HCl is within the normal residence time of tablets in the fasted stomach, suggesting that the tablet will disintegrate in the

50 40 30 20 10 0 0 50 100 150 200

250

300

Time (mins)
Fig. 5. Dissolution proles for karin tablets containing caffeine: distilled water (solid diamond), pH 1.3 (open triangle) and pH 6.8 (open square) showing: (a) dry karin tablet; (b) karin tablet after exposure to pH 6.8; (c) karin tablet after exposure to water; and (d) after exposure to pH 1.3 (tablet split into two to show the coloration) (scale bar: 10 mm).

164

A.E.O. Elkhalifa et al. / Journal of Cereal Science 50 (2009) 159165

Fig. 5. Drug release was slightly faster in distilled water than in the pH 6.8 phosphate buffer, and the release in 0.1 M HCl was signicantly faster than in either of the other two media. As an example, after 4.5 h, 77, 53, and 46% of caffeine had been released from the tablets in 0.1 M HCl, distilled water and pH 6.8 phosphate buffer, respectively. The difference in the drug release observed here between the two more biologically-relevant media, i.e. the 0.1 M HCl and the pH 6.8 phosphate buffer, may be attributed to a difference in the physical interaction of karin with the two media. Besides their pH values, these media differed also in their osmolality (pH 1.3: w100 mosmol, pH 6.8: w400 mosmol) and the ion type (pH 1.3: chloride, pH 6.8: phosphate), so the swelling behaviour of the karin tablet in these media may be expected to be different. Terebesi (2006) reported similar results with zein, when it was used as a tablet coating. There is also the possibility that deamidation might occur at pH 1.3 (Georget et al., 2008) leading to different results in the two media. In order to assess the release mechanism of caffeine from the karin tablet matrix, the dissolution data were tted mathematically to the power law previously detailed by Ritger and Peppas (1987) and shown here in Eq. (2).

4. Conclusion In this study, karin and protein bodies were extracted from sorghum our using a straightforward method. There was a change in the molecular weight prole of the two extracts, which could be explained by oligomerisation of the karin proteins during the preparation of the protein bodies. The secondary structure of karin after extraction was mainly governed by a helices with some b sheets as previously reported in the literature. A detail analysis of karin showed that upon tabletting there was a change in conformation of karin. Upon prolonged exposure to aqueous media, irrespective of pH, the conformation of karin was recovered. The potential of karin to act as a sustained-release tablet matrix was demonstrated using caffeine as a model drug, with drug release being observed over several hours. The rate of drug release showed some pH dependence, being faster at pH 1.3, and slower at pH 6.8, although given the expected residence time of the tablet in the stomach in fasted conditions, this may not be problematic. Acknowledgements The nancial support of The Royal Society, UK, for this work is gratefully acknowledged. Also the authors wish to thank Dr Allison Lewin (School of Chemical Sciences and Pharmacy, UEA) for the SDS-PAGE. References
Awika, J.M., Rooney, L.W., Waniska, R.D., 2004. Properties of 3-deoxyanthocyanins from sorghum. Journal of Agricultural and Food Chemistry 52, 43884394. Belton, P.S., Delgadillo, I., Halford, N.G., Shewry, P.R., 2006. Karin structure and functionality. Journal of Cereal Science 44, 272286. Berenji, J., Dahlberg, J., 2004. Perspectives of sorghum in Europe. Journal of Agronomy and Crop Science 190, 332338. British Pharmacopoeia. 2007. Medicines and Healthcare Products Regulatory Agency, London, UK. Bugs, M.R., Forato, L.A., Bortoleto-Bugs, R.K., Fischer, H., Mascarenhas, Y.P., Ward, R.J., Colnago, L.A., 2004. Spectroscopic characterization and structural modeling of prolamin from maize and pearl millet. European Biophysics Journal 33, 335343. Byaruhanga, Y.B., Emmambux, M.N., Belton, P.S., Wellner, N., Ng, K.G., Taylor, J.R.N., 2006. Alteration of karin and karin lm structure by heating with microwave energy and tannin complexation. Journal of Agricultural and Food Chemistry 54, 41984207. Byaruhanga, Y.B., Erasmus, C., Emmambux, M.N., Taylor, J.R.N., 2007. Effect of heating cast karin lms on their functional properties. Journal of the Science of Food and Agriculture 87, 167175. Carter, R., Reck, D.R., 1970. Low temperature solvent extraction process for producing high purity zein. US Patent 3,535,305. Duodu, K.G., Tang, H., Grant, A., Wellner, N., Belton, P.S., Taylor, J.R.N., 2001. FTIR and solid state 13C NMR spectroscopy of proteins of wet cooked and popped sorghum and maize. Journal of Cereal Science 33, 261269. Duodu, K.G., Nunes, A., Delgadillo, I., Parker, M.L., Mills, E.N.C., Belton, P.S., Taylor, J.R.N., 2002. Effect of grain structure and cooking on sorghum and maize in vitro protein digestibility. Journal of Cereal Science 35, 161174. Duodu, K.G., Taylor, J.R.N., Belton, P.S., Hamaker, B.R., 2003. Factors affecting sorghum protein digestibility. Journal of Cereal Science 38, 117131. El Nour, I.N.A., Peruffo, A.D.B., Curioni, A., 1998. Characterisation of sorghum karins in relation to their cross-linking behaviour. Journal of Cereal Science 28, 197207. Elkhalifa, A.O., Bernhardt, R., Bonomi, F., Iametti, S., Pagani, M.A., Zardi, M., 2006. Fermentation modies protein/protein and protein/starch interactions in sorghum dough. European Food Research Technology 222, 559564. Gao, C., Taylor, J., Wellner, N., Byaruhanga, Y.B., Parker, M.L., Mills, E.N.C., Belton, P.S., 2005. Effect of preparation conditions on protein secondary structure and biolm formation of karin. Journal of Agricultural and Food Chemistry 53, 306312. Gao, C., Stading, M., Wellner, N., Parker, M.L., Noel, T.R., Mills, E.N.C., Belton, P.S., 2006. Plasticization of a protein-based lm by glycerol: a spectroscopic, mechanical, and thermal study. Journal of Agricultural and Food Chemistry 54, 46114616. Gennadios, A., 2002. Protein Based Films and Coatings. CRC Press, Boca Raton, USA. Georget, D.M.R., Belton, P.S., 2006. Effects of temperature and water content on the secondary structure of wheat gluten studied by FTIR spectroscopy. Biomacromolecules 7, 469475.

MN kt n Mo

(2)

where MN/Mo is the cumulative drug release ratio, k a kinetic constant and n the release exponent, indicative of the mechanism of drug release. Table 1 shows the results of the tting. The values of n are between 0.5 and 0.6, indicating that the caffeine release is non-Fickian. In zein-based tablets (Georget et al., 2008), polymer relaxation and erosion were previously found to predominate in the drug release mechanism, and this also appears to be the case with karin-based tablets. These data suggest, therefore, that there is a common behaviour within the prolamin proteins. Fig. 5 also shows karin tablets before and after exposure to the different dissolution media for 4.5 h. It is interesting to note that the karin tablets appear intact after exposure to water or to pH 6.8 phosphate buffer, with some slight swelling and softening also being observed. However, after exposure to 0.1 M HCl, the tablets acquired a uniformly distributed pink colouration. This colour is most likely to be due to the reaction of HCl with the minor anthocyanin components of sorghum. It has been previously reported (Awika et al., 2004) that apigeninidin and luteolinidin in sorghum produce yellow and orange colours, respectively, in acidic solvents. Taken together, the dissolution and disintegration results suggest that drug release is greater in the acidic conditions of the stomach and lower in the more neutral conditions of the intestine. Given that the expected residence time of a tablet in an unfed stomach after dosing is less than 30 min, the total quantity of the drug released into the stomach is likely to be reasonably small, with the remainder of the drug showing a sustained-release prole in the intestine. Enteric coating of the karin tablet would prevent the dissolution of the drug into the stomach. Alternatively, dependent on the nature of the drug used and the disease state treated, such a bi-phasic release rate could be deemed to be benecial.
Table 1 Diffusional exponent (n) and kinetic constant (k) for the dissolution of karin tablets in water, HCl and buffer. Results are expressed as means (standard error of the mean) Within columns (n and k), means with the same superscript letter are not signicantly (p<0.05) different. Dissolution medium H2O pH 1.3 pH 6.8 n 0.621 (0.023) 0.536a (0.032) 0.633a (0.041)
a

(min-n) a

R2 0.997 0.985 0.998

1.718 (0.309) 4.132b (0.824) 1.421a (0.634)

A.E.O. Elkhalifa et al. / Journal of Cereal Science 50 (2009) 159165 Georget, D.M.R., Barker, S.A., Belton, P.S., 2008. A study on maize proteins as a potential new tablet excipient. European Journal of Pharmaceutics and Biopharmaceutics 69, 718726. Glennie, C.W., Harris, J., Liebenberg, N.V.D.W., 1983. Endosperm modication in germinating sorghum grain. Cereal Chemistry 60 (1), 2731. Hamaker, B.R., Mohamed, A.A., Habben, J.E., Huang, C.P., Larkins, B.A., 1995. Efcient procedure for extracting maize and sorghum kernel proteins reveals higher prolamin contents than the conventional method. Cereal Chemistry 72 (6), 583588. http://www.fao.org. http://www.fda.gov. Izquierdo, L., Godwin, I.D., 2005. Molecular characterization of a novel methioninerich delta-karin seed storage protein gene in sorghum (Sorghum bicolor L.). Cereal Chemistry 82, 706710. Katayama, H., Kanke, M., 1992. Drug release from directly compressed tablets containing zein. Drug Development and Industrial Pharmacy 18 (20), 21732184. Li, C.L., Martini, L.G., Ford, J.L., Roberts, M., 2005. The use of hypromellose in oral drug delivery. Journal of Pharmacy and Pharmacology 57 (5), 533546. Mazhar, H., Chandrashekar, A., 1995. Quantication and distribution of karins in the kernels of sorghum cultivars varying in endosperm hardness. Journal of Cereal Science 21, 155162. Momany, F.A., Sessa, D.J., Lawton, J.W., Selling, G.W., Hamaker, S.A.H., Willett, J.L., 2006. Structural characterization of a-zein. Journal of Agricultural and Food Chemistry 54, 543547. Nolasco, M.M., Amado, A.M., Ribeiro-Claro, P.J.A., 2006. Computationally-assisted approach to the vibrational spectra of molecular crystals: study of hydrogenbonding and pseudo-polymorphism. ChemPhysChem 7, 21502161. Oria, M.P., Hamaker, B.R., Axtell, J.D., Huang, C.-P., 2000. A highly digestible sorghum mutant cultivar exhibits a unique folded structure of endosperm protein bodies.

165

Proceedings of the National Academy of Sciences of the United States of America 97 (10), 50655070. Rahmelow, K., Hubner, W., 1996. Fourier self-deconvolution: parameter determination and analytical band shapes. Applied Spectroscopy 50 (6), 795804. Ritger, P.L., Peppas, N.A., 1987. A simple equation for description of solute release. II. Fickian and anomalous release from swellable devices. Journal of Controlled Release 5, 3742. Rom, D.L., Shull, J.M., Chandrashekar, A., Kirleis, A.W., 1992. Effects of cooking and treatment with sodium bisulte on in vitro protein digestibility and microstructure of sorghum our. Cereal Chemistry 69 (2), 178181. Shull, J.M., Watterson, J.J., Kirleis, A.W., 1991. Proposed nomenclature for the alcohol-soluble proteins (karins) of Sorghum bicolor (L. Moench) based on molecular weight, solubility, and structure. Journal of Agricultural and Food Chemistry 39, 8387. Surewicz, W.K., Mantsch, H.H., Chapman, D., 1993. Determination of protein secondary structure by Fourier transform infrared spectroscopy: a critical assessment. Biochemistry 32 (2), 389394. Taylor, J., Belton, P.S., 2002. Sorghum. In: Belton, P.S., Taylor, J.R.N. (Eds.), Pseudocereals and Less Common Cereals. Springer, Berlin, Germany, pp. 2538. Taylor, J.R.N., Schober, T.J., Bean, S.R., 2006. Novel food and non-food uses for sorghum and millets. Journal of Cereal Science 44, 252271. Terebesi, I., 2006. Alternative Polymers and Processes for Coating. PhD Thesis, Freie Universitat Berlin, Germany. Wellner, N., Mills, E.N.C., Brownsey, G., Wilson, R.H., Brown, N., Freeman, J., Halford, N.G., Shewry, P.R., Belton, P.S., 2005. Changes in protein secondary structure during gluten deformation studied by dynamic Fourier transform infrared spectroscopy. Biomacromolecules 6, 255261.

Journal of Cereal Science 50 (2009) 166174

Contents lists available at ScienceDirect

Journal of Cereal Science


journal homepage: www.elsevier.com/locate/jcs

Formation of grain chalkiness and changes in water distribution in developing rice caryopses grown under high-temperature stress
Tsutomu Ishimaru a, Akemi K. Horigane b, Masashi Ida a, Norio Iwasawa a, Yumiko A. San-oh a, Mikio Nakazono c, Naoko K. Nishizawa c, d, Takehiro Masumura e, Motohiko Kondo a, Mitsuru Yoshida b, *
a

National Institute of Crop Science, NARO, 2-1-18 Kannondai, Tsukuba, Ibaraki 305-8518, Japan National Food Research Institute, NARO, 2-1-12 Kannondai, Tsukuba, Ibaraki 305-8642, Japan Graduate School of Agricultural and Life Sciences, The University of Tokyo, Yayoi, Bunkyo, Tokyo 113-8657, Japan d Research Institute for Bioresources and Biotechnology, Ishikawa Prefectural University, 1-38 Suematsu, Nonoichi, Ishikawa 921-8836, Japan e Graduate School of Life and Environmental Science, Kyoto Prefectural University, Shimogamo, Sakyo, Kyoto 606-8522, Japan
b c

a r t i c l e i n f o
Article history: Received 25 December 2008 Received in revised form 15 April 2009 Accepted 17 April 2009 Keywords: Grain chalkiness High-temperature stress Magnetic resonance imaging Rice

a b s t r a c t
High-temperature stress during grain ripening facilitates the formation of chalky grains through loose packing of amyloplasts. In the present study, the changes in water distribution in the developing caryopses by high-temperature stress were investigated by magnetic resonance imaging. Milky-white and white-cored types of chalky grains, which had chalkiness around the centre of the endosperm, were formed in a high percentage in a high-temperature condition by the middle dry-matter accumulating stage. Magnetic resonance images of the early stage caryopses in the high-temperature condition showed a lower water content around the centre of the endosperm, where the disorganized development of amyloplasts was observed by scanning electron microscopy. Transcripts for a-amylase were, however, not detected there in the early and middle stage, indicating that starch degradation by a-amylase was not the cause of the formation of these types of chalky grains. In the middle stage, water content in the central chalky part became higher than in the lateral translucent part. Loose packing of amyloplasts in the chalky part may allow pooling of water in the free spaces leading to the higher water content after the middle stage. Together with the results, the physiological mechanism for the formation of white-cored and milky-white grains that occurred by high-temperature stress was discussed. 2009 Elsevier Ltd. All rights reserved.

1. Introduction Rice is one of the most important staple foods for more than three billion people in the world. Unlike wheat and corn, which are processed to yield our, rice is generally marketed as milled grains. Grain appearance is an important criterion in determining the quality of rice, and chalky grains have a negative impact on the quality (Kim et al., 2000). These authors demonstrated that when the proportion of chalky grains exceeded 15%, the rice had decreased eating quality. The presence of chalky grains decreases rices value in most of the world markets. High-temperature stress during grain ripening facilitates the formation of chalky grains.
Abbreviations: DAF, days after owering; MR, magnetic resonance; MRI, magnetic resonance imaging; NMR, nuclear magnetic resonance; T1, spinlattice relaxation time; T1W, T1 weighted; PDW, proton density weighted; T2, spinspin relaxation time. * Corresponding author. National Food Research Institute, NARO, Analytical Science Division, 2-1-12 Kannondai, Tsukuba, Ibaraki 305-8642, Japan. Tel.: 81 29 838 8033; fax: 81 29 838 7996. E-mail address: mitsuru@affrc.go.jp (M. Yoshida). 0733-5210/$ see front matter 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2009.04.011

Chalky grains are frequently found when the average temperature during the 20 day period after heading is above 2627  C (Wakamatsu et al., 2007). This climatic condition occurs more often recently than was the case before in Japan (Kobayashi et al., 2007; Tabata et al., 2007; Wakamatsu et al., 2007) possibly owing to global warming. Since global warming is forecasted to continue as a consequence of the greenhouse gas effect (IPCC, 2007), grain chalkiness caused by high-temperature stress will become a global problem for rice agriculture in the future. Chalky grains were categorized into white-cored, milky-white, white-back, white-based, and white-belly types depending on the presence of a chalky part in the grain (Nagato and Ebata, 1965; Tashiro and Wardlaw, 1991). White-cored grains have chalkiness in the centre of the endosperm, but milky-white grains have a broader part of chalkiness compared to white-cored grains. In common with these types of chalky grains, chalkiness is present in the inner part of the endosperm. White-based and white-back grains have chalkiness at the basal part and dorsal part, frequently including the basal part, of the endosperm, respectively. These types of chalkiness are present in the peripheral part of the endosperm.

T. Ishimaru et al. / Journal of Cereal Science 50 (2009) 166174

167

The time when the plant is exposed to high-temperature stress during ripening determines the part with chalkiness in the grain. Temporal exposure of rice plants to high-temperature from 4 to 12 days after heading induced the milky-white and white-back grains while exposure around 16 days after heading formed only white-back grains (Tashiro and Wardlaw, 1991). These differences in the development of the chalky part in grains are considered to be related to a difference in the pattern of starch accumulation (Nagato and Ebata, 1965). Starch is actively accumulated around the centre of the endosperm from early to middle stages and at the periphery at the late stage (Hoshikawa, 1968; Nagato and Kobayashi, 1959). The effect of grain chalkiness on the eating quality is different depending on the type of chalky grains. Milky-white grains negatively affect the eating quality, but white-back grains hardly affect it (Wakamatsu et al., 2007). The exposure of rice plants to high-temperature shortens the period for grain ripening (Sato and Inaba, 1973, 1976). The panicle is the organ that is highly sensitive to high-temperature stress during ripening (Morita et al., 2004; Sato and Inaba, 1973). Considerable effort has been expended to understand the underlying physiological mechanism of the high-temperature induced grain chalkiness from morphological (Zakaria et al., 2002), transcriptional (Jiang et al., 2003; Yamakawa et al., 2007), and proteomic (Lin et al., 2005) perspectives. Chalkiness was observed in the starchy endosperm, where the development of amyloplasts was decient (Lisle et al., 2000; Tashiro and Wardlaw, 1991). The physiological causes of the chalkiness are hypothesised to be an insufcient supply of nutrients to the developing endosperm (Sato and Inaba, 1976), reduced ability to synthesise starch in the endosperm (Jiang et al., 2003; Lin et al., 2005; Yamakawa et al., 2007), and the degradation of starch by a-amylase during ripening (Yamakawa et al., 2007). So far, these studies have been conducted using the whole caryopsis. Thus, the physiological mechanism for the formation of each type of chalkiness (ex. milky-white, white-back grains) has not yet been understood. During the ripening of cereal crops, a large amount of storage material is accumulated in the seeds concomitantly with a decrease in the water content. Nuclear magnetic resonance (NMR) has been used to investigate the water status of developing caryopses during ripening for several crops. The amount and the mobility of free water, detected by NMR, have been suggested as indices of biological activity in barley seeds (Kano et al., 1990). Funaba et al. (2006) rst studied the effect of temperature stress on the water status of developing rice grains using NMR. They showed that high/low temperature stress changed the NMR relaxation times, (spinlattice relaxation time (T1) and spinspin relaxation time (T2)) of water protons in the whole caryopsis. However, the relationship between change in spatial water distribution and the formation of grain chalkiness under high-temperature stress was not investigated in their study. Magnetic resonance imaging (MRI), a technique based on the phenomenon of NMR, has led to the understanding of the precise water distribution in developing grains of barley (Glidewell, 2006) and rice (Horigane et al., 2001). Horigane et al. (2001) used this technique to reveal that the low water content is rst manifested around the centre of the endosperm and then spreads to the periphery during ripening in rice. This change in water distribution corresponded to the shift of the site where starch was being actively accumulated (Hoshikawa, 1968). The aim of this study was to clarify the change in water distribution and expression of genes encoding a-amylase in the developing rice caryopsis under high-temperature stress in relation to the formation of milky-white and white-cored types of chalky grains. MRI investigated the change in the water distribution in developing rice caryopses by high-temperature stress. We further applied the laser microdissection technique and scanning electron

microscopy to the starchy endosperm, where milky-white and white-cored types of chalkiness could be frequently observed. 2. Experimental 2.1. Plant material Oryza sativa Koshihikari (Japonica rice variety) was grown in 2005 and 2007. Seeds were sown in a nursery box lled with soil at the beginning of May and June. One-month-old seedlings were transplanted into 0.02 m2 pots. Plants were grown out of doors until owering. As a basal dressing, 0.52.32.2 g of NP2O5K2O was applied, and 0.4 g of nitrogen per pot was applied as a top dressing approximately 2 weeks before heading. The heading date of rice plants was the beginning and the middle of August in the summer, respectively. The owering date for each caryopsis was recorded. At 3 or 4 days after owering (DAF), plants were transferred into a naturally illuminated temperature-controlled chamber. Day (13 h) and night (11 h) air temperatures were maintained at 26 and 20  C in the control conditions, and 33 and 27  C in the high-temperature conditions, respectively. Developing rice caryopses from the rst to the third primary rachis-branches counted from the top of the panicle were used for the following experiments. Samples were collected ve times from 7 DAF to harvest for each condition. Fully matured grains were harvested after 40 DAF from control plants and 35 DAF from the plants exposed to high-temperatures. Fully matured grains harvested in 2005 were used for scanning electron microscopy. Rice plants that had heading dates in the beginning of August in 2007 were used for measurement of grain dry weight and water content and for magnetic resonance imaging. Rice plants that had heading dates in the middle of August in 2007 were used for the expression analysis for a-amylase. 2.2. Grain dry weight and water content of whole grain Developing rice caryopses were dehulled and the fresh weight of single grain (mg) was measured. After grains were dried in a forced convection oven at 70  C for 3 days, grain dry weight (mg) was measured. The water content (% wet basis) was calculated by dividing weight loss by fresh weight. 2.3. Evaluation of grain appearance and median transversal sections of developing caryopsis The appearance of fully matured grains was evaluated visually and grains were classied into perfect, white-cored/milky-white, white-based/white-back, opaque and immature types according to the criterion of Tashiro and Wardlaw (1991). Median transversal sections of developing caryopsis were manually cut with a sharp razor, viewed under a stereo-microscope (SZX12, Olympus, Tokyo, Japan), and then immediately photographed with a digital camera (E-330, Olympus, Tokyo, Japan). 2.4. Scanning electron microscopy Perfect grains from the control condition and white-cored/ milky-white grains from the high-temperature condition were examined by scanning electron microscopy. They were transversally cut in half with a razor blade, the half-grains were attached to aluminium specimen stubs, and the specimens cut surface was coated with gold using an ion sputtering device (JFC-1100E, JEOL, Tokyo, Japan) under vacuum. The specimens were observed and photographed with a scanning electron microscope (Real Surface View VE-7800, KEYENCE, Osaka, Japan).

168

T. Ishimaru et al. / Journal of Cereal Science 50 (2009) 166174

2.5. Magnetic resonance imaging


1 H-MRI was performed using an NMR spectrometer (DRX300WB, Bruker, Karlsruhe, Germany) equipped with a microimaging accessory at a magnetic eld of 7.1 Tesla. The rice grain was placed in a folded thin plastic plate immediately after it was clipped from the panicle, held there with a rubber ring, and then placed in an NMR tube with a 5 mm o.d. Three or four grains from separate panicles from plants in each condition group and on each DAF were measured. The morphological structure of developing rice caryopsis was observed in T1 weighted (T1W) images of longitudinal sections, and the water distribution in transverse sections was estimated from proton density weighted (PDW) images. The positions for the transverse-sectional PDW images were determined using the longitudinal T1W images. All of the images were acquired using a spin echo pulse sequence on the ParaVision operating system (ver. 3.0.2, Bruker). The repetition time/echo time was set for 500/4.91 ms for the T1W images and 5000/ 3.86 ms for the PDW images, respectively. The eld of view was 10 5 mm2 for T1W images and 5 5 mm2 for PDW images, and the in-plane resolution and slice thickness for both images were 39 39 mm2 and 1 mm, respectively. The total scan times were 2 min 8 s with two accumulations for the T1W images and 21 min 20 s with two accumulations for the PDW images, and the images were acquired within 40 min after the grains had been clipped from a panicle.

Amplication System (NuGEN technologies Inc., San Carlos, CA) according to the manufacturers instructions. The PCR condition was as follows; 95  C: 15 s, 55  C: 15 s, 72  C: 30 s, repeated 35 cycles (GeneAmp PCR system 2700, Applied Biosystems, Foster City, CA). For the comparison, the RT-PCR for Actin was also conducted under the same condition (Ishimaru et al., 2007). Thereafter, the PCR products were separated on 2% agarose gel, stained with ethidium bromide, and photographed (Printgraph AE-6931FXCF, ATTO, Tokyo, Japan). 3. Results 3.1. Changes in dry weight and water content of whole grain during ripening Dry weight and water content of whole grain at 7 DAF did not differ between the grains from the control and high-temperature condition (Figs. 1A,B). Afterwards, dry-matter was accumulated but the water content of whole grain declined more rapidly in grains from the high-temperature condition than in those of the controls. Grain dry weight reached a maximum by 15 DAF in the high-temperature condition, but a few days later in controls. At full maturity, grain dry weight was signicantly lower in the hightemperature condition (P < 0.01 by Students t-test). As shown in Fig. 1, the dry weight and water content at 7, 11 and 15 DAF of grains from the high-temperature condition were quite similar to those at 7, 12 and 17 DAF of the controls, respectively. Based on the grain dry weight and water content of whole grain, the period for dry-matter accumulation was divided into three stages: early (7 DAF in both conditions), middle (12 DAF in control and 11 DAF

2.6. Expression analysis for a-amylase in the whole caryopsis Extraction of total RNA from the whole caryopsis (embryo, endosperm and maternal tissues) was based on the methods of Chang et al. (1993). First-strand cDNA mixture was prepared from 2.5 mg of total RNA in developing caryopses with oligo-dT13 primer and reverse transcriptase (SuperScript II, Invitrogen, Carlsbad, CA). An aliquot of the rst-strand cDNA mixture corresponding to 12.5 ng of total RNA was used as a template. For the quantitative RT-PCR, QuantiTect SYBER green PCR kit (Qiagen, Hilden, Germany) was used with a real time RT-PCR system (7500 Real Time PCR System, Applied Biosystems, Foster City, CA). The PCR condition was as follows; 95  C: 15 s, 55  C: 15 s, 72  C: 35 s, repeated 40 cycles. The analysed genes were Amy1A (AK101744), Amy3D (AK119761), Amy3E (AK064300) encoding a-amylase. The specic PCR primers for Amy1A (GCGCCTGGTGTCAATCAGAA and CGGATCGGATACAGCTCGTTG), Amy3D (GCCCTAAACTGAACGGGATA and CCAACGGTTACAAACTGCGTGA), Amy3E (TTTGCTGCGAGATGTGTACG and CTTGAGGATCGAAACGAACAG) were used. Transcript level of each gene was normalized with that of 18S rRNA (Cho et al., 2006: ATGATAACTCGACGGATCGC and CTTGGATGTGGTAGCCGTTT). 2.7. Expression analysis for a-amylase in the centre of the endosperm with laser microdissection Preparation of the specimen for laser microdissection was conducted with the method of Ishimaru et al. (2007). Briey, the developing rice caryopses were xed with the mixture of ethanolacetic acid (3:1), and embedded in 2% carboxymethyl cellulose. Transversal sections (8 mm thickness) were made at the median part of the developing caryopsis. The centre of the starchy endosperm was microdissected circularly approximately 300 mm in diameter with an AS LMD system (Leica Microsystems, Wetzlar, Germany). Total RNA was extracted with a Picopure RNA isolation kit (Molecular Devices, Sunnyvale, CA) using DNase I. Quantication of total RNA was determined by the uorescence based method, using a RiboGreen RNA Quantication kit (Molecular Probes, Eugene, OR). Total RNA (10 ng) was amplied with a WT-Ovation RNA

Fig. 1. Changes in dry weight (A) and water content (B) in a whole grain. Each value indicates the mean and bar represents SD of at least three caryopses during dry-matter accumulating stage and 18 grains at maturity.

T. Ishimaru et al. / Journal of Cereal Science 50 (2009) 166174

169

in high-temperature condition) and late (17 DAF in control and 15 DAF in high-temperature condition).

3.3. Changes in MR image of developing caryopses Grain morphology for each stage of dry-matter accumulation was observed in longitudinal T1W MR images. Representative images of the developing caryopses grown under control conditions are shown as in Figs. 4A,D,G,I, which are not so different from those of caryopses grown under high-temperature conditions. As the caryopses developed, signal intensities decreased in both T1W and PDW MR images, which were displayed with a gray scale and coloured scale, respectively. Signal intensity showing as black indicates below the detectable level. Signal intensity at the periphery of the endosperm in PDW images was the highest observed and decreased towards the inner part of the early stage grains from both temperature regimes (Figs. 4B,C). The signal intensity at the starchy endosperm was lower in grains from the high-temperature condition than in those from controls, in early stage grains, especially along the dorsiventral line (Fig. 4C). In the PDW image of the early stage, the pericarp had higher signal intensity in grains exposed to the hightemperature condition than did that of grains exposed to the control condition (Fig. 4C, arrowhead). From the middle stage, signal intensities in every part of the endosperm decreased in grains from both conditions compared to the signal intensities measured in the early stage (Figs. 4DK). Decreases in signal intensity of the lateral layer, between the peripheral and the central parts, were apparent in grains of the high-temperature condition in the middle stage (Fig. 4F). A similar decrease in signal intensity of the lateral layer was also observed in control grains at 14 DAF (Fig. 4H) and in the late stage grains from both conditions (Figs. 4J,K). The decrease of the signal from this area was faster in high-temperature treated grains than in control grains, although the signal intensities around the centre of the middle and late stage grains that had been exposed to hightemperature stress were higher than those of control grains examined at 14 DAF and in the late stage, respectively. 3.4. Expression analysis for a-amylase The developing caryopses in the early and middle stages were used. Transcript level of genes for a-amylase in the whole caryopsis was 24 folds higher in high-temperature condition than in the control in the early stage, but the difference in the transcript level was not clear for every gene in the middle stage (data not shown). With laser microdissection, the transcripts of a-amylase genes in the centre of the endosperm were also investigated. In the centre of the endosperm, mRNAs for a-amylases were not detected in both conditions for both stages at 35 cycles of RT-PCR, although PCR products for Actin at 35 cycles of RT-PCR conrmed successful amplication of mRNA and synthesis of cDNA in this system (Fig. 5). 4. Discussion

3.2. Grain appearance, development of amyloplasts, and formation of the chalky centre Table 1 presents the frequency data for the types of rice grain appearance. As many grains intermediate between the types whitecored and milky-white with a broad chalky part around the core were found, the data for these types were pooled. Grains intermediate between types white-back and white-based with chalky area along the back to base were also detected, so data for these types were also pooled. Grain appearance was quite different between grains from the two temperature conditions. Nearly all the control grains fell into the perfect appearance type (Table 1, Figs. 2A,B). In contrast, approximately 80% of the high-temperature treated grains had at least some areas of chalkiness. Most chalky grains were classied into the white-cored/milky-white grain and the white-based/ white-back types (Table 1). Especially, over a half of the grains had white-cored/milky-white types of chalkiness. The white-cored/ milky-white grains often also t the criteria of the white-based/ white-back type (Table 1, Figs. 2C,D). Chalky grains with the whitebelly appearance were hardly observed in this study. Scanning electron microscopic observation of fully matured grains revealed that the amyloplasts were packed without gaps both in the central and lateral parts of the endosperm in the grains from the control condition (Figs. 2E,F). In contrast, single and compound amyloplasts were loosely packed in the central chalky area of grains from the high-temperature condition, and numerous free spaces between amyloplasts were observed (Fig. 2G). The free spaces were also observed in the lateral part of the grain, but they were much smaller than those in the central chalky area (Fig. 2H). The stage when the formation of chalkiness in the central endosperm occurred was investigated in median transverse sections of grains from each dry-matter accumulating stage (Figs. 3AF). In the early stage, the endosperm was a milky-white colour, and there was no difference in its appearance between the two temperature conditions (Figs. 3A,D). In the middle stage, in grains from the control condition, the centre of the endosperm had become translucent (Fig. 3B), and in the late stage, the translucency had expanded into most parts of the endosperm, except in the periphery (Fig. 3C). However, in grains exposed to the high-temperature stress, chalkiness was observed around the centre of the endosperm in the middle stage (Fig. 3E) and it remained there through to the late stage (Fig. 3F).

Table 1 Frequency of types of rice grain by appearance. Heading date Condition Frequency of grain appearance (%) Perfect Whitecored Milkywhite Beginning of August Middle of August Control Hightemperature Control Hightemperature 98.2 11.3 97.4 16.0 0.4 65.8* 0.7 55.3** Whitebased Whiteback 0 60.2* 0 47.8** 0.7 4.9 0 0.7 0.7 3.0 1.9 4.8 Opaque Immature

Beginning of August: Control, 282 grains. High-temperature, 266 grains. Middle of August: Control, 153 grains. High-temperature, 293 grains. *45.2 and **24.6% of the grains had both white-cored/milky-white and white-based/white-back chalkiness.

The development of whole grains grown under control and high-temperature conditions was monitored in terms of their dry weights and water contents. Dry-matter accumulation was accelerated after 7 DAF by the high-temperature stress, and this was accompanied by a faster decrease in water content (Figs. 1A,B). This result indicated that physiological ripening was more advanced in the caryopses that had been exposed to the high-temperature condition than in the caryopses of controls. To eliminate temperature induced differences in the maturing rate of grains, we used the dry-matter accumulating stage estimated based on dry weight (mg) and water content (% wet basis) of whole grain rather than

170

T. Ishimaru et al. / Journal of Cereal Science 50 (2009) 166174

Fig. 2. Appearance and amyloplasts morphology of grains exposed to control (left column) and high-temperature (right column) conditions. A and C, typical appearance; B and D, median transverse sections. Arrow 15 indicate the type of chalkiness; 1, white-back; 2, white-based; 3, white-cored or milky-white; 4, milky-white; 5, white-cored. EH are the scanning electron micrographs of amyloplasts, E and G, centre part; F and H, lateral part; black arrowheads in H, free spaces between amyloplasts. Black and yellow rectangles on B and D indicate the positions for scanning electron micrographs of E, G and F, H, respectively.

DAF for comparison between control and high-temperature treated samples. 4.1. Changes in water distribution in the early stage Signals of the protons in biological samples detected by MRI are generally derived from water and lipid (Glidewell, 2006). Highly mobile molecules of free water, but not bound water and solid starch, which have an extremely short T2, can be detected by MRI, and the signal intensity of water in a PDW image generally reects a specimens water content (MacFall and Van As, 1996). Lipids accumulate

in aleurone cells located in the outermost endosperm tissue (Tanaka et al., 1995). Therefore, the distribution of proton signals detected in starchy endosperm indicates the water distribution. Water content estimated from the PDW image was higher at the periphery of endosperm compared to the centre in both temperature conditions in the early dry-matter accumulating stage (Figs. 4B,C) because water is supplied from the dorsal vascular bundle through the nucellar epidermis covering the surface of the endosperm (Hoshikawa, 1968). Around the centre of the endosperm, where starch was actively accumulated (Hoshikawa, 1968), water content was lower in the caryopses

T. Ishimaru et al. / Journal of Cereal Science 50 (2009) 166174

171

Fig. 3. Median transverse sections of developing rice caryopses grown under control (top) and high-temperature (bottom) condition. AC, Control; DF, high-temperature. A and D, early stage; B and E, middle stage; C and F, late stage.

exposed to the high-temperature condition than in those exposed to the control condition (Figs. 4B,C) while a high water content was detected in the lateral pericarp in high-temperature treated caryopses (Fig. 4C). Tashiro and Wardlaw (1991) reported that white-cored/milky-white grains were frequently observed in fully matured grain when rice plants had been exposed to a hightemperature from the early to middle dry-matter accumulating stages. Nagato and Ebata (1965) reported that white-back and white-based grains, which were chalky at the periphery of the endosperm, might be formed at the late stage, when the ability of starch synthesis had already declined. White-cored and milkywhite grains, which were chalky around the centre of the endosperm, might be formed from the early to middle stages, when the ability for starch synthesis was still active. In agreement with these reports, we showed the evidence that chalkiness around the centre of the endosperm was formed from the early to the middle stages by making the transverse section of developing caryopses (Fig. 3E). At maturity, single rounded as well as compound amyloplasts, were present at the chalky part (Fig. 2G) as reported previously (Lisle et al., 2000). The distribution of the chalkiness around the centre of the endosperm spatially coincided with the distribution of low water content areas that had been detected in the PDW image of early stage caryopses (Fig. 4C). These results suggested that the lower water content of the central endosperm

in the early stage of grain ripening in the high-temperature condition might affect plastid initiation (i.e. amyloplast number) or starch granule initiation. Yamakawa et al. (2007) suggests a relation of starch degradation by high-temperature induced a-amylase to the formation of grain chalkiness. We also observed the high-temperature induced transcription of a-amylase genes in the early stage in a whole caryopsis (data not shown). The mRNAs for a-amylase were, however, undetectable in the centre of the endosperm (Fig. 5). These results indicated that starch degradation by a-amylase was not the cause of the formation of the white-cored/milky-white type of chalkiness under high-temperature stress, and declined ability for starch synthesis may be the cause until the middle stage. According to the results of Yamakawa et al. (2007), the induction of a-amylase mRNA was most conspicuous after the middle stage. Thus, the formation of chalkiness through the starch degradation by a-amylase may have occurred at the later stage. The expression levels of many genes/proteins related to starch synthesis were reported to be altered in a whole caryopsis by high-temperature stress (Jiang et al., 2003; Lin et al., 2005; Yamakawa et al., 2007). It is possible that the exposure of rice plants to high-temperature caused the rapid decline of water content from the centre of the endosperm during the early stage, thus disorganizing amyloplast development there. Proling of gene expression in this part is further needed to

172

T. Ishimaru et al. / Journal of Cereal Science 50 (2009) 166174

Fig. 4. T1W images of longitudinal sections of grains from controls (A, D, G, I) and PDW images of transverse sections of developing rice caryopses of controls (B, E, H, J) and hightemperature treated plants (C, F, K) in the early (A, B, C), middle (D, E, F), 14 DAF (G, H) and late (I, J, K) stage. d, dorsal side; v, ventral side; dv, dorsal vascular bundle; g, glume; pe, pericarp; ce, central endosperm; arrowheads in C, area of high signal intensity in the lateral pericarp. The purple rectangles on the T1W images indicate the positions for the transverse-sectional PDW images. Bars indicate 1 mm.

elucidate the relationship between rapid decline of water and disorganized amyloplast development. The relation of hightemperature induced expression of a-amylase in a whole caryopsis to the other types of chalkiness (ex. white-back, white-based grains) should be also claried. 4.2. Change in water distribution after the middle stage Under control conditions, the site of active starch accumulation seemed to have shifted from the centre of the endosperm to the lateral part in the middle stage because the milky-white colour

had already been replaced by a translucent one around the centre with the decrease of water content (Figs. 3B and 4E). Although the decrease of water content in the whole grain was also observed in middle stage grains that had been exposed to the high-temperature condition (Fig. 1B), the distribution of proton signals in the endosperm was different between two temperature regimes. The water content in the central core of the grain was higher than that in the lateral part in the high-temperature treated caryopses (Fig. 4F). In the late stage, this pattern of water distribution was more clearly observed in the high-temperature treated caryopses (Fig. 4K) than in the controls (Figs. 4H,J). In the normally

T. Ishimaru et al. / Journal of Cereal Science 50 (2009) 166174

173

Fig. 5. Expression analysis for Amy1A, Amy3D and Amy3E in the centre of endosperm. For the comparison, PCR product for Actin was also electrophoresed. Numerals represent the numbers of biological replications in each treatment and dry-matter accumulating stage. Size marker (300, 200, 100 bp from the top to bottom) was shown in the leftmost lane (M). The PCR products from cDNA aliquot of whole caryopsis were shown in the rightmost lane (W).

developing caryopsis of these stages, water supplied from the dorsal vascular bundle ows on the dorsi-ventral axis in developing endosperm towards the ventral sides due to the degeneration of nucellar epidermis cells (Hoshikawa, 1968). As this cell degeneration is accelerated by the high-temperature stress (Zakaria et al., 2002), the water content at the central part may increase earlier in the high-temperature condition (Fig. 4F) than in the control condition (Figs. 4E,H). Besides, water from the dorsal vascular bundle was easily pooled in the gaps between the loosely packed amyloplasts in the central chalky area (Figs. 2G and 3E). On the other hand, gaps between amyloplasts were quite small in the lateral part (Fig. 2H), and water supply to this part from the nucellar epidermis and the central part was less, resulting in low water content (Fig. 4F). Chalkiness was observed around the centre of the endosperm by the middle stage (Fig. 3E), but it did not spread to the lateral part thereafter (Fig. 3F) even though the caryopses were kept under high-temperature stress during all dry-matter accumulating stages. The gaps between amyloplasts in the lateral part of grains were too small to give the chalky appearance even in the high-temperature condition (Fig. 2H). It is known that starch accumulates faster around the centre of the endosperm until the middle stage, producing large amyloplasts, whereas starch accumulation proceeds at a slower rate in the lateral part, resulting in smaller amyloplasts there (Hoshikawa, 1968; He and Suzuki, 1989). Horigane et al. (2006) revealed by MRI observation that the centre of the endosperm allowed more water to penetrate during soaking of the milled rice grain, suggesting that the packing of amyloplasts was intrinsically looser in the central endosperm compared to the lateral one, even in the normal grain. Our MR images showed that water content declined rst in the centre of the endosperm, then in the lateral endosperm (Figs. 4B,E). The initiation of amyloplast development might be hardly affected by high-temperature stress at the lateral endosperm because a decline of water content in the lateral endosperm was slower compared to the central endosperm in the early stage (Fig. 4C). Spatio-temporal analysis on difference in the development of amyloplast between the central and lateral parts of the endosperm with laser microdissection may provide

a clue for the identication of genes involved in the grain chalkiness around the centre of the endosperm. Acknowledgement This work was supported by a Grant-in-Aid from the National Agricultural and Food Research Organization (NARO), Japan. References
Chang, S., Puryear, J., Cairney, J., 1993. A simple and efcient method for isolating RNA from pine trees. Plant Molecular Biology Report 11, 113116. Cho, J.-I., Ryoo, N., Ko, S., Lee, S.-K., Lee, J., Jung, K.-H., Lee, Y.-H., Bhoo, S.K., Winderickx, J., An, G., Hahn, T.-R., Jeon, J.-S., 2006. Structure, expression, and functional analysis of the hexokinase gene family in rice (Oryza sativa L.). Planta 224, 598611. Funaba, M., Ishibashi, Y., Molla, A.H., Iwanami, K., Inoue-Iwaya, M., 2006. Inuence of low/high temperature on water status in developing and maturing rice grains. Plant Production Science 9, 347354. Glidewell, S.M., 2006. NMR imaging of developing barley grains. Journal of Cereal Science 43, 7078. He, G.C., Suzuki, H., 1989. Properties of starches from the outer layer and central core of the rice endosperm. Nippon Nogeikagaku Kaishi 63, 981989 (in Japanese with English summary). Horigane, A.K., Engelaar, W.M.H.G., Maruyama, S., Yoshida, M., Okubo, A., Nagata, T., 2001. Visualisation of moisture distribution during development of rice caryopses (Oryza sativa L.) by nuclear magnetic resonance microimaging. Journal of Cereal Science 33, 105114. Horigane, A.K., Takahashi, H., Maruyama, S., Ohtsubo, K., Yoshida, M., 2006. Water penetration into rice grains during soaking observed by gradient echo magnetic resonance imaging. Journal of Cereal Science 44, 307316. Hoshikawa, K., 1968. Studies on the development of endosperm in rice, 11. Development of amyloplasts in endosperm tissue. Proceedings of the Crop Science Society of Japan 37, 207216 (in Japanese with English summary). IPCC, 2007. Working Group II. 4th Assessment Report, Impacts, Adaptation and Vulnerability http://www.ipcc-wg2.org/. 7 Oct. 2008. Ishimaru, T., Nakazono, M., Masumura, T., Abiko, M., San-oh, A.Y., Nishizawa, K.N., Kondo, M., 2007. A method for obtaining high integrity RNA from developing aleurone cells and starchy endosperm in rice (Oryza sativa L.) by laser microdissection. Plant Science 173, 321326. Jiang, H., Dian, W., Wu, P., 2003. Effect of high temperature on ne structure of amylopectin in rice endosperm by reducing the activity of starch branching enzyme. Phytochemistry 63, 5359. Kano, H., Ishida, N., Kobayashi, T., Koizumi, M., 1990. 1H-NMR imaging analysis of changes of free water distribution in barley and soybean seeds during maturation. Japanese Journal of Crop Science 59, 503509.

174

T. Ishimaru et al. / Journal of Cereal Science 50 (2009) 166174 under different temperature. Proceedings of the Crop Science Society of Japan 42, 214219 (in Japanese with English summary). Sato, K., Inaba, K., 1976. High temperature injury of ripening in rice plant: V. On the early decline of assimilate storing ability of grains at high temperature. Proceedings of the Crop Science Society of Japan 45, 156161 (in Japanese with English summary). Tabata, M., Hirabayashi, H., Takeuchi, Y., Ando, I., Iida, Y., Ohsawa, R., 2007. Mapping of quantitative trait loci for the occurrence of white-back kernels associated with high temperatures during the ripening period of rice (Oryza sativa L.). Breeding Science 57, 4753. Tanaka, K., Kasai, Z., Ogawa, M., 1995. Accumulation of reserve substances and formation of subcellular organelles. In: Matsuo, T., Kumazawa, K., Ishii, R., Ishihara, K., Hirata, H. (Eds.), Science of the Rice Plant, Vol. 2. Physiology, Food and Agricultural Policy Research Centre, Tokyo, pp. 100116. Tashiro, T., Wardlaw, I.F., 1991. The effect of high temperature on kernel dimensions and the type and occurrence of kernel damage in rice. Australian Journal of Agricultural Research 42, 485496. Wakamatsu, K., Sasaki, O., Uezono, I., Tanaka, A., 2007. Effects of high air temperature during the ripening period on the grain quality of rice in warm regions of Japan. Japanese Journal of Crop Science 76, 7178 (in Japanese with English summary). Yamakawa, H., Hirose, T., Kuroda, M., Yamaguchi, T., 2007. Comprehensive expression proling of rice grain ripening-related genes under high temperature using DNA microarray. Pant Physiology 144, 258277. Zakaria, S., Matsuda, T., Tajima, S., Nitta, Y., 2002. Effect of high temperature at ripening stage on the reserve accumulation in seed in some rice cultivars. Plant Production Science 5, 160168.

Kim, S.S., Lee, S.E., Kim, O.W., Kim, D.C., 2000. Physicochemical characteristics of chalky kernels and their effects on sensory quality of cooked rice. Cereal Chemistry 77, 376379. Kobayashi, A., Genliang, B., Shenghai, Y., Tomita, K., 2007. Detection of quantitative trait loci for white-back and basal-white kernels under high temperature stress in japonica rice varieties. Breeding Science 57, 107116. Lin, S.K., Chang, M.C., Tsai, Y.G., Lur, H.S., 2005. Proteomic analysis of the expression of proteins related to rice quality during caryopses development and the effect of high temperature on expression. Proteomics 5, 21402156. Lisle, A.J., Martin, M., Fitzgerald, M.A., 2000. Chalky and translucent rice grains differ in starch composition and structure and cooking properties. Cereal Chemistry 77, 627632. MacFall, J.S., Van As, H., 1996. Magnetic resonance imaging of plants. In: ShacharHill, Y., Preffer, P.E. (Eds.), Nuclear Magnetic Resonance in Plant Biology. American Society of Plant Physiologists, pp. 3376. Morita, S., Shiratsuchi, H., Takahashi, J., Fujita, K., 2004. Effect of high temperature on grain ripening in rice plants analysis of the effects of high night and high day temperatures applied to the panicle and other parts of the plant. Japanese Journal of Crop Science 73, 7783 (in Japanese with English summary). Nagato, K., Ebata, M., 1965. Effects of high temperature during ripening period on the development and the quality of rice kernels. Proceedings of the Crop Science Society of Japan 34, 5966 (in Japanese with English summary). Nagato, K., Kobayashi, Y., 1959. On the development of the tissue of starch-cell in rice kernels. Proceedings of the Crop Science Society of Japan 27, 204206 (in Japanese with English summary). Sato, K., Inaba, K., 1973. High temperature injury of ripening in rice plant: II. Ripening of rice grains when the panicle and straw were separately treated

Journal of Cereal Science 50 (2009) 175183

Contents lists available at ScienceDirect

Journal of Cereal Science


journal homepage: www.elsevier.com/locate/jcs

Simulation of the factors affecting b-glucan levels during the cultivation of oats
Uma Tiwari*, Enda Cummins
Biosystems Engineering, UCD School of Agriculture, Food Science and Veterinary Medicine, University College Dublin, Beleld, Dublin 4, Ireland

a r t i c l e i n f o
Article history: Received 12 December 2008 Received in revised form 21 April 2009 Accepted 23 April 2009 Keywords: Oats b-Glucan Simulation Scenarios Sensitivity analysis

a b s t r a c t
A Monte Carlo simulation technique was employed to simulate the factors inuencing the level of bglucan content in both hulled (HO) and naked (NO) oat cultivars during cultivation of the crop. Probability density functions were used to model the uncertainty and variability in the input factors. A scenario analysis was subsequently developed to look at the impact of different model assumptions and input parameters. The simulated mean b-glucan level in harvested oats grain was 3.50 and 4.25 g/100 g for hulled (HO) and naked (NO), respectively. A sensitivity analysis highlighted that cultivar selection was the most important input parameter compared to other inputs in determining the nal b-glucan level (correlation coefcients of 0.64 and 0.79 for HO and NO, respectively). The analysis also indicated the positive effect of delayed sowing on b-glucan content (correlation coefcients of 0.32 and 0.25 for HO and NO, respectively). Germination and storage factors showed a negative impact on the nal b-glucan levels. The scenario analysis shows the applicability of the proposed model for various agronomic practices. This approach establishes a quantitative scientic ranking of factors inuencing b-glucan levels during the cultivation of oats. 2009 Elsevier Ltd. All rights reserved.

1. Introduction Oats (Avena sativa. L) remain an important cereal crop throughout the developing world, and for specialist uses in developed economies, being used for human food and livestock feed. In general, the use of oats grain as an animal feed has declined steadily (DEFRA, 2003), which may be associated with an emerging use and interest in oats as a human health food. The amount of oats used for human consumption has increased because of the dietary benets associated with phytochemicals, such as b-glucans, present in the grain (FDA, 1997). By incorporating oats b-glucan into various food products, including breakfast cereals, beverages, bread and infant foods (Flander et al., 2007; Yao et al., 2007) it improves the nutritive quality of food and may have positive health benets (FDA, 1997). b-Glucan content is predominately found in the endosperm of cereal grain

Abbreviations: CF, Climatic factor; FF, Fertiliser factor; Fu, Foliar urea; GF, Germination factor; GTemp, Germination temperature; GTime, Germination time; HO, Hulled oats; IF, Irrigation factor; L, Location; N, Nitrogen application; NO, Naked oats; Pr, Precipitation (growing period); SDays, Storage days; SDD, Sowing delay days; SDF, Sowing date factor; SF, Storage factor; SoF, Soil factor; SpH, Soil pH; SPr, Precipitation (spring oats); ST, Temperature (spring oats); STemp, Storage temperature; SMo, Storage moisture; T, Temperature (growing period); WPr, Precipitation (winter oats); WT, Temperature (winter oats). * Corresponding author. Tel.: 353 1 716 2163. E-mail address: uma.tiwari@ucd.ie (U. Tiwari). 0733-5210/$ see front matter 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2009.04.014

(Behall et al., 1997). Unlike other cereals such as barley, oats have a thick layer of cell wall in the sub-aleurone region of the kernel, which contains higher amounts of b-glucan (Fulcher and Miller,1993). Oat b-glucan is largely conned to the triploid tissues, and that relative to barley, its content is biased towards the aleurone, rather than being evenly spread through the endosperm. Oat b-glucan is a viscous and soluble dietary bre component, with potentially positive benets for human health and nutrition (Anttila et al., 2004; Tiwari and Cummins, 2009) in terms of lowering cholesterol and blood glucose levels (Maier et al., 2000). Oats can be broadly classied as hulled and naked (hull-less), the naked oats are nutritionally superior compared to conventional hulled oats (MAFF, 1990). Some scientists argue that the oats classications (i.e. hulled and naked oats) are only an articial one created by the oat breeders. However, by virtue of the fact there is a difference by weight because the hull contains negligible amounts of b-glucan compared to the endosperm of the grain (Wood, 1987) thus naked oats contains a higher percentage b-glucan compared to hulled oats (Longland and Valentine, 1993). Naked oats have a thin non-lignied husk on the outside of the grain which falls off during harvesting, resulting in a grain of higher energy, protein and lipid and lower bre content compared with conventional oats (Bhatty, 1995; Givens and Brunnen, 1987). To account this variation in b-glucan content, this study included both hulled and naked oats based on scientic literatures. However, it is noted these classications are an artefact of oat breeding studies.

176

U. Tiwari, E. Cummins / Journal of Cereal Science 50 (2009) 175183

Oat cultivars contain 3.05.0% of b-glucan (Chernyshova et al., 2007; Malkki et al., 2004). Variations in b-glucan content are inuenced by various cultivars, environmental and agronomic practices (Ajithkumar et al., 2005; Doehlert et al., 2001; Zhou et al., 1998). Furthermore, food processing profoundly inuences the physical nature of oat b-glucan, making it more available at the site of action in the gut. Thus, the measured grain b-glucan content at the point of utilisation for food product manufacture will be lower by a variable amount than that which nally reaches the site of action. In this study various cultivation and farm-level factors inuencing the level of b-glucan content in oats grain is assessed and a mathematical model developed to provide an estimate of bglucan levels in harvested grain. This process was modelled using Monte Carlo simulation techniques by considering both uncertainty and variability within the cultivation system, and probability of selecting various process criteria. The modelling approach involved three main components i.e. (a) modelling the level of b-glucan in grain, (b) modelling the change in b-glucan levels resulting from various environmental and agronomic practices, and nally (c) modelling b-glucan changes during storage and following harvesting. Model development was based on Irish agronomic practices followed for an oats crop. However, this approach and methodology can be adopted for any agronomic practice. The novelty of the approach used in this study lies in quantifying both the uncertainty and variability in factors inuencing b-glucan content for climatic, farm management and agronomic practices followed in Ireland, thus resulting in a quantitative scientic ranking of the factors inuencing b-glucans content. Thus, the objective of this study was to quantify the variability and uncertainty in the level of b-glucans in harvested oats and, through a sensitivity analysis, assess the relative importance various cultivation and management practices have on inuencing the nal bglucan levels in harvested oats. 2. Model methodology 2.1. Farm-level model development Various factors which may inuence b-glucan levels in oats as inferred from the scientic literature, include; cultivar factors, environmental conditions (location, soil types, precipitation, temperature), agronomic practices (sowing date, fertilisation, irrigation, harvesting) and storage conditions (i.e. time and temperature). These factors have been reported to affect both yield and chemical composition of the oats grain (Welch and Yong, 1980). Depending on growth environment, oat grain b-glucan content can vary markedly; but not to the same extent of barley b-glucan (Coles et al., 1991). Furthermore, varying agronomic and storage conditions of oats may also inuence the level b-glucan content in the harvest grain. These factors, in addition to varying agricultural practices, may result in varying b-glucan levels in harvested grain. The uniqueness of this study lies in quantifying both the uncertainty and variability in the input factors inuencing the oats bglucan content (using probability distributions), with specic emphases on Irish agricultural practices. Monte Carlo simulation was used to assess the effect of uncertainty and variability in the model input parameters on the predicted b-glucan levels by described inputs with probability distribution functions (PDFs). The probabilistic analysis can be used to propagate uncertainties in model inputs to estimate uncertainties and quantitative insights into model outputs (Cullen and Frey, 1999). The approach may aid in identifying key sources of uncertainty and variability that can be a focus of future model developments. A ow chart representation of the model is shown

in Fig. 1. The probability distributions used in the model are discussed in the following section and summarised in Table 1. 2.2. Input data description 2.2.1. Cultivar data Data on thirty hulled oat (HO) and naked oat (NO) cultivars were collated from existing scientic literature (Anon, 2006; Buerstmayr et al., 2007; Cowan et al., 2001; Givens et al., 2000; Hozova et al., 2007; Lim et al., 1992; Longland and Valentine,1993; Marshal, 2003; South et al., 1999; Wang et al., 2007) and were used to model the initial level of b-glucan in grain (Table 2). The variability in b-glucan content in the oat cultivars were modelled using a normal distribution with a mean and standard deviation of 3.65 and 0.47 g/100 g for HO and 4.43 and 0.92 g/100 g for NO cultivars, respectively. Fig. 2 shows the cumulative frequency distribution, supplemented by calculation of the AndersonDarling statistic. 2.2.2. Environmental conditions The b-glucan content of oats is inuenced by various environmental factors, including soil nitrogen level and precipitation (Sorrell and Simmons, 1987). Location (L) and growing area have only small effects within regions on b-glucan content, so a factor of 1 was assumed for the effect of location (L) in Ireland. Oats (winter and spring sown) is a versatile crop, which could be grown on many soil types around the world. In a recent study, Seling (2007) demonstrated that the b-glucan content of winter sown oat is similar to spring sown oats. In a study, OMahony (2001) reported that there are wide ranging soil types available in Ireland for oat crops, however recommended heavy soils for oat production. Hence, there is a relative consistency in soils which sow oat crops. Given there is little variation in the soil type on which oats are grown the authors feel the assumption that soil type will have negligible inuence is justied. A factor of 1 was assumed for the inuence of soil (SoF) as soil type was shown to have negligible effects on b-glucan content. The optimum soil pH for growth of oat crops is between 5.3 and 5.7 (Alam and Adam, 1979). Available literature does not indicate any effect of soil pH on b-glucan levels in oats. As a result, a factor of 1 for soil pH (SpH) was assumed in the model. Climate is the factor of primary importance governing oats growth, among the climate factors; temperature and precipitation are the most important, which inuence oat b-glucan levels. Based on several studies, temperatures ranging from 13 to 19  C are optimal for oat plants to produce high grain and straw yields

Selection of Genotypes

Environmental Conditions

Location Soil Types Climate (Temperature & Precipitation) Sowing Fertiliser Application Irrigation Harvesting

Agronomic Practices

Transportation

Storage Condition Days Temperature Moisture Content

Fig. 1. Flow chart representation used to model the cultivation of oats.

U. Tiwari, E. Cummins / Journal of Cereal Science 50 (2009) 175183 Table 1 Model input distributions to model the predicted b-glucan levels in oats. Symbol 1. Cultivar data HO NO 2. Environmental conditions L SoF SpH Climatic condition SPr WPr ST WT Pr T CF 3. Agronomic factors SDD SDF Fertiliser application N Fu FF Irrigation IF Germination GTime GTemp GF 4. Storage SDays STemp SMo SF Description Hulled oats Naked oats Mean 3.65 4.43 Distribution Normal (3.65, 0.47) Normal (4.43, 0.92) Units g/100 g g/100 g

177

Location Soil factor Soil pH Precipitation (spring oats) Precipitation (winter oats) Temperature (spring oats) Temperature (winter oats) Precipitation (growing period) Temperature (growing period) Climatic factor

1 1 1 60 95 15 7 95 7 0.951

Factor Versatile soil Factor Uniform (min 20, max 100) Uniform (min 40, max 150) Triangle (min 10, max 20) Triangle (min 5, max 10) Discrete ({WPr, SPr}, {0.51, 0.49}) Discrete ({WT, ST}, {0.51, 0.49}) Prediction line, Function of Pr and T mm/month mm/month  C  C mm/month  C

Sowing delay days Sowing date factor Nitrogen application Foliar urea Fertiliser factor Irrigation factor Germination time Germination temperature Germination factor

15 1.129 110 60 1.033 1 1 10 0.952

Uniform (min 0, max 30) Fixed factor Triangle (min 40, max 140) Fixed Prediction line, function of N and Fu Fixed factor Exponential, beta 1.33 Fixed value Prediction line, function of GTime, GTemp

days days kg/ha kg/ha kg/ha

days  C

Storage Storage Storage Storage

days temperature moisture factor

180 10 1225 0.730

Uniform (min 0, max 365) Triangle (min 5, max 20) Prediction line, function of STemp Prediction line, function of SDays, STemp

days  C %

(Sorrells and Simmons, 1992). The b-glucan content is dependent on the temperature of the growing period for the oats while high precipitation decreases the b-glucan content of oats (Saastamoinen, 1995). Givens et al. (2000) reported that in oat grain, b-glucan accumulation occurs late in grain development, suggesting that environmental stresses leading to an early end to grain development also reduce the b-glucan concentration. In a study, Coles et al. (2000) reported the b-glucan accumulation in barley grain occurs

Table 2 Descriptive statistics for initial oats cultivar.a Descriptive statistics Mean Standard error Median Mode Standard deviation Sample variance Kurtosis Skewness Range Minimum Maximum Count Hulled oats 3.648 0.086 3.608 2.800 0.472 0.223 0.343 0.265 1.860 2.800 4.660 30.000 Naked oats 4.426 0.169 4.100 3.500 0.923 0.852 4.377 1.662 4.469 3.271 7.740 30.000

a Lim et al., 1992; Longland and Valentine, 1993; South et al., 1999; Givens et al., 2000; Cowan et al., 2001; Marshal, 2003; Anon, 2006; Buerstmayr et al., 2007; Hozova et al., 2007; Wang et al., 2007.

after anthesis and rises sharply till the growth phase is completed. These studies show that environmental factors have a greater impact, while the authors suggest genetic background has a greater inuence on b-glucan compared to environmental factors. Additionally, Coles et al. (2000) also reported that the cell wall of barley co-opted to allows rapid sequestration and remobilisation to starch. Higher precipitation during grain ripening may reduce the bglucan content in oats (Brunner and Freed, 1994), whereas, low precipitation and the temperature ranging from 13 to 23  C for the growing season may inuence the b-glucan concentration (Manthey et al., 1999). In contrast, Doehlert et al. (2001) observed a signicant positive correlation (P < 0.01) between total precipitation in the last two months before harvest and the b-glucan concentration of ten oat cultivars. This apparent conict in results may be attributed to the variation in cultivars used. In order to model Irish climatic conditions, the mean monthly precipitation (Pr) and mean average temperature (T) was collated for the growing seasons from 1989 up to the year 2007 (KleinTank et al., 2008) for spring and winter grown oats. According to the Central Statistics Ofce (2007), spring and winter oats account for approximately 49% and 51% of the total acreage of oats in Ireland, respectively. According to Irish agronomic practices, spring oats are generally sown in late February or early March and harvested in late July whereas winter oats are usually sown primarily in mid November and harvested in late July. Hence, both spring and winter oats grown in Ireland will be subjected to precipitation and

178

U. Tiwari, E. Cummins / Journal of Cereal Science 50 (2009) 175183

1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1

Summary Statistic Mean (HO) 3.65 A-D 0.40

1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 2.5 3.0 3.5 4.0

Probability Density

Summary Statistic Mean (NO) 4.43 A-D 1.05

2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0

4.5

5.0

-glucan (g/100g)

-glucan (g/100g)

Fig. 2. Input distribution for the b-glucan level in hulled and naked oats cultivar.

temperatures experienced during their respective growing seasons. The growth period of spring oats are usually shorter than winter oats. In order to capture the uncertainty in the climatic inuence (CF) affecting the level of b-glucan content in spring oats (SO) and winter oats (WO); a uniform distribution was used to model the precipitation level with a minimum of 20 mm/month and maximum of 100 mm/month for the spring growing season and a minimum of 40 mm/month and maximum of 150 mm/month for the winter growing season. The mean average daily temperature for the spring growing season was modelled using a triangle distribution with a minimum of 10  C, most likely of 15  C and maximum of 20  C and the temperature for winter growing season was modelled using a triangle distribution with a minimum of 5  C, most likely of 7  C and maximum of 10  C in line with published data (KleinTank et al., 2008). A discrete distribution was then used to determine if spring or winter conditions would prevail for a simulation run with a probability of 49% for spring growing conditions and 51% for winter growing conditions based on seasonal acreage sown (CSO, 2007). The resulting distribution for precipitation and temperature are given in (Table 1). A discrete distribution was then used to model the precipitation and temperature during the growing season using the meteorological conditions for spring and winter oats with a probability of 49% spring growing conditions and 51% winter growing conditions. The resulting distribution for precipitation and temperature are given in equations (1) and (2) respectively.

a given study. The predicted model was signicant (P < 0.05) with R2 (coefcient of determination) of 0.57.

CF 0:6423 0:0349 T 0:00359 Pr 0:0004 Pr T (3)


where CF is the climatic factor inuence on b-glucan levels (factor increase/decrease) Therefore from equation (3), the CF can be evaluated from the corresponding effect on the b-glucan level due to climatic conditions. The model simulates precipitation (Pr in mm/month) and temperature (T in  C) using the distributions described earlier. 2.2.3. Agronomic practices The oats b-glucan content is signicantly (P < 0.05) inuenced by environmental, agronomic characteristics and grain quality traits (Doehlert et al., 2001). Buerstmayr et al. (2007) also showed the effect of genetic variations for yield, agronomic characters and grain quality traits were highly signicant (P < 0.001) in oats lines of worldwide origin. Farm management practices such as sowing date and nitrogen (N) application also have profound effects on grain quality (Zhou et al., 1998). For spring oats, sowing is usually carried out between February and March and for winter oats sowing is carried out in late July to mid November. Research has indicated that a crop sown in the latter half of the recommended sowing period may result in higher b-glucan levels. This delay in sowing positively affects the b-glucan content the degree of increase depends on the specic cultivar (Humphreys et al., 1994). Humphreys et al. (1994) reported that the sowing delay up to 20 days (1st21st May) and 29 days (24 April23 May) increased the b-glucan content signicantly (P < 0.01) and the datasets are shown in (Table 4). Further, oats protein content increased with delayed seeding (Nass et al., 1975) and this may be due to shortened grain lling period leading to reduce starch production (man and Graham, 1987) and subsequently increased b-glucan content (Humphreys et al., 1994). Protein and b-glucan are positively correlated (Brennan and Cleary, 2005). In order to capture the uncertainty in delayed sowing within a growing season a uniform distribution was used to model the sowing delay days (SDD) for both Irish spring and winter oats cultivation with a minimum of 0 day and maximum of 30 days as typical delay time (days) as reported by Humphreys et al. (1994). A b-glucan prediction line was created using the dataset of Humphreys et al. (1994) and thus the sowing delay factor (SDF) can be evaluated from the data points. A sowing delay is particularly relevant to Irish climatic conditions, which can occur frequently due to adverse weather conditions. Similarly, Mahdi et al. (1998) reported delayed sowing of wheat avoided the risk of crop failure but resulted in reduced average yields. The data used to model delayed sowing was based on a mixture of climatic conditions. This is a factor in affecting yield; however delays may be

Pr DiscretefWPr; SPrg; f0:51; 0:49g T DiscretefWT; STg; f0:51; 0:49g

(1) (2)

where WPr and SPr are the precipitation levels for winter and spring oats, respectively. WT and ST are the average temperature during the spring and winter grown periods, respectively. Available literature shows that the climatic conditions, precipitation and temperature (Pr and T) inuence b-glucan content in oats grain. Eighteen years of data were collated from 1989 to 2007 based on scientic literature and a correlation of 0.63 was found between precipitation and temperature. This was included in the model in the form of a correlation matrix. The correlation matrix shows that precipitation and temperature are negatively correlated. In order to compute a relationship between the climatic conditions (Pr and T) and b-glucan content, a non-linear estimation was used. Equation (3) shows the tted non-linear regression model based on non-linear regression analysis (performed in SAS, 9.1) of the data presented in Table 3. Factor increasing/decreasing b-glucan (Table 3) were calculated based on the optimum conditions for the highest b-glucan concentration from the results of each datasets presented by various authors (Doehlert et al., 2001; Manthey et al., 1999; South et al., 1999; Yao et al., 2007). For example, factor x/y i.e. x highest b-glucan content and y smallest b-glucan content in

U. Tiwari, E. Cummins / Journal of Cereal Science 50 (2009) 175183 Table 3 Effect of precipitation and temperature on b-glucan in oats.a Precipitation (mm/month)b 79.4 76.8 70.7 97.5 84.3 115.3 55.8 71.3 76.3 82.8 84.0 73.5 65.8 73.3 70.3 54.3 46.5
a b

179

Temperature ( C) 10.3 10.6 18.0 17.0 18.0 17.5 14.5 15.3 14.8 15.7 14.3 15.0 14.6 15.5 13.3 14.4 14.2

b-Glucan
(g/100 g) 2.45 2.62 4.20 5.60 6.45 5.53 4.33 4.28 4.22 4.26 3.00 3.46 3.67 3.36 3.00 3.01 3.01

Factor increase/ decrease b-glucan 0.935 1.000 1.000 0.868 1.000 0.857 1.010 1.000 0.985 0.994 0.867 1.000 1.061 0.972 1.000 1.004 1.002

Manthey et al., 1999; South et al., 1999; Doehlert et al., 2001; Yao et al., 2007. Average over growing for spring and winter season.

unavoidable due to adverse weather conditions particular in unpredictable climates such as Ireland. Oats are a low input crop, which require less fertiliser than other cereals, this may be due to the more prolic root structure enabling more efcient gathering of resources (Cowan et al., 2001). Nitrogen (N) is the major plant nutrient, which inuences the yield of cereal crops. Research has shown N fertilisation increases oats grain yield (Buerstmayr et al., 2007), but excess N leads to lodging and poor grain lling (Humphreys et al., 1994). However, foliar application of nitrogenous fertiliser enhances yield and b-glucan level in the grain (Anonymous, 2000). Foliar urea (FU) used to increase the protein content of grain and foliar application in combination with top dressing can help in better fertiliser utilisation by reducing immobilization of soil nitrogen (N) (Weightman et al., 2004). Compared with soil application, the foliar application of urea potentially reduces the risk of leaching or denitrication (Abad et al., 2004). Foliar urea application varies from 50 to 60 kg/ha (Abad et al., 2004; Weightman et al., 2001). In order to capture the

uncertainty in the rate of fertiliser application, a triangle distribution was used with a minimum of 40 kg/ha and maximum of 140 kg/ha, as recommended by Burke et al. (2001) with standard Fu dosage for oats 60 kg/ha. Foliar feeding of urea was found to enhance the b-glucan content of oats by 0.17%, giving an average bglucan yield increase of 20 kg/ha in the United Kingdom (Weightman et al., 2001). A fertiliser factor (FF) was calculated based on the nitrogen dosage and with standard FU 60 kg/ha as detailed in Table 5. The fertiliser factor was calculated and modelled (Table 5) by creating a prediction line from the data presented by Weightman et al. (2004). The model interpolated between the data points. Ireland receives sufcient rainfall throughout the year; thus, irrigation is not a common agronomic practice in Ireland. Hence, in the baseline model a factor of 1 (i.e. no effect) was assumed for irrigation (IF). Ireland climatic variability in rainfall during crop maturation, or during later stages of crop ripening, increases the potential for pre-harvest sprouting (germination) this could occur when the mature crop is left in the eld after maturity and rainfall induces the germination process in the grain (Doehlert and McMullen, 2003). Germinated oats exhibited much more breakage during dehulling and decreases b-glucan concentration in the grain. Doehlert and McMullen (2003) reported that there is an exponential decrease in b-glucan content during germination over a four-day germination period (Table 4). To capture the uncertainty in the germination time (GTime), an exponentially decreasing distribution was used with minimum of 0 days and beta value equal to 1.33 such that the 95th percentile of the distribution was equal to 4 days, thus remaining within the bounds given by Doehlert and McMullen (2003). The germination temperature (GTemp) temperature was assumed to be 10  C. The decrease in b-glucan due to the germination time (germination factor, GF) was modelled by creating a prediction line from the data presented in Table 4. 2.2.4. Storage Oats grains are stored on-farm (temporary storage) or in storage bins at a commercial location, for relatively long periods under ideal storage conditions. Storage temperature (15  C) and moisture (1012%) are of paramount importance in the safe storage of grain (Briggs et al., 1981). Storing the grain with optimum temperature and moisture has a negligible effect on b-glucan, whereas high moisture storage degrades the b-glucan content in the oats grain (Svihus et al., 1997) (Table 4). MacGregor et al. (1994) reported that degradation of malting barley b-glucan occurs during the early stages of germination, between 12 and 24 h after imbibition. Similarly, Gajdosova et al. (2007) observed greater loss of soluble bglucan compared to insoluble b-glucan during storage of oats and barley. This decrease may be due to the activity of endogenous enzymes (b-glucanase) naturally present in the grain, which hydrolyses the b-glucan. Data presented by Svihus et al. (1997) indicating a storage period of 3 months had an impact on b-glucan levels, as shown in Table 4. This data was used to model the underlying uncertainty in the effects of storage time (SDays) and

Table 4 Effect of agronomic factors on b-glucan content in oats.a Agronomic factor (days) Sowing delay daysa 0 29 0 20 Germination timeb 0 1 2 3 4 Storagec Storage (<11  C) 0 90 Storage (>11  C) 0 90
a b c

b-Glucan (g/100 g)
4.17 5.03 4.79 5.63

Factor increase/decrease b-glucan 1.000 1.206 1.000 1.176

6.65 6.09 5.45 5.16 4.20

1.000 0.952 0.813 0.829 0.670

Table 5 Effect of fertiliser on the level of b-glucan in oats.a Fertiliser application (kg/ha) FU N 0 40 100 140 Weightman et al., 2004. 3.05 3.12 3.14 3.20 1.000 1.024 1.028 1.048

b-Glucan (g/100 g)

4.57 4.32 3.06 2.64

1.000 0.945 1.000 0.863

Factor increase/ decrease b-glucan

Humphreys et al., 1994. Doehlert and McMullen, 2003. Svihus et al., 1997.

60 60 60 60
a

180 Table 6 Simulated model outputs for oats cultivar. Symbol BG-HO BG-NO Description

U. Tiwari, E. Cummins / Journal of Cereal Science 50 (2009) 175183

Formula HO L SoF SpH CF SDF FF IF GF SF NO L SoF SpH CF SDF FF IF GF SF

Units g/100 g g/100 g

b-Glucan in harvested hulled oats grain b-Glucan in harvested naked oats grain

storage temperature (STemp) and moisture (SMo). A uniform distribution was used to model the storage days with minimum of 0 days and maximum of 365 days and a triangle distribution was used to model storage temperature with a minimum of 5  C and maximum of 20  C (Svihus et al., 1997). Fuente et al. (1998) reported the importance of the temperature in degradation of b-glucan during storage. Uncertainty in the storage temperature along with moisture was modelled by using the dataset presented by Svihus et al. (1997), where temperature <11  C and >11  C was taken to capture the variability in the storage temperature along with moisture. Thus, when the model simulates storage days and storage temperature, the corresponding effect on the b-glucan level (SF) can be evaluated from the prediction line created using the dataset from Svihus et al. (1997). The model interpolated between the data points. 2.3. Scenario analysis and model run The input parameters were combined onto a spreadsheet (Microsoft Excel, 2003) running the @Risk add-on package (Palisade Software, Neweld, NY, USA) and the simulation was performed using Latin Hypercube sampling. The simulated outputs are presented in Table 6. The simulation was performed using the parameters and calculations presented and the model run for 10,000 iterations. A table summarising all the model inputs for the baseline model is also provided in Table 1. To capture the uncertainty and variability in the farm-level process, a scenario analysis was conducted. The scenarios developed used the same data as the baseline model except where specied. No fertiliser application was assumed for scenario 1, to assess the predicted inuence of fertilisation. Scenario 2 looked at the situation where oat grains are harvested on the day of maturity (i.e. growth stage 92) (White, 1995) hence, no on-farm germination with an assumption that there is no rainfall during harvesting. Scenario 3 looked in to a condition where there is no storage of the harvested grain. Table 7 details a summary of the three scenarios modelled and variation in inputs from the baseline model. 3. Results and discussion The simulated baseline model resulted in a number of output distributions, which can be used to predict the likely b-glucan content of harvested grain and the impact of different farm-level processes on b-glucan levels in oats. Outputs from the model allowed for a comparison of b-glucan content between HO and NO cultivars.

3.1. Comparison of b-glucan level in harvested oats grain For the baseline model, the simulated mean value for b-glucan content was 3.50 g/100 g for HO and 4.25 g/100 g for NO, respectively (Fig. 3), (95th percentile 4.68 and 6.15 g/100 g for HO and NO, respectively). Skewness is a measure of symmetry, or more precisely, the lack of symmetry and Kurtosis is a measure of whether the data are peaked or at relative to a normal distribution. Thus, the skewness and kurtosis value showed in (Fig. 3) are relative to the centre point of datasets. The simulated mean value of both HO and NO cultivars are within the range of the validation band shown in Fig. 2, representing the Irish/UK range for b-glucan content (ranging from 2.8 to 4.5 g/100 g for HO and 3.25.2 g/100 g for NO, respectively). This clearly indicating the varietal variation in b-glucan content of HO and NO as reported by Malkki et al. (2004) and is also consistent with the ndings of Gajdosova et al. (2007) who reported varietal variation in b-glucan content of oats ranging from 3.91 to 7.47% for naked oats and 1.974.09% for hulled oats, respectively. Thus the model predictions are within the expected range while validating the models estimates. 3.2. Farm-level model sensitivity analysis A sensitivity analysis is the study of how the uncertainty in the output can be apportioned qualitatively or quantitatively to the variation in the input parameters (Cacuci et al., 2005). A sensitivity analysis to assess the inuence of model inputs on b-glucan content in oats grain was carried out using the rank order correlation coefcient statistic and the analysis is given in Fig. 4. The analysis showed that the parameter having the greatest impact on model predictions was the initial level of b-glucan content of the oats grain (HO, NO) (with correlation coefcients of 0.64 and 0.79 for HO and NO cultivars, respectively). This analysis highlights the importance of cultivar selection in determining the b-glucan levels of harvested grain over any of the agronomical factors analysed. Researchers have reported that the effect of cultivars on b-glucan content of oats is signicant (Doehlert et al., 2001; Lim et al., 1992) and this model bears out this fact through the sensitivity analysis. Sowing date is the major factor inuencing rate of b-glucan development (correlation coefcient 0.32 and 0.25 for HO and NO, respectively). This is in agreement with the ndings of Humphreys et al. (1994), however, care needs to be exercised as delayed sowing may also have adverse effects on the grain yield and quality (Anon, 2006; White, 1995). Nitrogen fertilisation was shown to have a negligible effect on b-glucan concentration (correlation coefcient 0.01 for HO and NO, respectively). This can be substantiated by the study of Cowan et al. (2001) who state oat crops are low input crop, which require less fertiliser compared to other cereals. The limited effect of nitrogen fertilisation may be due to the fact that Irish soils are intrinsically high in nitrogen and agricultural practice (including grazing) leads to limited nitrogen depletion. Another critical factor for Irish conditions is the germination time (GTime), which had a negative effect on b-glucan (0.31, 0.22 for HO and NO, respectively). This may be due to greater probability of rain all year round and thereby a greater likelihood of germination could occur, especially if harvesting is delayed. This highlights the importance of harvesting the crop in dry conditions thus reducing the potential for germination.

Table 7 Different simulated scenarios in comparison with baseline model. Model/scenarios Input summary of scenario analysis Baseline model No fertiliser application Harvesting on physiological maturity No storage

b-Glucan level in harvested HO (g/100 g)


3.50 3.40 (3.0)a 3.73 (6.4) 3.88 (11)

b-Glucan level in harvested NO (g/100 g)


4.25 4.12 (3.0) 4.52 (6.4) 4.71 (11)

Baseline Scenario 1 Scenario 2 Scenario 3


a

Value in parentheses denotes the percentage change over baseline model.

U. Tiwari, E. Cummins / Journal of Cereal Science 50 (2009) 175183

181

90 80 70 60 50 40 30 20 10 0 1.0

2.0

3.0

4.0

5.0

6.0

7.0

50 45 40 35 30 25 20 15 10 5 0 0.0

Probability Density (10-2)

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

9.0

-glucan (g/100g)
Summary Statistic 3.65 Harvested Grain: Mean (HO) 3.50 Skewness Initial ~ harvested grain 0.00 ~ 0.26 Kurtosis Initial~ harvested grain 3.00 ~ 2.88
Initial Grain: Mean (HO)

-glucan (g/100g)
Summary Statistic
Initial Grain: Mean (NO)

4.43 Harvested Grain: Mean (NO) 4.25 Skewness Initial ~ harvested grain 0.00 ~ 0.27 Kurtosis Initial~ harvested grain 3.01 ~ 2.99

Fig. 3. Simulated uncertainty distribution for the b-glucan level in oats grain and nal harvested grain for the baseline model.

Storage days and Storage temperature were other factors predicted to have an inuence on the levels of b-glucan (correlation coefcient 0.40, 0.21 and 0.30, 0.15 for HO and NO, respectively). This indicates that storage time should be kept as short as possible with reduced storage temperature (<11  C). This is consistent with data reported by Svihus et al. (1997). 3.3. Scenario analysis results In addition to the baseline model a number of scenarios were simulated to highlight the impact of fertiliser application, harvest date and storage conditions. Results from the additional scenarios evaluated in the model are given in Table 7. By not applying fertiliser (scenario 1) b-glucan content was reduced by 3% for both HO and NO cultivars, respectively compared to the baseline model. This highlights the small, but positive effect, of fertiliser application on the b-glucan levels. The impact of harvesting date shows increased b-glucan by approximately 6.4% for both HO and NO cultivars, highlighting the importance of harvesting as close to growth stage 92 as possible to optimise the retention of b-glucan in the harvested grain and to reduce the possibility of germination and thereby

controlling the grain yield. By not storing oats grain after harvesting (scenario 3) b-glucan content signicantly increased by 11% for both HO and NO cultivars, highlighting the impact of storage on the level of b-glucan. This scenario analysis highlights the importance of agronomic practices such as fertilisation, harvesting date and storage of harvested grains, although they are not as effective as the initial cultivar selection, in effecting b-glucan content in oats. Similarly, Tiwari and Cummins (2008) in reported that the initial level of b-glucan in the cultivars plays a major role in nal b-glucan level in harvested barley and is more important than other agronomic factors. This study represents a preliminary effort to model various cultivation and storage conditions and their effects on b-glucan content in harvested oats grain. This is a novel approach to predict and assess the level of b-glucan content in oats grain. The model indicated a mean level of b-glucan in HO and NO of 3.50 g/100 g and 4.25 g/100 g, respectively, which is within the range of reported values found in the literature. A sensitivity analysis highlighted the impact of cultivar selection, which has a major role in determining the nal b-glucan level (correlation coefcients of 0.64 and 0.79 for HO and NO, respectively), followed by sowing delay,

Cultivar (HO, NO)

0.64 0.79

Sowing Delay Days (SDD)

0.32 0.25

Germination Time (GTime)

-0.40 -0.30

Storage Days (SDays)

-0.31 -0.22 Hulled Oats (HO) Naked Oats (NO)

Storage Temperature (STemp)

-0.21 -0.15

-1.0

-0.5

0.0

0.5

1.0

Correlation Coefficients
Fig. 4. Sensitivity analysis for the b-glucan levels in harvested oats cultivar.

182

U. Tiwari, E. Cummins / Journal of Cereal Science 50 (2009) 175183 Coles, G., Morrell, M., Coles, T.K., Abbot, R.C., 2000. Beta glucan extraction From cereals. In: International patent application number PCT/NZ00/00034 (New Zealand). Cowan, A.A., Valentine, J., Jones, D.M., Middleton, B.T., 2001. Assessment of Potential Varieties of Winter Oats for Improved Economic Competitiveness and End-use Characteristics (19982000). HGCA Project Report-241. Home Grown Cereals Authority, London. Cullen, A.C., Frey, C.H., 1999. Probabilistic techniques in exposure assessment. In: A Handbook for Dealing with Variability and Uncertainty in Models and Inputs. Plenum Publication Co-operation, pp. 1520. DEFRA, 2003. Department for Environment, Food and Rural Affairs, Statistics on Raw Materials Usage in the Production of Animal Feedstuffs by GB Retail Compounders. http://www.defra.gov.uk/esg/Excel/hstcomps.xls. Doehlert, D.C., McMullen, M., Hammond, J.J., 2001. Genotypic and environmental effects on grain yield and quality of oat grown in North Dakota. Crop Science 41, 10661072. Doehlert, D.C., McMullen, M.S., 2003. Identication of sprout damage in oats. Cereal Chemistry 80, 608612. Flander, L., Salmenkallio-Marttila, M., Suortti, T., Autio, K., 2007. Optimisation of ingredients and baking process for improved wholemeal oat bread quality. LWT Food Science and Technology 40, 860870. Food and Drug Administration (FDA), 1997. Final rule for food labelling: health claims; oats and coronary heart disease. Federal Regulations 7 (62), 35843681. Fuente, J.M., Perez De Ayala, P., Flores, A., Villamide, M.J., 1998. Effect of storage time and dietary enzyme on the metabolisable energy and digesta viscosity of barley-based diets for poultry. Poultry Science 77, 9097. Fulcher, G.R., Miller, S.S., 1993. Structure of oat bran and distribution of dietary bre components. In: Wood, P.J. (Ed.), Oat Bran. American Association of Cereal Chemists, USA.  Gajdosova, A., Petrulakova, Z., Havrlentova, M., Cervena, V., Hozova, B., Sturdk, E., Kogan, G., 2007. The content of water soluble and water-insoluble b-D-glucan in selected oats and barley varieties. Carbohydrate Polymers 70, 4652. Givens, D.I., Brunnen, J.M., 1987. Nutritive value of naked oats for ruminants. Animal Feed Science and Technology 18, 8387. Givens, D.I., Davies, T.W., Laverick, R.M., 2000. Dietary bre fractions in hulled and naked winter oat grain: effects of genotypes and various agronomic factors. Journal of the Science of Food and Agriculture 80, 491496. Hozova, B., Kuniak, L., Moravckova, P., Gajdosova, A., 2007. Determination of waterinsoluble b-D-glucan in the while-grain cereals and pseduocereals. Czech Journal of Food Science 25, 316324. Humphreys, D.G., Smith, D.L., Mather, D.E., 1994. Nitrogen fertiliser and seeding date induced changes in protein, oil and b-glucan contents of four oat genotypes. Journal of Cereal Science 20, 283290. KleinTank, A.M.G., Wijngaard, J.B., Konnen, G.P., Bohm, R., Demaree, G., Gocheva, A., Mileta, M., Pashiardis, S., Hejkrlik, L., Kern-Hansen, C., Heino, R., Bessemoulin, P., Muller-Westermeier, G., Tzanakou, M., Szalai, S., Palsdottir, T., Fitzgerald, D., Rubin, S., Capaldo, M., Maugeri, M., Leitass, A., Bukantis, A., Aberfeld, R., Van Engelen, A.F.V., Forland, E., Mietus, M., Coelho, F., Mares, C., Razuvaev, V., Nieplova, E., Cegnar, T., Lopez, A.J., Dahlstrom, B., Moberg, A., Kirchhofer, W., Ceylan, A., Pachaliuk, O., Alexander, L.V., Petrovic, P., 2008. Daily Dataset of 20th-century Surface Air Temperature and Precipitation Series for the European Climate Assessment & Dataset (ECA&D) Data and metadata available at. http://eca.knmi.nl. Lim, H.S., White, P.J., Frey, K.J., 1992. Genotypic effects on b-glucan content of oat lines grown in two consecutive years. Cereal Chemistry 69, 262265. Longland, A.C., Valentine, J., 1993. The mixed linked (1 / 3) (1 / 4)-b-D-glucan content of seven varieties of oat growth at three locations in the UK. Aspects of Applied Biology 36, 153157. MacGregor, A.W., Dushnicky, L., Schroeder, S.W., Balance, G.M., 1994. Changes in barley endosperms during early stages of germination. Journal of the Institute of Brewing 100, 8590. MAFF, 1990. UK Tables of Nutritive Value and Chemical Composition of Feedstuffs. Rowett Research Services Limited, Aberdeen, 420 pp. Mahdi, L., Bell, C.J., Ryan, J., 1998. Establishment and yield of wheat (Triticum turgidum L.) after early sowing at various in a semi-arid Mediterranean environment. Field Crop Research 58, 187196. Maier, S.M., Turner, N.D., Lupton, J.R., 2000. Serum lipids in hypercholesterolemic men and women consuming oat bran and amaranth product. Cereal Chemistry 77, 297302. Malkki, Y., Myllymaki, O., Teinila, K., Koponen, S., 2004. Method for preparing an oat product and a foodstuff enriched in the content of beta-glucan. US patent 6, 797, 307. Manthey, F.A., Hareland, G.A., Huseby, D.J., 1999. Soluble and insoluble dietary bre content and composition in oat. Cereal Chemistry 76, 417420. Marshal, D., 2003. www.ars.usda.gov/SP2UserFiles/ad_hoc/66452500Downloads/ 2002-03UniformWinterOatYieldTrialFinalReport.pdf. Nass, H.G., Johnston, H.W., Macleod, J.A., Sterling, J.D.E., 1975. Effects of seeding date, seed treatment, and foliar sprays on yield and other agronomic characters of wheat, oats and barley. Canadian Journal of Plant Science 55, 4147. OMahony, J., 2001. Growing quality of oats. In: Teagasc Tillage No: 3, pp. 414. Saastamoinen, M., 1995. Effects of environmental factors on the b-glucan content of two oat varieties. Acta Agriculturae Scandinavica, Section B Soil and Plant Science 45, 181187. Seling, S., 2007. Cultivation and processing properties of oat cultivars released between 2002 and 2005. Muhle Mischfutter 144, 1720.

(correlation coefcients of 0.32 and 0.25 for HO and NO, respectively). Although, the delayed sowing had a positive inuence on the nal b-glucan level, it should be noted that the delay sowing might affect the grain quality and yield which in turn affects the market price. Nitrogen fertilisation had a negligible impact on the model results; this could be due to the fact that oat crops are low input crops. The critical parameters such as germination time, storage days and temperature showed a negative inuence on the nal b-glucan level. These parameters highlight the importance of harvesting the crop at its physiological maturity and to minimise the storage time while maintaining a low storage temperature. The scenario analysis highlights the importance of agronomic practices such as harvesting date and storage of harvested grains, although they are not as effective as the cultivar selection in effecting bglucan content. This study also helps to assess the impact of uncertainty in the input parameters and thereby help to identify optimisation techniques. The proposed model could be further extended to include food processing operations to identify the key factors inuencing b-glucan in farm to fork approach. This study provides a qualitative scientic based ranking (viz the sensitivity analysis) of the effects of various factors on b-glucan levels in oats. This quantitative ranking would not have been possible without a comprehensive analysis of available quantitative data. Hence, this work contributes to current scientic knowledge and represents a new development in the understanding of the effects of various factors on b-glucan levels in oats. References
Abad, A., Lloveras, J., Michelena, A., 2004. Nitrogen fertilisation and foliar urea effects on durum wheat yield and quality and on residual soil nitrate in irrigated Mediterranean conditions. Field Crop Research 87, 257269. Ajithkumar, A., Andersson, R., man, P., 2005. Content and molecular weight of extractable b-glucan in American and Swedish oat samples. Journal of Agricultural and Food Chemistry 53, 12051209. Alam, S.M., Adam, W.A., 1979. Effect of soil pH on the growth and mineral content of oats. Pakistan Journal of Scientic and Industrial Research 22, 147151. man, P., Graham, H., 1987. Analysis of total and insoluble mixed-linked (1-3), (1-4)-b-D-glucan in barley and oats. Journal of Agricultural and Food Chemistry 35, 704709. Anon, 2006. Oat Production in Western Australia Available online. www.agric.wa. gov.au. Anonymous, 2000. Supervision 8/2000. Cholesterol was Claimed Control Guide. Food Labels and Other Marketing Used for Blood Cholesterol Suggestive Arguments Guide. National Food Agency Finland, Helsinki. Anttila, H., Sontag-Strohm, T., Salovaara, H., 2004. Viscosity of b-glucan in oat products. Agricultural and Food Science 13, 8087. Behall, K.M., Schoeld, D.J., Hallfrisch, J., 1997. Effect of b-glucan level in oat bre extracts on blood lipids in men and women. Journal American College of Nutrition 16, 4651. Bhatty, R.S., 1995. Laboratory and pilot plant extraction and purication of b-glucans from hull-less barley and oat brans. Journal of Cereal Science 22, 163170. Brennan, C.S., Cleary, L.S., 2005. The potential use of cereal (1 / 3, 1 / 4)-b-Dglucans as functional food ingredients. Journal of Cereal Science 42, 113. Briggs, D.E., Hough, J.S., Stevens, R., Young, T.W., 1981. Malting and Brewing Science. Malt and Sweet Wort, vol. 1. Chapman and Hall, London, UK, p. 387. Brunner, B.R., Freed, R.D., 1994. Oat grain b-glucan content as affected by nitrogen level, location and year. Crop Science 34, 473476. Buerstmayr, H., Krenn, N., Stephan, U., Grausgruber, H., Zechner, E., 2007. Agronomic performance and quality of oat (Avena sativa L.) genotypes of worldwide origin produced under Central European growing conditions. Field Crop Research 101, 343351. Burke, J.I., Browne, R.A., White, E.M., 2001. Factors Affecting Yield and Quality of Oats. Teagasc, Crop Research Centre, Oak Park Carlow. Project report 4493. No. 32. Cacuci, Dan G., Ionescu-Bujor, M., Navon, M., 2005. Sensitivity and uncertainty analysis. In: Applications to Large-scale Systems, vol. II. Chapman and Hall. Central Statistics Ofce (CSO), 2007. Statistical Product Agricultural Area Used and Crop Production. http://www.cso.ie/px/pxeirestat/Dialog/varval.asp?maAS HA1&tiCrop YieldbyTypeofCrop,YearandStatistic&path../Data base/Eirestat/Agriculture%20Area%20Used%20and%20Crop%20Production/ &lang1. Chernyshova, A.A., White, P.J., Scott, M.P., Jannink, J.-L., 2007. Selection for nutritional function and agronomic performance in oat. Crop Science 47, 23302339. Coles, G.D., Jamieson, P.D., Haslemore, R.M., 1991. Effects of moisture stress on the malting quality of Triumph barley. Journal of Cereal Science 14, 161177.

U. Tiwari, E. Cummins / Journal of Cereal Science 50 (2009) 175183 Sorrells, M.E., Simmons, S.R., 1992. Inuence of environment on the development and adaptation of oat. In: Marshall, H.G., Sorrells, M.E. (Eds.), Oat Science and Technology. American Society of Agronomy and the Crop Science Society of America, Madison, WI, pp. 115164. South, J.B., Townsley, C., Atkin, D., Alvey, M., Laverick, R.M., Webb, C., Campbell, G.M., Wang, R., Gray, D.A., 1999. Sourcing Added Value Food Ingredients from Home Grown Oats. Final Project Report Number 194. Home Grown Cereals Authority, Caledonia House, London. Svihus, B., Herstad, O., Newman, C.W., 1997. Effect of high moisture storage of barley, oats, and wheat on chemical content and nutritional value for broiler chickens. Acta Agriculturae Scandinavica, Section A Animal Sciences 47, 3947. Tiwari, U., Cummins, E., 2008. A predictive model of the effects of genotypic, preand post harvest stages on barley b-glucan levels. Journal of the Science of Food and Agriculture 88, 22772287. Tiwari, U., Cummins, E., 2009. Factors inuencing b-glucan levels and molecular weight in cereal-based products. Cereal Chemistry 86, 290301. Wang, R., Koutinas, A.A., Campbell, G.M., 2007. Dry processing of oats application of dry milling. Journal of Food Engineering 82, 559567.

183

Weightman, R.M., South, J.B., Laverick, R., 2001. Manipulation of oat b-glucan content using foliar urea. In: The Third European Conference on Functional Properties in Oats, 1718 May 2001, Uppsala. Final program and abstracts, p. 49. Weightman, R.M., Heywood, C., Wade, A., South, J.B., 2004. Relationship between grain (1 / 3) (1 / 4)-b-D-glucan concentration and the response of winter-sown oats to contrasting forms of applied nitrogen. Journal of Cereal Science 40, 8186. Welch, R.W., Yong, Y.Y., 1980. The effects of variety and nitrogen fertiliser on protein production in oats. Journal of the Science of Food and Agriculture 31, 541548. White, E.M., 1995. Oats in human nutrition and health. In: Welch, R.W. (Ed.), The Oat Crop Production and Utilization. Chapman and Hall, London, pp. 88118. Wood, P.J., 1987. Oat b-glucan potential value as dietary bre. In: Stivala, S.S., Crescenzi, V., Dea, I.C.M. (Eds.), Industrial Polysaccharides. Gordon & Breach, New York, pp. 215227. Yao, N., Jannink, J., White, P.J., 2007. Molecular weight distribution of (1 / 3) (1 / 4)b-glucan affects pasting properties of our from oat lines with high and typical amounts of b-glucan. Cereal Chemistry 84, 471479. Zhou, M.X., Roberts, G.L., Robards, K., Glennie-Holmes, M., Helliwell, S., 1998. Effects of sowing date, nitrogen application, and sowing rate on oat quality. Australian Journal of Agricultural Research 49, 845851.

Journal of Cereal Science 50 (2009) 184189

Contents lists available at ScienceDirect

Journal of Cereal Science


journal homepage: www.elsevier.com/locate/jcs

Differences in functional properties and biochemical characteristics of congenetic rice proteins


Xiaohong Cao, Huanbin Wen, Cuijuan Li, Zhenxin Gu*
College of Food Science and Technology, Nanjing Agricultural University, Nanjing, Jiangsu 210095, P.R. China

a r t i c l e i n f o
Article history: Received 22 January 2009 Received in revised form 21 April 2009 Accepted 23 April 2009 Keywords: Brown rice White rice Rice bran Functional properties Biochemical characteristics Congenetic rice protein

a b s t r a c t
Proteins in brown rice (BR), white rice (WR) and rice bran (RB) were extracted from the same paddy rice and investigated for their components, functional properties and chemical characteristics by SDS-PAGE methodology. BR and WR proteins possessed a poor solubility under weak acid conditions due to a high content of glutenin and richness in higher molecular-weight (MW) protein fractions. Rice bran protein contained signicantly lower molecular-weight components (MWs < 50 kDa) than those in WR and BR. Nitrogen solubility, foaming and emulsication properties of their rice protein preparations were affected not only by pH (311), but also by the concentrations of NaCl (0.42.0%) and sucrose (4.020.0%). All of them, particularly rice bran protein, had favorable functional properties in the medium with a high salt or sugar concentration. Therefore, they have a good potential for development in the food industry. 2009 Elsevier Ltd. All rights reserved.

1. Introduction Rice is one of the most important food crops all over the world. In China, 1.8 hundred million tons of rice are milled, which generate 10 million tons of rice bran as a byproduct each year (Kang and Wang, 2006). Rice is also used to produce starch, monosodium glutamate, pigments and rice wine, in addition to the staple food for two thirds of the population of China. Rice bran is generally extracted for oil or used as feed for livestock as well. The rice dregs and defatted rice bran residues, which contain abundant proteins, are usually discarded. Worldwide, many researches are now being carried out on various plant sources of proteins (Gorinstein et al., 2002; Rangel et al., 2003; Sogi et al., 2002; Tomotake et al., 2002), with the aim of increasing the nutritional value of food products at a low cost. In China, a very important application prospect is to explore the functional properties of rice proteins and make full use of them for minimizing the resource waste and environment pollution.

Abbreviations: BR, brown rice; BRP, brown rice protein; MWs, molecularweights; RB, rice bran; RBP, rice bran protein; SDS-PAGE, sodiumdodecylsulfatepolyacrylamide gel electrophoresis; WR, white rice; WRP, white rice protein. * Corresponding author. Tel./fax: 86 25 8439 6293. E-mail address: guzx@njau.edu.cn (Z. Gu). 0733-5210/$ see front matter 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2009.04.009

The quality of rice protein including white rice protein (WRP) and rice bran protein (RBP) is slightly inferior to that of oats, while it surpasses those of wheat and corn. In addition, RBP had a similar prole of essential amino acid requirements for 2 to 5-year-old children to those of casein and soy protein (Wang et al., 1999). It has already been recognized as a highly nutritious, hypoallergenic, and particularly healthy protein source for human consumption (Shih and Daigle, 2000). Whole rice grain contains many types of proteins, which have been isolated and characterized according to their solubility properties and biochemical characteristics using the Osborne extraction method (Marshall and Wadsworth, 1994). They have been named as albumin, globulin, glutelin, and prolamin fractions, respectively. Currently, studies on rice proteins have been aimed at improving the yield of extracted proteins from WRP and RBP. Various alkaline solutions or carbohydrate-cleaving enzymes are used to treat rice our (defatted rice bran) and to raise protein content (Shih and Daigle, 1997, 2000). In addition, proteolytic enzyme modication of protein has also been envisaged by Tang et al. (2003). Some researches have found that proteolytic enzyme modication of protein is an effective way not only to improve various functional properties and biochemical characteristics, but also to increase the application yield of proteins (Panyam and Kilara, 1996; Wu et al., 1998). The protein solubility, emulsifying and foaming properties can be improved with a limited degree of hydrolysis (Kim et al., 1990). Unfortunately, little is known about the functional properties and biochemical characteristics of the

X. Cao et al. / Journal of Cereal Science 50 (2009) 184189 Table 1 Proximate composition of BR, WR and RB. Index Content Moisture content BR WR RB 10.16 1.26 10.23 1.04 11.20 1.41 Ash 1.24 0.08 0.45 0.02 7.93 1.52 Starch 72.44 3.21 76.70 3.04 18.74 1.45 Protein 9.06 1.28 7.38 1.61 15.41 1.87 Fat 2.42 0.48 0.27 0.03 19.93 1.93 Crude ber 1.12 0.04 0.06 0.01 13.67 1.64

185

Soluble protein 1.04 0.03 0.63 0.02 4.73 0.51

Unit is g/100 g dry matter. The results are expressed as mean SD with three replications. BR, brown rice; WR, white rice; RB, rice bran.

three proteins coming from the same rice varieties. In this study, WRP, RBP and brown rice protein (BRP) were all extracted from the same paddy rice variety (Wuyunjing NO.2). Their sodium dodecylsulfate -polyacrylamide gel electrophoresis (SDS-PAGE) characteristics and solubility, foaming and emulsifying properties in different systems were investigated. 2. Materials and methods 2.1. Grain materials Paddy rice (Oryza sativa L.), var. Wuyunjing NO.2, was purchased from Jintan City in Jiangsu Province, P. R. China, and stored in a refrigerator at 4  C until use. After dehusking, the brown rice (BR) was polished to a 10% degree of milling (JLMZJ, Shanghai, China) and the rice bran (RB) was collected. BR and white rice (WR) were ground into our by an auto-pulverizer (FSD100A, Zhejiang, China). BR our, WR our, and RB were then screened to pass through an 80-mesh sieve. Fractions that passed through the sieve were defatted using four volumes of petroleum ether for 4 h in a Soxhlet apparatus. The defatted samples were dried at 30  C for 24 h prior to further treatments. 2.2. Determination of proximate composition of rice samples Moisture, total protein, crude fat, ash, crude ber and starch content were estimated by standard AOAC Methods (AOAC, 1990). Nitrogen conversion factor for protein was 5.95. Soluble protein was determined by the Bradford procedure (Bradford, 1976) using Coomassie Brilliant G-250 dye binding and bovine serum albumin as the standard. Proximate composition of three rice proteins was determined and their results are presented in Table 1. 2.3. Isolation of protein fractions Albumin, globulin, glutelin, and prolamin fractions in rice proteins were extracted from defatted powders based on solubility at room temperature (25 1  C) in a solvent of 0.5 mol/L NaCl, 0.1 mol/L NaOH and 70% ethanol, by the procedure of Agboola et al. (2005). In order to obtain the highest yield of protein, each extraction step was repeated twice. Albumin, globulin, glutelin and prolamin fractions were then puried by isoelectric precipitation at pH 4.1, 4.3, 4.8, and 4.6, respectively. They were then washed with distilled water and the pH value was adjusted to 7.0. The N content of each fraction was determined by the Kjeldahl method (AOAC, 1990). 2.4. Preparation of rice protein

After an adjustment of pH to 7.0, the samples were freeze-dried and stored at 4  C until use. 2.5. Determination of rice protein molecular weights SDS-PAGE analysis of the proteins was made in the Laemmli buffer system (Laemmli, 1970) under reducing conditions (addition of 2.5% [v/v] b- mercaptoethanol ME), and samples were separated on a pre-cast gel (420% gradient) in a Bio-Rad Mini PROTEANw3 system (Bio-Rad Laboratories, Hercules, CA, USA). Protein bands were stained with Coomassie Brilliant Blue R-250. Molecularweight markers (14.494 kDa) were also run on the same gel to estimate the molecular weights of the proteins in the sample lanes in the stained gels. 2.6. Determination of protein functional properties The rice protein functional properties were determined using the procedure of Bera and Mukherjee (1989); Gurpreet and Sogi (2007) and Kakali et al. (2008) with slight modications below. Solubility: Each protein sample was dispersed in deionized water at a concentration of 1% protein and adjusted to pH 310 with 0.1 mol/L HCl or 0.1 mol/L NaOH, magnetically stirred at ambient temperature for 30 min, then centrifuged at 10 000 g for 10 min. Protein content of the supernatants were determined by the Bradford process and calculated protein solubility. Water/fat absorption: Samples (0.5 g) were mixed with distilled water or rened peanut oil (5.0 mL) and the mixtures were centrifuged at 1200 g for 25 min. The pellet was allowed to drain for 30 min and the gain in volume per unit weight was recorded as water or oil absorption capacity. Foaming properties: Protein dispersions (1%) were prepared at various pH values (3.011.0), NaCl concentrations (0.42.0%, pH 7.0) or sucrose concentrations (4.020.0%, pH 7.0). The dispersions were then whipped in a mechanical homogenizer (Kinematic PT1200E, Swiss) at 10 000 rpm for 1 min. The volumes before and after whipping were recorded and the foam capacity calculated. The measurement of the foam volume at 30 min represented the foam stability. Emulsifying properties: Protein solutions (5%) were prepared at various pH values (3.011.0), NaCl concentrations (0.42.0%, pH 7.0) or sucrose concentrations (4.020.0%, pH 7.0). These were homogenized in a mechanical homogenizer at 10 000 rpm for 1 min. An equal quantity of rened peanut oil was added and the solutions were homogenized again at 20 000 rpm for 1 min, then centrifuged at 1200 g for 5 min. Emulsifying properties could then be deduced. The emulsifying stability was found by measuring the emulsifying volume after incubating at 80  C for 10 min and centrifuging at 1200 g for 5 min. 2.7. Statistical analysis

Three defatted samples were extracted with 8-fold (v/w) solvent (distilled water with 0.1 mol/L NaOH and HCl, adjusted to pH 8.5) at 30  C for 2 h, and then centrifuged at 4000 g for 20 min. The supernatants were precipitated by isoelectric point precipitation.

Statistical analyses of the variance were performed with the Statistical Analysis System software 8.2 (SAS, USA). All trials were carried out in triplicate and all data were reported as means SD

186 Table 2 Content of major fractions of rice proteins. Index Protein content Albumin BR WR RB 9.73 1.02 6.24 0.63 42.71 2.47 Globulin 7.48 1.17 5.98 0.32 12.5 1.84

X. Cao et al. / Journal of Cereal Science 50 (2009) 184189

Glutenin 74.95 4.04 78.76 3.78 40.25 2.55

Prolamin 5.53 0.62 6.91 1.00 3.24 0.07

Unit is g/100 g protein. The results are expressed as mean SD with three replications. BR, brown rice; WR, white rice; RB, rice bran.

(standard deviation). Differences among means were evaluated using Duncans multiple range tests at a signicance level of P < 0.05. 3. Results and discussion 3.1. Protein fractions of three rice materials Table 2 shows the protein content of the rice protein fractions. The major fraction (approximately 80% of the total) in BR and WR was glutelin, but this protein comprised only 42% of the protein in rice bran. The albumin fraction was the major fraction in RBP and its content was 3.39 and 5.84 times higher than that in BR and WR, respectively. The globulin fraction comprised 12.51% of RBP, 7.48% of BRP and 5.96% of WRP. Prolamin was a minor component of BRP (3%) and RBP (5%), but was the 2nd component with its content over 7% of the protein in WR. 3.2. Molecular-weight distribution of total protein in rice SDS-PAGE patterns of total protein in BR, WR and RB are shown in Fig. 1. The major polypeptides appeared in ve areas, i.e., >80 kDa, 6080 kDa, 4060 kDa, 2040 kDa, and <20 kDa (peak I V in Fig. 1). The densitometry measurements also showed that RB had the highest protein content among three materials. Proteins of molecular weight over 80 kDa (peak I) were higher in quantity in WRP than in BRP and were hardly existent in RBP. The molecular weight of peak II had a range of 4060 kDa and these proteins were notably more abundant in RBP than in BRP and WRP. Peak III and IV included those with a molecular weight of 2040 kDa and 14 20 kDa, which comprised 36.6%, 27.2% and 38.5%, and 33.3%, 44.1% and 25.6% of BRP, WRP and RBP, respectively. Proteins in peak V, with molecular weights below 14 kDa, were virtually absent in BRP
Fig. 2. Solubility of rice proteins at different pH values. BR, brown rice; WR, white rice; RB, rice bran. Data are shown as means. Error bars represent the SD.

and WRP, but comprised 17.3% of RBP. Borght et al. (2006) reported that rice albumins have major polypeptides with MWs of 18 20 kDa, globulins with major polypeptides appeared at 15, 25.5, 200 kDa and higher. Prolamins consisted of three polypeptide subunits with apparent MWs of 10, 13 and 16 kDa, and rice glutelin was composed of acidic subunits (3039 kDa) and basic subunits (1925 kDa) which came from a 57 kDa polypeptide precursor. Two subunits were covalently linked to each other by an intermolecular disulde bond resulting in glutelin molecules with MWs ranging from 64 to 500 kDa. The above major bands were clearly observed in lanes of BR, WR or RB in Fig. 1. 3.3. Rice protein properties 3.3.1. Solubility The protein solubility increased with an increase in alkalinity and acidity, as shown in Fig. 2. The rice proteins had an isoelectric point of pH 4.5, which is the main reason for their lower solubility at pH 5. Protein solubility in aqueous solutions is dependent on pH value. Under isoelectric pH, the electrostatic repulsion and ionic hydration are minimum and hydrophobic interactions between surface nonpolar patches are maximum (Damodaran and Paraf, 1997). RBP solubility was higher than that of BRP and WRP in pH

Fig. 1. SDS-PAGE pattern and scanning prole of protein fractions of three rice proteins. SDS-PAGE analysis of total proteins of BR, WR and RB. Total rice proteins extracted from equal amounts of powder from each type were separated by SDS-PAGE using a 12.5% acrylamide gel and stained with Coomassie Blue. MW, molecular mass marker in kilodaltons. Peak IV was relative content of subunit in different MW ranges. BR, brown rice; WR, white rice; RB, rice bran.

X. Cao et al. / Journal of Cereal Science 50 (2009) 184189

187

47. When the pH value was over 7 or below 4, WRP had a higher (P < 0.05) solubility than BRP or RBP. Glutelin was the main component of protein in both WRP and BRP, and it had a high MW protein composed of subunits bound by disulde linkages and is soluble only in dilute acid or alkali (Ilankovan et al., 2007). The solubility of three proteins all increased substantially under acidic (pH < 4) and alkaline (pH > 7) conditions. Alkaline and acidic solutions could accelerate rice protein denaturation and hydrolysis and therefore enhanced solubility (Wang et al., 1999). 3.3.2. Water/fat absorption BRP and WRP possessed a low and similar (p > 0.05) water binding ability (1.96 and 1.78 mL/g, respectively), followed by RBP (3.54 mL/g). RBP had the highest (3.83 mL/g) and WRP had the lowest (2.56 mL/g) oil absorption capacity which was 2.93 mL/g in BRP. Functional properties of protein may be related to protein solubility. In this experiment, the water absorption capacity was 1.78 mL/g and 3.54 mL/g in WRP and RBP, respectively, which is lower than that reported for soy protein (Campbell et al., 1992). Aletor et al. (2002) considered that water absorption capacity values ranging from 1.49 to 4.72 g/g could be used in viscous foods. The data from the present study indicated that rice protein, especially RBP, possessed good capacity for water absorption and could be used in products requiring high water retention. The oil

absorption of RBP was also the best among three rice proteins and was higher than casein and soy protein isolates (Gurpreet and Sogi, 2007). High oil absorption is essential in the formulation of many processed foods. It was a necessary condition that RBP consisted of more soluble protein components with low MWs. 3.3.3. Foaming properties BRP, WRP and RBP, although they were derived from the same rice, the different protein components had different properties. Three types of rice proteins possessed the lowest foaming capacity at pH 5 (Fig. 3). The change trend of three rice proteins in various pH solutions accordingly paralleled their solubility under the same conditions. Several functional properties of proteins were affected by protein solubility. As the pH was increased from 5 to 11, foaming capacity of BRP, WRP and RBP increased 3.65-, 5.52- and 4.21-fold, respectively. Agboola et al. (2005) obtained the same result using Australia brown rice glutelins extracted without or with enzymatic pre-processing. Among three protein types, RBP had the highest foam producing capacity under all of the tested salt and sucrose concentrations in this study. It could form an excellent base for high sugar food processing. With an increase in salt concentration from 0.4% to 2.0%, the foaming of RBP was enhanced signicantly (P < 0.01). However, for BRP and WRP, the foam only signicantly increased (P < 0.05) at low salt concentrations. When the salt concentration was over 0.8%, foaming capacity remained constant.

Fig. 3. Effects of pH, salt and sugar on foaming capacity and foaming stability of rice proteins. A & D show the changes of foaming capacity and stability of rice protein in different pH solution. B & E show the change of foaming capacity and stability of rice protein in salt solution. C & F show the change of foaming capacity and stability of rice protein in sugar solution. The rice protein solutions without NaCl and sucrose are regarded as the control. The control of foaming capacity in BR, WR and RB are 70.3%, 66.0% and 71.7%, respectively. The control of foaming stability in BR, WR and RB are 89.7%, 83.6% and 92.2%, respectively. Data are shown as means. Error bars represent the SD.

188

X. Cao et al. / Journal of Cereal Science 50 (2009) 184189

Fig. 4. Effects of pH, salt and sugar on emulsication capacity and emulsifying stability of rice proteins. A & D show the changes of emulsifying capacity and stability of rice protein in different pH solution. B & E show the change of emulsifying capacity and stability of rice protein in salt solution. C & F show the change of emulsifying capacity and stability of rice protein in sugar solution. The rice proteins solutions without NaCl and sucrose are regarded as the control. The control of emulsifying capacity in BR, WR and RB are 51.2%, 49.6% and 56.5%, respectively. The control of emulsifying stability in BR, WR and RB are 46.3%, 40.8% and 49.1%, respectively. Data are shown as means. Error bars represent the SD.

At a low ionic strength (<0.2 mol/L), salt increased the hydration capacity of protein (Damodaran and Paraf, 1997), and an increase in hydration capacity essentially came from the hydration shells of the bound ions. A low concentration of Na also improved formation and stabilization of the hydration shells. It had been reported to enhance the foaming of soy protein and egg protein (Bera and Mukherjee, 1989). The foam stability declined slowly as the pH value increased, but it rose as NaCl concentration increased. With increase in sucrose concentrations from 4.0% to 12.0%, foaming capacity of three proteins was unaffected. When sucrose concentration was over 12.0%, the foaming capacity signicantly declined (P < 0.05). Sucrose has a favorable solvency in water. It would contest H2O molecules with protein so that the foaming ability was decreased. However, the specic gravity of sucrose solution rose with an increase of sucrose concentration; a signicantly high (P < 0.05) foam stability was seen for three rice proteins with high sugar concentrations. 3.3.4. Emulsifying properties Three types of rice proteins were studied for their emulsifying properties under different conditions of pH, salt and sugar concentrations. Their emulsifying behavior and stability are shown in Fig. 4. The emulsifying properties of proteins were inuenced by charge adsorption and lm formation at the interface. Nakai (1983) indicated that protein solubility and its surface hydrophobicity were very important properties for determining protein-emulsifying capacity. WRP had the highest emulsifying capacity at pH 3

11 and was, on average, 18.52% higher than that of RBP. The emulsifying capacity of the protein preparations was at pH 35, and increased between pH 511. BRP, WRP and RBP produced a maximum emulsifying volume of 44.32%, 47.06% and 42.93%, respectively, at pH 11. Under this condition, the changes in their emulsion stability reected their emulsifying capacity. These results suggested that WRP (mainly glutelin) were better emulsiers than the other two in strong acid or alkali conditions probably because disulphide bonding was broken and protein solubility was improved. However, emulsifying capacity decreased signicantly (P < 0.05) with an increase in salt concentration. NaCl reduced charge adsorption at interfaces. The low hydrophobicity of protein would not facilitate the interaction between proteins and oils, resulting in the decrease of emulsifying properties (Wang et al., 1999). For all protein samples, the maximum emulsication was observed at 4% sugar (Fig. 4) and then decreased slowly with further increases in sugar content. RBP showed the highest emulsication (48.2752.94%) of three preparations in both salt and sugar solutions. Nakai (1983) reported that the emulsifying properties not only depended on the protein solubility but also on hydrophiliclipophilic balance of the particular protein. The protein with a higher solubility and smaller molecular size might facilitate the diffusion and spread at oilwater interfaces (Wu et al., 1998). The higher content of hydrophobic amino acids in RBP than in WRP and BRP improved its surface hydrophobicity (Kang and Wang, 2006). The exposed hydrophobic groups enhanced the interactions

X. Cao et al. / Journal of Cereal Science 50 (2009) 184189

189

between proteins and lipids. Moreover, available literature also showed that the foam stability and emulsifying power of rice bran protein was higher than that of casein at different pH, salt or sugar conditions (Gurpreet and Sogi, 2007). 4. Conclusions WRP and BRP have a lower solubility than RBP in neutral and weak acidic medium. A high pH enhances nitrogen solubility, thereby, considerably improves protein functional properties. Solubility is the primary determinant of emulsication and foaming properties. In salt solution, nitrogen solubility is reduced, which also altered the properties of foaming and emulsication. Three proteins have a good foam capacity, foaming and emulsifying stability in NaCl systems. A high sucrose concentration leads to a decrease in rice protein foaming and emulsifying properties and a signicant increase in emulsifying stability. RBP expresses satisfactory functional properties among three rice proteins. The rice proteins not only serve as basic nutritional supplements but are also suitable for a broad range of industrial food applications. They can be used for bakery goods, whipped toppings, and sausages etc. owing to their high water and oil binding capacities, which helped to reduce moisture loss and maintain soft mouth feel. They, especially RBP, can form an excellent base for the high sugar food systems like cake batters, frozen desserts and confections. Appendix. Supplementary information Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.jcs.2009.04.009. References
AOAC, 1990. Ofcial Methods of Analysis, 15th ed. Association of Ofcial Analytical Chemists, Washington, DC. Aletor, O., Oshodi, A.A., Ipinmoroti, K., 2002. Chemical composition of common leafy vegetables and functional properties of their leaf protein concentrates. Food Chemistry 78, 6368. Agboola, S., Ng, D., Mills, D., 2005. Characterisation and functional properties of Australian rice protein isolates. Journal of Cereal Science 41, 283290. Bera, M.B., Mukherjee, R.K., 1989. Solubility, emulsifying and foaming properties of rice bran protein concentrates. Journal of Food Science 54 (1), 142145. Borght, A.V.D., Vandeputte, G.E., Derycke, V., Brijs, K., Daenen, G., Delcour, J.A., 2006. Extractability and chromatographic separation of rice endosperm proteins. Journal of Cereal Science 44, 6874.

Bradford, M.M., 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Analytical Biochemistry 72, 248254. Campbell, N.F., Shih, F.F., Marshall, W.E., 1992. Enzymatic phosphorylation of soy protein isolate for improved functional properties. Journal of Agricultural and Food Chemistry 40, 403406. Damodaran, S., Paraf, A., 1997. Food Proteins and Their Applications. Marcel Dekker Inc., New York. Gorinstein, S., Pawelzlk, E., Licon, E.D., Haruenkit, R., Weisz, M., Trakhtenberg, S., 2002. Characterisation of pseudocereal and cereal proteins by protein and amino acid analyses. Journal of the Science of Food and Agriculture 82, 886891. Gurpreet, K.C., Sogi, D.S., 2007. Functional properties of rice bran protein concentrates. Journal of Food Engineering 79, 592597. Ilankovan, P., Hettiarachchy, N.S., Christian, S., Markus, I.B., 2007. Hydrophobicity, solubility, and emulsifying properties of enzyme-modied rice endosperm protein. Cereal Chemistry 84 (4), 343349. Kakali, B., Gautam, M., Santinath, G., 2008. Preparation and characterisation of protein hydrolysates from Indian defatted rice bran meal. Journal of Oleo Science 57 (1), 4752. Kang, Y.L., Wang, Z.C., 2006. Research status of rice bran protein. Journal of Cereal & Oils 3, 2224 (in Chinese). Kim, S.Y., Park, P.S.W., Rhee, K.C., 1990. Functional properties of proteolytic enzyme modication soy protein isolate. Journal of Agricultural and Food Chemistry 38, 651656. Laemmli, U.K., 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680685. Marshall, W.E., Wadsworth, J.I. (Eds.), 1994. Rice Science and Technology. Marcel Dekker Inc., New York. Nakai, S., 1983. Structure function relationship of food protein with emphasis on the importance of protein hydrophobicity. Journal of Agricultural and Food Chemistry 31 (4), 676683. Panyam, D., Kilara, A., 1996. Enhancing the functionality of food proteins by enzymatic modication. Trends in Food Science and Technology 7, 120125. Rangel, A., Domont, G.B., Pedrosa, C., Ferriera, S.T., 2003. Functional properties of puried vicilins from cowpea (Vigna unguiculata) and Pea (Pisum sativum) and cowpea protein isolate. Journal of Agricultural and Food Chemistry 51, 57925797. Shih, F.F., Daigle, K.W., 1997. Use of enzymes for the separation of protein from rice our. Cereal Chemistry 74, 437441. Shih, F.F., Daigle, K.W., 2000. Preparation and characterization of rice protein isolates. Journal of the American Oil Chemists Society 77 (8), 885889. Sogi, D.S., Garg, S.K., Bawa, A.S., 2002. Functional properties of seed meals and protein concentrates from tomato processing waste. Journal of Food Science 67, 29973001. Tang, S., Hettiarachchy, N.S., Eswaranandam, S., Crandall, P., 2003. Protein extraction from heat-stabilized defatted rice bran: II. the role of amylase, celluclast, and viscozyme. Food Chemistry and Toxicology 68 (2), 471475. Tomotake, H., Shimaoka, I., Kayashita, J., Nakajoh, M., Kato, N., 2002. Physicochemical and functional properties of buckwheat protein product. Journal of Agricultural and Food Chemistry 50, 21252129. Wang, M., Hettiarachchy, N.S., Qi, M., Burks, W., Siebenmogen, T., 1999. Preparation and functional properties of rice bran protein isolate. Journal of Agricultural and Food Chemistry 47, 411416. Wu, W.U., Hettiarachchy, N.S., Qi, M., 1998. Hydrophobicity, solubility, and emulsifying properties of soy protein peptides prepared by papain modication and ultraltration. Journal of the American Oil Chemists Society 75 (7), 845850.

Journal of Cereal Science 50 (2009) 190197

Contents lists available at ScienceDirect

Journal of Cereal Science


journal homepage: www.elsevier.com/locate/jcs

Role of oxidative cleavage and acid hydrolysis of oat beta-glucan in modelled beverage conditions
R. Kivela*, L. Nystrom, H. Salovaara, T. Sontag-Strohm
Department of Food Technology, P.O. Box 66, FIN-00014, University of Helsinki, Finland

a r t i c l e i n f o
Article history: Received 21 January 2009 Received in revised form 24 April 2009 Accepted 29 April 2009 Keywords: Beta-glucan Degradation Ascorbic acid Oxidative cleavage

a b s t r a c t
The effects of organic acids (ascorbic, citric and malic acids) associated with beverages were studied in an unpuried oat beta-glucan extract with the purpose of examining the stability of cereal beta-glucan in beverage conditions. Addition of ascorbic acid caused an immediate decrease in viscosity of the extract and the MW of beta-glucan. Citric and malic acid affected moderately and only after a heat treatement. This indicated a dominating role of ascorbic acid induced oxidative cleavage compared to the generally proposed acid hydrolysis of beta-glucan. The nature of ascorbic acid induced cleavage was studied with inhibitors (glucose, mannitol, catalase and phytic acid) and catalysts (Cu- and Fe-ions) of hydroxyl radical attacks. Glucose, mannitol and catalase inhibited and the metals effectively catalysed the viscosity decrease. These indicated that the degradation of beta-glucan in the ascorbic acid treated extract was induced by metal-catalysed hydroxyl radicals. Also, the beta-glucan extract used as a matrix lost its viscosity during storage (6  C) concurrently with MW decrease of beta-glucan. When added to the extract, mannitol, glucose and catalase each showed a slight stabilising trend and Fe2-ions caused an immediate decrease in viscosity. The oxidative cleavage appeared to be an important factor to consider in developing novel aqueous beta-glucan enriched products. 2009 Elsevier Ltd. All rights reserved.

1. Introduction Cereal beta-glucan ((1 / 3),(1 / 4)-b-D-glucan) is a major structural polysaccharide of the endospermic cell walls of oat and barley. Beta-glucan has a high water binding capacity, and an ability to impart high viscosity in solutions or form gels. The rheological properties correlate with product structure, and are closely related to claimed health benets of the oat bre (Braaten et al., 1994; Wood et al., 2000). Both the health claim of beta-glucan (FDA 1997) and increasing consumption of nutritionally rich snacks have stimulated the development of beta-glucan enriched shots and beverages. Food matrix is proposed to be a factor in the effectivity of health benets of beta-glucan, and the cholesterol lowering effect is shown to strengthen when beta-glucan is consumed as a drink (Kerckhoffs et al., 2003; Naumann et al., 2006). Overall stability of beta-glucan an enriched beverage relates to the stability of the macromolecules and the stability of the

Abbreviations: HPSEC, High pressure size exclusion chromatography; happ, apparent viscosity at a specic shear rate; MW, weight average molecular weight; CSLM, confocal scanning laser microscopy; H2O2, hydrogen peroxide; AH2, ascorbic acid. * Corresponding author. Tel.: 358 9 19158540; fax: 358 9 19158460. E-mail address: reetta.kivela@helsinki. (R. Kivela). 0733-5210/$ see front matter 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2009.04.012

suspension formed by the macromolecules. The molecular stability of beta-glucan is important, since the viscosity-related benets depend on the molecular weight. The molecular weight of betaglucan in oat ingredients has been extensively examined with HPSEC and the values reported are commonly in the range of 1000 3000 103 g/mol (Lazaridou and Biliaderis, 2007). Beta-glucan is known to degrade in food processes, and the MW of oat bran betaglucan is reported to decrease signicantly in a beverage manufacturing process (man et al., 2004). The degradation of beta-glucan in foods has been considered to originate from enzymatic or acid hydrolysis (Lazaridou and Biliaderis, 2007). Oat ingredients are typically heat-treated and beverages are commonly pasteurized in order to eliminate microbes and enzymatic activity. Since fruit juices are acidic, the acid hydrolysis is the mechanism generally considered to induce a degradation of beta-glucan in beverages. A total hydrolysis of beta-glucan by acid treatment requires a very low pH (pH 12) and high temperature (70  C) (Johansson et al., 2006). Partial hydrolysis is generally examined with different inorganic acids, such as hydrochloric, phosphoric, triuoroacetic and sulfuric acids in suitable heating conditions (Tosh et al., 2004; Vaikousi and Biliaderis, 2005). To our knowledge no clear evidence exists on the hydrolytic role of organic carboxylic acids present in juices and acidic foods. Recently, we have shown that ascorbic acid can depolymerise puried beta-glucan in

R. Kivela et al. / Journal of Cereal Science 50 (2009) 190197

191

an aqueous solution in the presence of ferrous ions (Kivela et al., 2009). The suggested mechanism for this was hydroxyl radicaldriven cleavage, initiated by the reduction potential of ascorbic acid. Ascorbic acid, the common antioxidative food additive, is associated with OH-radical production by the Fenton reaction (Eqn. (1)) in polysaccharide oxidation (Buettner and Jurkiewicz, 1996; Fry, 1998). The oxidative as well as the antioxidative nature of ascorbic acid is due to its excellent reducing capacity. In small amounts, ascorbic acid behaves as a pro-oxidant rather than an antioxidant. Ascorbic acid has an ability to initiate the oxidation reaction by reducing the catalytic metals (Eqn. (2)) and the dissolved oxygen (Eqn. (3)), but at higher concentrations it terminates the reaction chains by forming sufcient amounts of relatively inactive and hence protective ascorbate radicals as intermediates. The switch in the reactivity has been proposed to depend also upon the ratio of ascorbic acid and catalytic metals of the oxidative chain reactions (Buettner and Jurkiewicz, 1996). Other food-related reducing ` ` agents may also behave as pro-oxidants. Valles-Pamies et al. (1997) showed the degradation of starch polymers in the presence of each sulphite (a food additive), glutathione and ascorbic acid. Also, phenolic compounds including ferulic acid, an antioxidant and copassenger of cereal dietary bers, has a moderately high reduction potential and is evidenced to behave as a pro-oxidant in certain aqueous conditions (Wee et al., 2003). CuD =Fe2D DH2 O2 /OH DOHL D2H2 ODCu2D =Fe3D Fenton


(J.T. Baker, The Netherlands), citric and malic acid (anhydrous, Fluka BioChemika), D()-Glucose (BDH Laboratory Supplies, England), phytic acid (50% solution, ICN Biomedicals Inc., USA), catalase (from bovine liver 23 000 U/mg protein (SigmaAldrich Chemie GmbH, Germany), D()-mannitol, iron(II)sulfate heptahydrate (FeSO4$7H2O), copper(II)sulfate pentahydrate (CuSO4$5H2O), sodium hydroxide (NaOH) and hydrochloric acid (HCl) (E. Merck, Germany). 2.2. Preparation and characterisation of beta-glucan extract Oat bran (6 g) was extracted with puried water (100 ml, Millipore system) at 37  C for 30 min with shaking. The dispersion was centrifuged (10 min with 16 000 g) and the supernatant was collected and placed in a boiling water bath for 10 min to denature proteins. After boiling, the solution was centrifuged and the supernatant was used as the sample extract. Sodium azide (0.02%) was added to the sample to prevent microbial growth and enzyme activity during storage. The samples containing catalase were prepared without azide as it acts as a catalase inhibitor (Mueller et al., 1997). 2.2.1. Beta-glucan concentration Quantication of b-glucan was performed with an enzymatic method (Approved Method 32-23, AACC 2000) using a Megazyme kit BBG (Megazyme, Bray Business Park, Bray, Co. Wicklow, Ireland). Moisture content of the extracts was measured according to Approved Method 44-60 (AACC 2000) with minor modications (Degutyte-Fomins et al., 2002). The quantication was performed in triplicate from the oat bran extracts of all series prior to sample preparation. 2.2.2. Weight average molecular weight of beta-glucan Weight average molecular weights (MW) of stored extracts (1 day, 1, 2 and 3 weeks, ascorbic acid treated and ascorbic acid-Fe2 treated extract) were determined by size exclusion chromatography (HPSEC). The system consisted of Alliance 2690 separation module, M-474 uorescence detector (lex 410 nm\lem 430 nm) and Empower software. Three mHydrogel columns (2000, 500, and 250) connected in series were utilised at 60  C. The eluent was 50 mM NaOH, ow rate 0.5 ml/min and injection volume was 50 ml. For specic detection of beta-glucans, calcouor reagent (120 mg/l) was delivered through a T-union at a ow rate of 0.4 ml/min. Samples were diluted with 50 mM NaOH, containing 0.1% NaBH4, to a nal concentration of 100500 mg/l of beta-glucan. The calibration was based on the commercial beta-glucan standard (Megazyme, medium viscosity) and the calibration curve for the molecular weight determinations was based on a beta-glucan standard of known molecular weight, determined by dual angle laser light scattering. 2.2.3. Monosaccharide composition Beta-glucan extract was methanolysed for the monosaccharide analysis and analysed with a gas chromatographic method according to Rantanen et al. (2007) with similar instruments. The samples and standards were analysed in triplicate. 2.2.4. Protein concentration Nitrogen content of the extract was determined in triplicate using the Dumas combustion method, and reported as protein Nx6.25. 2.2.5. Reducing capacity The total reducing capacity of the extract was determined spectrophotometrically by the FolinCiocalteu method (Satue ym., 1995) with ferulic acid as standard. FolinCiocalteau reagent (1 ml,

(1) AH2 D2Cu


2D

=Fe

3D

/AD2H D2Cu =Fe


D D

2D

(2) (3)

AH2 DO2 /ADH2 O2

Like most other macromolecules, cell wall polysaccharides are also susceptible to degradation by active oxygen species including OH-radicals produced in the Fenton-type reaction (Fry et al., 2002). In the Fenton reaction, hydrogen peroxide (H2O2) is reduced by a reduced transition metal (Cu, Fe2, Zn). These radicals have a very short lifetime, which makes the reaction site-specic and highly dependent on the location of the catalytic metals. Beta-glucan as well as many other high molecular weight polysaccharides have been reported to form metal-complexes (Platt and Clydesdale, 1984). These beta-glucan associated metals can make beta-glucan readily susceptible to the oxidative cleavage, as the catalyst is located in the vicinity of the target molecule (Chevion, 1988): Polysacch: MenL1D DH2 O2 /Polysacch: MenD .OH DOHL D2H2 O Fenton Polysacch: MenD .OH /Degraded Polysaccharide (Me: Cu, n 2; Fe, n 3). (4)

The aim was to study 1) the structural stability of unpuried beta-glucan extract and a beverage model based on the extract, 2) the occurence of oxidative cleavage of beta-glucan in the extract, 3) the relevance of acid hydrolysis and oxidative cleavage of betaglucan in modelled beverage conditions and 4) the protective effect of pH and potential Fenton reaction inhibitors on the structural stability of the extract in the beverage conditions. 2. Materials and methods 2.1. Materials Oat bran (Oat Well 14%) was supplied by Swedish Oat Fibre (Vrobacka, Sweden). Other reagents were L() Ascorbic acid

192

R. Kivela et al. / Journal of Cereal Science 50 (2009) 190197

Table 1 Effect of the potential Fenton reaction inhibitors (PA phytic acid, glc glucose, mannitol and catalase) on the apparent viscosity (happ, at 10 s1) of beta-glucan extract with and without ascorbic acid (AH2, 2 mM) at different time points (tx) at pH 4.8 0.2, and on the apparent viscosity of untreated extract (pH 6.5). The average happ of each sample is compared to the happ of control at t0 (shown as percents).

happ(tx) (mPas)
x 2h 1 day 1 week 2 weeks 4 weeks Untr. 145 17 140 15 122 20 107 20 73 37 Control 142 11 (95%) 135 18 (91%) 107 22 (72%) 86 25 (58%) 64 32 (43%) PA 137 5 (92%) 124 10 (83%) 75 3 (50%) 61 8 (41%) 40 9 (27%) Glucose 135 4 (91%) 130 0.3 (87%) 106 4 (71%) 93 2 (63%) 70 3 (47%) Mannitol 130 0.2 (87%) 133 2 (89%) 105 3 (71%) 94 2 (63%) 76 5 (51%) Catalase 156 31 (105%) 143 28 (96%) 126 25 (85%) 108 33 (72%) 90 28 (61%) AH2 42 9 (28%) 6 0.4 (4%) 2 0.02 (2%) 3 0.2 (2%) n.d. PA AH2 18 6 (12%) 51 (3%) 2 0.04 (2%) 2 0.1 (1%) n.d. Glc AH2 139 13 (93%) 70 9 (47%) 15 11 (10%) 91 (6%) n.d. Mann. AH2 120 14 (80%) 62 7 (42%) 15 3 (10%) 12 3 (8%) n.d. Catal. AH2 123 29 (82%) 71 5 (48%) 42 (3%) 21 (2%) n.d.

diluted to 1:10) was added to the sample extract (0.2 ml) with 7.5% sodium carbonate (0.8 ml). After 30 min, absorbance of the solution was read at 765 nm. Triplicate analyse were carried out in the dark. 2.2.6. Mineral analysis Mineral content of the extract was analysed after wet-digestion (HNO3) using an inductively coupled plasma-mass spectrometer (ICP-MS) with external standard method. Phytic acid (myoinositolhexaphosphate) content was calculated by dividing phosphorous content by the proportion of phosphorous atoms in the myoinositol molecule (32.3%). 2.2.7. Beta-glucanase activity The beta-glucanase activity of the extracts was determined in a beta-glucanagar-matrix by treatment with congo-red dye, which stains high molecular weight beta-glucan. The matrix was prepared from a commercial beta-glucan (High viscosity barley, Megazyme International) by wetting it with ethanol and then reuxing with puried water at 80  C for 3 h. The concentration of dissolved beta-glucan was 0.7% (w/v). Agar (2%, w/v) was dissolved in the beta-glucan solution bringing the system to boil and then heating at 50  C for 10 min. Fresh or 2 or 4 weeks old extract, puried water (negative control) and licchenase solution (50 U/ ml, positive control) was added to the cooled gel-matrix. After 24 h reaction at room temperature, congo-red solution (1% w/v) was pipetted on the samples. After 1 h, the dye was rinsed with puried water and the betaglucanase activity of the samples was visualised, comparing with the color intensity of the negative (no color depletion) and positive (bleached red color) control.

samples were prepared in each series (overall number of replicates n 6). All the series included untreated extract with pH 6.5 and adjusted volume, and all other than pH series included control sample with pH 4.8 0.2 (Fig. 5 and Table 1) or 4.3 0.2 (Fig. 3). 2.4. Rheology Viscosity properties of the samples were characterised using a ThermoHaake RheoStress 600 rheometer (Thermo Electron GmbH, Dreieich, Germany). A ow curve was obtained using a cone and plate geometry (60 mm, 1 ) over a shear rate range of 0.3300 0.3 s1. All the rheological experiments were performed at 10  C. The equipment was controlled with standard oils during the study and the method-related variation was <15 mPa. 2.5. Confocal scanning laser microscopy (CSLM) The fresh extract of oat bran and 2 weeks stored (6  C) extract were each mixed with 20 ml/l calcouor (100 mM sodium carbonate buffer, pH 10) at a ratio 1:1 (V/V). The mixtures were pipetted onto microscope glass slides and imaged with confocal scanning laser microscopy (Leica TCS SP2 AOBS, Germany) with excitation wavelength 350 nm and emission 420 nm. Triplicate analyses were performed. 2.6. Statistical analysis The rheological measurements were carried out in three independent series and the statistical analysis was performed from the apparent viscosity values (happ) at 10 s1. Signicant differences between the samples were determined by using the analysis of variance and Duncans multiple range test. Results were processed by Excel 2003 (Microsoft Corporation, Washington, USA) and Statgraphics Plus 4.0 (Statistical Graphics Corporation, Maryland, USA). A condence level of 95% was required for differentiation of samples. 3. Results 3.1. Composition of the oat bran extract The aqueous oat bran extract contained 0.15 0.01% (w/v) beta-glucan and 0.07 0.01% (w/v) of protein; the dry matter content of beta-glucan was 30 1% and protein 14.1 0.3%. A small amount (<0.5% of d.m.) of xylan, arabinose and galactose were found after hydrolysis of the extract. The amount of reducing

2.3. Sample preparation Glucose, mannitol, ascorbic acid, citric acid and malic acid powders were added to the extract by mixing gently at 30  C for a minute. Phytic acid and catalase were added as solutions. Glucose, mannitol, phytic acid and catalase were added before ascorbic acid, when used simultaneously. After dissolving the reagents, the pH was adjusted to 4.8 0.2 with HCl or NaOH (in the carboxylic acid series pH was 4.4 0.4) The sample volumes were equalized with water and quickly cooled down to 6  C for viscosity measurements. Samples were rst analysed immediately after cooling (time zero). Samples were stored at 6  C for 2, 24 (1 day), 48, 72, 168 (1 week), 336 (2 weeks), 504 (3 weeks) and 672 (4 weeks) hours. Insoluble precipitate released during storage was removed before analysis. Each sample was analysed once in a series and the whole series wasrepeated three times. In the pH studies, duplicate

R. Kivela et al. / Journal of Cereal Science 50 (2009) 190197

193

compounds of the extract was 1.0 0.2% (d.m), when determined by the FolinCiocalteau method as ferulic acid equivalents. Phytic acid content of the extract was 6% (d.m.) when calculated from the determined total phosphorous content. No beta-glucan hydrolysing enzyme activity was observed in fresh, 2 or 4 week old extracts compared to the water and licchenase samples, when analysed in beta-glucan agar matrix with congo-red dying. During storage (at 6  C), precipitate formation and consistent decrease in turbulence of the extract were observed. The insoluble precipitate of 2 week old extract contained 51% proteins and 26% phytic acid. The formation of a similar precipitate was obtained in treated extracts with simultaneous viscosity decrease. The concentration of reducing compounds in the extract did not change during 2 week storage. 3.2. Viscosity of the stored extract The viscosity of the beta-glucan extract decreased during the storage (6  C). Zero shear viscosity decreased approximately 2%, 15%, 25% and 40% after 1 day, 1, 2 and 3 week storage, respectively (Fig. 1a). The ow curves of only one series is shown in Fig. 1a, but signicant variation between the three independent series (standard deviation of 1%, 10%, 20% and 25% at time points 1 day, 1 week, 2 and 3 weeks, respectively) needs to be noted. The average MW of beta-glucan was 1400 103 g/mol (polydispersion index pd 5.6), 1300 g/mol (pd 5.6), 1100 103 g/mol (pd 5.1) and 1000 103 (pd 5.0) g/mol in the 1 day, 1, 2 and 3 week old extracts analysed in 50 mM M NaOH solution. Plotting log(MW*c), where c is concentration (constant), against loghapp at one shear rate (30 s1) resulted in a linear curve (R2 0.9565), which was very similar to the one based on the data and curve presented by Wood et al. (2000) (Fig. 1b). The fresh and 2 weeks stored extract were stained with calcouor and visualised by CSLM for studying the aggregate formation during the storage. An aggregate and cluster formation was observed, which appeared to grow with storage time (Fig. 2a and b). 3.3. Citric, malic and ascorbic acids Carboxylic acids that commonly occur and are added in beverages were studied for their effects on beta-glucan stability. Addition of ascorbic acid (4 mM, pH 4.7 0.1) decreased the happ (10 s1) of the extract approximately 74% in 2 h, while citric acid (4 mM, pH 4.1 0.1) and malic acid (4 mM, pH 4.3 0.2) did not affect the viscosity (Fig. 3). Neither did the happ of citric and malic acid samples differ from the viscosities of untreated (pH 6.5 0.1) and control (4 mM, pH 4.3 0.2 adjusted with HCl) samples after 1 day, but the viscosity of the ascorbic acid treated sample decreased over 95% (Fig. 3). The MW of beta-glucan decreased from 1400 103 g/mol to 50 103 g/mol (Pd 2) in 1 day in the ascorbic acid treated extract. Role of pasteurizing was studied with a heat treatment (90  C, 5 min). The heat treatment accelerated the viscosity decrease (2 h) in the ascorbic acid treated extract from 74% to 90%. The viscosities were decreased also in other samples after the heat treatment so that in untreated extract the happ was decreased after 2 h approximately 6%, 13% in control and malic acid treated samples and 27% in citric acid treated sample (Fig. 3). The viscosity decrease of heattreated citric acid sample was signicantly higher than the viscosity decrease of the malic acid, control and untreated extract samples. Concentration of ascorbic acid affected the rate of the viscosity decrease of the extract at pH 4.8 0.2. The concentration of 2 mM, 4 mM and 10 mM of ascorbic acid induced a 30 8%, 75 6% and

a 1000
Viscosity (mPas)
1 day 1 week 2 weeks 3 weeks 100

10 0.1

10

100

1000

Shear rate (1/s)

b
Log (viscosity at 30 1/s)

2 1.9 1.8 1.7 1.6 1.5 1.4 1.3 1.2 1.1 1 5.1 5.2 5.3 5.4 5.5 y = 2.6014x - 12.322 R2 = 0.9674 y = 2.0781x - 9.1208 R2 = 0.9565

Log (Mw*c)
Fig. 1. The viscosity of untreated beta-glucan extract (pH 6.5) after 2 h, 1, 2 and 3 week storage (6  C) (a). Plot showing the relationship between log of apparent viscosity (30 s1) of the untreated extract, and log of concentration (c, % w/v) times weight average molecular weight (MW, g/mol) of beta-glucan. The grey curve is adapted from the data of Wood et al. (2000) (b).

90 1% decrease of the happ in 2 h when compared to the pH adjusted control. 3.4. pH Effect of six different pH conditions (2.5, 3.5, 4.5, 5.5, 6.6 and 7.5) was studied in the extract containing ascorbic acid (2 mM, 35 mg/100 ml). The different pH conditions did not affect the rate of viscosity decrease signicantly in 2 h (Fig. 4). After 1 day storage, the viscosity of the extract was signicantly higher at pH 2.5 and 7.5 compared to pH 3.5, 4.5, 5.5 and 6.5. However, after 1 week, at all pH conditions, less than 5% of the initial viscosity remained (Fig. 4). Without ascorbic acid, pH had no signicant effect on the happ of the extract. 3.5. Ferrous and cupric ions The metals related to the Fenton reaction as catalysts were studied in respect to the extract viscosity with and without 2 mM ascorbic acid (Fig. 5). Ferrous ions (Fe2) alone without ascorbic acid decreased the happ immediately by 50%, while cupric ions (Cu2) had no signicant effect on the viscosity. When the metalions were each added together with 2 mM ascorbic acid, the viscosity of the Fe2-sample decreased immediately by 60% compared to 50% without ascorbic acid, and with Cu2 ascorbic acid, caused over 95% immediate decrease in the viscosity (Fig. 5). The MW of beta-glucan decreased (in 1 day) from 1400 103 g/mol to 150 103 g/mol (pd 2.3) in the sample containing ferrous ions

194

R. Kivela et al. / Journal of Cereal Science 50 (2009) 190197

Fig. 2. CSLM images of a typical aggregate in a fresh (A) and two weeks old (B) beta-glucan extract. Image sizes are 85 mm 85 mm.

and was <30 103 g/mol after 1 day when both Fe2-ions and ascorbic acid were added. 3.6. Potential Fenton reaction inhibitors The role of phytic acid as an inhibitor of the oxidative cleavage was studied due to its evidenced antioxidant activity by its metalchelating ability. No inhibition effect, but rather a signicant catalysing effect on the spontaneous viscosity decrease of the extract was obtained after phytic acid addition. After 1 day, 83% of the viscosity was sustained in the phytic acid (1.5 mM) treated extract compared to 95% in the case of the control sample (Table 1). Addition of phytic acid also accelerated the viscosity decrease of the ascorbic acid treated extract (Table 1). The phytic acid solution used for sample preparation was analysed to contain 0.002 mM Fe as an impurity. Glucose and mannitol were studied as potential Fenton reaction inhibitors, since they can inhibit the attack of OH-radicals. Addition of glucose and mannitol showed a slight stabilising effect on the viscosity of the extract when compared to the control, although the effect was not statistically signicant (Table 1). The ascorbic acid induced viscosity decrease of the extract signicantly slowed
160 140 120 100 80 60 40 20 0 d h a f a

down with glucose and mannitol addition (Table 1). Glucose was added in common concentration of juices and snack products (0.5 M, 9 g/100 ml) and mannitol at the same concentration. Catalase is an enzyme, which represents the Fenton reaction inhibitor by its ability to decompose the reagent, H2O2. Catalase (3000 U/ml) could moderately slow down the viscosity decrease of the extract, but the inhibition was not statistically signicant (Table 1). The viscosity decreased signicantly slower when ascorbic acid was added to the extract after catalase addition than without catalase treatment (Table 1). 4. Discussion 4.1. Stability of the beta-glucan extract The viscosity of the beta-glucan extract (0.15% of beta-glucan) decreased spontaneously during storage (Fig. 1a), suggesting instability of the beta-glucan molecule. Weight average molecular weight (MW) also showed a decreasing trend, and log(happ) as a function of log(MW*c) was linear (Fig. 1b) indicating that the viscosity decrease depends on the decreasing MW (c 0.15%, constant). The result is consistent with the results of Wood et al.

Apparent viscosity at 10 1/s (mPas)

f a f a f b g b g c 0 hour g b g 2 hours 1 day

e
t) t.) d (p as ac i .) .) .)

5)

3)

id

id

(p as

st

H 6.

ac i

4.

ac

ac

(p a

st (p a id ac

tro l( pH

C itr ic

ic

(p

ic

co rb

ed

al

ro l

ed

nt re

on

As

on t

at

ac al ic M

at

id

(p a As c.

Fig. 3. The effect of carboxylic acids (4 mM, pH 4.14.7) with or without pasterurization (past.) on the viscosity of oat bran extract in storage. Different letters within one time point denote a statistically signicant difference (p < 0.05).

C itr ic

nt re

st

R. Kivela et al. / Journal of Cereal Science 50 (2009) 190197

195

Apparent Viscosity at 10 1/s (mPas)

100 %

a b b

App. viscosity (sample/untreated extract)

90 % 80 % 70 % 60 % 50 % 40 % 30 % 20 % 10 % 0% 2.5 f c

0 hour 1 day 2 weeks b b

170 150 130 110 90 70 50 30 10 -10 -10 90 190 290

Control Cu2+ Fe2+

AH2 AH2+Fe2+ AH2+Cu2+

d 3.5

d 4.5

d f d 5.5 f 6.5 f 7.5 f

Fig. 4. The effect of pH conditions (2.5, 3.5, 4.5, 5.5, 6.5 and 7.5) on the rate of viscosity decrease of the oat bran extract in presence of ascorbic acid (2 mM). Different letters within one time point denote a statistically signicant difference (p < 0.05).

390

490

Time (hours)
(2000) as shown in Fig. 1b. The MW-determinations were performed with NaOH as eluent and sample solvent to release molecules from aggregates (Li et al., 2006). However, the differences between the MW values (10001400 103 g/mol) in the four samples were small and may be in the exclusion limit of the HPSECmethod. The precipitate released from the extract matrix with decreasing viscosity was analysed to form mostly of proteins (51%) and phytic acid (24%). Temelli et al. (2004) observed precipitate formation in barley beta-glucan beverages during storage and they suggested that the precipitate consisted of protein. However, the precipitate was shown here to consist not only of proteins, but also of phytic acid and minerals in signicant amounts. The role of oxidative cleavage induced by endogenous minor compounds in the storage stability of the extract was studied with the effect of the potential Fenton reaction inhibitors and catalysts. Catalase decomposes H2O2, which usually occurs in natural solutions, and thus can inhibit the OH-radical formation (Fry, 1998) Catalase seemed to stabilise the extract so that after four weeks, 61% of the viscosity remained while 43% remained in the control sample (Table 1). However, due to the variation between series, the effect was not statistically signicant. Also, glucose and mannitol altered moderately the viscosity decrease of the extract, though not signicantly (Table 1). Polyols and simple sugars are known to alter the kinetics of the OH-radical attacks (Li et al., 1996; Morelli et al., 2003), and are known to be needed in relatively high concentration for effective inhibition (typical for competitive substrate mechanism). They can effectively inhibit metal-mediated OH-radicals, but not the radicals induced by radiation (metal-complexation mechanism), though both inhibition mechanisms have been suggested to occur (Li et al., 1996). In the present study, mannitol altered the viscosity decrease more effectively than glucose (Table 1). Generally, mannitol was shown to inhibit the OH-radical attacks more effectively than glucose with an unknown mechanism (Morelli et al., 2003). The inhibition effect of catalase, glucose and mannitol on the spontaneous viscosity reduction was moderate and not necessarily signicant, but it may indicate the occurrence of metaland H2O2-mediated OH-radical production and oxidative cleavage of beta-glucan in the untreated extract. The added phytic acid rather catalysed the viscosity decrease of the extract, though inhibition was expected. After 4 weeks, 72% of the initial viscosity was lost compared to 57% for the control sample (Table 1). Phytic acid is associated with cereal bre and is an effective metal chelator, which can bind the catalysing metals covalently (without any coordination sites for water) and thus keep them in an oxidized state. Due to the excellent metal binding capacity, phytic
Fig. 5. The effect of ferrous and cupric ions on the storage viscosity of the beta-glucan extract with and without ascorbic acid (AH2).

acid has an antioxidant activity in the Fenton-like oxidation mechanism (Graf et al.,1987). However, the antioxidant capacity of metalchelators is controversial, and it has, for example, been shown that a common metal-chelator EDTA catalyses the Fenton/HaberWeiss reactions, while DETAPAC, another metal-chelator, effectively inhibits the reactions (Buettner et al., 1983). The catalysing effect of EDTA, which is explained by its water coordination site, was also obtained in the present study as the viscosity decrease of the oat bran extract (data not shown). Burkitt and Gilbert (1990) showed that in the presence of phytate-Fe2 -complex, oxygen radicals are formed at a low but signicant level and in the presence of phytateFe3-complex at a very low level. They concluded that the Fenton reaction is actually catalysed by phytate, but the Haber Weiss cycle (which includes two reactions: the reduction of iron or copper with superoxide and the Fenton reaction) as a whole is blocked by this chelator, creating discrepancies in the results for the antioxidative capacity of the phytic acid in natural systems. According to the results of Burkitt and Gilbert (1990), the moderate catalysing effect of phytic acid observed in the present study indicates the presence of endogenous catalytic metals and other Fenton reagents. These may initiate the OH-radical production and affect the viscosity by beta-glucan degradation. The effects of adding metal ions to the extract support the idea of endogenous H2O2 occurring in the extract. The catalytic metals in different states, copper in its oxidized (Cu2) and iron in its reduced (Fe2) form, were studied. The OH-radical formation requires transition metals to be in the reduced form. However, in the presence of ascorbic acid the oxidized state may also be reduced and thus act as a catalyst. Both metals catalysed the beta-glucan degradation in the presence of ascorbic acid (Fig. 5), copper being more effective. The more effective catalysing capacity of copper has been shown earlier by Buettner and Jurkiewicz (1996). However, when the Fe2- and Cu2-ions were added solely to the beta-glucan extract, only reduced ferrous ions (Fe2) had an effect and oxidized cupric ions (Cu2) did not. The addition of Fe2-ions immediately and clearly decreased the viscosity of the extract and the trend was similiar to the ascorbic acid treated extract, while oxidizied Cu2-ions had only a moderate, insignicant effect on the viscosity (Fig. 5). Since in solution of puried beta-glucan, the addition of ionic iron did not signicantly affect the viscosity (Kivela at al., 2009), the decrease indicates that the extract contains minor compounds, which have potential to initiate oxidative cleavage in the presence of added free catalyst. The extract was analysed to have reduction potential, but

196

R. Kivela et al. / Journal of Cereal Science 50 (2009) 190197

obviously the potential was not high enough to reduce the oxidized state of copper (E (Cu2/Cu) 0.159) to initiate the OH-radical production. Thus it is likely that the extract contains H2O2, which can initiate the oxidative degradation with Fe2-ions and the reactions may accelerate in the presence of the free catalyst (Eqn.(1)). In addition to molecular degradation, aggregation and further occulation can also affect the viscosity by packing molecules and thus interfering with the entangled structure and water binding capacity of the polymers. Beta-glucan is shown to form micellelike aggregates in water (Grimm et al., 1995; Vrum et al., 1992; Wu et al., 2006). In this study, an increase in the size of the aggregate or cluster of aggregates with time was observed (Fig. 2a and b). The average diameter of the clusters was 510 mm in the fresh extract and 3060 mm in two weeks old extract with a betaglucan concentration of 0.1%. The two week old clusters seemed to have a dense core, with diameter about 1015 mm. The aggregation behaviour of beta-glucan in aqueous environment and the observed cluster formation is consistent with the results of Wu et al. (2006), who reported over 3 mm lengths of oat betaglucan aggregates in a 0.01% solution of beta-glucan with MW 440 000 g/mol. The aggregation in this study may explain the time dependent increase in variance of the viscosities of untreated extract between the different series, since homogenity of the sample solutions decreased. The occulation behaviour of oat beta-glucan needs to be studied properly in natural solutions in further investigations. 4.2. Stability of beta-glucan in modelled beverage conditions The addition of ascorbic acid caused a drastic viscosity reduction of the extract viscosity, whereas the addition of malic acid or citric acid did not affect the viscosity of the beta-glucan extract in similar conditions (Fig. 3). The MW of beta-glucan in the ascorbic acid treated extract decreased from 1400 103 to 50 103 g/mol in a day, thus indicating depolymerisation. Ascorbic acid is known to participate in oxidative cleavage by reducing dissolved oxygen, and catalysing metals to be reagents in Fenton-type reactions (Buettner and Jurkiewicz, 1996). The ascorbic acid induced viscosity decrease of beta-glucan has been studied previously with puried beta-glucan (Kivela et al., 2009). Beta-glucan was shown to depolymerise in these conditions as well, but an exogenic metal catalyst was needed to initiate the reaction in the solution of puried beta-glucan. Further, the decrease in viscosity was slower in the solution of puried betaglucan when compared to the extracted beta-glucan used in this study. Addition of 10 mM ascorbic acid reduced the viscosity 65% immediately and 90% after 2 h in the extract, whereas it was 15% and 80% in the pure solution, respectively. Beta-glucan is known to associate with metals (Platt and Clydesdale, 1984), and the complexes may be dissociated during the purication processes. The metal-complexes which likely occur in the extract may target beta-glucan more effectively to the oxidative cleavage, as the radicals are formed in its immediate vicinity. When a pasteurization process was applied, citric acid and malic acid also caused a viscosity decrease in the extract, citric acid being more effective than malic acid (Fig. 3). The effect of malic acid was equal to the effect of the HCl-treated control sample. The equation of Fig. 1b gives an estimation of MW, when the happ(30 s1) and concentration are known. Using the equation for the pasteurized citric and malic acid treated samples, estimated MW is 1100 103 g/ mol and 1200 103 g/mol, respectively, compared to the 1400 103 g/ mol of the untreated extract. This means that pasteurization at 90  C for 5 min caused slight, approximately 21% decrease in the MW in the citric acid and 14% in the malic acid and HCl-treated samples. This may indicate that a rapid pasteurization process could alsoinduce acid

hydrolysis in the presence of common organic acids, which has been proposed by man et al. (2004) in the case of pasteurized (90  C/ 5 min) apple juice based oat bran beverage. Vaikousi and Biliaderis (2005) found, using phosphoric acid, that the viscosity of beta-glucan solution started to decrease at pH 44.5 after approximately 30 min at 82.5  C by acid hydrolysis. In the present study, the heat treatment caused a slight viscosity decrease also in the untreated (pH 6.5) extract. This suggests that other mechanisms, such as the oxidative cleavage or changes in the physical state of beta-glucan polymers or their aggregates may also contribute to the observed heat treatment induced viscosity decrease. The more signicant effect of citric acid on the viscosity decrease may originate from the slightly lower pH of the citric acid sample. Citric acid has also been shown to catalyse OHradical production and hence the oxidative cleavage by metalcomplexation in biological systems (Burkitt and Gilbert, 1990). 4.3. Effect of pH and potential Fenton reaction inhibitors on the stability of the extract in presence of ascorbic acid Different pH conditions were tested to show the effect of pH on extract viscosity and its stability. The viscosity loss was greatest in pH 3.5, 4.5 5.5 and 6.5, whereas in low pH (2.5) and in high pH (7.5) the ascorbic acid induced viscosity loss appeared to be less prominent (Fig. 4). The results are consistent with those of Fry (1998), who used pure xyloglucan and found a maximum at pH 4.7 and the reaction rate decreasing in order of pH 4.5 > 5.5 > 3.5 [ 6.6 [ 7.8. They explained the maximum by a higher reducing power of the monoanion form of ascorbic acid (pKa 4.2). In the present study, ascorbic acid had a similar effect in the pH range 3.56.5, which may be due not only to the different polysaccharide but also to impurities present in our extract. In any case, at each pH studied, the viscosity was lost in a week, and it was concluded that pH adjustment in a beverage product is not a suitable technological method to protect beta-glucan from oxidative cleavage. The effects of potential inhibitors of the Fenton-type oxidative cleavage, namely phytic acid, catalase, glucose and mannitol, were also studied in the presence of asorbic acid. When phytic acid was added to the oat bran extract (also containing endogenous metals and phytate), the ascorbic acid induced viscosity decrease was catalysed rather than inhibited (Table 1). This is partially consistent with Lee and Hendricks (1997) who observed total inhibition in lipid peroxidation by 0.5 and 5 mM phytic acid in the presence of 0.1 mM Fe3 only when the ascorbic acid concentration was lower than 1 mM. The addition of catalase with ascorbic acid decreased the rate of beta-glucan degradation (Table 1). At this level, catalase was not able protect beta-glucan, but the experiment showed the role of H2O2 in the ascorbic acid induced degradation of beta-glucan. Glucose, when used in a concentration typical of sweetened juices, and mannitol, also had a retarding effect on the viscosity decrease of the ascorbic acid containing extract. The inhibition effect was however signicantly less effective than found with solutions of puried polysaccharides at similar concentrations of glucose and mannitol (Fry, 1998; Kivela et al., 2009; Schweikert et al., 2000). The difference may originate from the metalbeta-glucan complexes of the natural extract, which can prevent both the competitive inhibition process and the possible weak metalglucose/mannitol-complexation. In this study, we have demonstrated that a beta-glucan extract in the presence of ascorbic acid (2 mM, 35 mg/100 g) lost its initial viscosity with storage time of 1 h at 6  C. At the same time there was a clear decrease in molecular weight of betaglucan. Citric and malic acid decreased the viscosity of the extract only moderately after a heat treatment. Inhibitors of the oxidative cleavage (glucose, mannitol and catalase) slightly inhibited and

R. Kivela et al. / Journal of Cereal Science 50 (2009) 190197

197

metal ions (Cu2 and Fe2) catalysed the action of ascorbic acid in the extract. These ndings indicated that OH-radical driven oxidative cleavage is the major factor causing beta-glucan degradation in beverages and acid hydrolysis may have a minor role. The viscosity of the matrix used was not stable during the storage. The viscosity decrease appeared to be linear with the decrease of MW of beta-glucan and the possible occurence of oxidative cleavage was studied with the inhibitors. Mannitol, glucose and catalase showed a stabilising trend and phytic acid could moderately accelerate the viscosity decrease. Thus Fenton-type oxidative cleavage of beta-glucan might occur also in the untreated extract. The Fenton-reaction catalyst, Fe2-ions, caused a clear decrease in the viscosity when added solely to the untreated extract. This suggested the occurrence of the endogenous reactions catalysed by the free Fe2-ions. The observed aggregation/occulation of beta-glucan during storage may also alter the storage stability of the extract. In conclusion, the use of ascorbic acid and mineral fortication favor the oxidative cleavage of beta-glucan, and they may be a serious threat for the molecular stability of beta-glucan in aqueous environment. Control of the oxidative cleavage may be a challenge in natural aqueous systems such as beverages and shots. However, oxidative cleavage could also offer a cheap and easy way to fractionate and concentrate beta-glucan for modifying its functional properties in aqueous foods. Acknowledgements The authors are thankful to D.Sc. (Tech.) Tapani Suortti (VTT, Finland) for performing the molecular weight measurement and data analysis, and to the Finnish Ministry of Agriculture and Forestry and the University of Helsinki Funds for the nance. Also technician Outi Brinck is thanked for her contribution in the labo ratory work quality and M.Sc. Paivikki Timonen for her observations in the study eld. References
Braaten, J.T., Wood, P.J., Scott, F.W., Wolynetz, M.S., Lowe, M.K., Bradley-White, P., Collins, M.W., 1994. Oat beta-glucan reduces blood cholesterol concentration in hypercholesterolemic subjects. European Journal of Clinical Nutrition 48, 465474. Buettner, G.R., Doherty, T.P., Patterson, L.K., 1983. The kinetics of the reaction of superoxide radical with Fe(II1)complexes of EDTA, DETAPAC and HEDTA. FEBS Letters 158, 134136. Buettner, G.R., Jurkiewicz, B.A., 1996. Catalytic metals, ascorbate and free radicals: combinations to avoid. Radiation Research 145, 532541. Burkitt, M.J., Gilbert, B.C., 1990. Model studies of the iron-catalysed HaberWeiss cycle and the ascorbate-driven Fenton reaction. Free Radical Research 10, 265 280. Chevion, M., 1988. A site-specic mechanism for free radical induced biological damage: the essential role of redox-active transition metals. Free Radical Biology 5, 2737. Degutyte-Fomins, L., Tuula Sontag-Strohm, T., Salovaara, H., 2002. Oat bran Fermentation by rye Sourdough. Cereal Chemistry 79, 345348. Fry, S.C., 1998. Oxidative scission of plant cell wall polysaccharides by ascorbateinduced hydroxyl radicals. Biochemical Journal 332, 507515. Fry, S.C., Miller, J.G., Dumville, J.C., 2002. A proposed role for copper ions in cell wall loosening. Plant and Soil 247, 5761.

Graf, E., Empson, K.L., Eaton, J.W., 1987. Phytic acid. A natural antioxidant. The Journal of Biological Chemistry 262, 1164711650. Grimm, A., Kruger, E., Burchard, W., 1995. Solution properties of b-D-(1, 3) (1, 4)glucan isolated from beer. Carbohydrate Polymers 27, 205214. Johansson, L., Virkki, L., Anttila, H., Esselstrom, H., Tuomainen, P., Sontag-Strohm, T., 2006. Hydrolysis of b-glucan. Food Chemistry 97, 7179. Kerckhoffs, D., Hornstra, G., Mensink, R.P., 2003. Cholesterol-lowering effect of betaglucan from oat bran in mildly hypercholesterolemic subjects may decrease when beta-glucan is incorporated into bread and cookies. The American Journal of Clinical Nutrition 78, 221227. Kivela, R., Gates, F., Sontag-Strohm, T., 2009. Degradation of cereal beta-glucan by ascorbic acid induced oxygen radicals. Journal of Cereal Science 49, 13. Lazaridou, A., Biliaderis, C.G., 2007. Molecular aspects of cereal b-glucan functionality: physical properties, technological applications and physiological effects. Journal of Cereal Science 46, 101119. Lee, B.J., Hendricks, B.J., 1997. Metal-catalyzed oxidation of ascorbate, deoxyribose and linoleic acid as affected by phytic acid in a model system. Journal of Food Science 62, 935938. Li, S., Patapoff, T.W., Nguyen, T.H., Borchardt, R.T., 1996. Inhibitory effect of sugars and polyols on the metal-catalyzed oxidation of human relaxin. Journal of Pharmaceutical Sciences 85, 868872. Li, W., Wang, Q., Cui, S.W., Huang, X., Kakuda, Y., 2006. Elimination of aggregates of (1 / 3) (1 / 4)-b-D-glucan in dilute solutions for light scattering and size exclusion chromatography study. Food Hydrocolloids 20, 361368. Morelli, R., Russo-Volpe, S., Bruno, N., Lo Scalzo, R., 2003. Fenton-dependent damage to carbohydrates: free radical scavenging activity of some simple sugars. Journal of Agricultural and Food Chemistry 51, 74187423. Mueller, S., Riedel, H.-D., Stremmel, W., 1997. Determination of catalase activity at physiological hydrogen peroxide concentrations. Analytical Biochemistry 245, 5560. Naumann, E., van Rees, A.B., Onning, G., Oste, R., Wydra, M., Mensink, R.P., 2006. Beta-glucan incorporated into a fruit drink effectively lowers serum LDLcholesterol concentrations. The American Journal of Clinical Nutrition 83, 601605. Platt, S.R., Clydesdale, F.M., 1984. Binding of iron by cellulose, lignin, sodium phytate and beta-glucan, alone and in combination, under simulatedgastrointestinal pH conditions. Journal of Food Science 49, 531535. Rantanen, H., Virkki, L., Tuomainen, P., Kabel, M., Schols, H., Tenkanen, M., 2007. Preparation of arabinoxylobiose from rye xylan using family 10 Aspergillus aculeatus endo-1,4-b-d-xylanase. Carbohydrate Polymers 68, 350359. Schweikert, C., Liszkay, A., Schopfer, P., 2000. Scission of polysaccharides by peroxidase-generated hydroxyl radicals. Phytochemistry 53, 565569. Temelli, F., Bansema, C., Stobbe, K., 2004. Development of an Orange-Flavored barley beta-glucan beverage. Cereal Chemistry 81, 499503. Tosh, S.M., Wood, P.J., Wang, Q., Weisz, J., 2004. Structural characteristics and rheological properties of partially hydrolyzed oat b-glucan: the effects of molecular weight and hydrolysis method. Carbohydrate Polymers 55, 425436. Vaikousi, H., Biliaderis, C.G., 2005. Processing and formulation effects on rheological behavior of barley b-glucan aqueous dispersions. Food Chemistry 91, 505516. ` ` Valles-Pamies, B., Barclay, F., Hill, S.E., Mitchell, J.R., Paterson, L.A., Blanshard, J.M.V., 1997. The effects of low molecular weight additives on the viscosities of cassava starch. Carbohydrate Polymers 34, 3138. Vrum, K.M., Smidsrod, O., Brant, D.A., 1992. Light scattering reveals micelle-like aggregation in the (1/3),(1/4)-b-D-glucans from oat aleurone. Food Hydrocolloids 5, 497511. Wee, L.M., Long, L.H., Whiteman, M., Halliwell, B., 2003. Factors affecting the ascorbate- and phenolic-dependent generation of hydrogen peroxide in dulbeccos modied eagles medium. Free Radical Research 37, 11231130. Wood, P.J., Beer, M.U., Butler, G., 2000. Evaluation of role of concentration and molecular weight of oat beta-glucan in determining effect of viscosity on plasma glucose and insulin following an oral glucose load. The British Journal of Nutrition 84, 1923. Wu, J., Zhang, Y., Xie, B, Wang, H., Deng, S., 2006. Visualization of single and aggregated hulless oat (Avena nuda L.) (1 / 3), (1 / 4)-b-d-glucan molecules by atomic force microscopy and confocal scanning laser microscopy. 54, 925934. man, P., Rimsten, L., Andersson, R., 2004. Molecular weight distribution of bglucan in oat-based foods. Cereal Chemistry 81, 356360.

Journal of Cereal Science 50 (2009) 198204

Contents lists available at ScienceDirect

Journal of Cereal Science


journal homepage: www.elsevier.com/locate/jcs

Effect of particle size on kinetics of starch digestion in milled barley and sorghum grains by porcine alpha-amylase
Ghaid J.S. Al-Rabadi, Robert G. Gilbert, Michael J. Gidley*
Centre for Nutrition and Food Sciences, School of Land Crop & Food Sciences, The University of Queensland, St Lucia, Brisbane, Qld 4072, Australia

a r t i c l e i n f o
Article history: Received 19 February 2009 Received in revised form 14 May 2009 Accepted 15 May 2009 Keywords: Starch digestion alpha-Amylase Grain particle size Diffusion

a b s t r a c t
The inuence of milled grain particle size on the kinetics of enzymatic starch digestion was examined. Two types of cereals (barley and sorghum) were ground, and the resulting grounds separated by size using sieving, with sizes ranging from w0.1 to w3 mm. In vitro enzymatic digestion was performed, using pancreatic alpha-amylase, amyloglucosidase and protease, to determine fractional-digestion rates over 24 h. The resulting glucose production rate data were well tted by simple rst-order kinetics. For each sieve screen size, the digestion rate of barley was always higher than that of sorghum. The rate coefcients for digestion showed a decrease with increasing size, and could be well tted by an inverse square relationship. This is consistent with the supposition that starch digestion in these systems is controlled by diffusion of enzyme through the grain fragment. Apparent diffusion coefcients of alphaamylase obtained by tting the size dependence were 0.76 (sorghum) and 1.7 (barley) 107 cm2 s1, 9 (sorghum) and 4 (barley) times slower than predicted for a molecule of the size of alpha-amylase in water. 2009 Elsevier Ltd. All rights reserved.

1. Introduction Enzymatic digestion of starch can be affected by many factors such as granule structure (Chiotelli and Le Meste, 2002), the presence of lipids (Baldwin et al., 1997), the presence of protein (Williamson et al., 1992), the presence of minerals (Escarpa et al., 1997), amylose to amylopectin ratio (Sajilata et al., 2006), digestion conditions (Boisen and Fernandez, 1997) and particle size (Blasel et al., 2006). The effect of particle size on starch digestion has typically been related to the available surface area for enzymatic action by alpha-amylase. For example, for each 100 mm increase in particle size in ground maize grain, the degree of starch access by alpha-amylase was found to decrease by 26.8 g/kg starch (Blasel

Abbreviations: A, proportionality constant in dependence of digestion rate coefcient on particle size; Ci, starch fraction digested at a given time in the ith sieve screen size; dgw, geometric mean diameter; DMSO, dimethyl sulfoxide; D, diffusion coefcient; ki, rst-order rate coefcient for fractional digestion (subscript b for barley, s for sorghum); kB, Boltzmann constant; ln, natural logarithm; R, radius of the diffusing species; r2, coefcient of correlation; rpm, revolutions per minute; SEM, standard error of the mean; Sgw, geometric standard deviation; Si, ith screen sieve size; Si,av, the average particle size for the ith and (i 1)th sieves; T, absolute temperature; t, incubation or digestion time; h, viscosity of the medium. * Corresponding author. Tel.: 61 7 3365 2145; fax: 61 7 3365 1177. E-mail address: m.gidley@uq.edu.au (M.J. Gidley). 0733-5210/$ see front matter 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2009.05.001

et al., 2006). Several studies have been conducted to investigate and to characterize the kinetics of starch digestion for different ground grains by alpha-amylase (Goni et al., 1997; Kong et al., 2003; Weurding et al., 2001, 2003; Wilfart et al., 2007, 2008). However, these did not characterize the kinetics of starch digestion for ground grain fragments after segregation by size which will be the rst purpose of this study. Microscopic examination has been used to suggest that starch digestion by alpha-amylase is controlled partially by diffusion processes (Helbert et al., 1996). Information that quanties and relates the effect of starch origin, exemplied here with barley and sorghum, on alpha-amylase diffusion within grain fragments will be the second purpose of this study. Previous studies of protein and other factors responsible for characteristic digestion properties of e.g. sorghum (Ezeogu et al., 2004; Zhang and Hamaker, 1998) point to major grain-specic effects controlling alpha-amylase activity. The importance of quantifying the diffusion rates and consequent kinetics of starch digestion by alpha-amylase, is that this determines the extent of starch digestion within a certain time. This has practical applications in estimating glucose ux that is considered as an important food characteristic in human nutrition (inuencing glycemic and insulinemic responses), maximizing animal production performance, and for industrial applications such as fermentation for ethanol production. The data for digestion rate as a function of grain size will be treated with the simplest approach: assuming rst-order kinetics

G.J.S. Al-Rabadi et al. / Journal of Cereal Science 50 (2009) 198204

199

(Goni et al., 1997), without invoking any mechanistic assumptions. It will be seen that this approach in fact provides a good t for the data. First-order kinetic ts to digestion data have hitherto been performed for barley (Weurding et al., 2001) and for sorghum (Weurding et al., 2001; Wilfart et al., 2008) but these were for samples without size differentiation. In the present work, rstorder ts to digestion of particles as a function of their size will yield size-dependent rate coefcients for this process. This size dependence will then be used to make mechanistic inferences that are quantitatively consistent with these observations. 2. Experimental 2.1. Materials 2.1.1. Cereal grinding Barley (Binalong) and sorghum (Buster) were obtained from the Queensland Department of Primary Industry and Fisheries, Australia. Cereals were milled using a hammer mill (Australian Agriculture Machinery Group, Australia) at 1140 rpm using a 4 mm screen size for each cereal type. Grinding was conducted in duplicate at the feed processing facility of the Queensland Department of Primary Industry and Fisheries (Alexandra Hill, Brisbane, Australia). Ground barley and sorghum were collected at a constant motor load during milling. Moisture content was determined by drying the ground material in a hot oven at 135  C for 3 h. The ground cereals were segregated by size using vertical multiple sieving under gravity with mechanical agitation using a sieve shaker (Endecotts shaker, ExTech Pty. Ltd., Victoria, Australia). Eight screen sieve (Endecotts Ltd, London, England) sizes were selected to give a broad spectrum of particle size ranging from 4.76 mm to 0.045 mm (pan) as shown in Table 1. 2.2. Methods 2.2.1. Sieving analysis Ground barley and sorghum were sieved and the geometric mean diameter (dgw) and geometric standard deviation (Sgw) for each cereal type were determined with 100 g sample according to

(ASAE, 2003) using a sieve shaker for 29 min. The sieved material so obtained, for each sieve fraction size, was kept at 4  C until further use. The selected sieves, fraction yield on each sieve after sieving and the average particle sizes for ground barley and sorghum are listed in Table 1. 2.2.2. In vitro digestibility analysis The in vitro starch digestion method used was adapted, with modications, from the enzymatic method described by Muir et al. (1995) and Htoon et al. (2009). This method consists of a three-step enzymatic hydrolysis mimicking digestion in the mouth, in the stomach and the small intestine in a closed system, followed by the measurement of glucose after the following incubation times: 0, 0.5, 1, 2, 6 and 24 h. Zero hour digestion was started at the beginning of the small intestinal simulation step, as this is the main site of starch digestion and alpha-amylase activity. The enzymatic digestion was carried out in three successive steps. The rst step was an enzymatic hydrolysis with a solution of articial porcine salivary alpha-amylase (Sigma, reference A3176) with 5 mL carbonate buffer at pH 7 containing 10.87 mg/mL alphaamylase (enzyme activity of 25 units/mg). Approximately 0.5 g of grain sample was weighed to accuracy 0.01 mg in a 250 mL bottle. Each sample was mixed gently with salivary alpha-amylase solution at ambient temperature for 1015 s. This step was followed by enzymatic hydrolysis with 5 mL porcine pepsin (Sigma, reference P6887 (enzyme activity of 2490 units/mg)) solution at pH 2 using 0.02 M HCl and incubated in a shaking water bath (Rateck Instrument Pty Ltd, Borniva, Victoria, Australia) at 37  C and 85 rpm for 30 min. The third digestion step simulating small intestinal conditions was carried out using pancreatin (Sigma, P1750 (activity of 4 U.S.P specication)), a mixture of primarily protease and alpha-amylase, and amyloglucosidase from Aspergillus niger (Sigma, Reference A7420 (enzyme activity of 31.2 units/mg)). Following 30 min solubilisation of pepsin, 25 mL of sodium acetate buffer, containing 0.2% sodium azide, at pH 6.0 and 5 mL of 0.02 M NaOH was added. Sodium azide was used to prevent any microbial activity during the digestion periods. Five mL of pancreatin and amyloglucosidase solution (prepared by adding 2 mg pancreatin and 0.8975 mg amylogucosidase per mL acetate buffer containing 0.2% sodium azide) was added to the mixture, after which the bottles were placed in a 37  C water shaking bath (zero hour). Hydrolysis with pancreatin was carried out for six time periods (0, 0.5, 1, 2, 6 and 24 h). Blank samples were included at the second and third steps. At the end of each incubation period, bottles were incubated in ice for 5 min to stop enzyme activity. After centrifugation (for 10 min at 3000 g at 25  C), the glucose concentration in the supernatant was determined by using a glucose oxidase colorimetric analysis (Reference TR-1511-200 Thermo Electron Noble Park, Victoria, Australia). Colour absorption was measured at a wavelength of 505 nm (Pharmacia LKB-Ultrospec III, England) and glucose concentration was converted into starch content by applying the factor 162/180 (0.9) to correct for the loss of one water molecule (molecular weight 18) for every linkage of glucose molecule (molecular weight 180). The digestibility of each particle size level and enzyme incubation time was analysed in a randomised order and in duplicate. Starch digestibility was presented as grams per 100 g starch. 2.2.3. Total starch analysis Particle sizes above 0.5 mm were ground nely, using mortar and pestle, until the whole sample passed 0.5 mm sieve size. Subsequently, 50 mg of ground samples, controls (5 and 20% (w/w) regular maize starch (Sigma, reference S-5269) mixed with cellulose on a dry matter basis) and reagent blank (empty tubes) were wetted using 0.4 mL of 80% ethanol and then stirred on

Table 1 Geometric mean particle size diameter (dgw), geometric standard deviation (Sgw), fraction yield (SEM) and starch content (SEM) for material retained on each of the selected sieves. Item Grain Barley dgw (mm) Sgwa Sieve size (mm) 4.760 2.800 1.700 1.000 0.500 0.250 0.125 0.045 (pan) Total
a b

Sorghum 0.74 2.02 Fraction yield (g) (SEM) 0.0 1.3 6.6 29.1 38.6 16.2 5.2 1.9 0.0 0.0 0.1 0.2 0.4 1.0 0.4 0.3 Starch content (%) (SEM) c 59.0 65.7 65.3 57.0 57.8 63.9 65.0 0.6 1.0 1.6 0.6 0.7 0.4 0.4

1.09 2.09 Average particle size (mm) c 3.8 2.3 1.4 0.75 0.38 0.19 0.045 Fraction yield (g) (SEM) 0.0b 2.4 25.5 39.8 19.7 7.2 2.8 2.3 0.0 0.1 1.1 0.1 0.7 0.2 0.2 0.1 Starch content (%) (SEM) c 51.1 51.0 51.7 37.3 30.9 28.6 40.2 1.0 2.5 3.3 0.6 1.9 0.1 1.4

99.6 0.1

99.0 0.1

Log normal standard deviation (ASAE, 2003). Values are duplicate measurements of 100 g sample retained on top of sieves after 29 min shaking. c No material was retained on the 4.76 mm sieve.

200

G.J.S. Al-Rabadi et al. / Journal of Cereal Science 50 (2009) 198204

Fig. 1. Time dependence of fraction digestion (proportion of total starch) of in vitro starch digestion for different particle sizes for (a) barley and (b) sorghum (duplicate measurements for each time).

a vortex mixer. Immediately, 2 mL dimethyl sulfoxide (DMSO) was added, stirred on a vortex mixer and then placed in boiling water for 5 min. Immediately, 0.1 mL of thermostable alpha-amylase (enzyme activity 3000U/mL) from Bacillus licheniformis (Megazyme, Ireland, E-BLAAM) in 3 mL MOPS (3-(N-morpholino)propanesulfonic acid) buffer was added. Samples were vigorously stirred on a vortex mixer in a boiling water bath for 12 min (stirring in vortex every 3 min). The MOPS buffer had been previously prepared by adding 11.55 g MOPS (Sigma, reference M9381) to about 900 mL of water and then adjusting to pH 7.0 using 1 M HCl. Calcium chloride dihydrate (0.74 g) and sodium azide (0.2 g) were added and the nal volume was adjusted to 1 L and stored at room temperature. Samples (5 mL) were then incubated in a 50  C water bath and 4 mL of sodium acetate buffer (200 mM, pH 4.5) was added for each sample and 2 min allowed for thermal equilibration before adding 0.1 mL amyloglucosidase (enzyme activity of 3260 U/mL) (Megazyme E-AMGDP). Samples then were stirred in a vortex and incubated at 50  C for 30 min with gentle mixing at 10 min intervals. Starch content was measured in the same way as described for in vitro digestibility analysis. Starch content for each particle size, controls and reagent blank were analysed at random and in

duplicate. Starch content (standard error) for each particle size on each sieve screen size is listed in Table 1. 2.2.4. Data tting For the sake of simplicity, rst-order kinetics was used to investigate the kinetics of starch digestion under simulated small intestine conditions by alpha-amylase for different particle sizes of each cereal type. This is given by equation (1):

Ci 1 eki t

(1)

where Ci is the starch fraction digested at time t in the ith sieve screen size (and 1 Ci is the undigested fraction); ki is the fractional-digestion rate coefcient (h1), t is the incubation or digestion time (h). For the purpose of data tting, the value of ki was obtained by linear-least-squares t of the solution of eq. (1), viz., as the slope of a plot of ln(1 Ci) against t for each screen size. The linearity of such a plot will provide a measure of the applicability of rst-order kinetics. The average particle size for the ith sieve, Si,av, was taken as the average sizes of the two consecutive sieves (Si, Si1), as shown in equation (2):

G.J.S. Al-Rabadi et al. / Journal of Cereal Science 50 (2009) 198204

201

Fig. 2. Fit of rst-order kinetics of starch digestion at different average particle sizes for barley at 0.045 mm (A), 0.188 mm (B), 0.375 mm (C), 0.75 mm (D), 1.35 mm (E), 2.25 mm (F) and 3.87 mm (G). The slope (rate coefcient (kb)), intercept and coefcient of correlation (r2) are indicated.

Si;av

Si Si1 2

(2)

coefcients so obtained for each particle size are listed in Table 2. In all cases, the rate coefcients decrease with increasing sieve size. This observation was quantied by tting to a power law:

3. Results and discussion Fig. 1(a) and (b) shows digestibility data for barley and for sorghum, given as the fraction of total starch digested at different times. Figs. 2 and 3 show the results of tting all data with rst-order kinetics for barley and sorghum, respectively. All data show a good t to rst-order behaviour, with r2 values from 0.90 to 0.99 and from 0.87 to 0.99 for barley and sorghum, respectively. The rate

n ki A Si;av

(3)

for some exponent n. It was found that n 2 gave a good t to all data, as shown in Fig. 4 (the 95% condence intervals are also shown); no other integer or half-integer exponent could provide an adequate t to the data. The exponent n 2 suggests diffusion control, based on the relation between time and Fickian diffusion to an average distance from a point:

202

G.J.S. Al-Rabadi et al. / Journal of Cereal Science 50 (2009) 198204

Fig. 3. Fit of rst-order kinetics of starch digestion at different average particle sizes for sorghum at 0.045 mm (A), 0.188 mm (B), 0.375 mm (C), 0.75 mm (D), 1.35 mm (E), 2.25 mm (F), and 3.87 mm (G). The slope (rate coefcient (ks)), intercept and coefcient of correlation (r2) are indicated.

S2 6Dt i;av

(4)

where D is the diffusion coefcient and t the time at which diffusion has occurred over a mean-squared distance S2 . The observed i,av exponent is explained if this diffusion time is equated to the inverse of the rst-order rate coefcient:

ki

6D S2 i;av

(5)

This is admittedly a simplistic rationalization: solution of Ficks equation for diffusion of, for example, a penetrant (in this case an enzyme) entering the surface of a sphere (the particle) and diffusing into the interior shows a complex dependence of concentration on

time (e.g. Carslaw and Jaeger,1959; Crank,1979). However, the use of eq. (5) should be correct to within a factor of order unity. A related explanation is that the inverse square relationship between rate constant and particle size could be due to specic surface area effects (which also scale with the square of size for spherical particles). In this model, the rate-determining step would be the initial entry of the enzyme into the particle, rather than the proposed model which postulates that movement of the enzyme within the particle is rate determining. In the absence of any obvious barrier at the exterior of particles, we propose that diffusion within the particle is a more likely rate-determining step. Closer inspection of Table 2 and Fig. 4 shows that for both barley and sorghum, the fractional-digestion rate coefcient can be classied into two categories for different screen sieve sizes. Those in

G.J.S. Al-Rabadi et al. / Journal of Cereal Science 50 (2009) 198204 Table 2 Fractional-digestion rate coefcient (h1) (standard deviation) characterising the in vitro enzymatic digestion by alpha-amylase and amyloglucosidase of starch in barley and sorghum for different sieve size openings (mm) at six incubation time periods (0, 0.5, 1, 2, 6 and 24 h). Sieve opening (mm) 4.76 2.8 1.7 1.0 0.5 0.250 0.125 0.045
a

203

Average sieve size (mm) a 3.78 2.25 1.35 0.75 0.375 0.188 0.045

kb (Barley) a (4.3 (1.2 (4.0 (2.3 (3.9 (4.3 (4.7

ks (Sorghum) a (1.9 (7.5 (1.6 (2.9 (7.0 (1.1 (1.1

0.5) 0.7) 0.1) 0.1) 0.2) 0.3) 0.2)

103 102 102 101 101 101 101

0.2) 103 0.6) 103 0.1) 102 0.2) 102 0.3) 102 0.06) 101 0.1) 101

absence of cell wall barriers. Scanning electron microscopy analysis of the smallest fraction (<45 mm) shows that starch granules are essentially separated from cell wall fragments and protein bodies. Remaining intra-cellular barriers to starch digestion (e.g. granule surface structure and/or internal architecture) would then be the factors that contribute to the enzyme digestion rate, with sorghum having a slower overall digestion rate than barley for the varieties studied. The relationships between fractional-digestion rate coefcient and average size for large (second category) particle sizes for barley and sorghum (Fig. 4) were tted by equations (6) and (7):

2 1=kb 16:26 Si;av 4:15 2 1=ks 36:57 Si;av 0:36

barley sorghum
1

(6) (7)

No sieved material was present at this sieve size.

the rst category, for sieve sizes < 0.5 mm for barley and <0.25 mm for sorghum, exhibit high fractional-digestion rate coefcient values, and do not show a large variation with sieve size: eq. (5) does not provide a good t for these smallest sizes. The other category, for larger sieve sizes, showed a much larger dependence on sieve size: Eq. (5) provides a good t. Based on the size cut-off for this second category of behaviour (0.250.5 mm), we suggest that a major factor in enzyme diffusion is passage across endosperm cell walls. Individual endosperm cells are typically 0.050.1 mm; therefore starch granules in particles of >0.5 mm will be predominantly within intact cell walls whereas particles < 0.25 mm would be expected to have many broken cells with exposed intra-cellular contents. For these smaller particles, the rate of starch digestion is close to constant, but different for sorghum and barley. We suggest that this corresponds to the rates at which starch is digested in the

Here kb and ks are the fractional-digestion rate coefcients (h ) for barley and sorghum, respectively, and Si,av is in mm. These equations are based on eq. (5), with the non-zero intercept included to take into account the observed slight deviation from this relation for smaller sizes. The apparent diffusion coefcients obtained using these ts and eq. (5) are D 0.062 and 0.027 mm2 h1, or 1.7 107 and 0.76 107 cm2 s1, for barley and sorghum respectively. The lower diffusion coefcient in the sorghum sample studied compared to the barley sample studied could be attributed to differences in cell wall architecture, or to the fact that the protein matrix that surrounds starch granules in most cereals is a particularly effective barrier in sorghum (Hamaker et al., 1987; Kotarski et al., 1992; Rooney and Pugfelder, 1986) and may limit the mobility of alphaamylase. An alternative or additional reason is the presence of anti nutritional factors such as tannin (Bjorck and Nyman, 1987), which can bind to the alpha-amylase enzyme and reduce its mobility through the grain fragments. Lower fractional-digestion rate coefcients for ground, un-fractionated, sorghum compared to barley have been previously reported by Weurding et al. (2001). A diffusion coefcient can be related to the radius R of the diffusing species through the StokesEinstein relationship:

kB T 6phR

(8)

Fig. 4. Correlation between average particle size squared and fractional-digestion rate coefcient (as the reciprocal 1/k) for (a) barley and (b) sorghum. The 95% condence intervals of the reciprocals of the rate coefcients, obtained by conventional regression analysis, are also shown.

where kB is the Boltzmann constant, T the absolute temperature, and h the viscosity of the medium. The diffusion coefcients deduced above can be used with eq. (8) to estimate an apparent equivalent radius R of the diffusing species, using the bulk viscosity of water. The result is apparent radii of 12 nm for diffusion in a barley particle, and 27 nm for sorghum. These values are ca 4 and 9 times larger than the true radius of approximately 3 nm for porcine pancreatic alpha-amylase (Payan et al., 1980). This difference provides an expression of the relative barriers to alphaamylase movement in barley and sorghum. Within the same cereal type, the strong dependence of digestion time on particle size, as shown in Table 2, shows that the distribution of particle size after grinding is important in determining the rate of digestion of ground (non-fractionated) grain. This is especially important for coarse milling applications as used for e.g. animal feeds where a wide range of hammer mill screen sizes are used that would each result in wide distribution of particle sizes. This aspect will be developed further in a subsequent publication. From a nutritional point of view, overall conversion of starch into glucose energy is a result of two complementary processes: digestion rate and passage rate (Wilfart et al., 2007). The fractionaldigestion rate coefcient is important in this context because it governs glucose absorption (Weurding et al., 2001). This means for

204

G.J.S. Al-Rabadi et al. / Journal of Cereal Science 50 (2009) 198204 Blasel, H.M., Hoffman, P.C., Shaver, R.D., 2006. Degree of starch access: an enzymatic method to determine starch degradation potential of corn grain and corn silage. Animal Feed Science and Technology 128, 96107. Boisen, S., Fernandez, J.A., 1997. Prediction of the total tract digestibility of energy in feedstuffs and pig diets by in vitro analyses. Animal Feed Science and Technology 68, 277286. Carslaw, H.S., Jaeger, J.C., 1959. Conduction of Heat in Solids, second ed. Clarendon Press, Oxford. Chiotelli, E., Le Meste, M., 2002. Effect of small and large wheat starch granules on thermomechanical behavior of starch. Cereal Chemistry 79, 286293. Crank, J., 1979. The Mathematics of Diffusion. Oxford University Press, Oxford. Escarpa, A., Gonzalez, M.C., Morales, M.D., SauraCalixto, F., 1997. An approach to the inuence of nutrients and other food constituents on resistant starch formation. Food Chemistry 60, 527532. Ezeogu, L.I., Duodu, K.G., Taylor, J.R.N., 2004. Effects of endosperm texture and cooking conditions on the in vitro starch digestibility of sorghum and maize ours. Journal of Cereal Science 42, 3344. Goni, I., Garcia-Alonso, A., Saura-Calixto, F., 1997. A starch hydrolysis procedure to estimate glycemic index. Nutrition Research 17, 427437. Hamaker, B.R., Kirleis, A.W., Butler, L.G., Axtell, J.D., Mertz, E.T., 1987. Improving the in vitro protein digestibility of sorghum with reducing agents. Proceedings of the National Academy of Sciences of the United States of America 84, 626628. Helbert, W., Schulein, M., Henrissat, B., 1996. Electron microscopic investigation of the diffusion of Bacillus licheniformis a-amylase into corn starch granules. International Journal of Biological Macromolecules 19, 165169. Htoon, A., Shrestha, A.K., Flanagan, B.M., Lopez-Rubio, A., Bird, A.R., Gilbert, E.P., Gidley, M.J., 2009. Effects of processing high amylose maize starches under controlled conditions on structural organisation and amylase digestibility. Carbohydrate Polymers 75, 236245. Kong, B.W., Kim, J.I., Kim, M.J., Kim, J.C., 2003. Porcine pancreatic alpha-amylase hydrolysis of native starch granules as a function of granule surface area. Biotechnology Progress 19, 11621166. Kotarski, S.F., Waniska, R.D., Thurn, K.K., 1992. Starch hydrolysis by the ruminal microora. Journal of Nutrition 122, 178190. Muir, J.G., Birkett, A., Brown, I., Jones, G., Odea, K., 1995. Food processing and maize variety affects amount of starch escaping digestion in the small intestine. American Journal of Clinical Nutrition 61, 8289. Payan, F., Haser, R., Pierrot, M., Frey, M., Astier, J.P., 1980. The three-dimensional structure of alpha-amylase from porcine pancreas at 5 . Resolution of the active-site location. Acta Crystallogr Sect B 36, 416421. Rooney, L.W., Pugfelder, R.L., 1986. Factors affecting starch digestibility with special emphasis on sorghum and corn. Journal of Animal Science 63, 16071623. Sajilata, M.G., Singhal, R.S., Pushpa, R.K., 2006. Resistant starchda review. Comprehensive Reviews in Food Science and Food Safety 5, 117. Weurding, R.E., Enting, H., Verstegen, M.W.A., 2003. The relation between starch digestion rate and amino acid level for broiler chickens. Poultry Science 82, 279284. Weurding, R.E., Veldman, A., Veen, W.A.G., van der Aar, P.J., Verstegen, M.W.A., 2001. In vitro starch digestion correlates well with rate and extent of starch digestion in broiler chickens. Journal of Nutrition 131, 23362342. Wilfart, A., Jaguelin-Peyraud, Y., Simmins, H., Noblet, J., van Milgen, J., Montagne, L., 2007. A step-wise in vitro method to estimate kinetics of hydrolysis of feeds. Livestock Science 109, 179181. Wilfart, A., Jaguelin-Peyraud, Y., Simmins, H., Noblet, J., van Milgen, J., Montagne, L., 2008. Kinetics of enzymatic digestion of feeds as estimated by a stepwise in vitro method. Animal Feed Science and Technology 141, 171183. Williamson, G., Belshaw, N.J., Self, D.J., Noel, T.R., Ring, S.G., Cairns, P., Morris, V.J., Clark, S.A., Parker, M.L., 1992. Hydrolysis of A-type and B-type crystalline polymorphs of starch by alpha-amylase, beta-amylase and glucoamylase. Carbohydrate Polymers 18, 179187. Zhang, G., Hamaker, B.R., 1998. Low alpha-amylase starch digestibility of cooked sorghum ours and the effect of protein. Cereal Chemistry 75, 710713.

example that nutritive content may not be the only factor that determines growth and performance in production animals. For example, Weurding et al. (2003) reported that diets with similar amount of digestible nutrients but different digestion kinetics can result in different growth performance in broiler chickens. In humans, fractional-digestion rates can inuence both glycemic scores (Goni et al., 1997) and the fermentation level in the lower digestive tract; both have health impacts in human nutrition. The outcomes of this study suggest that the extent of starch digestion from uncooked ground grains in vivo can be to some extent controlled and predicted based on the milling size distribution, for both animal and human nutrition. In conclusion, enzyme digestion of barley and sorghum grain fragments of different sizes can be well tted by simple rst-order kinetics, as has been found previously for other starch-based systems (Goni et al., 1997). The fractional-digestion rate, as a function of the fragment size, suggests that digestion is controlled by diffusion of enzyme through the grain fragment. The extent of alpha-amylase diffusion through the grain fragment depends on the grain (barley vs. sorghum). These data showed that particle size and the distribution of particle size are a major factor in starch digestion from ground uncooked grains, with potentially major health and nutritional implications for both animals and humans. More generally, this study suggests that amylase activity in industrial applications can be manipulated (inhibited or accelerated) by controlling factors that affect amylase diffusion and its mobility in any system. Acknowledgements The authors thank the Queensland Department of Primary Industries and Fisheries (Alexandra Hill, Brisbane, Australia) for supplying grains and for assistance in milling. We also thank John Black, Peter Sopade, Barbara Williams, Wayne Bryden, Honest Madziva, Lesleigh Force, Sharon Nielsen, Deirdre Mikkelsen and Peter Torley for their technical advice. Funding was provided by the Co-operative Research Centre for an Internationally Competitive Pork Industry and the Australian Research Council (Discovery Project DP0985694). References
ASAE, 2003. Methods of Determining and Expressing Fineness of Feed Materials by Sieving. Standard no. S319.3. American Society of Agricultural and Biological Engineers, pp. 202205. Baldwin, P.M., Melia, C.D., Davies, M.C., 1997. The surface chemistry of starch granules studied by time-of-ight secondary ion mass spectrometry. Journal of Cereal Science 26, 329346. Bjorck, I.M., Nyman, M.E., 1987. In vitro effects of phytic acid and polyphenols on starch digestion and ber degradation. Journal of Food Science 52, 15881594.

Journal of Cereal Science 50 (2009) 205209

Contents lists available at ScienceDirect

Journal of Cereal Science


journal homepage: www.elsevier.com/locate/jcs

Characteristics of enzymatic hydrolysis of thermal-treated wheat gluten


Jin-shui Wang a, *, Zhi-yan Wei a, Lu Li a, Ke Bian a, Mou-ming Zhao b
a b

School of Food Science and Technology, Henan University of Technology, Zhengzhou 450052, PR China College of Light Industry and Food Sciences, South China University of Technology, Guangzhou 510640, PR China

a r t i c l e i n f o
Article history: Received 19 December 2008 Received in revised form 18 May 2009 Accepted 20 May 2009 Keywords: Wheat gluten Enzymatic hydrolysis Thermal treatment Molecular mass distribution

a b s t r a c t
Effect of thermal treatment at 5090  C on wheat gluten hydrolysis by papain was evaluated in this study. Thermal treatment decreased the amount of sodium dodecyl sulfate (SDS) extractable protein. The treatments at 80 and 90  C had a strong impact on protein extractability. Thermal treatment for 30 min resulted in a signicant reduction in SDS extractable glutenin level in wheat gluten. A signicant drop in free sulphydryl level was found in wheat gluten treated at 70  C for 30 min. It indicated that cross-linking of glutenin through SS occurred during thermal treatment. The treatments at 7090  C led to signicant decreases in soluble and nitrogen level, while signicant increases in peptide nitrogen amount in the hydrolysates from treated gluten were found. A time-dependent effect was observed for the changes in soluble forms of nitrogen and PN. Thermal treatment resulted in molecular mass distribution change according to gel permeation chromatography analysis. Thermal treatment signicantly increased the amount of fractions with molecular mass beyond 10 K (67.2%) in the hydrolysates and greatly decreased the amounts of fractions with MM of 105 K and below 5 K in hydrolysates. 2009 Elsevier Ltd. All rights reserved.

1. Introduction The increasing human population over the last decades has greatly inuenced the demand for food products. Plant proteins are an economic and versatile substitute for animal proteins as functional food ingredients. Wheat gluten, an economically important co-product in the recovery of wheat starch in wet processing of wheat our, is an abundant plant protein source. Wheat gluten consists of two components, glutenins and gliadins. Gliadins are single-chained polypeptides with a molecular mass (MM) between 30,000 and 80,000, whereas glutenins are multi-chained polypeptides with MM ranging from 80,000 to several millions (Payne et al., 1987). In the food industry, wheat gluten is used mainly as an additive for improvement of baking quality of our and is readily available in large amounts at low cost. The utilization of wheat gluten protein resources would be economically interesting for wheat starch factories. Several attempts have been made to produce more valuable and useful

Abbreviations: CAN, acetonitrile; DH, degree of hydrolysis; DTNB, 5,5-dithiobis(2-nitrobenzoic acid); E/S, enzyme to substrate; SH, sulphydryl; GPC, gel permeation chromatography; MM, molecular mass; PN, peptide nitrogen; SDS, sodium dodecyl sulfate; SDS-EG, SDS extractable glutenin; SDS-EP, SDS extractable protein; TFA, triuoroacetic acid. * Corresponding author. Tel.: 86 371 677 899 70; fax: 86 371 677 898 17. E-mail address: jinshuiw@163.com (J.-s. Wang). 0733-5210/$ see front matter 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2009.05.004

products. Enzymatic hydrolysis is one of the main methods for altering food proteins (Franek et al., 2000; Hrckova et al., 2002). During the enzymatic hydrolyzation process, proteins are cleaved to smaller molecules, namely smaller peptides and free amino acids. Hence, the nutritional quality and safety of the products are improved. Enzymatic hydrolysis offers an advantage over the chemical hydrolysis process of plant protein. Protein hydrolysates have properties that make them attractive as a source of protein in human nutrition. They are physiologically better than intact proteins because their intestinal absorption appears to be more effective due to the increase of solubility and peptide content (Panyam and Arun, 1996). The protein hydrolysates are widely used as nutritional supplements, functional ingredients, and avor enhancers in foods, coffee whiteners, cosmetics, personal care products, and confectionery, and in the fortication of soft drinks and juices. Protein hydrolysates are also used in soups, sauces, gravies, snacks, meat products, and other savory applications (Giese, 1994). Therefore, enzymatic hydrolysis can be used to improve or modify the physicochemical, functional, and/or sensory properties of native proteins without losing nutritional value (Kristinsson and Rasco, 2000a). Enzymatic hydrolysis of proteins, especially partial hydrolysis, is often employed to improve the functionalities of food proteins (Kristinsson and Rasco, 2000b; Kuipers et al., 2005). However, investigation of enzymatic hydrolysis of food proteins has been limited by complexity and versatility of protein structure, as well as randomness of the enzymatic

206

J.-s. Wang et al. / Journal of Cereal Science 50 (2009) 205209

hydrolysis process, and lability, versatility and complexity of enzymatic protein hydrolysates. In the previous study, enzymatically modied wheat gluten hydrolysates with enhanced functional properties were obtained (Wang et al., 2006). However, some factors such as slow and incomplete hydrolysis, hydrolysates with bitter taste, low yield of the desired hydrolysates and difculty of separation for hydrolysates exist. These are choke points for restricting the development of the hydrolysis industry for food proteins. Hence, it is very important to facilitate the enzymatic hydrolysis of food proteins by means of modication or pretreatment before hydrolysis. Thermal treatment is a physical technique frequently used in the food industry to modify the properties of proteins (Plumb et al., 1995). The degree of denaturation is highly related to the treatment time and temperature applied (Torreggiani et al., 2008). Thermal modication is often used to facilitate the enzymatic hydrolysis of proteins. Till now, the effect of thermal treatment on the properties of wheat gluten protein hydrolysates is still unknown. In this work, wheat gluten proteins were subjected to thermal treatment. The change of free sulphydryl (SH), total SDS extractable protein, and glutenin level in wheat gluten proteins was investigated. After enzymatic hydrolysis by papain, the levels of soluble nitrogen, formaldehyde nitrogen and peptide nitrogen in wheat gluten hydrolysates were determined. The degree of hydrolysis of wheat gluten hydrolysates was analyzed. The molecular weight distribution of hydrolysates were also evaluated. 2. Materials and methods 2.1. Materials Commercial wheat gluten with 71.5% (w.b) crude protein and 6.8% moisture was used. Papain (food-grade enzyme, 6 105 U g1) was kindly provided by Guangzhou Enzyme Co.(Guangzhou, China). Other reagents used in this study were of analytical grade. 2.2. Thermal treatment and enzymatic hydrolysis of wheat gluten An 8% (w/v) aqueous dispersion of wheat gluten was incubated in a water bath at specic temperature (50, 60, 70, 80, 90  C) for 10, 20, 30, 40, 50 and 60 min. After thermal treatment, the dispersions of the thermal-treated wheat gluten proteins were cooled to 45  C, and then were hydrolysed using papain for 6 h at the optimal condition (pH 6.5, 45  C). During hydrolysis, the pH of the broths was adjusted every 15 min to the optimal value with 1 N NaOH. The ratio of enzyme to gluten proteins [E/S] was 1500 U/g. After hydrolysis, the resultant hydrolysates were heated in boiling water for 10 min to inactivate the enzyme, and then were rapidly cooled to 25  C. The resultant hydrolysates were centrifuged for 20 min at 4000g. The supernatant was used for analysis. The control was the hydrolysate prepared using wheat gluten without thermal treatment. 2.3. Free sulphydryl (SH) determination Free SH groups were determined colorimetrically after reaction with 5,5-dithio-bis(2-nitrobenzoic acid) (DTNB). Samples (1.0 2.0 mg of protein/ml) were shaken for 60 min in 0.05 M sodium phosphate buffer (pH 6.5) containing 2.0% (v/v) SDS, 3.0 M urea and 1.0 mM tetrasodium ethylenediamine tetraacetate. DTNB reagent (0.1% w/v in sample buffer, 100 ml) was mixed with 1.0 ml sample and the extinction at 412 nm was determined after 45 min and centrifugation (3 min, 11,000g). Absorbance values were converted to amounts of free sulphydryl using a calibration. Samples

(100.0 mg) were extracted three times with reduced glutathione (Veraverbeke et al., 2000). 2.4. Size-exclusion HPLC(SE-HPLC) of the thermal-treated wheat gluten SE-HPLC was performed using an HP 1100 system with automatic injection (Agilent Technologies Inc., Palo Alto, CA, USA). All samples (1.0 mg/ml) were extracted with a 0.05 M sodium phosphate buffer (pH 6.8) containing 2.0% sodium dodecyl sulfate (SDS) and loaded (60 ml) on a Biosep-SEC-S4000 column (Phenomenex, Torrance, USA). The elution solvent was (1:1, v/v) acetonitrile (ACN)/water containing 0.05% (v/v) triuoroacetic acid (TFA). The ow rate was 1.0 ml/min at a temperature of 30.8  C (Veraverbeke et al., 2000) and eluted protein was detected at 214 nm. The elution proles were divided into two fractions using the lowest absorbance reading between the two peaks as the cutoff point. The rst fraction corresponds to the amount of SDS extractable glutenin, the second can be assigned to the amount of SDS extractable gliadin. Total SDS extractable protein and glutenin were calculated from the peak areas and expressed as percentage of the peak area of unheated gluten extracted with the SDS buffer in the presence of 1.0% dithiothreitol (DTT). 2.5. Evaluation of soluble nitrogen, formaldehyde nitrogen and peptide nitrogen levels After hydrolysis by papain, the soluble nitrogen content of the supernatant was determined by the semi-micro Kjeldahl method. The formaldehyde nitrogen content of the supernatant was determined by the formaldehyde titration method (Klomklao et al., 2006). The peptide nitrogen content was calculated as soluble nitrogen content formaldehyde nitrogen content. The levels of soluble nitrogen, formaldehyde nitrogen and peptide nitrogen were expressed as a percentage on the basis of total nitrogen content in wheat gluten proteins, respectively. 2.6. Molecular mass distribution of gluten hydrolysates The molecular mass distributions of supernatants in the hydrolysates were estimated by gel permeation chromatography on a Superdex peptide 10/300 GL column (Amersham Biosciences, Corp., NJ, USA) with UV detection at 214 nm. Elution was achieved at 0.5 ml min1 by 0.25 M phosphate buffer (pH 7.2). The column was calibrated with Globin _II (MW 2512 D), Globin _I (MW 6214 D), Globin _ (MW 8519 D), Globin _ _II (MW 10700 D) and Globin _ _I (MW 14404 D). 2.7. Statistical analysis Amounts of free SH-, SDS extractable gluten protein and glutenin, soluble nitrogen, formaldehyde nitrogen and peptide nitrogen were quadruplicate determinations while molecular mass distributions were measured in duplicate. Consequently, a variance analysis (ANOVA) was performed on each experiment to determine the effect of hydrolysis at 95% or 99% level. 3. Results and discussion 3.1. Effect of thermal treatment on 2% SDS extractability Fig. 1 shows the change in 2% SDS extractability of wheat gluten treated at 50, 60, 70, 80 and 90  C for different times. Obviously, thermal treatment decreased the amount of SDS extractable protein. The treatment at 80 and 90  C had a strong impact on

J.-s. Wang et al. / Journal of Cereal Science 50 (2009) 205209

207

Fig. 1. Change in 2% SDS extractability calculated according to wheat gluten proteins treated at 50, 60, 60, 80, 90  C for different intervals. SDS-EP: SDS extractable protein; SDS-EG: SDS extractable glutenin. The values in this Figure are means of triplicates.

protein extractability. Treatment at 80 and 90  C for 30 min led to reduction of protein extractability by 55.04% and 63.1%, respectively. No signicant (P < 0.01) decreases were found in the thermal-treated samples for longer time. Moreover, most of the glutenin became unextractable during thermal treatment (Fig. 1). Thermal treatment for 30 min resulted in a signicant (P < 0.01) reduction of the amounts of SDS extractable glutenin in wheat gluten proteins. 3.2. Change in free SH-content of wheat gluten during thermal treatment The free SH-content of wheat gluten remained constant at 50 and 60  C. A signicant drop (P < 0.05) in the content of free SHgroups was found in the wheat gluten treated at 70  C for 30 min (Fig. 2). Simultaneously, SDS extractability of glutenin (Fig. 1) signicantly (P < 0.01) decreased, indicating cross-linking of glutenin through disulphide bonding. Then, the free SH-content continued to decrease. This indicated a further association of glutenin through sulphydryl/disulphide interchange reactions, leading to larger protein aggregates, reected in a lower SDS extractability as shown in Fig. 1. The thermal treatment at higher temperature (80, 90  C) for 10 min began a signicant (P < 0.05) drop. Further treatment resulted in a sharp decrease in the amounts of free SHgroups. It was accompanied by a sharp decrease in gliadin SDS extractability and a further decrease in glutenin extractability in SDS (Fig. 1). This led to the proposal that gliadin cross-links with glutenin through formation of disulphide bonds. With longer holding times the free SH-content remained constant. The changes induced by heat eventually lead to large gluten protein aggregates

with formation of gliadinglutenin bonds through SS cross-linking in the process (Lagrain et al., 2007, 2005; Morel et al., 2002; Singh and MacRitchie, 2004).

Fig. 2. Changes in free SH-contents in the wheat gluten treated at 50, 60, 70, 80, 90  C. The values in this Figure are means of triplicates.

208

J.-s. Wang et al. / Journal of Cereal Science 50 (2009) 205209

3.3. Changes in the nitrogen levels in the wheat gluten hydrolysates Figs. 35 present the changes in soluble forms of nitrogen (soluble nitrogen and formaldehyde nitrogen) and peptide nitrogen (PN) during enzymatic hydrolysis of the thermal-treated wheat glutens. The levels of soluble nitrogen, formaldehyde nitrogen and PN of untreated wheat gluten hydrolysates were 62.3%, 18.7% and 43.6%, respectively. The treatments at 50 and 60  C gave slight decreases in soluble nitrogen level and formaldehyde nitrogen level, while the treatments at 70, 80 and 90  C led to signicant (P < 0.05) decreases in formaldehyde nitrogen level, compared to untreated wheat gluten hydrolysates (Figs. 3 and 4). The PN amounts in the hydrolysates from wheat gluten treated at lower temperature (50, 60  C) keep a constant increase with increase of treatment time (Fig. 5). Signicant (P < 0.05) increases in the PN amounts in the hydrolysates from wheat gluten treated at higher temperature (70, 80, 90  C) were found. Moreover, a time-dependent effect was also observed for the changes in soluble forms of nitrogen and PN. Formaldehyde nitrogen level was generally used to evaluate total free amino groups in the protein hydrolysates and the degree of hydrolysis (DH) (Chaveesuk et al., 1993). The efciency of enzymatic hydrolysis of food proteins could be determined according to the levels in soluble nitrogen, formaldehyde nitrogen and PN in the hydrolysates. Thermal treatment led to the exposure of active sites in the protein and increased accessibility of protein to enzyme attack (Rehman and Shah, 2005). Meanwhile, the formation of more SS groups had more impact on protein structure. The active sites were hidden inside and hard to be hydrolysed by protease. In the present work, the soluble nitrogen and formaldehyde nitrogen levels of wheat gluten hydrolysates were decreased by thermal treatment. Both treatment temperature and time showed negative effects on the levels of soluble nitrogen and formaldehyde nitrogen. The changes in free SH-level conrmed that thermal treatment induced a more compact structure of wheat gluten. This could explain the lower efciency of enzymatic hydrolysis due to thermal treatment. However, the hydrolysates were rich in peptide nitrogen. Hence, thermal treatment before enzymatic hydrolysis is benecial to production and preparation of the peptides with desired functional properties.
64

Fig. 4. Change in formaldehyde nitrogen amounts in the wheat gluten hydrolysates. The values in this Figure are means of triplicates. The conditions of enzymatic hydrolysis of wheat gluten are as described in Fig. 3.

62

Soluble nitrogen content (%)

60

58 50C 56 60C 70C 54 80C 90C 52

50

10

20

30

40

50

60

Treatment time (min)


Fig. 3. Change in soluble nitrogen amounts in the wheat gluten hydrolysates. The values in this Figure are means of triplicates. Wheat gluten proteins were hydrolysed using papain for 6 h at the optimal condition (pH 6.5, 45  C). Fig. 5. Change in peptide nitrogen amounts in the wheat gluten hydrolysates. The values in this Figure are means of triplicates. The conditions of enzymatic hydrolysis of wheat gluten are as described in Fig. 3.

J.-s. Wang et al. / Journal of Cereal Science 50 (2009) 205209

209

molecular mass beyond 10 K in the hydrolysates from thermaltreated glutens compared with the control. Acknowledgements The authors thank the Henan Provinces Project for Scientic and Technological Innovation Talent of University (No.2008IRTSTHN006), and Talent Fund of Henan University of Technology, for their nancial support of this study. The authors are grateful to Dr. F. MacRitchie for editorial assistance with the manuscript. References
Chaveesuk, R., Smith, J.P., Simpson, B.K., 1993. Production of sh sauce and acceleration of sauce fermentation using proteolytic enzymes. Journal of Aquatic Food Product Technology 2, 5977. Franek, F., Hohenwarter, O., Katinger, H., 2000. Plant protein hydrolysates: preparation of dened peptide fractions promotion growth and production in animal cells cultures. Biotechnology Progress 16, 688692. Giese, J., 1994. Proteins as ingredients: types, functions, applications. Food Technology 48, 5060. Hrckova, M., Rusnakova, M., Zemanovic, J., 2002. Enzymatic hydrolysis of defatted soy our by three different proteases and their effect on the functional properties of resulting protein hydrolysates. Czech Journal of Food Science 2, 714. Klomklao, S., Benjakul, S., Visessanguan, W., Kishimura, H., Simpson, B.K., 2006. Effects of the addition of spleen of skipjack tuna (Katsuwonus pelamis) on the liquefaction and characteristics of sh sauce made from sardine (Sardinella gibbosa). Food Chemistry 98, 440452. Kristinsson, H.G., Rasco, B.A., 2000a. Fish protein hydrolysates: production, biochemical, and functional properties. Critical Reviews in Food Science and Nutrition 40, 4381. Kristinsson, H.G., Rasco, B.A., 2000b. Biochemical and functional properties of Atlantic salmon (Salmo salar) muscle proteins hydrolyzed with various alkaline proteases. Journal of Agricultural and Food Chemistry 48, 657666. Kuipers, B.J.H., van Koningsveld, G.A., Alting, A.C., Driehuis, F., Gruppen, H., Voragen, A.G.J., 2005. Enzymatic hydrolysis as a means of expanding the cold gelation conditions of soy proteins. Journal of Agricultural and Food Chemistry 53, 10311038. Lagrain, B., Thewissen, B.G., Brijs, K., Delcour, J.A., 2007. Mechanism of gliadin glutenin cross-linking during hydrothermal treatment. Food Chemistry 107, 753760. Lagrain, B., Brijs, K., Veraverbeke, W.S., Delcour, J.A., 2005. The impact of heating and cooling on the physico-chemical properties of wheat gluten-water suspensions. Journal of Cereal Science 42 (3), 327333. Morel, M.H., Redl, A., Guilbert, S., 2002. Mechanism of heat and shear mediated aggregation of wheat gluten protein upon mixing. Heating gluten dispersions at high temperature. Journal of Cereal Science 39 (2), 297301. Panyam, D., Arun, K., 1996. Enhancing the functionality of food protein by enzymatic modication. Trends in Food Science and Technology 7, 120125. Payne, P.I., Nightingale, M.A., Krattiger, A.F., Holt, L.M., 1987. The relationship between HMW glutenin subunit composition and the breadmaking quality of British-grown wheat varieties. Journal of the Science of Food and Agriculture 40, 5165. Plumb, G., Mills, E., Tatton, M., DUrsel, C., Lambert, N., Morgan, M., 1995. Effect of thermal and proteolytic processing on glycinin, the 11S globulin of soy (Glycine max): a study utilizing monoclonal and polyclonal antibodies. Journal of Agricultural and Food Chemistry 42, 834840. Rehman, Z., Shah, W.H., 2005. Thermal heat processing effects on antinutrients, protein and starch digestibility of food legumes. Food Chemistry 91, 327331. Singh, H., MacRitchie, F., 2004. Changes in proteins induced by heating gluten dispersions at high temperature. Journal of Cereal Science 39 (2), 297301. Torreggiani, A., Di Foggia, M., Manco, I., De Maio, A., Markarian, S.A., Bonora, S., 2008. Effect of sulfoxides on the thermal denaturation of hen lysozyme: a calorimetric and Raman study. Journal of Molecular Structure 891 (13), 115122. Veraverbeke, W.S., Larroque, O.R., Bekes, F., Delcour, J.A., 2000. In vitro polymerization of wheat glutenin subunits with inorganic oxidizing agents. I. Comparison of single-step and stepwise oxidations of high molecular weight glutenin subunits. Cereal Chemistry 77, 582588. Wang, J.S., Zhao, M.M., Yang, X.Q., Jiang, Y.M., 2006. Improvement on functional properties of wheat gluten by enzymatic hydrolysis and ultraltration. Journal of Cereal Science 44, 93100.

Fig. 6. Effect of thermal treatment on molecular mass distribution of gluten hydrolysates. Control untreatment sample, treatment gluten treated at 70  C for 30 min, MM molecular mass. The conditions of enzymatic hydrolysis of wheat gluten are as described in Fig. 3.

3.4. Effect of thermal treatment on the molecular mass distribution of gluten hydrolysates As mentioned above, signicant (P < 0.05) changes in structure and enzymatic hydrolysis characteristics of wheat gluten treated at 70  C for 30 min were found according to the free SH-level, SDS extractability of gluten proteins, and amounts of soluble forms of nitrogen and peptide nitrogen. The thermal treatment at 70  C for 30 min was the important turning point for wheat gluten proteins. Hence, change in molecular mass distribution of hydrolysates from the untreated wheat gluten and those treated at 70  C for 30 min is shown in Fig. 6. The amount of fractions with molecular mass (MM) beyond 10 K (67.2%) in the hydrolysate from the thermal-treated gluten signicantly (P < 0.05) increased compared with the hydrolysate from the untreated sample (28.7%). Moreover, the thermal treatment greatly decreased the amounts of fractions with MM of 105 K and below 5 K in the hydrolysates (Fig. 6). This was consistent with change in amounts of peptide nitrogen as shown in Fig. 5. 4. Conclusion From the results of our study, we concluded that the thermal treatment at a range of temperatures of 5090  C led to denaturation of wheat gluten proteins. Amounts of free SH- in wheat gluten proteins decreased after the thermal treatment. Thermal treatment caused a reduction in SDS extractability of gluten proteins. The content of peptide nitrogen in the hydrolysates from treated glutens signicantly increased compared to the untreated samples. Decrease in contents of soluble nitrogen and formaldehyde nitrogen in the hydrolysates from thermal-treated glutens was found. This resulted in a predominance of soluble peptides with

Journal of Cereal Science 50 (2009) 210218

Contents lists available at ScienceDirect

Journal of Cereal Science


journal homepage: www.elsevier.com/locate/jcs

Genetic variability in yellow pigment components in cultivated and wild tetraploid wheats
` A.M. Digesu a, C. Platani a, L. Cattivelli a, G. Mangini b, A. Blanco b, *
a b

CRA Cereal Research Centre, S.S. 16 km 6757, 1100 Foggia, Italy Department of Environmental and Agro-Forestry Biology and Chemistry, sect. Genetic and Plant Breeding, University of Bari, Via Amendola 165/A, 70126 Bari, Italy

a r t i c l e i n f o
Article history: Received 26 January 2009 Received in revised form 1 May 2009 Accepted 15 May 2009 Keywords: Carotenoids Yellow pigments HPLC T. turgidum

a b s t r a c t
Yellow pigment concentration (YPC) in durum wheat is an important criterion in the assessment of semolina quality, particularly in determining the commercial and nutritional quality of end-products. Genetic variability of YPC and carotenoid components was analysed in 102 wild and cultivated tetraploid wheat accessions in two trials. Overall, modern cultivars showed signicantly higher values of YPC compared to old cultivars and wild ssp. dicoccum and ssp. dicoccoides accessions. Total carotenoid concentration varied between 1.178 and 4.416 mg/g with an average of 2.460 mg/g. The portion of carotenoids amounted to 33.2% of the YPC in 80 wheat accessions examined in the 2006 trial. Lutein was the main component of carotenoids, followed by zeaxanthin and b-carotene. a-carotene and b-cryptoxanthin were minor components. Pigment concentration was negatively correlated with kernel weight and grain protein concentration. Signicant positive correlations were found between b* index and YPC. Knowledge of the carotenoid composition and concentration is useful for wheat breeders in the development of cultivars with high yellow colour and enhanced phytochemical concentrations, and provides valuable information for evaluating contributions to health benets from the consumption of durum wheat end-products. 2009 Elsevier Ltd. All rights reserved.

1. Introduction Yellow pigment concentration in durum wheat is a criterion in the assessment of semolina quality and is of particular importance in determining the commercial and nutritional quality of endproducts such as pasta. Yellow to amber colour is generally preferred by consumers rather than a brown or cream. Semolina colour is the result of carotenoid pigment concentration of the grain and any losses during storage of the grain or semolina due to carotenoid oxidative degradation by lipoxygenase during processing, as well as processing conditions (for a review see Pagnotta et al., 2005; Troccoli et al., 2000). Carotenoids are among the most important natural pigments, having wide distribution, different structure and numerous biological functions. In plants, they are part of the light harvesting

Abbreviations: HPLC, high-performance liquid chromatography; YPC, yellow pigment concentration. * Corresponding author. Tel.: 39 0039 080 5442992; fax: 39 0039 080 5442200. E-mail address: blanco@agr.uniba.it (A. Blanco). 0733-5210/$ see front matter 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2009.05.002

complexes, involved in photo-oxidative protection, serve as precursors of the hormone abscisic acid and provide the yellow, orange and red colours in many owers and fruits (Della Penna and Pogson, 2006; Hirschberg, 2001). Two distinct classes of carotenoids are recognised: carotenes, which are tetraterpenoid hydrocarbons, and xanthophylls, which contain one or more oxygen groups (Van den Berg et al., 2000). In humans, the nutritional importance of carotenoids comes mainly from the pro-vitamin A activity of b-carotene, a-carotene, b-cryptoxanthin and others, with at least one intact non-oxygenated b-ionone ring (Zile, 1998). In addition to their role as precursors of vitamin A, carotenoids have been associated with a reduced risk of cancer, decrease in cardiovascular diseases, protection of the macula region of the retina and prevention of cataracts, and increased levels of iron absorption (Garcia-Casal, 2006; Granado et al., 2003; Le Marchand et al., 1993; Mares-Perlman et al., 2002). Several analytical procedures have been developed to evaluate yellow pigment concentration in durum semolina and pasta (see review by Fratianni et al., 2005). The most popular colour measurement instruments in the food industry are based on the colour-space system L*, a*, b* (CIE, Commission Internationale de lEclairage, 1986) due to the safety and rapidity of the technique

` A.M. Digesu et al. / Journal of Cereal Science 50 (2009) 210218

211

and good correlation between chemical determinations and reectance measurements (Fratianni et al., 2005; Johnston et al., 1980). Yellow pigments are usually analysed by using ICC method 152 (ICC, 1990) or AACC method 14-50 (AACC, 2000), both based on the extraction of pigments from semolina or pasta with watersaturated n-butyl alcohol and subsequent spectrophotometrical determination. High-performance liquid chromatography (HPLC) is widely accepted as an accurate and sensitive technique to allow the separation and identication of carotenoid components (Panli et al., 2004). Specic HPLC procedures to detect carotenoids in durum grain and end-products have been developed by Burkhardt and Bohm (2007), Fratianni et al. (2005), Hentschel et al. (2002). There is little literature related to the HPLC analysis of carotenoids in durum wheat and data have been reported on carotenoid composition of small numbers of samples (from one to ten cultivars). The concentration of carotenoids is higher in durum (from 1.50 to 5.00 mg/g) than in bread wheat (from 0.50 to 2.00 mg/g) (Fratianni et al., 2005; Hentschel et al., 2002; Hidalgo et al., 2006; Leenhardt et al., 2006) as a result of genetic selection for high yellow pigment concentration in durum wheat breeding programs. Inheritance of grain or semolina pigment concentration is complex, but overall heritability of pigment concentration was found to be reasonably high (from 0.34 to 0.95) (Clarke et al., 2006). The objective of this study was to explore the genetic variability of yellow pigment concentration and carotenoid components of a tetraploid wheat collection including 102 accessions of wild and cultivated wheats belonging to different sub-species of Triticum turgidum L. This will provide information on the existing variability and genetic resources available to breeders to improve colour and nutritional value of pasta and other durum wheat endproducts.

2.2. Yellow pigment concentration The evaluation of yellow pigment concentration (YPC) was made according to AACC Approved Method 14-50 (AACC, 2000) with slight modications, as described by Fares et al. (1991): 1 g of each sample was extracted with 5 mL of water-saturated n-butyl alcohol on an orbital incubator (Gallenkamp INR-200) for 3 h at 260 oscillations per minute. Samples were centrifuged for 7 min at 2417 xg (Beckman JA21) and absorbance of n-butyl alcohol extracts was measured by a Beckman DU-65 UV-spectrophotometer at 435.8 nm (the wavelength of maximum absorption of lutein, taking into account that this xanthophyll is the dominant carotenoid in durum wheat). To improve reliability of the analysis, calibration of the instrument was carried out by reading a blank solvent extraction. Pigment concentration was calculated using the extinction coefcient of lutein and the LambertBeer Law (Rodriguez-Amaya and Kimura, 2004). The reported data are the mean of three replications. 2.3. Brown and yellow indexes Brown and yellow indexes were determined using the reectance colorimeter Chroma Meter CR-300 (Minolta) equipped with a pulsed xenon arc lamp. Absolute measurement of L*, a*, b* (CIE, 1986) coordinates in the Munsell colour system were taken using D65 lightning. Samples to be analysed were placed in a granular material support. Brown index was calculated as 100-L* (Borrelli et al., 2003). Results were the average of three replications. 2.4. Carotenoid components Lutein, zeaxanthin, b-cryptoxanthin, a-carotene and b-carotene standards were obtained from LGC Promochem (LGC Promochem, Milano, Italy). All other reagents were of analytical or HPLC grade and were purchased form Levanchimica (Bari, Italy). Stock solutions of carotenoid standards were dissolved in ethanol, degassed to remove oxygen, and stored at 20  C. Concentrations were determined spectrophotometrically using the LambertBeer Law (Rodriguez-Amaya and Kimura, 2004). Solutions were diluted in methanol: di-chloromethane (MeOH: DCM 45:55 v/v) to make calibration curves. Carotenoid pigment extraction was performed according to Konopka et al. (2005) with some modications. Briey, 2 g of each sample was extracted under nitrogen in a screwcapped tube by adding 4 mL of extraction buffer (hexane/acetone, 80:20 v/v) and 600 mL of butyled hydroxytoluene (BHT) (0.1% w/v) as an anti-oxidant. Samples were stirred for a few minutes and kept in the dark for 16 h. Samples were then centrifuged for 10 min at 3000 rpm (Heraeus Sepatech), supernatants placed in glass tubes and residues extracted once again. The organic layer was collected and ltered with a gyroscope lter for syringe PTFE (porosity of 0.45 mm). Then, 5 mL of extract was evaporated to dryness under a nitrogen stream. Finally, the dry residues were re-dissolved in 200 mL of MeOH: DCM (45:55 v/v). A sample volume of 20 mL was injected into an Agilent Technologies 1100 HPLC system equipped with an automatic sampler and a diode array detector (DAD). Separation was done on a YMC C30 column (250 4.6 mm i.d., particle size 5 mm). The mobile phase was methanol and methyltert-butyl ether (89:11 v/v) at a constant ow rate of 1 mL/min. The mobile phase was previously degassed by sonication for 10 min. Spectrophotometric detection was achieved by means of a diode array detector in the range 400600 nm. Peaks were detected at 450 nm. Carotenoids were identied through their characteristic spectra and comparison of retention times with known standard solutions. Calibration curves were constructed using ve different concentrations of lutein (between 0.125 and 22.05 mg/mL),

2. Experimental 2.1. Plant materials A collection of 102 tetraploid wheat accessions (T. turgidum L.), including 52 old and modern cultivars of durum wheat (ssp. turgidum var. durum), 16 accessions of ssp. turgidum var. turanicum, 27 accessions of spp. dicoccum and 7 wild accessions of ssp. dicoccoides, were evaluated for yellow pigment concentration, yellow and brown indexes, some grain yield traits (yield per spike, kernel weight, grain protein concentration) over two years (2006 and 2007). Eighty accessions grown in 2006 were analysed by HPLC for chemical determination of carotenoid components. The wheat collection was grown in the eld at the University of Bari at Valenzano (Bari, Italy), in a randomized complete block design with three replications. In both experiments, each accession was planted at a rate of 30 seeds per row 1 m long with 30 cm between rows. During the growing season, standard cultivation practices were used. Plots were hand harvested at maturity and grain yield per spike was calculated as grain yield per row/number of spikes per row (about 5060 spikes). A 15 g seed sample per plot was used to determine 1000-kernel weight. Grain protein concentration, expressed as percentage of protein on a dry weight basis, was directly obtained on a 2 g sample of whole-meal our using nearinfrared reectance spectroscopy. Yellow pigments were analysed by AACC method 14-50, the colorimeter system and high-performance liquid chromatography (HPLC). To minimize the degradation of pigments by oxidative enzymes, the samples were ground in a laboratory mill (Cyclotec Sample Mill, Tecator) with a 1 mm sieve and the resulting whole our stored at 4  C for a maximum of 24 h before analysis.

212

` A.M. Digesu et al. / Journal of Cereal Science 50 (2009) 210218

zeaxanthin (between 0.6 and 6.45 mg/mL), b- cryptoxanthin (between 0.1 and 1 mg/mL), a-carotene (between 0.08 and 0.85 mg/ mL) and b-carotene (between 0.05 and 0.6 mg/mL). Carotenoid concentrations were calculated using a linear regression (concentration versus area) of the ve-point standard curve. Results were the average of three replications. 2.5. Statistical analysis Standard ANOVA (mixed model analysis of variance with genotype as xed factor and blocks as random factor) was performed using the software MSTAT-C. When signicant differences were detected, Fishers least signicant difference (LSD) was computed. Genetic variance (s2G) and environmental variance (s2E) were obtained by using the standard ANOVA. Heritability (h2) was estimated by the ratio s2G/s2P, where s2P is phenotypic variance (s2P s2G s2E). Pearson phenotypic correlation coefcients were calculated between yellow pigment concentration, yellow and brown indexes, carotenoid components and grain yield components in each environment. 3. Results 3.1. Yellow pigment concentration and colorimetric indexes Yellow pigment concentration (YPC) and yellow (b*) and brown (100-L*) indexes of the tetraploid wheat collection evaluated over two years (2006 and 2007) are reported in Table 1. Wheat accessions are grouped by sub-species and cultivated durums grouped by year of release. YPC varied between 3.8 and 10.6 mg/g in 2006 and between 3.9 and 12.3 mg/g in 2007, with average values of 6.7 and 7.1 mg/g, respectively. Durum wheat cultivars (ssp. turgidum var. durum) showed the highest mean values (7.2 mg/g in 2006 and 7.8 mg/g in 2007). The cvs. Primadur, Svevo, Grecale, Normanno, Duetto, Exeldur and Brindur had signicantly higher YPC (>10 mg/g) than the others. Notably, all of the most popular cultivars in Italy (Simeto, Iride, Duilio, Ciccio, Claudio and Creso) showed a yellow pigment concentration higher than 6.5 mg/g, a value sufcient to meet the durum market needs. When the durum cultivars are grouped by year of release, the results indicated that most recent varieties had signicantly higher YPC than old cultivars: mean values of YPC were 6.5 mg/g (2006) and 6.8 mg/g (2007) for cultivars released before the 1970s, and 8.1 mg/g (2006) and 8.5 mg/g (2007) for cultivars released in the last decade. This observation suggests that grain and semolina colour has become a sign of quality worthy of note in durum, and that, in the last two decades, breeders have focused particular attention on high YPC during selection of new cultivars. YPC of var. turanicum accessions was comparable to the value detected in the oldest durum wheat cultivars, with mean values of 5.8 mg/g (range 4.7 8.5 mg/g) and 6.8 mg/g (range 4.910.9 mg/g) in 2006 and 2007, respectively. A relatively high value was found in the accession PI 306665 (8.5 mg/g) in 2006 and in PI 278350 (10.9 mg/g) in 2007. The ssp. dicoccum and ssp. dicoccoides accessions showed mean values signicantly lower than the durum wheat cultivars (5.6 and 6.1 mg/g in 2006 and 2007, respectively, for ssp. dicoccum, and 5.7 and 5.9 mg/g in 2006 and 2007, respectively, for ssp. dicoccoides), with small ranges of variation. Yellow index (b*) showed the same trend as yellow pigment concentration, with a range of variation of 13.319.1 in 2006 and 11.517.9 in 2007. Durum wheat cultivars showed the highest average value (16.4 in 2006 and 14.9 in 2007). High b* values were observed for the cvs. Cosmodur, Exeldur, Svevo, Grecale and Ambral in 2006, and for the cvs. Exeldur, Svevo, Normanno and Primadur in 2007. Accessions of var. turanicum exhibited b* values similar to

those of old durum wheat cultivars (range 14.118.7 in 2006, and 12.9-17.0 in 2007), while accessions of ssp. dicoccum (range 11.5 15.0 in 2007) and ssp. dicoccoides (range 11.715.2 in 2007) had the lowest b* values. Brown index (100 L*) is a qualitative parameter that provides an indication about the polyphenol oxidase activity in wheat grain and semolina. The brown index had a generally opposite trend compared to yellow index, with range of variations between 14.4 and 19.6 in 2006 and 14.0 and 18.5 in 2007. Accessions of ssp. dicoccum had the highest mean values for both trials, while the lowest values were detected in the durum cultivars. 3.2. Carotenoid components Eighty wheat accessions grown in 2006 were analysed by HPLC to determine the total amount of carotenoids and the relative quantities of the main carotenoid components: lutein, zeaxanthin, b-cryptoxanthin, a-carotene and b-carotene (Table 2). Fig. 1 shows a typical HPLC chromatogram for carotenoids of a durum wheat sample (cv. Primadur). Total carotenoid concentration varied from 1.178 mg/g (Timilia) to 4.416 mg/g (Svevo) with an average of 2.460 mg/g. The highest mean value was found in durum cultivars (2.728 mg/g), while the lowest mean value was found in the sp. dicoccum accessions (1.515 mg/g). Durum cultivars released after 1991 had a signicantly higher mean value (3.109 mg/g) than cultivars released in the period 19711991 (2.560 mg/g) or before 1971 (2.327 mg/g). HPLC identied lutein as the main carotenoid component in the examined tetraploid wheat accessions (average 1.605 mg/g), followed by zeaxanthin (0.133 mg/g) and b-carotene (0.014 mg/g), while a-carotene (0.011 mg/g) and b-cryptoxanthin (0.008 mg/g) were minor components, in several cases present only in traces. HPLC analysis showed the presence of signicant amounts of other non-identied carotenoid compounds (0.688 mg/g) that were grouped into a single class named as other components. Most of these compounds could be geometric isomers of lutein and zeaxanthin (cistrans) or traces of violaxanthin and neoxanthin resulting from epoxidation of zeaxanthin. Wide ranges of variation for all identied carotenoid components were observed in the wheat accessions. Lutein ranged from 0.699 mg/g (Farvento) to 3.078 mg/g (Svevo). Durum cultivars had the highest mean value (1.852 mg/g), while var. turanicum, ssp. dicoccum and ssp. dicoccoides accessions showed average values much lower than durum (1.296, 0.876 and 1.068 mg/g, respectively). Zeaxanthin had a range of variation from 0.850 to 0.230 mg/g. Among the tetraploid wheat sub-species, ssp. dicoccum and ssp. dicoccoides had the highest mean values (0.164 and 0.148 mg/g, respectively). The accession MG3533/52 of ssp. dicoccoides showed the highest value (0.230 mg/g) of the whole collection. The range of variation of b-carotene concentration was 0.006 0.033 mg/g. Durum cultivars showed the highest mean value (0.015 mg/g), the commercial cultivars Meridiano, Normanno, Grecale and Dylan showing interesting values higher than 0.027 mg/g. Var. turanicum, ssp. dicoccum and ssp. dicoccoides accessions had lower values than durum, with a mean value of 0.013 mg/g and range from 0.006 to 0.022 mg/g. However, taking into account the total carotenoid concentration, wild accessions of ssp. dicoccum and ssp. dicoccoides showed higher percentages of b-carotene (mean value 0.84%) than durum (mean value 0.54%). Interestingly, the accessions MG 5473 and MG 5444/235 showed 1.0 and 1.2% of b-carotene, respectively. a-carotene and b-criptoxanthin were found in small amounts, below the detection limit in several wheat accessions. Mean values of 0.011 mg/g (range 0.0020.030 mg/g) and 0.008 mg/g (range 0.0020.014 mg/g) were observed for a-carotene and b-criptoxanthin, respectively, in the accessions where the two carotenoid components were detected.

` A.M. Digesu et al. / Journal of Cereal Science 50 (2009) 210218

213

Table 1 Average values, least signicant difference (LSD) and heritability (h2B) of yellow pigment content (AACC method 14-50) and colorimetric indexes (brown index and yellow index) in a T. turgidum collection evaluated at Valenzano (Bari, Italy) over two years (2006 and 2007). Sub-species Cultivar or accession Brindur Cannizzo Ceedur Ciccio Claudio Cosmodur Exeldur Iride Agridur Meridiano Parsifal Svevo Vesuvio Cirillo Duetto Dylan Fiore Grecale Normanno Tiziana Virgilio Mean Range Appulo Creso Valnova Valgerardo Karel Produra Sansone Tito Ambral Primadur Arcangelo Duilio Grazia Latino Neodur Ofanto Simeto Tresor Messapia Mean Range Timilia Russello Aziziah Grifoni Capeiti Cappelli Hymera Trinakria Kiperounda Polesine Taganrog Shram 5 Mean Range Before 1971 19711990 Year of release Yellow pigment content (mg/g) 2006 19912008 8.7 6.1 8.3 7.4 7.4 9.7 9.3 7.5 7.3 9.3 6.8 10.5 7.0 7.1 9.2 8.3 5.7 10.6 9.3 6.0 8.0 8.1 (5.710.6) 6.7 6.5 7.9 5.2 5.8 5.4 4.8 6.4 8.5 9.9 6.5 6.4 6.5 5.4 7.9 7.1 7.4 8.4 4.0 6.7 (4.09.9) 3.8 7.2 6.9 7.4 8.2 5.6 7.9 6.5 7.0 4.3 5.4 7.8 6.5 (3.88.2) 2007 10.1 6.9 6.9 8.0 6.5 9.3 11.2 7.4 7.8 9.4 6.6 10.5 8.3 8.0 11.1 9.4 5.9 10.1 10.4 6.1 8.1 8.5 (5.911.2) 6.4 8.2 8.6 6.9 8.1 7.8 5.9 8.3 9.5 12.3 7.8 6.0 7.8 5.3 9.6 8.5 7.8 8.5 4.4 7.8 (4.412.3) 4.1 8.1 6.6 8.1 8.3 6.5 8.8 6.5 6.7 4.1 6.8 7.5 6.8 (4.18.8) Brown index (100-L*) 2006 16.0 16.5 15.6 17.7 15.0 16.3 17.1 15.4 15.9 15.8 16.0 16.1 16.4 16.7 14.5 16.0 15.8 16.1 16.5 15.5 16.2 16.1 (14.517.7) 16.6 16.4 16.8 17.1 15.8 15.9 16.4 16.0 16.5 15.4 16.2 14.4 16.9 15.3 16.4 16.1 17.2 15.4 15.0 16.1 (14.417.2) 17.4 17.0 16.0 15.9 16.8 15.9 15.6 16.4 15.9 17.2 15.1 15.5 16.2 (15.117.4) 2007 14.9 15.7 17.1 15.1 14.0 15.2 16.5 15.1 15.2 15.0 14.7 15.3 15.6 15.1 15.5 14.8 14.7 14.8 15.8 15.1 15.2 15.3 (14.017.1) 15.3 15.1 15.3 14.9 15.6 15.4 16.5 15.8 15.1 15.6 16.0 14.4 15.3 14.4 15.5 15.2 15.5 15.8 14.3 15.3 (14.316.5) 16.8 15.9 15.4 15.7 14.8 15.8 15.3 15.4 14.9 14.8 15.2 15.4 15.5 (14.816.8) Yellow index (b*) 2006 17.2 15.7 16.6 17.2 16.2 18.5 18.4 16.2 16.8 17.8 15.8 19.1 15.6 17.2 16.8 17.4 15.4 18.7 17.5 15.6 16.9 17.0 (15.419.1) 16.4 16.1 17.3 16.2 14.9 15.2 13.7 15.9 18.5 17.4 15.7 15.2 17.9 15.1 17.8 16.9 16.6 16.7 13.3 16.1 (13.318.5) 13.0 16.2 17.0 16.7 17.8 14.5 16.5 16.0 15.5 14.2 14.5 16.6 15.7 (13.017.8) 2007 16.1 14.2 13.8 14.9 13.8 16.1 16.9 14.6 15.5 15.9 13.6 16.9 14.9 15.2 16.4 15.6 13.5 16.1 16.6 14.2 14.9 15.2 (13.516.9) 13.9 14.5 14.9 14.9 15.5 15.5 13.3 15.5 16.1 17.9 15.3 13.8 15.8 13.2 16.4 15.4 14.8 15.8 12.3 15.0 (12.317.9) 12.0 15.4 14.2 15.6 15.0 14.6 15.6 14.7 13.7 12.0 14.4 15.4 14.4 (12.015.6)

ssp. turgidum var. durum

(continued on next page)

214 Table 1 (continued ) Sub-species Cultivar or accession Kamut Cltr 11390 PI 68287 PI 113392 PI 113393 PI 167481 PI 254206 PI 278350 PI 306665 PI 576854 PI 623629 PI 623656 PI 624209 PI 254204 PI 624429 PI 290530 Mean Range ssp. dicoccum Lucanica Farvento MG 4375 MG 5473 MG 5350 MG 3521 MG 4387 MG 5323 MG 5416/1 MG 5471/1 MG 5471/2 MG 5471/3 MG 15516/1 MG 15516/3 MG 15516/4 MG 15516/5 MG 15516/6 MG 15516/7 MG 15516/8 MG 15516/9 MG 15516/10 MG 15516/11 MG 5344/1 MG 5344/2 MG 5300/1 MG 5300/3 MG 5293/1 Mean Range

` A.M. Digesu et al. / Journal of Cereal Science 50 (2009) 210218

Year of release

Yellow pigment content (mg/g) 2006 6.5 7.1 5.9 5.0 5.5 6.8 5.4 6.7 8.5 6.2 5.4 4.9 4.7 4.8 4.9 4.9 5.8 (4.78.5) 5.5 5.6 5.6 (5.55.6) 2007 6.9 7.8 8.1 5.6 5.1 6.8 5.6 10.9 9.1 7.9 4.9 5.6 6.1 6.9 6.3 5.1 6.8 (4.910.9) 6.8 4.9 7.7 5.3 4.4 5.9 6.6 5.5 8.3 5.5 4.5 5.1 5.9 6.1 6.2 6.7 6.4 7.3 5.6 7.5 5.9 6.1 5.7 6.4 5.2 5.7 7.4 6.1 (4.48.3)

Brown index (100-L*) 2006 15.4 15.6 15.9 16.0 16.5 14.9 16.3 16.9 15.8 16.8 15.9 16.6 16.6 15.8 16.5 16.8 16.1 (14.916.9) 19.6 18.5 19.1 (18.619.6) 2007 14.9 14.9 14.8 14.6 16.3 14.5 15.3 15.0 15.4 15.8 15.0 14.9 16.3 16.5 16.1 14.1 15.3 (14.116.5) 18.5 16.5 18.0 17.4 16.1 16.5 18.5 16.8 17.2 16.1 16.6 16.7 18.0 17.5 17.2 17.2 17.2 18.4 17.9 16.7 17.7 17.0 16.5 17.3 16.5 17.0 18.0 17.2 (16.118.5)

Yellow index (b*) 2006 15.7 16.7 16.1 15.5 15.5 15.5 15.3 14.1 18.7 16.0 15.0 15.0 14.4 14.5 14.2 15.4 15.5 (14.118.7) 14.0 14.3 14.2 (14.014.3) 2007 15.5 14.9 15.2 13.6 13.8 14.3 14.0 17.0 16.2 15.0 13.0 13.6 14.2 15.8 14.6 12.9 14.6 (12.917.0) 12.8 11.5 13.3 12.4 11.9 13.5 13.3 13.1 14.5 12.5 12.1 12.2 13.8 14.3 14.1 13.9 13.9 15.0 13.9 14.8 14.2 14.1 12.0 12.9 12.9 13.5 12.7 13.3 (11.515.0)

ssp. turgidum var. turanicum

ssp. dicoccoides

MG MG MG MG MG MG MG

4328/61 29896 5510/7 5444/235 4337/198 5445 4330/66

5.3 5.3 6.6 5.4 5.7 (5.36.6) 6.7 (3.810.6) 0.6 0.94

3.9 7.6 8.3 6.3 5.0 5.3 5.1 5.9 (3.98.3) 7.1 (3.912.3) 0.9 0.91

18.0 16.7 16.2 18.9 17.5 (16.218.9) 16.3 (14.419.6) 1.1 0.59

17.4 14.6 15.6 17.2 14.2 17.6 16.1 16.1 (14.217.6) 15.9 (14.018.5) 0.9 0.79

18.0 16.7 16.2 18.9 17.5 (16.218.9) 16.0 (13.319.1) 1.1 0.81

11.7 15.2 14.9 12.1 12.7 12.7 13.4 13.2 (11.715.2) 14.3 (11.517.9) 0.6 0.94

Mean Range Mean Range LSD (0.05) h2B not evaluated.

` A.M. Digesu et al. / Journal of Cereal Science 50 (2009) 210218

215

Table 2 Average values (mg/g), least signicant difference (LSD), and heritability (h2B) of carotenoid components in a T. turgidum collection evaluated at Valenzano (Bari, Italy) in 2006. Sub-species ssp.turgidum var. durum Cultivar or accession Year of release Lutein Brindur Cannizzo Ceedur Ciccio Claudio Cosmodur Exeldur Iride Agridur Meridiano Parsifal Svevo Vesuvio Cirillo Duetto Dylan Fiore Grecale Normanno Tiziana Virgilio Mean Range Appulo Creso Valnova Valgerardo Karel Produra Sansone Tito Ambral Primadur Arcangelo Duilio Grazia Latino Neodur Ofanto Simeto Tresor Messapia Mean Range Timilia Russello Aziziah Grifoni Capeiti Cappelli Hymera Trinakria Kiperounda Polesine Taganrog Shram 5 Mean Range ssp. turgidum var. turanicum Kamut Cltr 11390 PI 68287 PI 113392 PI 113393 PI 167481 PI 254206 PI 278350 PI 306665 PI 576854 Before 1971 19711990 19912008 2.487 1.332 1.939 1.843 2.164 2.843 2.606 2.037 1.980 2.367 1.611 3.078 1.625 1.869 2.421 2.510 1.559 2.947 2.822 1.352 2.105 Zeaxanthin 0.112 0.085 0.138 0.108 0.113 0.126 0.160 0.118 0.116 0.110 0.100 0.125 0.093 0.096 0.156 0.122 0.092 0.129 0.129 0.176 0.114

b-crypto-xanthin a-carotene
0.013 0.007 0.006 0.012 0.009 0.011 0.011 0.009 0.013 0.013 0.013 0.009 0.006 0.008 0.006 0.010 tr 0.013 0.008 tr 0.008 0.018 0.010 0.012 0.012 0.013 0.022 0.020 0.007 0.020 0.026 0.013 0.023 0.019 0.012 0.011 0.017 tr 0.029 0.016 0.008 0.009

b-carotene
0.024 0.014 0.012 0.024 0.012 0.023 0.019 0.013 0.013 0.033 0.013 0.023 0.015 0.014 0.018 0.029 0.008 0.027 0.029 0.014 0.010

Other componentsa Total carotenoids 0.799 0.596 0.653 0.654 0.729 0.904 0.800 0.660 0.872 1.060 0.623 1.159 0.827 0.523 0.850 0.849 0.198 1.060 1.153 0.749 0.631 3.453 2.044 2.760 2.653 3.039 3.929 3.617 2.844 3.014 3.608 2.373 4.416 2.585 2.523 3.462 3.536 1.857 4.205 4.157 2.299 2.877 3.109 (1.8574.416) 2.353 2.160 3.366 2.032 1.924 2.236 1.819 2.236 3.179 3.532 2.727 2.170 2.777 2.134 3.281 2.770 3.087 3.472 1.352 2.560 (1.3523.532) 1.178 2.420 2.166 2.954 3.345 1.769 2.808 2.362 2.471 1.458 2.201 2.737 2.327 (1.1783.345) 2.509 2.836 2.381 1.705 1.729 2.351 1.829 2.826 4.329 2.432 (continued on next page)

2.167 0.120 0.010 (1.3323.078) (0.8050.176) (0.0060.013) 1.655 1.385 2.331 1.394 1.365 1.270 1.301 1.588 2.094 2.394 1.843 1.412 1.985 1.299 2.087 2.038 1.920 2.405 0.843 0.091 0.128 0.101 0.154 0.140 0.096 0.123 0.141 0.122 0.159 0.097 0.098 0.122 0.104 0.123 0.088 0.100 0.178 0.122 0.006 0.002 0.004 0.007 tr tr 0.002 0.004 0.007 0.008 0.008 0.014 0.004 0.008 0.009 0.008 0.006 0.011 tr

0.016 0.018 0.779 (0.0070.029) (0.0080.033) (0.1981.159) 0.007 0.011 0.011 0.008 0.003 0.011 0.007 0.007 0.010 0.011 0.007 0.018 0.013 0.010 0.012 0.013 0.010 0.015 tr 0.008 0.016 0.013 0.014 0.007 0.012 0.007 0.017 0.017 0.021 0.009 0.010 0.011 0.013 0.018 0.016 0.018 0.020 0.008 0.586 0.617 0.907 0.455 0.409 0.847 0.380 0.479 0.929 0.939 0.763 0.619 0.642 0.700 1.032 0.606 1.033 0.844 0.379

1.716 0.120 0.007 (0.8432.405) (0.0880.178) (0.0020.014) 0.721 1.421 1.358 2.012 2.395 1.176 1.861 1.561 1.563 0.861 1.381 1.898 0.134 0.184 0.126 0.170 0.126 0.150 0.131 0.124 0.131 0.109 0.163 0.126 tr 0.006 0.007 0.006 0.009 tr 0.010 0.011 tr tr 0.006 tr

0.010 0.013 0.693 (0.0030.018) (0.0070.021) (0.3791.032) tr 0.008 0.007 0.014 0.010 0.008 0.015 0.012 0.018 tr 0.006 0.011 0.010 0.012 0.009 0.013 0.013 0.013 0.015 0.010 0.020 0.007 0.013 0.010 0.313 0.789 0.660 0.740 0.793 0.422 0.776 0.644 0.739 0.481 0.633 0.692

1.517 0.140 0.008 (0.7212.395) (0.1090.184) (0.0060.011) 1.586 1.977 1.521 0.955 1.097 1.367 1.006 1.837 2.135 1.290 0.152 0.141 0.154 0.134 0.132 0.157 0.113 0.113 0.197 0.130 0.006 0.010 0.005 0.005 0.008 tr tr 0.014 0.008 0.004

0.011 0.012 0.640 (0.0060.018) (0.0070.020) (0.3130.793) 0.018 0.030 0.007 0.002 tr 0.007 tr 0.008 0.010 0.007 0.020 0.020 0.015 0.009 0.006 0.016 0.013 0.022 0.015 0.014 0.727 0.658 0.679 0.600 0.487 0.804 0.697 0.832 1.964 0.988

216 Table 2 (continued ) Sub-species

` A.M. Digesu et al. / Journal of Cereal Science 50 (2009) 210218

Cultivar or accession Year of release Lutein PI PI PI PI PI PI 623629 623656 624209 254204 624429 290530 1.010 0.848 0.940 1.034 1.097 1.036

Zeaxanthin 0.135 0.141 0.130 0.160 0.148 0.137

b-crypto-xanthin a-carotene
tr 0.005 tr tr tr 0.007 tr 0.007 0.002 tr tr tr

b-carotene
0.009 0.014 0.009 0.007 0.009 0.009

Other componentsa Total carotenoids 0.581 0.595 0.547 0.567 0.563 0.387 1.735 1.609 1.628 1.768 1.817 1.576 2.198 (1.5764.329) 1.265 1.552 1.508 1.583 1.669 1.515 (1.2651.669) 1.276 2.083 2.114 1.424 1.856 1.748 2.491 1.510 1.823 (1.2762.491) 2.460 (1.1784.416) 0.300 0.94

Mean Range ssp. dicoccum Farvento MG 5473 MG 5350 MG 3521 MG 5323 Mean Range ssp. dicoccoides MG MG MG MG MG MG MG MG 4328/61 29896 5510/7 5444/235 5445 4330/66 3533/52 4343

1.296 0.142 0.007 (0.8482.315) (0.1130.197) (0.0040.014) 0.699 0.891 0.935 0.960 0.894 0.173 0.176 0.130 0.166 0.173 0.004 tr tr tr 0.004

0.010 0.013 0.730 (0.0020.030) (0.0060.022) (0.3871.964) tr 0.006 tr 0.002 0.004 0.010 0.016 0.014 0.009 0.014 0.379 0.463 0.429 0.446 0.580

0.876 0.164 (0.6990.960) (0.1300.176) 0.710 1.391 1.157 0.773 1.111 0.892 1.570 0.943 0.143 0.098 0.129 0.166 0.113 0.118 0.230 0.189 tr 0.010 tr tr tr tr tr tr

0.004 0.013 0.459 (0.0020.006) (0.0090.016) (0.3790.580) 0.006 0.008 0.011 0.007 0.011 0.010 tr tr 0.009 0.014 0.016 0.017 0.015 0.016 0.017 0.009 0.408 0.561 0.801 0.461 0.606 0.712 0.674 0.368

Mean Range Mean Range LSD (0.05) h2B

1.068 0.148 (0.7101.570) (0.0980.230) 1.605 (0.6993.078) 0.265 0.93 0.133 (0.8500.230) 0.041 0.48 0.008 (0.0020.014) 0.004 0.72

0.009 0.014 0.574 (0.0060.011) (0.0090.017) (0.3680.801) 0.011 (0.0020.030) 0.006 0.79 0.014 (0.0060.033) 0.007 0.57 0.688 (0.1981.964) 0.279 0.63

tr traces. a Obtained as difference between total carotenoid and the sum of the identied components.

3.3. Heritability and correlation between yellow pigments and grain yield components Broad-sense heritability, as estimated on an accession mean basis, was found to be high for yellow pigment concentration (0.94 and 0.91 in 2006 and 2007, respectively) and yellow index (0.81 and 0.94 in 2006 and 2007, respectively), and relatively high for brown index (0.59 and 0.79 in 2006 and 2007, respectively) (Table 1). High heritability was also found for total carotenoid concentration (0.94) and lutein (0.93), intermediate values for b-cryptoxanthin (0.72) and a-carotene (0.79), and relatively low values for b-carotene (0.57) and zeaxanthin (0.48) (Table 2). Correlation analysis revealed highly signicant and positive relationships (P 0.001) among single carotenoid components and

total carotenoid concentration determined by HPLC (data not shown). Total carotenoid concentration was strongly correlated with yellow pigment concentration (0.94; P 0.001) and with yellow index (0.88; P 0.001) (Table 3). A signicant correlation was also observed between the last two parameters (0.88 and 0.90 in 2006 and 2007 trials, respectively; P 0.001). Signicant negative correlations were observed for YPC and yellow index with grain protein concentration (P 0.010.001). These correlations were signicant for each of the two trials, suggesting consistent relationships. Total carotenoid concentration, yellow pigment concentration and yellow index were found to be positively correlated with grain yield per spike (P 0.010.001). The three colour parameters showed negative correlation values with 1000kernel weight, but only the correlations between yellow pigment concentration and 1000-kernel weight in 2006, and between yellow index and 1000-kernel weight in 2007 were signicant at P 0.05. Brown index was positively correlated with grain protein concentration (P 0.01 and P 0.001 in 2006 and 2007, respectively) and negatively correlated with grain yield per spike (P 0.01 and P 0.001 in 2006 and 2007, respectively), total carotenoid concentration (P 0.05) in 2006, yellow pigment concentration (P 0.05) and yellow index (P 0.01) in 2007. 4. Discussion and conclusions Wheat breeding has generally focused on improving yield and yield stability, disease resistances and technological properties, mostly gluten quality and quantity, and to a lesser extent, colour. Yellow pigment of durum wheat semolina has recently received considerable attention to improve the yellow colour in end-products such as pasta, and the nutritional value of carotenoid

Fig. 1. High-performance liquid chromatogram of carotenoids extracted from durum wheat cultivar Primadur.

` A.M. Digesu et al. / Journal of Cereal Science 50 (2009) 210218

217

Table 3 Correlations between yellow pigment content (AACC method 14-50), total carotenoid content (HPLC), colorimetric indexes (brown and yellow indexes), grain protein content, kernel weight and grain yield per spike in a T. turgidum collection grown at Valenzano (Bari, Italy) in 2006 and 2007. Trait Yellow pigment content Yellow index Brown index Grain protein content Kernel weight Grain yield per spike Trial 2006 2006 2007 2006 2007 2006 2007 2006 2007 2006 2007 Total carotenoid pigments 0.94*** 0.88*** 0.88*** 0.90*** 0.22 0.23* 0.34** 0.32** 0.29* 0.17 0.34** 0.24** 0.16 0.33** 0.41*** 0.31** 0.21 0.24* 0.38** 0.17 0.39** 0.61*** 0.01 0.13 0.31** 0.37*** Yellow pigment content Yellow index Brown index

0.24* 0.40** 0.08 0.49***

*, **, ***: Signicant differences at 0.05P, 0.01P and 0.001P, respectively.

components. Thus, increasing yellow colour is a valuable goal for improving the commercial and nutritional values of durum wheat products. Yellow pigment analysis of 102 accessions showed highly signicant differences (P < 0.001) in the grain pigment concentration among the wheat genotypes in two trials. The results showed wide variation of YPC, ranging from the highest value of the elite cultivar Svevo (10.5 mg/g) of durum wheat to the lowest of the old cultivar Timilia (3.8 mg/g). Overall, modern commercial cultivars showed signicantly higher values compared to traditional cultivars and wild ssp. dicoccum and ssp. dicoccoides accessions. This can be explained by the fact that modern durum breeding programs have selected genotypes with high pigment concentrations because yellow colour has been considered in the last two decades to be an important quality parameter with regard to pasta production. Although environmental factors can inuence yellow pigment concentration, the genetic component is predominant because heritability values detected in 2006 and 2007 trials were more than 0.90, conrming previous studies (Clarke et al., 2006; Eloua et al., 2001; Veronico et al., 2001) that underlined the polygenic nature but also the strong genotypic component of yellow pigment concentration in durum wheat. Total carotenoid concentration determined by HPLC varied between 1.178 and 4.416 mg/g (average 2.460 mg/g), in agreement with that found by others authors (Fortmann and Joiner, 1978; Fratianni et al., 2005; Hentschel et al., 2002; Hidalgo et al., 2006; Leenhardt et al., 2006) in durum and bread wheats. Higher average value (8.41 mg/g) was found in a collection of 54 accessions of einkorn wheat (T. monococcum L.) by Hidalgo et al. (2006), in part due to the lower kernel weight of einkorn in comparison to durum and to the negative relationship between carotenoid concentration and kernel weight (see later) found in cereals. Lutein (65.2%) was the main component of carotenoids, followed by zeaxanthin (5.4%) and b-carotene (0.6%); a-carotene (0.5%) and b-cryptoxanthin (0.3%) were minor components, in several cases present only in traces. HPLC analysis also showed the presence of signicant quantities of other not identied carotenoid compounds (28.0%). Lutein was always found to be the major carotenoid in diploid, tetraploid and hexaploid wheats, even if the proportion varied greatly in different assessments (Adom et al., 2003; Fortmann and Joiner, 1978; Fratianni et al., 2005; Hentschel et al., 2002; Hidalgo et al., 2006; Leenhardt et al., 2006; Serpen et al., 2008). The relative proportion of zeaxanthin, b-cryptoxanthin, a-carotene and bcarotene reported by these authors varied from traces or not detectable quantities to signicantly high proportion of a-carotene (6.3%) in einkorn wheat (Hidalgo et al., 2006), and of b-carotene (3.7%) (Fratianni et al., 2005) and zeaxanthin (8.5%) (Panli et al.,

2004) in durum. The components with higher pro-vitamin A activity (b-carotene, a-carotene, b-cryptoxanthin) showed wide variation (from 0.015 to 0.072 mg/g) indicating the need for screening cultivated and wild germplasm to identify accessions with higher concentration of specic carotenoids, and the possibility of improving the pro-vitamin A activity of commercial cultivars via traditional breeding and selection. Noteworthy, the durum cultivar Meridiano and the var. turanicum accession Cltr11390 had the highest values of pro-vitamin A components expressed as absolute amount (0.072 and 0.060 mg/g, respectively) and as percentage of the total carotenoid concentration (2.0 and 2.1%, respectively). The apparently high differences in the relative proportion of carotenoid components reported in the literature may be attributed to genetic effects, environmental conditions and agronomic practices, but also to different extraction methods, internal standards, milling fractions, particle size of semolina and presence of peripheral parts, HPLC procedures, etc. Total carotenoid concentration, as determined by HPLC, compared with yellow pigment concentration, measured by AACC method 14-50, showed that the portion of carotenoids amounted to 33.2% of the yellow pigments in the 80 tetraploid wheat accessions examined in the 2006 trial. This implies that there are unknown colour-producing compounds in the durum extracts absorbing light at 435.6 nm. This observation is in accordance with results provided by Hentschel et al. (2002) and by Leenhardt et al. (2006). Hentschel et al. (2002) found that the carotenoid fraction of yellow pigment concentration amounted to only 3050% in eight durum cultivars, and concluded that there are other compounds not yet identied that contribute considerably to the yellow colour of the grain extracts. In contrast, Burkhardt and Bohm (2007) and Fratianni et al. (2005) showed close correspondence between reectance measurements and total carotenoid concentration determined on durum cultivars by both the standard AACC 14-50 or ICC 152 methods and the HPLC procedure. Pigment concentration demonstrated a slightly negative correlation with kernel weight, as did yellow index in one trial. This negative correlation was also found by Alvarez et al. (1999), Clarke et al. (2006), Marais (1992), Veronico et al. (2001) and Zhang et al. (2008), and could be partly attributed to dilution effects, with increases in grain starch in large kernels reducing the proportion of pigments. The highly signicant negative correlations observed between yellow pigment parameters with grain protein concentration are particularly important for durum breeding programs. The genetic basis of trait correlations may include single genes with pleiotropic effects or the tight linkage of several genes controlling different traits. Combining quantitative trait loci (QTLs) for high pigment concentration with those for high grain protein

218

` A.M. Digesu et al. / Journal of Cereal Science 50 (2009) 210218 Eloua, I., Nachit, M.M., Martin, L.M., 2001. Identication of a microsatellite on chromosome 7B showing a strong linkage with yellow pigment in durum wheat (Triticum turgidum L. var. durum). Hereditas 135, 255261. Fares, C., Platani, C., Tamma, G., Leccese, F., 1991. Microtest per la valutazione del colore in genotipi di frumento duro. Molini dItalia 12, 1921. Fortmann, K.L., Joiner, R.R., 1978. Wheat pigments and our colour. In: Pomeranz, Y. (Ed.), Wheat Chemistry and Technology. American Association of Cereal Chemists Inc., St. Paul, MN, USA, pp. 493522. Fratianni, A., Irano, M., Panli, G., Acquistucci, R., 2005. Estimation of color of durum wheat. Comparison of WSB, HPLC, and reectance colorimeter measurements. Journal of Agricultural and Food Chemistry 53, 23732378. Garcia-Casal, M.N., 2006. Carotenoids increase iron absorption from cereal-based food in the human. Nutrition Research 26, 340344. Granado, F., Olmedilla, B., Blanco, I., 2003. Nutritional and clinical relevance of lutein in human health. British Journal of Nutrition 90, 487502. Hentschel, V., Kranl, K., Hollmann, J., Lindhauer, M.G., Bohmand, V., Bitsch, R., 2002. Spectrophotometric determination of yellow pigment content and evaluation of carotenoids by high-performance liquid chromatography in durum wheat grain. Journal of Agricultural and Food Chemistry 50, 66636668. Hidalgo, A., Brandolini, A., Pompei, C., Piscozzi, R., 2006. Carotenoid and tocols of einkorn wheat (Triticum monococcum ssp. monococcum L.). Journal of Cereal Science 44, 182193. Hirschberg, J., 2001. Carotenoid biosynthesis in owering plants. Current Opinion in Plant Biology 4, 210218. ICC (International Cereal Chemistry), 1990. ICC Method 152. Standard Methods of the International Association for Cereal Chemistry and Technology. Verlag, Moritz Schafer, Detmold, Germany. Johnston, R.A., Quick, J.S., Donnelly, B.J., 1980. Note on comparison of pigment extraction and reectance colorimeter methods for evaluating semolina colour. Cereal Chemistry 57, 447448. Konopka, I., Czaplicki, S., Rotkiewicz, D., 2005. Differences in content and composition of free lipids and carotenoids in our of spring and winter wheat cultivated in Poland. Food Chemistry 95, 290300. Leenhardt, F., Lyan, B., Rock, E., Boussard, A., Potus, J., Chanliaud, E., Remesy, C., 2006. Genetic variability of carotenoid concentration, and lipoxygenase and peroxidase activities among cultivated wheat species and bread wheat varieties. European Journal of Agronomy 25, 170176. Le Marchand, L., Hankin, J.H., Kolonel, L.N., Breecher, G.R., Wilkens, L.R., Zhao, L.P., 1993. Intake of specic carotenoids and lung cancer risk. Cancer Epidemiology Biomarkers Prevention 2, 183187. Marais, G.F., 1992. Gamma irradiation induced deletions in an alien chromosome segment of the wheat Indis and their use in gene mapping. Genome 35, 225229. Mares-Perlman, J.A., Millen, A.E., Ficek, T.L., Hankinson, S.E., 2002. The body of evidence to support a protective role for lutein and zeaxanthin in delaying chronic disease. Overview. Journal of Nutrition 132, 518S524S. Pagnotta, M.A., Blanco, A., Gadaleta, A., Fares, C., 2005. Functional determinants of grain quality. In: Royo, C., Naschit, M.N., Di Fonzo, N., Arus, J.L., Pfeiffer, W.H., Slafer, G.A. (Eds.), Durum Wheat Breeding: Current Approaches and Future Strategies. The Haworth Press, Inc., Binghamton, NY, pp. 483527. Panli, G., Fratianni, A., Irano, M., 2004. Improved normal-phase high-performance liquid chromatography procedure for the determination of carotenoids in cereals. Journal of Agricultural and Food Chemistry 52, 63736377. Rodriguez-Amaya, D., Kimura, M., 2004. HarvestPlus Handbook for Carotenoids Analysis. In: HarvestPlus Technical Monographs Series, 2. IFPRI CIAT, Washington DC, USA and Cali, Colombia. Serpen, A., Gokmen, V., Karagoz, A., Koksel, H., 2008. Phytochemical quantication and total antioxidant capacities of emmer (Triticum dicoccon Schrank) and einkorn (Triticum monococcum L.) wheat landraces. Journal of Agricultural and Food Chemistry 56, 72857292. Troccoli, A., Borrelli, G.M., De Vita, P., Fares, C., Di Fonzo, N., 2000. Durum wheat quality: a multidisciplinary concept. Journal of Cereal Science 32, 99113. Van den Berg, H., Faulks, R., Granado, H.F., Hirschberg, J., Olmedilla, B., Sandmann, G., Southon, S., Stahl, W., 2000. The potential for the improvement of carotenoid levels in foods and the likely systemic effects. Journal of the Science of Food and Agriculture 80, 880912. ` Veronico, P., Somma, S., Blanco, A., 2001. Variabilita del contenuto in pigmenti ` carotenoidi nelle moderne varieta di frumento duro. Atti IV Convegno Nazio` ` nale Biodiversita: Opportunita di sviluppo sostenibile. IAM, Valenzano (Bari), Italy, 67 September 2001, vol. 3, 225230. Zhang, W., Chao, S., Manthey, F., Chicaiza, O., Brevis, J.C., Echenique, W., Dubcovsky, J., 2008. QTL analysis of pasta quality using a composite microsatellite and SNP map of durum wheat. Theoretical and Applied Genetics 117, 13611377. Zile, M.H., 1998. Vitamin A and embryonic development: an overview. The Journal of Nutrition, 455S458S.

concentration will require detailed information on the chromosomal position and effects of each QTL controlling all of these traits in the genetic material under study. In agreement with results published by Fratianni et al. (2005), highly signicant correlations were found between reectance measurement (b* index) and yellow pigment concentration determined by both standard AACC method and HPLC procedure, thus conrming the usefulness of yellow index as a fast and safe method for screening yellow pigments in large-scale wheat breeding programs. Indeed, the development of a rapid screen for kernel and semolina colour has facilitated the increased grain yellow colour of new durum cultivars developed in the last two decades, while marker-assisted selection may allow additional improvement in the future. Yellow index, although strongly related to pigment concentration, provides only relative and not absolute values of grain or semolina colour and does not provide information on the proportional composition of carotenoids. The HPLC technique, although expensive, resulted in a sensitive, selective and reliable method for determining the qualitative and quantitative distribution of carotenoid compounds in cereals (Hentschel et al., 2002; Panli et al., 2004). In view of the increasing importance of nutritional aspects in the denition of food quality, complete information on the estimation of colour and nutritional value of grain, semolina and end-products of durum and bread wheats may be obtained by combining the b* index and HPLC techniques. The present survey of 102 accessions of wild and cultivated tetraploid wheats demonstrated marked variations in the carotenoid composition and concentration. Such information is useful for plant breeders in screening and selecting cultivars with high yellow colour and enhanced phytochemical concentrations, and provides valuable information for evaluating potential contributions to health from the consumption of durum wheat end-products. Acknowledgements This research was supported by grants from the Ministero ` dellUniversita e della Ricerca, Italy, projects FISR and AGROGEN, and by the EU grant DEVELONUTRI. References
AACC (American Association of Cereal Chemists), 2000. AACC Ofcial Method 14-50. In: Approved Methods of the American Association of Cereal Chemists, tenth ed. St. Paul, MN, USA. Adom, K.K., Sorrells, M.E., Liu, R.H., 2003. Phytochemical proles and antioxidant activity of wheat varieties. Journal of Agricultural and Food Chemistry 51, 78257834. Alvarez, J.B., Martin, L.M., Martin, A., 1999. Genetic variation for carotenoid pigment content in the amphiploid Hordeum chilense Triticum turgidum conv. durum. Plant Breeding 118, 187189. Borrelli, G.M., De Leonardis, A.M., Fares, C., Platani, C., Di Fonzo, N., 2003. Effects of modied processing conditions on oxidative properties of semolina dough and pasta. Cereal Chemistry 80, 225231. Burkhardt, S., Bohm, V., 2007. Development of a new method for the complete extraction of carotenoids from cereals with special reference to durum wheat (Triticum durum Desf.). Journal of Agricultural and Food Chemistry 55, 8295 8301. CIE (Commission Internationale de lEclairage), 1986. Publication 15.2. Colorimetry, second ed. CIE Central Bureau Kegelgasse, Wien, Austria. 27-A-1030. Clarke, F.R., Clarke, J.M., McCaig, T.N., Knox, R.E., DePauw, R.M., 2006. Inheritance of yellow pigment concentration in four durum wheat crosses. Canadian Journal of Plant Science 86, 133141. Della Penna, D., Pogson, B.J., 2006. Vitamin synthesis in plants: tocopherols and carotenoids. Annual Review of Plant Biology 57, 711738.

Journal of Cereal Science 50 (2009) 219226

Contents lists available at ScienceDirect

Journal of Cereal Science


journal homepage: www.elsevier.com/locate/jcs

The gluten protein and interactions between components determine mixograph properties in an F6 recombinant inbred line population in bread wheat
Yong Zhang a, Jianwei Tang a, Jun Yan a, b, Yelun Zhang a, Yan Zhang a, Xianchun Xia a, Zhonghu He a, c, *
a

Institute of Crop Science, National Wheat Improvement Centre/The National Key Facility for Crop Gene Resources and Genetic Improvement, Chinese Academy of Agricultural Sciences (CAAS), 12 Zhongguancun South Street, Beijing 100081, PR China Cotton Research Institute, Chinese Academy of Agricultural Sciences (CAAS), Huanghedadao, Anyang 455000, Henan Province, PR China c CIMMYT China Beijing Ofce, C/O CAAS, Beijing 100081, PR China
b

a r t i c l e i n f o
Article history: Received 6 February 2009 Received in revised form 19 May 2009 Accepted 27 May 2009 Keywords: Triticum aestivum L. Gluten protein Mixograph 1BL.1RS HPLC Quantity of gluten protein fraction

a b s t r a c t
One hundred and sixty-eight F6 recombinant inbred lines (RILs) derived from Chinese wheat cultivars, PH82-2 and Neixiang188, were used to determine the cumulative effects of HMW-GS and LMW-GS composition and quantity of gluten protein fractions on dough mixograph properties. A wide range of variation for all parameters in the RILs was detected. Major gene loci of HMW-GS were associated with variation in mixograph characters, but accounted for no more than 25.3% of the phenotypic variations. Glu-D1, together with Glu-B3, played the most important role in determining the properties. Additive effects of HMW-GS and LMW-GS showed major contributions to most of the variation of mixograph parameters, and epistatic effects were also important and could be counter to additive effects of individual loci. The quantity of gluten protein fractions, especially the quantity of glutenin, LMW-GS, and Glu-B3, showed highly signicant correlations with most of the quality parameters, but the correlation coefcients were inuenced by grain hardness, protein content, or both. Protein quality could be greatly improved through increasing the quantity of glutenin, while holding desirable composition of HMW-GS and LMW-GS alleles, with an appropriate ratio of quantity of glutenin to gliadin. 2009 Elsevier Ltd. All rights reserved.

1. Introduction China is the largest wheat (Triticum aestivum L.) producer and consumer in the world, with an annual production of around 110 million tons harvested from about 23 million hectares. Generally, Chinese wheats show acceptable protein content, but poor protein quality and thus, inferior pan bread, mechanized steamed bread, and dry white noodle quality (Zhang et al., 2005), which all require dough with at least medium strong to strong gluten strength (He

Abbreviations: HMW-GS, high molecular weight glutenin subunit; HPLC, high performance liquid chromatography; LMW-GS, low molecular weight glutenin subunit; MPH, mixograph peak height; MPT, mixograph peak time; MPW, mixograph peak width; MRS, mixograph right slope; MTxH, mixograph band height at 8 min; MTxW, mixograph band width at 8 min; RILs, recombinant inbred lines. * Corresponding author. Institute of Crop Science, National Wheat Improvement Centre/The National Key Facility for Crop Gene Resources and Genetic Improvement, Chinese Academy of Agricultural Sciences (CAAS), 12 Zhongguancun South Street, Beijing 100081, PR China. Tel.: 86 10 82105691; fax: 86 10 82108547. E-mail address: zhhe@public3.bta.net.cn (Z. He). 0733-5210/$ see front matter 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2009.05.005

et al., 2004). Therefore, improvement in gluten strength has become a major priority in the current wheat breeding programs. Differences in gluten strength among genotypes are mainly attributed to the different combinations of storage protein variants including two major fractions, namely glutenin and gliadin, with roughly equal amounts (Cornish et al., 2006; Payne et al., 1984). These two fractions are the main determinants of dough viscoelasticity, with the major function from glutenin composition and minor or modifying effects from gliadins (Bekes et al., 2001; Cornish et al., 2006; Payne et al., 1984). The glutenin molecules dissociate into high molecular weight (HMW) and low molecular weight (LMW) glutenin subunits (DOvidio and Masci, 2004; Payne and Coreld, 1979). The genes coding for HMW-GS in wheat are located at three complex loci (Glu-A1, Glu-B1, and Glu-D1) on the long arms of group 1 homoeologous chromosomes (Payne and Lawrence, 1983), whereas LMW-GS at Glu-A3, Glu-B3, and Glu-D3 loci are located on the short arms of group 1 homoeologous chromosomes (DOvidio and Masci, 2004; Jackson et al., 1996). Studies using segregating populations (Payne et al., 1984), cultivars (He et al., 2005), biotypes developed from a cultivar

220

Y. Zhang et al. / Journal of Cereal Science 50 (2009) 219226

(Lawrence et al., 1987), near isogenic lines (Carrillo et al., 1990), lines decient in certain glutenin subunits (Lawrence et al., 1988), and single chromosome substitution lines (Payne et al., 1987) have indicated that both HMW-GS and LMW-GS are signicant determinants of dough strength, with the total effect of Glu-1 loci being relatively larger than that of Glu-3 loci (DOvidio and Masci, 2004); 2*, 17 18, 5 10, Glu-A3d, and Glu-B3d are correlated with high protein quality, and Glu-B3j shows a strong negative effect on all protein quality traits (Gobaa et al., 2008; He et al., 2005; Liu et al., 2005; Nieto-Taladriz et al., 1994). Moreover, signicant effects of quantity of gluten protein fractions on dough properties and processing quality have also been observed. The quantity of glutenin, including HMW-GS and LMW-GS, is signicantly correlated with dough strength-related traits and bread-making quality, and the ratio of quantity of glutenin to gliadin (Glu/Gli) shows signicant positive correlation with dough properties (Cornish et al., 2006; Larroque et al., 2001; Zhang et al., 2007a,b). Therefore, both desirable allelic combination of glutenin subunits and quantity of gluten protein fractions could lead to the improvement of dough properties. However, information on the effect of LMW-GS on dough properties is much less available than that of HMW-GS (DOvidio and Masci, 2004). Thus, further investigation is needed to establish the relationship between the LMW-GS alleles and dough properties, so that LMW-GS and HMW-GS can be considered together for dough quality improvement. Moreover, to date, little research has been conducted to explain the effects of changes in quantity and composition of glutenin together on dough properties. The 1BL.1RS chromosomal translocation has been used extensively in wheat breeding programs in many countries including China (Rajaram et al., 1990; Zhou et al., 2004), due to its resistance to several important diseases including powdery mildew, leaf rust, stripe rust, and its signicant contribution to agronomic performance. However, signicant and negative effect of the 1BL.1RS translocation on dough properties and quantity of glutenin and LMW-GS has also been reported by Gobaa et al. (2008), Liu et al. (2005), and Zhang et al. (2007a). In the Chinese Yellow and Huai Valley Wheat Region, up to 59% of the leading cultivars carried 1BL.1RS (Zhou et al., 2004). Therefore, 168 recombinant inbred lines (RILs) derived from a cross of PH82-2 (non-1BL.1RS translocation line) and Neixiang 188 (1BL.1RS translocation line), which differ greatly in glutenin subunits and dough properties, were used to assess the relative roles of HMW-GS and LMW-GS, and the quantity of gluten protein fractions on the variation in mixograph properties. The mixograph has been used widely in wheat breeding programs in various countries including CIMMYT, USA, Australia, and China, due to its close correlations with baking quality (Khatkar et al., 1996). 2. Experimental 2.1. Wheat samples Two hundred and forty F2-derived F6 recombinant inbred lines (RILs) were developed from a cross of PH82-2 and Neixiang 188 in Anyang Station of the Institute of Crop Science, Chinese Academy of Agricultural Science (CAAS), which is located in Henan province in the Yellow and Huai Valley Wheat Region. It is the largest wheat region in China, occupying more than 60% of the wheat area and contributing about 70% of the production (He et al., 2004). PH82-2 and Neixiang 188 were the two leading cultivars in this region from the early 1990s to 2003, with large variations in gluten strength and differing at most of the HMW-GS and LMW-GS loci. PH82-2 has tall plant height and medium grain yield, with hard grain, good steamed bread and dry white Chinese noodle quality (He et al.,

2004), and glutenin alleles 1, 14 15, 2 12, Glu-A3d, and Glu-B3d, respectively. Neixiang 188 has medium to short plant height and high grain yield, with soft grain, but some sticky dough, inferior Chinese steamed bread and dry white noodle quality, and glutenin alleles 1, 7 9, 5 10, Glu-A3a, and Glu-B3j, respectively. These two cultivars were crossed in 1998, with the objective to combine grain yield potential and good quality together. Field trials were conducted in a latinized alpha lattice design (Zhang et al., 2006) with partial three replications at Anyang and Jiaozuo in Henan Province, and Taian in Shandong Province in the 20052006 and 20062007 seasons. In total, 390 plots were assigned into a 13 row 30 column array at each location, among which 60 RILs were randomly selected and then planted with three replications across locations and years, while the other 180 RILs were planted with single replication. The two parents were also included as check cultivars with 15 replications in each eld trial. The seeds were planted at normal eld sowing rate by a semiautomatic seeder into six-row plots that are 4 m long and 0.20 m apart. Each plot was given a standard fungicide and hand-weeded protection to ensure optimal grain yield and quality. Fertilizer was applied at each location according to soil test recommendations for optimal grain yield. In the population of 240 RILs, 72 were heterozygous at one or more loci as conrmed by reversed-phase high performance liquid chromatography (RP-HPLC) (unpublished data from our lab) and were deleted from the dataset. Thus, only the data from 168 lines were used for subsequent analysis in this study. 2.2. Quality testing Grain hardness was tested on 300-kernel samples using a Perten 4100 Single Kernel Characterization System (SKCS, Perten Instruments North America Inc., Reno, NV). Hard, medium hard, and soft samples were tempered to around 16%, 15.5%, and 14% moisture content, respectively, and milled into our using a Brabender Quadrumat Junior mill according to AACC approved method 26-21A (AACC, 2000) to quality testing with extraction rate around 65%. Flour protein content was determined with a near infrared transmittance (NIT) analyzer Foss-Tecator 1241 (Foss, Hoganas, Sweden), calibrated by the Kjeldahl method (AACC approved method 46-12). Flour moisture content, falling number, and mixograph properties were determined according to AACC approved methods 44-15A, 56-81B, and 54-40A, respectively. A 35-g mixograph (National Mfg. Co., Lincoln, NE) was used to evaluate the mixing properties of our samples. Mixograph peak time (MPT), peak height (MPH), peak width (MPW), right slope (MRS), and band height and width at 8 min (MTxH and MTxW) were used as the quality parameters. 2.3. Electrophoresis HMW-GS and LMW-GS allelic variants were determined by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDSPAGE) using a vertical 13% polyacrylamide gel and Coomassie blue staining (Liu et al., 2005; Singh et al., 1991). HMW-GS were classied using the nomenclature of Payne and Lawrence (1983), and LMW-GS according to Singh et al. (1991). The presence of the Glu-B3j allele was considered as an indicator of the presence of the 1BL.1RS translocation (Gupta and Shepherd, 1992). The allelic variations at Glu-D3 could not be identied in this procedure and thus, there was no correlated report about the contribution of allelic variation at Glu-D3 to the quality parameters in this study, due to the less important effect compared with the other loci on quality traits (Eagles et al., 2002) and the complexity of this locus (Appelbee et al., 2009).

Y. Zhang et al. / Journal of Cereal Science 50 (2009) 219226

221

2.4. HPLC Only our samples from Jiaozuo in Henan Province through the two seasons were analyzed to quantify the amount of protein fractions, because HPLC analysis is very costly and time consuming. The procedures for gluten protein extraction and RP-HPLC analysis were conducted following the method of Larroque et al. (2001) and Zhang et al. (2007a). The procedures for total protein extraction and SE-HPLC analysis for the ratio of quantity of glutenin to gliadin were conducted according to Larroque et al. (2000). Glu-D3 was quantied only as the remaining AU area of Glu-A3 and Glu-B3 from the LMW-GS evaluation for the correlation analysis with the quality parameters. 2.5. Statistical analysis All data were combined and analyzed by tting an appropriate spatial model with rows and columns through locations and years (Gilmour et al., 1997). The best linear unbiased predictions from the best-t model for all traits were used as raw data for subsequent analysis. For the purposes of estimating variance and covariance components, parents were deleted from the dataset. HMW-GS and LMW-GS variants were treated as xed effects, while lines nested in the HMW-GS and LMW-GS allelic variants as random. A reasonable strategy for assessing the relative amount of a term is to estimate its sum of squares in a row of comparable terms. The F-test for variance signicance was derived from the type 3 sum of squares according to a mixed model. The differences between the pairs of alleles least-square means were tested using the contrast command. The relationships between the quality parameters and the quantity of gluten protein fractions were examined by Pearsons correlation coefcients using SAS PROC CORR (SAS Institute, 2000), only from the dataset from Jiaozuo over the two seasons. 3. Results The occurrence of alleles at Glu-B1, Glu-D1, Glu-A3, and Glu-B3 loci among the 168 RILs indicated that signicantly more than half of the lines possessed 14 15 (58.3%) at Glu-B1 and Glu-B3d (64.3%) at Glu-B3 locus, and a little more lines had 5 10 (53.0%) at Glu-D1 and Glu-A3d (56.5%) at Glu-A3. In detail, the frequencies of the two parent types, 1, 14 15, 2 12, Glu-A3d, Glu-B3d and 1, 7 9, 5 10, Glu-A3a, Glu-B3j, were 11.9% and 6.5%, respectively. Thirteen recombinant types were found, with the highest frequent type of 1, 14 15, 5 10, Glu-A3d, Glu-B3d (18.5%), followed by the type of 1, 14 15, 2 12, Glu-A3d, Glu-B3d (14.3%), and the type of 1, 14 15, 2 12, Glu-A3a, Glu-B3d (11.9%), whereas the type of 1, 14 15, 2 12, Glu-A3a, Glu-B3j showed the lowest frequency (0.6%). It indicated a clear linkage between 7 9 and Glu-B3j, and 14 15 and Glu-B3d. 3.1. Characterization of quality parameters of two parents and RILs The MPH, MPW, and MTxH for PH82-2 and Neixiang 188 were 59.9, 30.5, 39.3 mm and 38.1, 14.3, 37.0 mm, respectively, and the mean values of the RILs performed normal distributions and slightly biased to high value parent PH82-2 (Table 1). MPT, MRS, and MTxW for PH82-2 and Neixiang 188 were 1.9 min, 4.7, 4.1 mm and 7.2 min, 1.8 , 12.9 mm, respectively, and the mean values of the RILs performed normal distributions and slightly biased to low value parent PH82-2. For protein fractions, the quantity of glutenin, LMW-GS, Glu-A3, Glu-B3, and Glu/Gli for PH82-2 and Neixiang 188 were 33.2, 24.8, 2.5, 14.1, and 0.9 AU and 32.0, 21.3, 0.6, 11.9, and 0.7 AU, respectively, and the mean values of the RILs performed normal distributions and slightly biased to high

Table 1 Mean and range of dough-related quality properties in the RILs of PH82-2/Neixiang 188 cross. Trait Parent PH82-2 Glutenina HMW-GS Glu-A1 Glu-B1 Glu-D1 x-HMW y-HMW LMW-GS Glu-A3 Glu-B3 Glu-D3 Glu/Gli MPTb MPH MPW MRS MTxH MTxW 33.2 8.4 1.7 3.6 3.1 7.0 1.4 24.8 2.5 14.1 8.2 0.9 1.9 59.9 30.6 4.7 39.3 4.1 Neixiang 188 32.0 10.7 2.2 4.7 3.9 8.6 2.2 21.3 0.6 11.9 8.8 0.7 7.3 38.1 14.4 1.8 37.0 12.9 RIL Mean 34.0 9.7 1.8 4.3 3.5 7.8 1.8 24.3 1.8 14.0 8.5 0.8 3.3 52.1 24.6 3.6 38.7 7.2 SD 3.9 1.4 0.3 0.7 0.5 1.1 0.3 3 1.1 1.9 0.9 0.1 1.4 5.9 4.9 1.1 3.8 4.3 Range 23.946.1 6.314.2 1.22.6 2.76.6 2.25.0 5.211.3 1.03.0 16.733.6 0.44.1 9.519.0 6.511.1 0.61.0 1.48.6 38.767.4 14.835.9 6.50.6 27.848.0 3.125.8

a Quantity of protein fractions was shown in 106 absorbance units of HPLC corresponding to 1 mg of our, abbreviated as AU. b MPT, mixograph peak time (min); MPH, mixograph peak height (mm); MPW, mixograph peak width (mm); MRS, mixograph right slope ( ); MTxH, mixograph band height at 8 min (mm); MTxW, mixograph band width at 8 min (mm).

value parent PH82-2. The quantity of HMW-GS, Glu-A1, Glu-B1, GluD1, x type HMW-GS, y type HMW-GS, and Glu-D3 for PH82-2 and Neixiang 188 were 8.4, 1.7, 3.6, 3.1, 7.0, 1.4, and 8.2 AU and 10.7, 2.2, 4.7, 3.9, 8.6, 2.2, and 8.8 AU, respectively, and the mean values of the RILs performed normal distributions. The mean value of the quantity of glutenin for the RILs was higher than that of high value parent PH82-2. The ranges observed in the RILs were all much larger than those of the variations between the parents. 3.2. Effect of allelic variation at Glu-1 and Glu-3 loci on mixograph parameters Allelic variation at Glu-B1 had highly signicant effects on MPTand MTxW, accounting for 8.1% and 6.7% of the phenotypic variation, respectively, followed by MPW, MRS, and MTxH (Table 2). Lines possessing 7 9 showed signicantly higher MPT, MRS, MTxH, and MTxW than those with 14 15, whereas contrary results were observed for MPH (Table 3). Allelic variation at Glu-D1 had highly signicant effects on MPT, MRS, and MTxW, accounting for 17.2%, 6.7%, and 6.2% of the phenotypic variation, respectively, followed by MPW and MTxH (Table 2). Lines possessing 5 10 showed signicantly higher MPT, MRS, MTxH, and MTxW than those with 2 12, whereas contrary results were observed for MPW (Table 3). Allelic variation at Glu-A3 had a signicant effect on MPT (Table 2). Lines possessing Glu-A3d exhibited signicantly longer MPT than those with Glu-A3a (Table 3). Allelic variation at Glu-B3 exhibited highly signicant effects on MTxH and MTxW, accounting for 11.7% and 8.1% of the phenotypic variation, respectively, followed by MPH and MPW (Table 2). Lines possessing Glu-B3d showed signicantly higher MPH, MPW, MTxH, and MTxW than those with Glu-B3j (Table 3). 3.3. Effects of interactions among Glu-1 and Glu-3 allelic variants on mixograph parameters The allelic interactions at glutenin subunit loci also inuenced mixograph parameters (Table 2). Interactions between Glu-B1 and Glu-D1 had signicant effects on MPT and MPH. Lines possessing 5 10 showed signicantly higher MPH and longer MPT than

222

Y. Zhang et al. / Journal of Cereal Science 50 (2009) 219226

Table 2 Sum of square for dough-related quality parameters in the RILs of PH82-2/Neixiang 188 cross. Source DF MPTa SS Glu-B1 Glu-D1 Glu-A3 Glu-B3 Glu-B1 Glu-D1 Glu-B1 Glu-A3 Glu-B1 Glu-B3 Glu-D1 Glu-A3 Glu-D1 Glu-B3 Glu-A3 Glu-B3 Glu-B1 Glu-D1 Glu-A3 Glu-B1 Glu-D1 Glu-B3 Glu-B1 Glu-A3 Glu-B3 Glu-D1 Glu-A3 Glu-B3 Error
a * ** ***

MPH % 8.1 17.2 2.5 0.2 4.5 0.6 0.3 1.4 2.9 0.1 0.2 0.1 0.2 0.3 61.2 SS 45 46 26 144** 98* 4 7 6 133** 1 4 30 1 0 2898 % 1.3 1.3 0.8 4.2 2.8 0.1 0.2 0.2 3.9 0.0 0.1 0.9 0.0 0.0 84.2

MPW SS 50* 74* 31 96** 31 0 0 6 59* 1 0 3 0 0 1753 % 2.4 3.5 1.5 4.6 1.5 0.0 0.0 0.3 2.8 0.0 0.0 0.1 0.0 0.0 83.4

MRS SS 8** 9*** 1 1 0 0 1 1 1 0 0 0 0 0 116 % 6.0 6.7 0.8 0.6 0.0 0.1 0.8 0.9 0.8 0.0 0.0 0.2 0.3 0.1 82.7

MTxH SS 38* 92** 0 217*** 7 3 3 28 10 0 2 9 27 0 1416 % 2.1 5.0 0.0 11.7 0.4 0.2 0.1 1.5 0.6 0.0 0.1 0.5 1.4 0.0 76.5

MTxW SS 135*** 126*** 15 164*** 4 14 7 91** 7 18 0 4 14 0 1426 % 6.7 6.2 0.7 8.1 0.2 0.7 0.3 4.5 0.3 0.9 0.0 0.2 0.7 0.0 70.4

1 1 1 1 1 1 1 1 1 1 1 1 1 1 153

10*** 22*** 3* 0 6*** 1 0 2 4** 0 0 0 0 0 78

The same abbreviations as those used in Table 1. Indicates signicant difference at P 0.05 level. Indicates signicant difference at P 0.01 level. Indicates signicant difference at P 0.001 level.

those with 2 12 when 14 15 was present, whereas signicantly lower MPH and longer MPT than those with 2 12 when 7 9 was present (Table 3). Interactions between Glu-D1 and Glu-A3 had a signicant effect on MTxW (Table 2). Lines possessing GluA3d showed signicantly lower MTxW than those with Glu-A3a when 2 12 was present, whereas signicantly higher MTxW than those with Glu-A3a when 5 10 was present (Table 3). Interactions between Glu-D1 and Glu-B3 exhibited signicant effects on MPT, MPH, and MPW (Table 2). Lines possessing GluB3d had signicantly higher MPH and MPW, but shorter MPT than those with Glu-B3j when 2 12 was present, whereas signicantly lower MPH and MPW, but longer MPT than those with Glu-B3j

when 5 10 was present (Table 3). Two-way interactions including Glu-B1 by Glu-A3, Glu-B1 by Glu-B3, Glu-A3 by Glu-B3, and all three-way interactions among Glu-1 and Glu-3 allelic variants did not show any signicant effects on mixograph parameters (Table 2). 3.4. Effect of quantity of gluten protein fractions on mixograph parameters Associations between mixograph parameters and quantity of gluten protein fractions in the RILs segregated between the lines with Glu-B3d and Glu-B3j, as indicated by the correlation analysis between MTxH and glutenin quantity (Fig. 1). The correlations between mixograph parameters and quantity of gluten protein fractions for lines with Glu-B3d are presented in Table 4, and correlations between mixograph parameters and quantity of gluten protein fractions for lines with Glu-B3j are shown in Table 5. Low to medium correlation coefcients between quantity of Glu/Gli and all mixograph parameters were observed in the RILs with Glu-B3d (Table 4) or Glu-B3j (Table 5) when using their data alone. The highest correlation coefcient was observed for MTxW

Table 3 Mean value of quality parameters associated with allelic variation at HMW-GS and LMW-GS in the RILs of PH82-2/Neixiang 188 cross. Common allele Allele Number MPTa MPH 2.8b 4.0a 2.6b 4.2a 3.1b 3.6a 3.5a 3.3a 2.4b 3.1b 2.8b 5.3a 2.5a 2.7a 3.8a 4.6a 52.2a 49.6a 52.1a 49.7a 51.6a 50.2a 53.2a 48.6b MPW MRS 25.2a 22.5b 25.3a 22.4b 24.6a 23.1a 25.7a 21.9b 4.3b 3.2a 4.3b 3.2a 3.9a 3.6a 3.6a 3.9a MTxH MTxW 36.4b 38.8a 35.9b 39.2a 37.7a 37.5a 40.5a 34.7b 4.5b 9.0a 4.8b 8.7a 6.2a 7.3a 9.3a 4.3b 2.9a 6.1a 6.7a 11.3a 5.1b 4.5b 7.3b 10.1a 6.9a 2.8a 11.6a 5.7a

14 15 98 79 70 2 12 79 5 10 89 Glu-A3a 73 Glu-A3d 95 Glu-B3d 108 Glu-B3j 68 14 15 2 12 5 10 2 12 5 10 Glu-A3a Glu-A3d Glu-A3a Glu-A3d Glu-B3d Glu-B3j Glu-B3d Glu-B3j 45 53 34 36 40 39 33 56 53 26 55 34

51.6ab 25.7a 4.9a 34.3a 52.8a 24.8a 3.8a 38.5a 52.6a 46.7b 53.0a 51.2a 50.3a 49.2a 25.0a 3.7a 37.6a 20.0a 2.7a 40.0a 26.3a 4.3a 36.5a 24.3a 4.2a 35.4a 22.9a 3.5a 38.8a 21.8a 3.0a 39.7a 28.6a 4.3a 39.4a 22.0b 4.3a 32.5a 22.9b 2.8a 41.5a 21.9b 3.6a 37.0a

79

2 12

5 10

2 12

2.3c 56.5a 2.8bc 47.7b 4.7a 3.7b 50.0b 49.5b

5 10

Different letters in the same column indicate signicant difference at P 0.05. a The same abbreviations as those in Table 1.

Fig. 1. Associations between mixograph band height at 8 min and glutenin quantity in the RILs with Glu-B3d or Glu-B3j.

Y. Zhang et al. / Journal of Cereal Science 50 (2009) 219226

223

Table 4 Correlation and partial correlation analysis for quality parameters with quantity of gluten protein fractions in the RILs with Glu-B3d allele. Trait Glu/Gli MPTe 0.36***,a 0.25**,b 0.40***,c 0.32**,d 0.50*** 0.81*** 0.72*** 0.85*** 0.38*** 0.71*** 0.61*** 0.75*** 0.19 0.56*** 0.43*** 0.60*** 0.36*** 0.63*** 0.59*** 0.69*** 0.45*** 0.76*** 0.63*** 0.77*** 0.34*** 0.69*** 0.59*** 0.74*** 0.47*** 0.73*** 0.63*** 0.75*** 0.53*** 0.81*** 0.73*** 0.84*** 0.43*** 0.45*** 0.45*** 0.46*** 0.50*** 0.81*** 0.72*** 0.85*** 0.32*** 0.65*** 0.53*** 0.68*** MPH 0.27** 0.03 0.49*** 0.25* 0.09 0.47*** 0.36*** 0.69*** 0.22* 0.32*** 0.21* 0.53*** 0.37*** 0.16 0.02 0.37*** 0.21* 0.24* 0.23* 0.47*** 0.11 0.44*** 0.24* 0.59*** 0.26** 0.28** 0.18 0.51*** 0.07 0.42*** 0.27** 0.57*** 0.03 0.51*** 0.41*** 0.73*** 0.28** 0.42*** 0.43*** 0.52*** 0.09 0.47*** 0.34*** 0.68*** 0.21* 0.34*** 0.19 0.53*** MPW 0.23* 0.01 0.40*** 0.24* 0.01 0.42*** 0.42*** 0.60*** 0.15 0.27** 0.26** 0.44*** 0.29** 0.17 0.12 0.35*** 0.16 0.18 0.24* 0.37*** 0.04 0.38*** 0.31** 0.50*** 0.19* 0.23* 0.24* 0.42*** 0.02 0.35*** 0.30** 0.47*** 0.05 0.46*** 0.47*** 0.64*** 0.31** 0.39*** 0.44*** 0.47*** 0.01 0.43*** 0.42*** 0.60*** 0.15 0.28** 0.23* 0.42*** MRS 0.38*** 0.26** 0.43*** 0.34*** 0.35*** 0.68*** 0.58*** 0.72*** 0.21* 0.55*** 0.44*** 0.60*** 0.03 0.41*** 0.27** 0.45*** 0.21* 0.50*** 0.45*** 0.56*** 0.26** 0.58*** 0.44*** 0.60*** 0.18 0.54*** 0.43*** 0.59*** 0.28** 0.55*** 0.44*** 0.57*** 0.39*** 0.70*** 0.60*** 0.73*** 0.38*** 0.41*** 0.41*** 0.42*** 0.34*** 0.67*** 0.56*** 0.71*** 0.23* 0.58*** 0.45*** 0.62*** MTxH 0.17 0.45*** 0.17 0.32** 0.87*** 0.83*** 0.86*** 0.85*** 0.85*** 0.80*** 0.82*** 0.81*** 0.81*** 0.73*** 0.72*** 0.72*** 0.82*** 0.78*** 0.78*** 0.77*** 0.79*** 0.73*** 0.78*** 0.77*** 0.85*** 0.80*** 0.80*** 0.80*** 0.80*** 0.75*** 0.80*** 0.79*** 0.83*** 0.80*** 0.83*** 0.82*** 0.30** 0.29** 0.29** 0.30** 0.87*** 0.84*** 0.87*** 0.87*** 0.82*** 0.76*** 0.78*** 0.77*** MTxW 0.46*** 0.42*** 0.46*** 0.42*** 0.64*** 0.83*** 0.72*** 0.83*** 0.53*** 0.73*** 0.61*** 0.73*** 0.35*** 0.59*** 0.44*** 0.59*** 0.55*** 0.72*** 0.64*** 0.72*** 0.51*** 0.68*** 0.56*** 0.68*** 0.51*** 0.73*** 0.60*** 0.73*** 0.56*** 0.71*** 0.61*** 0.71*** 0.66*** 0.83*** 0.73*** 0.82*** 0.46*** 0.47*** 0.46*** 0.47*** 0.60*** 0.79*** 0.68*** 0.79*** 0.52*** 0.72*** 0.59*** 0.72***

Table 5 Correlation and partial correlation analysis for quality parameters with quantity of gluten protein fractions in the RILs with Glu-B3j allele. Trait Glu/Gli MPTe 0.37**,a 0.32*,b 0.36**,c 0.32*,d 0.54*** 0.87*** 0.71*** 0.87*** 0.45*** 0.75*** 0.61*** 0.75*** 0.43*** 0.65*** 0.57*** 0.65*** 0.31* 0.57*** 0.43*** 0.57*** 0.52*** 0.75*** 0.64*** 0.75*** 0.38** 0.68*** 0.52*** 0.68*** 0.61*** 0.79*** 0.72*** 0.79*** 0.54*** 0.83*** 0.69*** 0.83*** 0.47*** 0.50*** 0.48*** 0.50*** 0.42*** 0.80*** 0.62*** 0.80*** 0.30* 0.53*** 0.41** 0.53*** MPH 0.34** 0.03 0.40** 0.15 0.20 0.63*** 0.32* 0.73*** 0.28* 0.44*** 0.19 0.54*** 0.24 0.36** 0.22 0.46*** 0.36** 0.28* 0.04 0.35** 0.14 0.51*** 0.27* 0.58*** 0.34** 0.36** 0.09 0.45*** 0.03 0.54*** 0.41** 0.64*** 0.15 0.63*** 0.35** 0.73*** 0.26* 0.53*** 0.43*** 0.55*** 0.33** 0.50*** 0.19 0.62*** 0.32* 0.29* 0.06 0.36** MPW 0.27* 0.06 0.27* 0.13 0.00 0.56*** 0.41** 0.62*** 0.05 0.48*** 0.34** 0.54*** 0.02 0.40** 0.34** 0.47*** 0.12 0.37** 0.23 0.42*** 0.04 0.48*** 0.36** 0.52*** 0.09 0.43*** 0.28* 0.49*** 0.12 0.48*** 0.43*** 0.54*** 0.02 0.54*** 0.40** 0.59*** 0.30* 0.42*** 0.39** 0.43*** 0.17 0.38** 0.23 0.45*** 0.11 0.34** 0.21 0.38** MRS 0.44*** 0.37** 0.43*** 0.40** 0.27* 0.61*** 0.47*** 0.61*** 0.13 0.42** 0.31* 0.42*** 0.12 0.35** 0.28* 0.35** 0.16 0.46** 0.34** 0.46*** 0.07 0.27* 0.19 0.27* 0.09 0.38** 0.26* 0.38** 0.23 0.41** 0.36** 0.41** 0.31* 0.62*** 0.50*** 0.62*** 0.33* 0.37** 0.35** 0.37** 0.16 0.54*** 0.39** 0.54*** 0.18 0.46*** 0.35** 0.46*** MTxH 0.33** 0.58*** 0.44*** 0.51*** 0.86*** 0.86*** 0.80*** 0.86*** 0.82*** 0.80*** 0.74*** 0.80*** 0.81*** 0.77*** 0.73*** 0.76*** 0.73*** 0.68*** 0.62*** 0.67*** 0.75*** 0.70*** 0.68*** 0.71*** 0.79*** 0.77*** 0.70*** 0.77*** 0.76*** 0.71*** 0.69*** 0.71*** 0.82*** 0.80*** 0.75*** 0.80*** 0.40** 0.37** 0.40** 0.40** 0.81*** 0.81*** 0.71*** 0.79*** 0.71*** 0.64*** 0.60*** 0.64*** MTxW 0.40** 0.45*** 0.42*** 0.41** 0.64*** 0.81*** 0.67*** 0.80*** 0.54*** 0.68*** 0.55*** 0.66*** 0.54*** 0.62*** 0.54*** 0.60*** 0.49*** 0.62*** 0.49*** 0.60*** 0.49*** 0.56*** 0.48*** 0.55*** 0.50*** 0.63*** 0.50*** 0.62*** 0.61*** 0.67*** 0.61*** 0.66*** 0.64*** 0.78*** 0.66*** 0.77*** 0.44*** 0.45*** 0.44*** 0.45*** 0.54*** 0.73*** 0.57*** 0.72*** 0.46*** 0.56*** 0.45*** 0.54***

Glutenin

Glutenin

HMW-GS

HMW-GS

Glu-A1

Glu-A1

Glu-B1

Glu-B1

Glu-D1

Glu-D1

x-HMW

x-HMW

y-HMW

y-HMW

LMW-GS

LMW-GS

Glu-A3

Glu-A3

Glu-B3

Glu-B3

Glu-D3

Glu-D3

Coefcient of correlation using the data alone. Coefcient of one order partial correlation with protein content effect removed. c Coefcient of one order partial correlation with grain hardness effect removed. d Coefcient of two order partial correlation with both protein content and grain hardness effects removed. e The same abbreviations as those used in Table 1. * Indicates correlation coefcient signicantly different from zero at P 0.05 level. ** Indicates correlation coefcient signicantly different from zero at P 0.01 level. *** Indicates correlation coefcient signicantly different from zero at P 0.001 level.
b

Coefcient of correlation using the data alone. Coefcient of one order partial correlation with protein content effect removed. c Coefcient of one order partial correlation with grain hardness effect removed. d Coefcient of two order partial correlation with both protein content and grain hardness effects removed. e The same abbreviations as those used in Table 1. * Indicates correlation coefcient signicantly different from zero at P 0.05 level. ** Indicates correlation coefcient signicantly different from zero at P 0.01 level. *** Indicates correlation coefcient signicantly different from zero at P 0.001 level.
b

224

Y. Zhang et al. / Journal of Cereal Science 50 (2009) 219226

(0.46) in the RILs with Glu-B3d, and for MPW (0.27) and MRS (0.44) in the RILs with Glu-B3j. For MPT (0.40 and 0.36) and MPH (0.49 and 0.40) in the RILs with Glu-B3d or Glu-B3j, it was observed by one order partial analysis with grain hardness effect removed. For MTxH (0.45 and 0.58) in the RILs with Glu-B3d or GluB3j, it was observed by one order partial correlation analysis with our protein content effect removed. For MPW (0.40) and MRS (0.43) in the RILs with Glu-B3d, it was observed by one order partial correlation analysis with grain hardness effect removed. For MTxW (0.45) in the RILs with Glu-B3j, it was observed by one order partial correlation analysis with our protein content effect removed. 3.4.1. Correlations between quantity of gluten protein fractions and mixograph parameters in the RILs with Glu-B3d allele MPT, MPH, MPW, and MRS had low to medium correlations with quantity of all gluten protein fractions when using their data alone, and higher correlation coefcients were observed by one order partial correlation analysis with our protein content or grain hardness effect removed, whereas the highest correlation coefcients were observed for quantity of all gluten protein fractions including glutenin (0.85, 0.69, 0.60, and 0.72), HMW-GS (0.75, 0.53, 0.44, and 0.60), Glu-A1 (0.60, 0.37, 0.35, and 0.45), Glu-B1 (0.69, 0.47, 0.37, and 0.56), Glu-D1 (0.77, 0.59, 0.50, and 0.60), x type HMW-GS (0.74, 0.51, 0.42, and 0.59), y type HMW-GS (0.75, 0.57, 0.47, and 0.57), LMW-GS (0.84, 0.73, 0.64, and 0.73), Glu-A3 (0.46, 0.52, 0.47, and 0.42), Glu-B3 (0.85, 0.68, 0.60, and 0.71), and Glu-D3 (0.68, 0.53, 0.42, and 0.62) by two order partial correlation analysis with both our protein content and grain hardness effects removed (Table 4). MTxH had the highest and signicantly positive correlations with quantity of all gluten protein fractions including glutenin (0.87), HMW-GS (0.85), Glu-A1 (0.81), Glu-B1 (0.82), Glu-D1 (0.79), x type HMW-GS (0.85), y type HMW-GS (0.80), LMW-GS (0.83), Glu-A3 (0.30), Glu-B3 (0.87), and Glu-D3 (0.82) when using their data alone, whereas less signicant correlations were obtained by one or two order partial correlation analysis with our protein content or grain hardness effect, or both our protein content and grain hardness effects removed. MTxW had medium and signicantly positive correlations with quantity of all gluten protein fractions when using their data alone, and higher correlation coefcients were found by one or two order partial correlation analysis with grain hardness effect, or both our protein content and grain hardness effects removed, whereas the highest and signicantly positive correlations were found for quantity of all gluten protein fractions including glutenin (0.83), HMW-GS (0.73), Glu-A1 (0.59), Glu-B1 (0.72), Glu-D1 (0.68), x type HMW-GS (0.73), y type HMW-GS (0.71), LMW-GS (0.83), Glu-A3 (0.47), Glu-B3 (0.79), and Glu-D3 (0.72) by one order partial correlation analysis with our protein content effect removed. 3.4.2. Correlations between quantity of gluten protein fractions and mixograph parameters in the RILs with Glu-B3j allele MPT, MRS, and MTxW had low to medium correlations with quantity of all gluten protein fractions when using their data alone, and higher correlation coefcients were detected by one order partial correlation analysis with our protein content or grain hardness effect removed, whereas the highest coefcients were detected for quantity of all gluten protein fractions including glutenin (0.87, 0.61, and 0.81), HMW-GS (0.75, 0.42, and 0.68), Glu-A1 (0.65, 0.35, and 0.62), Glu-B1 (0.57, 0.46, and 0.62), Glu-D1 (0.75, 0.27 and 0.56), x type HMW-GS (0.68, 0.38, and 0.63), y type HMW-GS (0.79, 0.41, and 0.67), LMW-GS (0.83, 0.62, and 0.78), Glu-A3 (0.50, 0.37, and 0.45), Glu-B3 (0.80, 0.54, and 0.73), and Glu-D3 (0.53, 0.46, and 0.56) by one order partial correlation analysis with our protein content effect removed (Table 5).

MPH and MPW had low correlations with quantity of all gluten protein fractions when using their data alone, and higher correlation coefcients were obtained by one order partial correlation analysis with our protein content or grain hardness effect removed, whereas the highest coefcients were obtained for quantity of all gluten protein fractions including glutenin (0.73 and 0.62), HMW-GS (0.54 and 0.54), Glu-A1 (0.46 and 0.47), Glu-B1 (0.35 and 0.42), Glu-D1 (0.58 and 0.52), x type HMW-GS (0.45 and 0.49), y type HMW-GS (0.64 and 0.54), LMW-GS (0.73 and 0.59), Glu-A3 (0.55 and 0.43), Glu-B3 (0.62 and 0.45), and Glu-D3 (0.36 and 0.38) by two order partial correlation analysis with both our protein content and grain hardness effects removed. MTxH had the highest and signicantly positive correlations with quantity of all gluten protein fractions including glutenin (0.86), HMW-GS (0.82), Glu-A1 (0.81), Glu-B1 (0.73), Glu-D1 (0.75), x type HMW-GS (0.79), y type HMW-GS (0.76), LMW-GS (0.82), Glu-A3 (0.40), Glu-B3 (0.81), and Glu-D3 (0.71) when using their data alone, whereas less signicant correlations were detected by one or two order partial correlation analysis with our protein content or grain hardness effect, or both our protein content and grain hardness effects removed. 4. Discussion Allelic variation at Glu-B1 and Glu-D1 totally accounted for from only 2.6% of the phenotypic variation for MPH to 25.3% of the phenotypic variation for MPT in the present study, with Glu-D1 contributing more for almost all mixograph parameters. Lines possessing subunit 7 9 had a signicantly better MPT, MRS, MTxH and MTxW than those with 14 15, lines possessing 5 10 had a significantly better MPT, MRS, MTxH, and MTxW than those with 2 12, which were largely in agreement with those of He et al. (2005) and Liu et al. (2005). Only part variation of mixograph parameters could be attributed to the variation in HMW-GS, and the importance of HMW-GS might be overestimated before. Kolster et al. (1991) pointed out that the response to selection obtained by techniques estimating the quality of protein, such as the SDS sedimentation test, equals the response obtained by selection for HMW-GS. The introduction of high quality HMW-GS into a breeding program is only an important rst step to increase processing quality. Meanwhile, it would be of interest to study the effect of LMW-GS on wheat quality in order to justify the use of this group of proteins in selection, in agreement with DOvidio and Masci (2004). Allelic variation at Glu-A3 had no signicant effect on the quality parameters except for MPT, which accounted for 2.5% of the phenotypic variation. Signicant effect of allelic variation at Glu-B3 occurred for most of the mixograph parameters including MPH, MPW, MTxH, and MTxW, accounting for from 4.2% for MPH to 11.7% for MTxH of the phenotypic variation. Consequently, LMW-GS, especially Glu-B3, made large contributions to mixograph properties, which is in agreement with Brites and Carrillo (2001), who pointed out that some LMW-GS made large positive contribution to MPT. Cornish et al. (2006) indicated that for dough strength, the Glu-1 alleles generally made a greater contribution than the Glu-3 alleles, and of the Glu-3 loci, Glu-B3 locus made the largest contribution, whereas LMW-GS was more important to dough extensibility (Dr. F. Bekes, personal communication; Cornish et al., 2006). Therefore, both HMW-GS and LMW-GS, especially allelic variation at Glu-D1 and Glu-B3, play important roles in determining dough-related properties and processing quality (Cornish et al., 2006; Eagles et al., 2002; He et al., 2005; Liu et al., 2005). The effects of HMW-GS and LMW-GS on most mixograph parameters investigated appeared to be partly additive. The lines containing two alleles associated with good quality, namely, Glu-B1c

Y. Zhang et al. / Journal of Cereal Science 50 (2009) 219226

225

and Glu-D1d, Glu-B1c and Glu-B3d, or Glu-D1d and Glu-B3d, had on average superior MPT, MRS, MTxH, and MTxW to lines with only one favorable allele. Such additive effects among HMW-GS and LMW-GS on processing quality were also reported by Gupta et al. (1989) and Kolster et al. (1991). Furthermore, effects of interactions between Glu-B1 and Glu-D1, Glu-D1 and Glu-A1, and Glu-D1 and Glu-B3 on most of the mixograph parameters were present in the RILs, clearly showing that epistatic effects have an important contribution to the variation, and should therefore be taken into consideration in quality improvement. This agrees with Carrillo et al. (1990), Cornish et al. (2006), and Nieto-Taladriz et al. (1994). Therefore, the results here led to the following four important observations: (1) the major gene loci of HMW-GS were associated with variation in mixograph characters, but accounted for no more than 25.3% of the variations in the RILs; (2) the associations were not randomly distributed among HMW and LMW glutenin genes on Glu-B1, Glu-D1, Glu-A3, and GluB3, and Glu-D1 and Glu-B3 played the most important roles in determining mixograph properties; (3) additive effects of HMW-GS and LMW-GS showed big contributions to most variation of mixograph parameters; and (4) epistatic effects, expressed as interactions among homozygous loci, are important for some characters and can be counter to additive effects of individual loci. Because many loci are involved in wheat quality, taking the additive and epistatic effects into account, breeding programs would be more efcient if improved populations are created through recurrent selection, in which numerous quality alleles are pyramided cycle by cycle with electrophoresis, rather than by selecting high quality lines from unimproved material. High and signicant correlations between quantity of gluten protein fractions and mixograph parameters were indicated by correlation and partial correlation analysis. Quantity of glutenin had the highest and signicant correlations with most of the mixograph parameters for both groups of lines with Glu-B3d or GluB3j, followed by quantity of LMW-GS and Glu-B3. Increasing the order of correlation from raw to one or two could largely improve the coefcients for both groups of the lines with Glu-B3d or Glu-B3j for most of the mixograph parameters, including MTxW by one order partial correlation analysis with our protein content effect removed, MPH and MPW by two order partial correlation analysis with grain hardness and our protein content effects removed. Therefore, the correlations between most of the mixograph parameters and quantity of gluten protein fractions were inuenced by grain hardness, protein content, or both. Cornish et al. (2006) also indicated that besides glutenin, protein content and puroindoline genes inuence most grain quality traits to varying degrees. Zhang et al. (2007b) indicated that the quantity of glutenin, HMW-GS, and LMW-GS were highly and signicantly correlated with farinograph development time and stability, extensograph maximum resistance and extension area, and quantity of LMW-GS could explain 82.8% of the total variation of dough maximum resistance. The correlations or partial correlations between quantity of LMW-GS and mixograph parameters were much larger than that of HMW-GS in this study, among which Glu-B3 contributed the largest. However, the results obtained from analyzing genetic lines decient in varying numbers or amounts of LMW-GS or HMW-GS indicated that HMW-GS, on a constant weight basis, have more pronounced effects on dough resistance than LMW-GS, whereas extensibility was equally affected by both of them (Cornish et al., 2006; Gupta et al., 1991). Since LMW-GS are quantitatively the major group of glutenin subunits, contributing considerably more than that of HMW-GS to total glutenin, they are expected to play quite an important role in determining protein quality of bread wheats (Cornish et al., 2006; Gupta et al., 1991). It seems that greater amounts of LMW-GS in total glutenin might account for

their generally stronger correlation with dough properties, as observed in this study. The high and signicant correlations between the quantity of Glu/Gli and the mixograph parameters indicated the importance of balance of quantity of gliadin and glutenin to wheat quality. Cornish et al. (2006) indicated that a balanced dough with good strength and extensibility is mainly inuenced by the ratio of quantity of glutenin to gliadin (Glu/Gli). Therefore, both desirable allelic variation of glutenin and quantity of storage protein fractions could lead to the improvement of protein quality, illustrating the difculties for breeders in selecting high quality wheat by only analyzing the storage proteins through electrophoresis. The results clearly indicated the value of continuing to study quantity of gluten protein fractions for obtaining more comprehensive results for use in predicting dough properties. More knowledge about quantity of gluten protein fractions and effects of them on dough properties are still needed to be studied. It is necessary to determine the precise effects of quantitative trait loci (QTL) for quantity of gluten protein fractions, especially for quantity of glutenin, LMW-GS, and Glu-B3, and to validate markers associated with quantity of these fractions for use in marker assisted selection (MAS). Acknowledgement This study was supported by the National Science Foundation of China (30600393 and 30830072), the National Basic Research Program (2009CB118300), the international collaboration project on wheat improvement from the Ministry of Agriculture of China (2006-G2), and Core Research Budget of the Non-prot Governmental Research Institutions (ICS, CAAS). References
AACC, 2000. Approved Methods of the AACC, tenth ed. American Association of Cereal Chemists, St. Paul, MN. Appelbee, M.J., Mekuria, G.T., Nagasandra, V., Bonneau, J.P., Eagles, H.A., Eastwood, R.F., Mather, D.E., 2009. Novel allelic variants encoded at the Glu-D3 locus in bread wheat. Journal of Cereal Science 49, 254261. Bekes, F., Gras, P.W., Anderssen, R.S., Appels, R., 2001. Quality traits of wheat determined by small-scale dough testing methods. Australian Journal of Agricultural Research 52, 13251338. Brites, C., Carrillo, J.M., 2001. Inuence of high molecular weight (HMW) and low molecular weight (LMW) glutenin subunits controlled by Glu-1 and Glu-3 loci on durum wheat quality. Cereal Chemistry 78, 5963. Carrillo, J.M., Rousset, M., Qualset, C.O., Kasarda, D.D., 1990. Use of recombinant inbred lines of wheat for study of associations of high-molecular-weight glutenin subunit alleles to quantitative traits. 1. Grain yield and quality prediction tests. Theoretical and Applied Genetics 79, 321330. Cornish, G.B., Bekes, F., Eagles, H.A., Payne P.I., 2006. Prediction of dough properties for bread wheats. In: Wrigley, C., Bekes, F., Bushuk, W. (Eds.), Gliadin and Glutenin, the Unique Balance of Wheat Quality. AACC Internal, St. Paul, Minnesota, USA, pp. 243-280 DOvidio, R., Masci, S., 2004. The low-molecular-weight glutenin subunits of wheat gluten. Journal of Cereal Science 39, 321339. Eagles, H.A., Hollamby, G.J., Gororo, N.N., Eastwood, R.F., 2002. Estimation and utilization of glutenin gene effects from the analysis of unbalanced data from wheat breeding programs. Australian Journal of Agricultural Research 53, 367377. Gilmour, A.R., Cullis, B.R., Verbyla, A.P., 1997. Accounting for natural and extraneous variation in the analysis of eld experiments. Journal of Agricultural, Biological, and Environmental Statistics 2, 269293. Gobaa, S., Brabant, C., Kleijer, G., Stamp, P., 2008. Effects of the 1BL.1RS translocation and of the Glu-B3 variation on fteen quality tests in a doubled haploid population of wheat (Triticum aestivum L.). Journal of Cereal Science 48, 598603. Gupta, R.B., Singh, N.K., Shepherd, K.W., 1989. The cumulative effect of allelic variation in LMW and HMW glutenin subunits on dough properties in the progeny of two bread wheats. Theoretical and Applied Genetics 77, 5764. Gupta, R.B., McRitchie, F., Shepherd, K.W., Ellison, F., 1991. Relative contribution of LMW and HMW glutenin subunits to dough strength and dough stickiness of bread wheat. In: Bushuk, W., Tkachuk, R. (Eds.), 4th Int Workshop on Gluten Proteins. AACC, Winnipeg, St. Paul, MN, pp. 7180. Gupta, R.B., Shepherd, K.W., 1992. Identication of rye chromosome 1R translocations and subunits in hexaploid wheats using storage proteins as genetic markers. Plant Breeding 109, 130140. He, Z.H., Yang, J., Zhang, Y., Quail, K.J., Pena, R.J., 2004. Pan bread and dry white Chinese noodle quality in Chinese winter wheats. Euphytica 139, 257267.

226

Y. Zhang et al. / Journal of Cereal Science 50 (2009) 219226 Payne, P.I., Coreld, K.G., 1979. Subunit composition of wheat glutenin proteins, isolated by gel ltration in a dissociating medium. Planta 145, 8388. Payne, P.I., Lawrence, G.J., 1983. Catalogue of alleles for the complex loci, Glu-A1, Glu-B1, and Glu-D1, which code for high-molecular-weight subunits of glutenin in hexaploid wheat. Cereal Research Communications 11, 2935. Payne, P.I., Holt, L.M., Jackson, E.A., Law, C.N., 1984. Wheat storage proteins: their genetics and their potential for manipulation by plant breeding. Philosophical Transactions of the Royal Society of London, Series B 304, 359371. Payne, P.I., Seekings, J.A., Worland, A.J., Jarvis, M.G., Holt, L.M., 1987. Allelic variation of glutenin subunits and gliadins and its effect on breadmaking quality in wheat: analysis of F5 progeny from Chinese Spring Chinese Spring (Hope 1A). Journal of Cereal Science 6, 103118. Rajaram, S., Villareal, R., Mujeeb-Kazi, A., 1990. Global Impact of 1B/1R Spring Wheat. Agronomy abstracts 105. American Society of Agronomy, Madison, WI. SAS Users Guide, 2000. Statistics. SAS Institute, Inc., Cary, NC. Singh, N.K., Shepherd, K.W., Cornish, G.B., 1991. A simplied SDS-PAGE procedure for separating LMW subunits of glutenin. Journal of Cereal Science 14, 203208. Zhang, Y., Zhang, Y., He, Z.H., Ye, G.Y., 2005. Milling quality and protein properties of autumn-sown Chinese wheats evaluated through multi-location trials. Euphytica 143, 209222. Zhang, Y., He, Z.H., Zhang, A.M., van Ginkel, M., Pena, R.J., Ye, G.Y., 2006. Pattern analysis on protein properties of Chinese and CIMMYT spring wheat cultivars sown in China and CIMMYT. Australian Journal of Agricultural Research 57, 811822. Zhang, P.P., He, Z.H., Chen, D.S., Zhang, Y., Larroque, O.R., Xia, X.C., 2007a. Contribution of common wheat protein fractions to dough properties and quality of northern-style Chinese steamed bread. Journal of Cereal Science 46, 110. Zhang, P.P., He, Z.H., Zhang, Y., Zhang, Y., Xia, X.C., 2007b. Pan bread and Chinese white salted noodle qualities of Chinese winter wheat cultivars and their relationship with gluten protein fractions. Cereal Chemistry 84, 370378. Zhou, Y., He, Z.H., Zhang, G.S., Xia, L.Q., Chen, X.M., Gao, Y.C., Jing, Z.B., Yu, G.J., 2004. Utilization of 1BL/1RS translocation in wheat breeding in China. Acta Agronomica Sinica 30, 531535 (in Chinese).

He, Z.H., Liu, L., Xia, X.C., Liu, J.J., Pena, R.J., 2005. Composition of HMW and LMW glutenin subunits and their effects on dough properties, pan bread, and noodle quality of Chinese bread wheats. Euphytica 82, 345350. Jackson, E.A., Morel, M.H., Sontag-Strohm, T., Branlard, G., Metakovsky, E.V., Redaelli, R., 1996. Proposal for combining the classication systems of alleles of Gli-1 and Glu-3 loci in bread wheat (Triticum aestivum L.). Journal of Genetics and Breeding 50, 321336. Khatkar, B.S., Bell, A.E., Schoeld, J.D., 1996. A comparative study of the inter-relationships between mixograph parameters and bread-making qualities of wheat ours and glutens. Journal of the Science of Food and Agriculture 72, 7185. Kolster, P., Eeuwijk, F.A., van Gelder, W.M.J., 1991. Additive and epistatic effects of allelic variation at the high-molecular-weight glutenin subunit loci in determining the bread-making quality of breeding lines of wheat. Euphytica 55, 277285. Larroque, O.R., Gianibelli, M.C., Gomez Sanchez, M., MacRitchie, F., 2000. Procedure for obtaining stable protein extracts of cereal our and whole meal for sizeexclusion HPLC analysis. Cereal Chemistry 77, 448450. Larroque, O.R., Bekes, F., Wrigley, C.W., Rathmell, W.G., 2001. Analysis of gluten proteins in grain and our blends by RP-HPLC. In: Shewry, P.R., Tatham, A.S. (Eds.), Wheat Gluten. Royal Society of Chemistry, Cambridge, UK, pp. 136139. Lawrence, G.J., Moss, H.J., Shepherd, K.W., Wrigley, C.W., 1987. Dough quality of biotypes of eleven Australian wheat cultivars that differ in high-molecularweight glutenin subunit composition. Journal of Cereal Science 6, 99101. Lawrence, G.J., MacRitchie, F., Wrigley, C.W., 1988. Dough and baking quality of wheat lines decient in glutenin subunits controlled by the Glu-A1, Glu-B1, and Glu-D1 loci. Journal of Cereal Science 7, 109112. Liu, L., He, Z.H., Yan, J., Zhang, Y., Xia, X.C., Pena, R.J., 2005. Allelic variation at the Glu-1 and Glu-3 loci, presence of 1B.1R translocation, and their effects on mixographic properties in Chinese bread wheats. Euphytica 142, 197204. Nieto-Taladriz, M.T., Perretant, M.R., Bouguennec, A., 1994. Effect of gliadins and HMW and LMW subunits of glutenin on dough properties in the F6 recombinant inbred lines from a bread wheat cross. Theoretical and Applied Genetics 88, 8188.

Journal of Cereal Science 50 (2009) 227234

Contents lists available at ScienceDirect

Journal of Cereal Science


journal homepage: www.elsevier.com/locate/jcs

Importance of extensional rheological properties on ber-enriched corn extrudates


Dhananjay A. Pai a, c, Orane A. Blake a, c, Bruce R. Hamaker a, c, Osvaldo H. Campanella b, c, *
a

Department of Food Science, Purdue University, West Lafayette, IN 47907, USA Department of Agricultural and Biological Engineering, Purdue University, West Lafayette, IN 47907, USA c Whistler Carbohydrate Research Center, Purdue University, West Lafayette, IN 47907, USA
b

a r t i c l e i n f o
Article history: Received 7 April 2008 Received in revised form 17 May 2009 Accepted 30 May 2009 Keywords: Corn bran Extrusion Shear rheology Extensional rheology

a b s t r a c t
Incorporation of insoluble bers into extruded cereals severely limits their expansion and reduces crispness. The specic objective of this work was to investigate how corn bran and its fractions mixed with cornmeal affect rheology and extrudate expansion. Alkali-treated bran (ATB) and alkaline-soluble bran (ASB) were prepared from unmodied corn bran (UMB). Mixtures of the different corn bran fractions and degermed cornmeal having a 26% (w/w) of total dietary bers were extruded in a twin screw extruder and expansion of the extrudates was determined. Melt shear rheology of mixtures of cornmeal and different corn bran fractions having a 20% total dietary ber was determined using a capillary rheometer at 120  C while the extensional rheology was determined using lubricated squeezing ow. The expansion of extrudates containing ATB was larger than those containing UMB, while extrudates containing ASB showed a greater expansion that was close to that of the control. Addition of UMB to cornmeal increased shear and extensional viscosity signicantly as compared to the control. ATB addition increased the shear viscosity of the mixture to a small extent but showed the highest extensional viscosity amongst the samples. Addition of ASB resulted in mixtures having lower shear and extensional viscosities than the control. 2009 Elsevier Ltd. All rights reserved.

1. Introduction Dietary ber has a number of health benets. However, its incorporation into extruded puffed snack foods and breakfast cereals limits pufng, reduces crispness and, in the latter case, decreases bowl life. Almost invariably, it has been found that increasing ber concentration in the extrudate formulations reduces the expansion volume of extruded foods. Mendonca et al. (2000) reported that co-extrusion of corn bran and cornmeal resulted in less radial expansion and undesirable product textural characteristics with increased bran content. Ahmed (1999), Chinnaswamy and Hanna (1991), Hseih et al. (1991), Hu et al. (1993), Jin et al. (1994, 1995), Lue et al. (1991), Ozboy and Koksel (2000) and Onwulata et al. (2000) have reported similar results on extrusion of cornmeal, corn our or corn grits mixed with various

Abbreviations: ASB, Alkali soluble bran; ATB, Alkali-treated bran; IDF, Insoluble dietary ber; SDF, Soluble dietary ber; SEI, Sectional expansion index; TDF, Total dietary ber; UMB, Unmodied bran. * Corresponding author at: Department of Agricultural and Biological Engineering, Purdue University, 745 Agriculture Mall Drive, West Lafayette, IN 47907, USA. Tel.: 1 765 496 6330; fax: 1 765 496 1115. E-mail address: campa@purdue.edu (O.H. Campanella). 0733-5210/$ see front matter 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2009.05.007

types of ber. In general, it has been observed that reduction in expansion during extrusion results in products that are dense, tough, and non-crispy (Lue et al., 1991). There is abundant literature describing the effects of ber incorporation on extruded product microstructure and physicochemical properties. Chinnaswamy and Hanna (1991) reported that for extrudates produced by co-extrusion of corn starch with cellulose bers, water solubility of the extrudates, degradation level of amylopectin, and iodine binding abilities of the gel permeation chromatography fractions changed signicantly when compared with extrudates solely based on corn starch. Lue et al. (1991) reported that decreasing the particle size of sugar beet ber improved both radial and longitudinal expansion. Jin et al. (1995), using scanning electron microscopy, found that increasing ber content resulted in less expanded extrudates with smaller air cells and thicker cell walls. In a number of studies, addition of ber made the product darker and the darkness increased with ber content (Ahmed, 1999; Lin et al., 2002) and ber particle size (Lue et al., 1994). Moraru and Kokini (2003) hypothesized that, above a critical concentration, bers may disrupt the continuous structure of the melt, impeding its elastic deformation during expansion. However, no literature has been found showing studies that investigate mechanisms by which ber polymers affect expansion during extrusion.

228

D.A. Pai et al. / Journal of Cereal Science 50 (2009) 227234

Feeds used for the extrusion of foods consist of biopolymers such as starch and proteins and some minor constituents (e.g. salt, sugars, lipids), each having a specic functionality. Researchers who have studied the mechanism for extrudate expansion have noted that expansion is produced primarily by two effects, one is known as the die swell phenomenon, and the other is produced by the sudden vaporization of water in the food melt (Fan et al., 1994; Padmanabhan and Bhattacharya, 1989). The latter phenomenon is more predominant in foods and it is considered as the main factor affecting the structure of extruded foods (Guy, 2001). Expansion phenomenon has been related to the shear viscosity of the melt by a few researchers (Chinnaswamy and Hanna, 1991; Della Valle et al., 1997; Fan et al., 1994; Kokini et al., 1992; Mercier and Feillet, 1975), though actual experimental data relating rheology to extrusion has been limited to very few studies such as the ones conducted by Della Valle et al. (1997), Kokini et al. (1992), Lai and Kokini (1990) and Li et al. (2004). The expansion of a melt as it emerges from the extruder die can be expressed as a function of many of its physicochemical properties including properties like glass transition temperature and more specically rheological properties under shear and extensional ow. In order to facilitate the bubble expansion process as the superheated water is quickly vaporized at atmospheric pressure, the melt lm must ow easily in the bubble walls (Guy, 2001). In rheological terms, this ow is closely related to the shear viscosity of the melt. Extensional properties of the expanding melt are important at later stages of the expansion process, i.e. when the bubble growth is resisted by the elasticity of thin lms separating the extrudate bubbles (Guy, 2001; Steffe, 1996). After expansion, the temperature of the product falls rapidly due to evaporative cooling and to a lesser extent, convective cooling. As a consequence of this rapid decrease of temperature, and also moisture loss, there is a large increase in the extrudate viscosity. The temperature at which transformations occur is very close to the glass transition temperature of the extrudate. For temperatures below the glass transition, i.e. when the material is in the glassy state, the material becomes rigid and forms a solid foam having an open structure. Starches are biopolymers that favor the expansion in an extrusion process and, in general, good and well-expanded cellular structures can be produced from starches of different sources. Increasing ber content in the form of bran, results in premature rupture of gas cells, which reduces overall expansion (Mendonca et al., 2000; Moore, 1994). In microscopic sections of ber-enriched extrudates, it is easy to observe a continuous phase formed by starch polymers as well as the presence of a dispersed ber phase (Guy, 2001). It has been shown that bers are not much affected by the harsh environment existing inside the extruder and largely retain their macrostructure (Guy, 2001). Thus, their physical presence will reduce the expansion potential of the air cells forming in the extrudate matrix due to disruption of the continuous starch phase. This effect is observed during the extrusion of whole meal wheat our with bran concentrations of approximately 89% (Guy, 2001). The complex extrudate expansion process takes place under high temperature and low moisture conditions. As discussed, it encompasses various structural changes and phase transitions, such as extrudate swell, bubble growth, and bubble collapse contributing to the expansion (Chang, 1992). A comprehensive review was carried out by Moraru and Kokini (2003) on the nucleation and expansion of cereals by direct extrusion and expansion of extruded pellets in microwaves. Most of the reviewed works, along with model development, like for example those described by Alavi et al. (2003), Fan et al. (1994), Schwartzberg et al. (1995), Wang et al. (2005), have indicated that an understanding of the rheology of the extrudate melt during extrusion is of

paramount importance, mainly because of its inuence on the expansion process and the quality of the nal product. Initial extrusion trials conducted in our laboratory, which consisted of co-extrusion of cornmeal with different corn ber fractions, showed vast differences in extrusion expansion. Unmodied corn bran at 26% total dietary ber signicantly reduced the expansion of extrudates, but modication of its molecular structure using an alkaline treatment enhanced the sectional expansion index of the extrudates. Incorporation of the alkaline-soluble fraction of corn bran at 26% into the extrudate formulation facilitated signicant cross-sectional expansion of extrudates, such that their degree of expansion was similar to a control sample containing no added ber. These extrusion experiments clearly suggested that the disparity in the expansion behavior was due to the effect of physicochemical characteristics of the ber on the extrudate expansion process. An understanding at a mechanistic level of how ber polymers affect melt rheology leading to extrudate expansion is unavailable in the literature. Acknowledging the importance of rheology of the melt under two types of ow, which are prevalent in the expansion process was one of the aims of this research. Specically, the objective of this work was to investigate how corn bran and its various fractions, mixed with cornmeal, affect melt rheology, characterized by shear and extensional ow, and extrudate expansion. 2. Materials and methods 2.1. Materials Corn bran (particle size < 50 mm) and degermed cornmeal were obtained from Bunge Ltd. (Danville, IL). Sodium hydroxide pellets and concentrated hydrochloric acid were purchased from Mallinckrodt Baker Ltd. (Philipsburg, NJ). Absolute ethanol was obtained from Aaper Alcohol and Chemical Products (Shelbyville, KY). Dow Corning 4 compound silicone oil from Dow Corning (Midland, MI) and vital wheat gluten (Hodgson Mill Brand) were obtained for the lubricated squeezing ow experiments. 2.2. Methods 2.2.1. Fractionation of corn bran Corn bran was mixed with 5% NaOH in a 1:10 ratio for 2 h at room temperature to hydrolyze ferulic acid ester cross-linkages (Doner et al., 2000). The mixture was neutralized with hydrochloric acid. 95% ethanol was then added to the neutralized extract in a ratio of 2:1. The precipitate, termed alkali-treated bran (ATB), was dried at 50  C and milled using a Cyclotec cyclone mill equipped with a screen of mesh size 40. Separation of the alkaline-soluble hemicellulosic fraction from the alkaline-insoluble fraction was done by centrifuging the alkaline mixture (before the neutralization step to produce alkalitreated bran) at 11,159g for 8 min. The supernatant was neutralized with HCl and 80% ethanol was added to the supernatant in a ratio of 2:1 to obtain the alkaline-soluble bran (ASB). The precipitate was dried at 50  C and milled using the cyclone mill equipped with a screen of mesh size 40. Total dietary ber content (soluble and insoluble) of the above samples was determined using AOAC Method 991.43 (AOAC, 1995). 2.2.2. Twin screw extrusion Each ber type [i.e. ATB, ASB, and unmodied bran (UMB)] was mixed with cornmeal in a suitable ratio in order to produce a mixture containing 26% total dietary ber. A total dietary ber content of 26% was selected because it would enable manufacturers to claim an excellent source of ber in their nal product. The

D.A. Pai et al. / Journal of Cereal Science 50 (2009) 227234

229

mixture was subsequently extruded using a Krupp Werner and Peiderer ZSK-25 (WP-25) twin screw extruder (Ramsey, NJ), 25 mm screw diameter, L/D equal to 25 and a standard geometry to produce snacks consisting of two sets of mixing disks and one set of reverse screw elements located close to the die. All the other elements were forward screws. The die head consisted of two circular dies of 4 mm diameter. Extrusion was done under the following conditions: water input ow rate of 9.3 g/min, feed input ow rate of 200 g/min, zone temperatures of 25, 75, 75, 100, and 125  C, and at 250 rpm. Based on the moisture content of the ingredients and the amount of water injected into the extruder, the moisture content of the extruded material ranged between 18 and 20%. Sectional expansion index (SEI) was measured where SEI (De/Dd)2 [Dd and De are the diameters (m) of the die insert and extrudate, respectively] (Alvarez-Martinez et al., 1988). LEI was determined as described by Alvarez-Martinez et al. (1988), i.e. as LEI SEI VEI where VEI is the volumetric expansion index dened as VEI rsolid density =rbulk . rbulk and rsolid density are the bulk and solid densities of the extrudate, respectively. Bulk density was calculated by the equation given by Ali et al. (1996), as rbulk 4=pD2 L where L is the length per gram of extrudate (cm/g). The solid density of the extrudate, rsolid density , was calculated by dividing the mass of the sample by the true solid volume (i.e. the volume of the sample that does not include its pores), which was measured by a stereopycnometer (Quantachrome Instruments, Boynton Beach, Fl). SEI and LEI provide expansion of the sample in two different directions, the rst is perpendicular to the ow whereas the second is in the ow direction. The expansion patterns dened by these two parameters are strongly inuenced by the product formulation (Guy, 2001) but, in most extrusion expanded products, an anisotropic expansion is observed. Although the mechanisms of the expansion phenomenon are not totally elucidated it is believed that the observed anisotropy is due to the elasticity of the melt (Guy, 2001). 2.2.3. Capillary rheometry (shear ow) A twin-bore capillary rheometer (Rosand RH 2000 Capillary Rheometer, Malvern Instruments, Southborough, MA) was used to measure the shear viscosity of the melt. This instrument offers the surest way to prevent evaporation and is one of the simplest extrusion and die ow simulators (Macosko, 1994). 100 g of sample was prepared by mixing cornmeal, ber fractions and water in proportions such that each sample contained approximately 20% of total dietary ber and a total moisture content of 25%. Due to limitations in the capillary viscometer pressure transducer, the total amount of dietary ber had to be reduced to 20% and the water content increased to 25% as compared to 26% total dietary ber and 1820% moisture content during the extrusion process. Moisture content of all samples was determined as per standard AACC Approved Methods (1991). The ingredients were mixed in a Kitchen Aid mixer (Troy, Ohio, USA) for 4 min. The samples thus obtained were then milled using the cyclone mill equipped with a screen of mesh size 40. Cornmeal with 25% moisture and no added ber served as the control. Steady shear experiments were performed at 120  C because this temperature was the highest that could be attainable in the viscometer without scorching the samples during the test, mainly due to the longer residence of the sample in the capillary viscometer as compared to the residence time in the melting zone of the extruder. Samples (100 g) were loaded in one of the capillary bores. A die of L/D equal to 8 (length 32 mm, diameter 4 mm) was attached to the end of the measurement bore. Use of long capillary dies minimizes the contribution of the non-linear portion of the pressure drop (Byler and Kwei, 1971). Steady shear viscosities were measured at shear rates of 100, 150, 175 and 200 s1 which are

typical during an extrusion process. The speed of the moving piston was varied by the software program resulting in different calculated shear rates. No end correction was applied due to the high sample requirement. The software, Flowmaster, was programmed to measure the viscosity readings only at steady pressures. All experiments were performed in triplicate. 2.2.4. Lubricated squeezing ow rheometry (extensional ow) Converging ow into the die of a twin screw extruder involves a combination of shear and extensional ow and bubble growth arising from the evaporation of water from the extrudate matrix (Steffe, 1996). Characterization of the melt extensional viscosity requires sophisticated equipment. Entrance pressure drop measured from ow into a converging die has been used to evaluate the extensional viscosity of cornmeal dough (Bhattacharya et al., 1994; Padmanabhan and Bhattacharya, 1993; Seethamraju and Bhattacharya, 1994). Although this may be the best method for approximating extrusion conditions, the amount of sample needed for each trial is very high. Hence, lubricated squeezing ow between two plates was chosen to characterize the extensional viscosity of the mixed ber samples. Apart from the simplicity of this method, it offers a practical way to avoid the extensive structural disruption that occurs at the narrow space of the die. By using lubricated squeezing ow, this damage can be almost completely eliminated and the specimen can be tested as practically intact by carefully placing the material on the lower plate with a wide spatula or spoon (Corradini et al., 2000; Suwonsichon and Peleg, 1999). Huang and Kokini (1993) and Wikstrom and Bohlin (1999) have used this technique for characterizing extensional viscosity of wheat doughs. A dough, or at least a cohesive mass of the sample, is required to obtain good data from the lubricated squeezing ow experiments. However, apart from cornmeal and sample of cornmeal with ASB, it was found that forming a cohesive mass with UMB or ATB was not possible at the testing temperature (room temperature). Hence, gluten had to be added at an 8.3% level to produce cohesive doughs. The possibility that gluten may affect the rheology of the dough was considered since van Vliet et al. (1992) reported strain hardening of wheat doughs possibly due to the presence of gluten. In our work the mixture of cornmeal and gluten served as a base to observe the effect of the different ber components on the extensional rheology of the dough. The samples thus prepared consisted of cornmeal mixed with gluten and each of the ber components ATB, ASB and UMB. Water was added to each of the samples and mixed well so as to form a cohesive dough. 1 g of sample was prepared with resultant composition of approximately 20% total dietary ber, 25% cornmeal, 8.3% gluten and the rest, water. Accounting for the presence of water in all components of the dough, the sample total moisture content varied between 46 and 47%. Cornmeal with added gluten and no added ber at the same moisture content served as the control. Although the moisture content was much higher than typical moisture contents encountered in extrusion, the objective of performing these experiments was to observe the effects of various bers on extensional viscosity during biaxial extensional ow in order to relate the results to the expansion during extrusion. Lubricated squeezing ow experiments were conducted with a Sintech 1/G Universal Testing Machine (MTS, Eden Prairie, MN). Teon plates, 25.4 mm in diameter, were used for the tests. The plates were lubricated with silicone oil and 7 g of each sample was initially compressed by moving the crosshead at extremely low speeds until the sample reached an initial height of 10 mm. All experiments were conducted with the constant area method in which the initial sample diameter was equal to the diameter of the plates. The crosshead was then moved at constant downward

230

D.A. Pai et al. / Journal of Cereal Science 50 (2009) 227234 Table 2 SEI, LEI and bulk density of extrudates obtained by mixing 26% of various bers with cornmeal. Key: alkali-treated bran (ATB), alkali soluble bran (ASB), unmodied bran (UMB). Sample SEI LEI Bulk density (g/cc) 0.01 K (Pa sn) n 89,119 37,138 106,182 150,178 0.223 0.148 0.227 0.214

velocities of 10 and 20 mm/min until the point where the force reached a value of 100 lbf. A load cell of 100 lbf was used to measure the force exerted on the sample. Experiments were conducted in duplicate. To ensure measurements under true equilibrium conditions, a constant strain rate has to be maintained. The use of a constant deformation speed does not imply the presence of constant strain rate, an aspect that could be challenged in this test. However tests were performed at different deformation velocities and the trends observed were similar. In other words, sufciently low deformation speeds were used to assure that the extensional viscosities were determined at conditions close to equilibrium. 3. Results 3.1. Corn bran fraction The results of analysis of the different ber components obtained from the fractionation procedure are shown in Table 1. After alkali treatment, there was a signicant reduction in insoluble ber from 79 to 52%, and an increase in soluble ber from 1.6 to 30%. The water soluble ber component of the alkali-treated bran is termed hemicellulose B or corn ber gum (Whistler, 1993). The main components of the water insoluble portion of alkali-treated bran consist of hemicellulose A and cross-linked insoluble hemicellulose, such as hemicellulose C (15% w/w of bran), which is a linear xylan polysaccharide with a few arabinose branches, and cellulose (16% w/w) (Sugawara et al., 1994). The main dietary ber component of the alkaline-soluble bran fraction (ASB) contains 64% of soluble dietary ber, which is the branched hemicellulose B. Results given in Table 1 illustrate that alkali treatment of corn bran disrupts its supramolecular structure, producing dietary ber that can be separated into various components, which are hemicelluloses A, B, and an alkali-insoluble fraction that consists of cellulose and cross-linked hemicellulose (hemicellulose C), and starch. 3.2. Extrusion trials with 26% ber components The sectional expansion index (SEI) of extrudates containing 26% ASB (SEI 27) was equivalent to the SEI of the control (no ber added, SEI 27), and both ASB and ATB have SEIs larger than UMB (Table 2). LEI of the control as well as of the mixture containing alkaline-soluble ber (ASF) was similar whereas a slight increase was observed for mixtures containing alkali-treated ber (ATB). Conversely samples containing unmodied bran (UMB) exhibited large values of LEI, that translated into dense and elongated extrudates (Table 2). 3.3. Capillary viscometry Fig. 1 shows apparent viscosities obtained from steady shear experiments at different shear rates plotted against the apparent wall shear rate. As shown by the linearity of the plot in logarithmic coordinates, control and ber-containing (UMB, ATB, ASB) samples showed a pseudoplastic behavior and obeyed the power law model

Control (no added ber) 27a Cornmeal ASB 27a Cornmeal ATB 21b Cornmeal UMB 11d

0.41de 9.3c 0.41e 5.4a 0.49cd 7.2b 0.79a 14.6d

Values with the same letter are not signicantly different (P < 0.05).

in the selected shear rate range. The power law model is described by the following equation:

_ h K gn1
Or in terms of the natural logarithms as:

(1)

_ ln h ln K n 1 ln g

(2)

_ where, h is the apparent viscosity and g is the shear rate in the die. K and n are parameters of the power law model known as consistency and ow indexes, respectively. Fig. 1 shows differences in melt shear rheology based on addition of bers. With the addition of UMB to the cornmeal at the 20% level, the shear viscosity increased substantially compared to the control. However, addition of ATB resulted in a signicant increase in shear viscosity while a decrease was noted with the addition of ASB.

3.4. Extensional rheometry using lubricated squeezing ow The values of stress growth function measured by extensional ow were plotted against the biaxial extensional strain as shown in Fig. 2. The extensional strain can be calculated as shown in Eq. (3).

_ 3

  1 V t 2 Ht

(3)

The extensional viscosity in lubricated squeezing ow is given by the following equation (Campanella and Peleg, 2002).

mb

2FtHt pR2 V

(4)

8000 6000

Apparent Viscosity (Pa.s)

4000 2000 1000 800 600 400 200 100 80 Corn Meal Corn Meal + UMB Corn Meal + ATB Corn Meal + ASB 100 120 140 160 180 200 220 240

Table 1 Dietary ber compositions of corn bran preparations. Key: alkali-treated bran (ATB), alkali soluble bran (ASB), unmodied bran (UMB). Sample ATB ASB UMB IDF, % 52 1 9.4 0.3 79 5 SDF, % 30 1 64 0 1.6 0.3 TDF, % 82 1 73 0 83 5

Shear Rate (1/s)


Fig. 1. Results of capillary rheometry of cornmeal with various bers. Apparent shear viscosity as a function of shear rate. Key: alkali-treated bran (ATB), alkali soluble bran (ASB), unmodied bran (UMB).

D.A. Pai et al. / Journal of Cereal Science 50 (2009) 227234

231

4. Discussion
Corn Meal + ATB

Extensional Viscosity (Pa.s)

1000000 Corn Meal + UMB Corn Meal 100000 Corn Meal + ASB

4.1. Effect of rheology on expansion Cornmeal with ASB extrudate exhibited a much lower shear viscosity and equal SEI and LEI compared to the control cornmeal extrudate with no added ber. Fan et al. (1994) suggested that, by neglecting the effect of surface tension, the rate of bubble growth during extrusion expansion can be estimated as:

DP dR=dt ma R

(5)

10000

Strain
Fig. 2. Results of lubricated squeezing ow rheometry of cornmeal with various bers. Extensional viscosity (stress growth function) as a function of stain. Key: alkali-treated bran (ATB), alkali soluble bran (ASB), unmodied bran (UMB).

where mb is the extensional viscosity, F(t) is the force at time t after commencement of the experiment, H(t) is the momentary sample height, R is the radius of the sample, which is the same as the radius of the plate (constant area method) and V is the crosshead velocity or velocity of deformation. All samples showed strain thinning behavior which is different to the strain hardening observed in doughs prepared with wheat our. As a consequence, the effect of the addition of gluten needs to be discussed. Stading et al. (2006) showed that the extensional viscosity of wheat our dough having protein contents of 9.512% increased with increases in protein (gluten). In contrast, sorghum dough, which did not have gluten, exhibited a much lower extensional viscosity as compared to wheat dough and the strain hardening effect was not observed. The authors concluded that the presence of gluten was responsible for the higher extensional viscosities and the strain hardening effect observed in wheat dough. Although no research studying the effects of gluten addition on the extensional viscosity of cornmeal was found, our experiments showed that the absence of gluten in cornmeal dough resulted in doughs having lower extensional viscosities (data not shown) but no strain hardening was observed. Similar trends were observed for mixtures containing gluten in quantities up to 10%. The addition of gluten to cornmeal in small quantities (8.3%) increased the extensional viscosity of the mixtures and was able to form a cohesive mass that allowed for squeezing ow testing, but it did not change the rheological trends observed in samples without gluten. Thus, the addition of gluten served to improve the squeezing ow testing and its reproducibility without qualitatively changing the behavior of the samples. Differences in extensional rheology were observed with the type of ber used, cornmeal with ATB composite showing the highest extensional viscosity. Incorporation of UMB in the cornmeal resulted in an extensional viscosity higher than that for cornmeal with no added ber, and ASB had an extensional viscosity slightly lower than the control cornmeal. However, no strain hardening phenomena was observed. Samples of ASB with cornmeal and no gluten showed lower extensional viscosity (not shown here). Hence, gluten addition probably exerted its effects through increased extensional viscosity of the system, while also forming a dough-like structure that did not show the strain hardening effect shown in wheat our dough.

where, dR/dt is the rate of increase in the radius, R is the radius of the bubble at any given time t, DP is the pressure difference, and ma is the apparent shear viscosity of the melt. A direct inference of this model is that a high melt viscosity may reduce bubble growth and expansion. The model was tested by others (Lai and Kokini, 1990) during extrusion of single starches and with values of viscosity measured by an online viscometer attached to the extruder die. Likewise, Della Valle et al. (1997) showed that, at a given moisture content and temperature, volumetric expansion increased as the melt viscosity decreases. However, an excessively low viscosity resulted in a decrease in volumetric expansion due to shrinkage and collapse after maximum expansion (Della Valle et al., 1997; Kokini et al., 1992). With an appropriate or optimum low shear viscosity, presumably the rate of bubble growth is high and bubbles can grow to larger sizes. Fan et al. (1994) have reported increased bubble diameter as well as increasing collapse with higher moisture contents but in this case both the control and cornmeal with ASB were adjusted to the same moisture content. It could be possible that the lower shear viscosity might have led to a higher degree of collapse in the sample with ASB. However, cross sections of extrudates prepared with cornmeal and ASB did not show any collapse but a very uniform bubble size distribution as opposed to cornmeal alone, which showed collapse and bubble coalescence. Both samples showed a similar radial and longitudinal expansion, suggesting that the shear viscosity could not be the sole rheological parameter governing the extrudate expansion and that extensional viscosity probably has a role to play in expansion. As the material exits the die, bubble growth is initiated. The expansion mechanism in food extrusion is as schematically shown in Figs. 3 and 4. The material expands due to a combination of die swell, mainly contributed by the elasticity of the melt, and bubble expansion driven by the pressure difference between the melt exiting the extruder and the atmospheric pressure. Chang (1992) worked on the effects on dough elasticity of the die swell. However, in foods, due to the presence of a relatively large amount of water, bubble growth driven by vapor pressure differences contributes to

Bubble Moisture Steam Air Heat

Die

Puff

Collapse

Set

Harden

Fig. 3. Mechanism of extrusion expansion.

232

D.A. Pai et al. / Journal of Cereal Science 50 (2009) 227234

Fig. 4. Cross-sectional image of extrudates containing modied bers. Extrudates from left to right contain no added ber (CTRL), alkali soluble bran (ASB), alkali-treated bran (ATB), unmodied bran (UMB).

expansion to a larger extent than die swell. Hence in this discussion, our focus is on both shear and extensional melt rheology affecting the growth of the bubbles, and neglecting the contribution of elasticity to die swell. The elasticity of the sample, as suggested, could affect the longitudinal expansion measured by LEI. It can be considered that both shear and extensional viscosities are of importance in exerting their inuence on extrudate expansion. Following the schematic expansion mechanism illustrated in Fig. 3, once bubble growth is initiated, the bubbles have to overcome the resistance created by the shear viscosity in order to grow. After the bubbles have grown sufciently, the lm of melt material between the bubbles undergoes biaxial extension (Gendron and Daigneault, 2000; Steffe, 1996). Hence, consideration of extensional viscosity becomes an essential part in the prediction of extrusion expansion from the melt rheology. The bubble expansion created by moisture vaporization increases and causes the melt lm between the bubbles to get thinner. Once a certain normal stress in the melt lm has been exceeded due to further expansion, the melt lm ruptures causing bubbles to collapse. For the bubbles to remain with the same expanded size, the melt should have an extensional viscosity high enough to withstand the extensional stress caused by bubble expansion. Guy (2001) has reported that bran particles may act as nucleating agents. In the present study, extrudates with bran had a larger number of bubbles compared to the cornmeal control. Addition of the ASB fraction to cornmeal resulted in the highest expansion which was contributed by a large number of bubbles that were smaller than for the cornmeal alone, but larger than those from the other treatments. Thus, a number of factors can be related to the product characteristics. First, ASB incorporation showed a signicantly lower shear viscosity than the control cornmeal. Although low shear viscosity tends to favor bubble growth, it may also cause a faster water loss on exiting the die due to higher diffusion of water vapor. This phenomenon has been investigated by Sun (2004) and Stange and Munstedt (2006) who showed that higher shear viscosity causes a slower diffusion of the gas in polymers resulting in better foam expansion. Second, although the extensional viscosity was lower with the ASB treatment than for the cornmeal control, the reason for similar expansion probably cannot be explained by the reduced extensional viscosity alone. The presence of high amount of bubbles of large sizes may have increased the heat transfer area of the extrudate containing ASB as compared to cornmeal. This would cause a more rapid heat and moisture loss, which promotes the attainment of the glass transition sooner than the control sample, thus preventing further bubble growth. The combination of reasons presented above, i.e. higher water vapor diffusion due to low shear viscosity and enhanced heat and mass transfer area, both leading to faster heat and water loss, may explain why cornmeal with ASB has the same

expansion and more bubbles of smaller size compared to the control despite showing a much lower shear and slightly lower extensional viscosity. Excluding the sample of cornmeal with ASB, shear viscosity increased in the order of cornmeal, cornmeal with ATB, and cornmeal with UMB whereas extensional viscosity increased in the order of cornmeal, cornmeal with UMB, and cornmeal with ATB. Comparing the rheology of cornmeal with UMB to that of cornmeal with ATB and results of the corresponding extrusion treatments, it can be concluded that the noted increase in extensional viscosity and decrease in shear viscosity is concomitant with increased extrusion expansion. This trend suggests that a high relative extensional viscosity and a low shear viscosity are important for good expansion. On the other hand, addition of both UMB and ATB to cornmeal increased shear and extensional viscosities when compared to cornmeal alone, thus increasing overall resistance to expansion, and causing lower expansion. Gendron and Daigneault (2000) have reported that for homogeneous synthetic polymers, the initial growth of bubbles requires a low extensional viscosity. Subsequently, a high extensional viscosity is advantageous such that the cell walls the melt membranes between the bubbles may withstand the stretching force experienced during the later stages of bubble growth. This non-linear increase in the extensional viscosities with increasing strain is known as a strain hardening phenomenon (Gendron and Daigneault, 2000). Although no strain hardening phenomena has been found in our study, ATB and UMB exhibited high extensional viscosities at high strains, but also had high extensional viscosities at low strains which may impede expansion due to an increased initial resistance. When looking at the overall picture involving the tested bers, that exhibiting the lowest shear and extensional viscosities, i.e. cornmeal with ASB, gave the best expansion suggesting that low viscosities create a favorable rheology for expansion. 4.2. Effect of structure of bers on rheology Solubilization, or in this case molecular dispersal, of arabinoxylan in the ATB ber caused a decrease in shear viscosity and an increase in extensional viscosity. On the other hand, ASB with the highest soluble ber content exhibited the lowest shear viscosity and the lowest extensional viscosity as well. This nding suggests that the interaction of ASB with the starch in the cornmeal is different from the interactions of UMB or ATB with the starch. The major component of ASB is hemicellulose B which is a branched arabinoxylan (Whistler, 1993). It is probable that due to its high degree of branching, ASB is able to interact with the starch in a better manner than the other two bers. Based on these results, it is postulated that ASB is more molecularly dispersed in the starch. This interaction would result in a behavior of the ber in the melt

D.A. Pai et al. / Journal of Cereal Science 50 (2009) 227234

233

similar to starch, without disrupting the melt structure and resulting in a melt rheology which favors bubble growth and expansion similar to cornmeal. The main components of the water insoluble portion of the ATB ber consist of cross-linked insoluble hemicellulose, such as hemicellulose C (15% w/w of bran, linear xylan polysaccharide with few arabinose branches), and cellulose (16% w/w) (Sugawara et al., 1994). UMB had 79% insoluble dietary ber as compared to ATB which had 52%. The high proportions of cross-linked and linear polymers that compose UMB are likely to break the continuity of the melt forming a multiphase melt. Thus, ATB forms a melt that is less discontinuous than the melt formed by UMB. In addition, ASB would be well dispersed in the melt forming a single-phase, continuous melt which has a behavior rheologically similar to starch. A similar phenomenon is found during the extrusion of synthetic polymer, where the use of highly branched polymers such as LDPE exhibits a much superior foaming and expansion behavior than the linear HDPE (Gendron and Daigneault, 2000). Similarly, Moraru and Kokini (2003) have speculated that amylopectin-rich starches expand more than amylose based starches since amylose chains align themselves in the shear eld and thus are difcult to pull apart during expansion. The linear structure of amylose allows formation of entanglements and thus increased viscosity (Moraru and Kokini, 2003). Li et al. (2004) measured the rheological properties of high amylose and amylopectin starches and their mixtures during extrusion in a modied slit viscometer and found similar results. These observations suggest that a melt containing highly branched polymers is favorable for expansion. Thus, the branched arabinoxylan fraction of hemicellulose B, which is present most predominately in the ASB ber, would lead to higher expansion. In conclusion, the results presented support the view that, to obtain good extrudate expansion, the melt should have a shear viscosity which is low enough to promote bubble growth and expansion, but high enough to prevent bubble collapse. Moreover, the melt should exhibit an extensional viscosity high enough to withstand the stretching forces during expansion, but low enough such that the melt can be stretched easily. Conditions where shear and extensional viscosities are either too low to maintain bubble integrity or too high to make bubbles grow, and resulting in difculty in melt stretching, can lead to a lower expansion. Thus, extrudate expansion depends on the interplay of shear and extensional rheologies. Among the bers tested, the molecularly dispersed and relatively highly branched ASB ber likely best interacted with the starch in cornmeal, resulting in a melt with appropriate shear and extensional viscosities similar to cornmeal and resulting in a low resistance to bubble growth and expansion. Both ATB and UMB bers performed poorly when compared to ASB due to the presence of relatively high amounts of non-dispersed cross-linked molecules, particularly in the UMB. These results indicate that the molecular dispersion of bers results in a homogenous melt that is likely to enhance expansion of ber with cornmeal. Shear and extensional rheological properties of the dough produced with these components are both important to predict the expansion of ber-enriched puffed, extruded snack and breakfast products.

References

AACC, 1991. Method 44-40 for Moisture Analysis, Approved Methods, American Association of Cereal Chemists International, St. Paul, MN. Ali, Y., Hanna, M.A., Chinnaswamy, R., 1996. Expansion characteristics of extruded corn grits. Lebensmittel-Wissenschaft und-Technologie 29, 702707. Ahmed, Z.S., 1999. Physico-chemical, structural and sensory quality of corn-based ax-snack. Die Nahrung 43, 253258.

Alavi, S.H., Rizvi, S.S.H., Harriott, P., 2003. Process dynamics of starch-based microcellular foams produced by supercritical uid extrusion. I: model development. Food Research International 36, 309319. Alvarez-Martinez, L., Kondury, K.P., Harper, J.M., 1988. A general model for expansion of extruded products. Journal of Food Science 53, 609615. AOAC, 1995. Total, Soluble and Insoluble Dietary Fiber in Foods. Ofcial Methods of Analysis, 16th ed. AOAC, Arlington, VA. Bhattacharya, M., Seethamraju, K., Padmanabhan, M., 1994. Entrance pressure-drop studies of corn meal dough during extrusion-cooking. Polymer Engineering and Science 34 (15), 11871195. Byler, L.L., Kwei, T.K., 1971. Flow behavior of polyethylene melts containing dissolved gases. Journal of Polymer Science 35 (Part C), 165176. Campanella, O.H., Peleg, M., 2002. Squeezing ow viscometry for nonelastic semiliquid foods theory and applications. Critical Reviews in Food Science and Nutrition 42 (3), 241264. Chang, C.N., 1992. Study of the Mechanism of Starchy Polymer Extrudate Expansion. Ph.D. thesis. Rutgers, The State University of New Jersey, New Brunswick. Corradini, M.G., Stern, V., Suwonsichon, T., Peleg, M., 2000. Squeezing ow of semi-liquid foods between parallel Teon coated plates. Rheologica Acta 39, 452460. Chinnaswamy, R., Hanna, M.A., 1991. Physicochemical and macromolecular properties of starchcellulose ber extrudates. Food Structure 10 (3), 229239. Della Valle, G., Vergnes, B., Collona, P., Patria, A., 1997. Relations between rheological properties of molten starches and their expansion behavior in extrusion expansion. Journal of Food Engineering 31 (3), 277296. Doner, L.W., Sweeney, G.A., Hicks, K.B., 2000. Isolation of Hemicellulose from Corn Fiber. U.S. Patent 6147206. Fan, J., Mitchell, J., Blanshard, J., 1994. A computer simulation of the dynamics of bubble growth and shrinkage during extrudate expansion. Journal of Food Engineering 23, 337356. Gendron, R., Daigneault, L., 2000. Rheology of thermoplastic foam extrusion process. In: Lee, S.T. (Ed.), Foam Extrusion. Technomic Publishing Company, Lancaster, Pennsylvania, pp. 3580. Guy, R., 2001. Raw materials for extrusion cooking. In: Guy, R. (Ed.), Extrusion Cooking: Technologies and Applications. Woodhead Pub., Cambridge, England, pp. 528. Hseih, F., Huff, H.E., Lue, S., Stringer, L., 1991. Twin-screw extrusion of sugar beet and corn meal. Lebensmittel-Wissenschaft und-Technologie 24, 495500. Hu, L., Hsieh, F., Huff, H.E., 1993. Corn meal extrusion with emulsier and soybean ber. Lebensmittel-Wissenschaft und-Technologie. (Food Science and Technology) 26 (6), 544551. Huang, H., Kokini, J., 1993. Measurement of biaxial extensional viscosity of wheat our doughs. Journal of Rheology 37 (5), 879891. Jin, Z., Hsieh, F., Huff, H.E., 1994. Extrusion-cooking of corn meal with soy ber, salt, and sugar. Cereal Chemistry 71 (3), 227234. Jin, Z., Hsieh, F., Huff, H.E., 1995. Effects of soy ber, salt, sugar and screw speed on physical-properties and microstructure of corn meal extrudate. Journal of Cereal Science 22 (2), 185194. Kokini, J.L., Chang, C.N., Lai, L.S., 1992. The role of rheological properties on extrudate expansion. In: Kokini, J.L., Ho, C.-T., Karwe, M.V. (Eds.), Food Extrusion Science and Technology. Marcel Drekker Inc., New York, N.Y., pp. 631653. Lai, L.S., Kokini, J.K., 1990. The effect of extrusion operating conditions on the online apparent viscosity of 98% amylopectin (amioca) and 70% amylose (hylon 7) corn starches during extrusion. Journal of Rheology 34, 12451266. Li, P.X., Campanella, O.H., Hardacre, A.K., 2004. Using an in-line slit-die viscometer to study the effects of extrusion parameters on corn melt rheology. Cereal Chemistry 81 (1), 7076. Lin, G.-G., Shih, H.-H., Chai, P.-C., Hsu, S.-J., 2002. Inuence of side-chain structures on the viscoelasticity and elongation viscosity of polyethylene melts. Polymer Engineering and Science 42, 22132221. Lue, S., Hsieh, F., Huff, H.E., 1991. Extrusion cooking of corn meal and sugar beet ber: effects on expansion properties, starch gelatinization, and dietary ber content. Cereal Chemistry 68, 227234. Lue, S., Hsieh, H., Huff, H.E., 1994. Modeling of twin-screw extrusion-cooking of cornmeal and sugar-beet ber mixtures. Journal of Food Engineering 21 (3), 263289. Macosko, C.W., 1994. Shear rheometry: pressure driven ows. In: Macosko (Ed.), Rheology: Principles, Measurements and Applications. Wiley-VCH Inc., New York, N.Y., pp. 237283. Mendonca, S., Grossmann, M.V.E., Verhe, R., 2000. Corn bran as a ber source in expanded snacks. Lebensmittel-Wissenschaft und-Technologie. (Food Science and Technology) 33 (1), 28. Mercier, C., Feillet, P., 1975. Modication of carbohydrate components by extrusion cooking of cereal components. Cereal Chemistry 52, 283. Moore, G., 1994. Snack food extrusion. In: Frame, N. (Ed.), The Technology of Extrusion Cooking. Blackie, A&P, London, UK, pp. 110143. Moraru, C.I., Kokini, J.L., 2003. Nucleation and expansion during extrusion and microwave heating of cereal foods. Comprehensive Reviews in Food Science and food Safety 2, 120138. Ozboy, O., Koksel, H., 2000. Effects of sugar beet ber on the quality and dietary ber content of extrusion products. Zuckerindustrie 125 (11), 903905. Onwulata, C.I., Konstance, R.P., Strange, E.D., Smith, P.W., Holsinger, V.H., 2000. High-ber snacks extruded from triticale and wheat formulations. Cereal Foods World 45 (10), 470473. Padmanabhan, M., Bhattacharya, M., 1989. Extrudate expansion during extrusion cooking of foods. Cereal Foods World 34 (11), 945949.

234

D.A. Pai et al. / Journal of Cereal Science 50 (2009) 227234 Sun, S.F., 2004. Physical Chemistry of Macromolecules: Basic Principles and Issues, second ed. Wiley, New York. Suwonsichon, T., Peleg, M., 1999. Imperfect squeezing ow viscometry of mustards with suspended particulates. Journal of Food Engineering 39, 217226. van Vliet, T., Janssen, A.M., Bloksma, A.H., Walstra, P., 1992. Strain hardening of dough as a requirement for gas retention. Journal of Texture Studies 23 (4), 439460. Wang, L., Ganjyal, G.M., Jones, D.D., Weller, C.L., Hanna, M.A., 2005. Modeling of bubble growth dynamics and nonisothermal expansion in starch-based foams during extrusion. Advances in Polymer Technology 24 (1), 2945. Whistler, R.L., 1993. Hemicelluloses. In: Whistler, R.L., BeMiller, J.N. (Eds.), Industrial Gums, third ed. Academic Press Inc., San Diego, USA. Wikstrom, K., Bohlin, L., 1999. Extensional ow studies of wheat our dough. II. Experimental method for measurements in constant extension rate squeezing ow and application to ours varying in breadmaking performance. Journal of Cereal Science 29 (3), 227234.

Padmanabhan, M., Bhattacharya, M., 1993. Effect of extrusion processing history on the rheology of corn meal. Journal of Food Engineering 18 (4), 335349. Schwartzberg, H.G., Wu, J.P.C., Nussinovitch, A., Mugerwa, J., 1995. Modeling deformation and ow during vapor-induced pufng. Journal of Food Engineering 25 (3), 329372. Seethamraju, K., Bhattacharya, M., 1994. Effect of ingredients on the rheological properties of extruded corn meal. Journal of Rheology 38 (4), 10291044. Stading, M., Oom, A., Pettersson, A., Edrud, S., 2006. Extensional rheology of cereal protein systems. In: Annual Transactions of the Nordic Rheology Society, vol. 14. Stange, J., Munstedt, H., 2006. Rheological properties and foaming behavior of polypropylenes with different molecular structures. Journal of Rheology 50 (6), 907923. Steffe, J.F., 1996. Rheological Methods in Food Process Engineering, second ed. Freeman Press, East Lansing, MI. Sugawara, M., Tetsuya, S., Totsuka, A., Takeuchi, M., Ueki, K., 1994. Composition of corn hull dietary ber. Starch/Staerke 46, 335337.

Journal of Cereal Science 50 (2009) 235240

Contents lists available at ScienceDirect

Journal of Cereal Science


journal homepage: www.elsevier.com/locate/jcs

Impact of the baking kinetics on staling rate and mechanical properties of bread crumb and degassed bread crumb
Alain Le-Bail*, Kahina Boumali, Vanessa Jury, Fadhel Ben-Aissa, Ruben Zuniga
` ENITIAA, UMR GEPEA CNRS 6144, UNAM, Rue de la Geraudiere, 44322 Nantes Cedex 3, France

a r t i c l e i n f o
Article history: Received 5 February 2009 Received in revised form 26 May 2009 Accepted 29 May 2009 Keywords: Bread baking Crumb Texture Staling Amylopectin

a b s t r a c t
This paper presents a study on the impact of baking conditions on crumb staling. Breads were baked at 220  C, 200  C and 180  C corresponding to 6, 8 and 10 min to rise the temperature to 98  C in the crumb (heating rates 13, 9.8 and 7.8  C/min respectively with an initial temperature of 20  C). A new protocol has been developed, consisting in baking a slab of degassed dough in a miniaturized oven to mimic the baking conditions of conventional bread making. Texture tests were done during staling on degassed crumb and on conventional crumb. Calorimetry tests showed that during storage, amylopectin recrystallisation occurred before crumb stiffening. A rst order kinetics model was used to t the evolution of the crumb texture (Youngs modulus) and of the recrystallisation of amylopectin. The results showed that the hardening of the crumb during staling occurred after retrogradation of amylopectin. In addition, the staling rate was faster for faster baking kinetics. A mechanical model showed that the relative Young modulus is proportional to the square of the relative density of the crumb. 2009 Published by Elsevier Ltd.

1. Introduction Texture is maybe the most studied physical parameter of bread. Texture of the crumb, is linked to the mechanical properties of the material constituting the foam and to its structure (Liu and Scanlon, 2003). The cellular structure of the crumb can be characterized by different parameters as presented by Scanlon and Zghal (2001). These authors proposed that two macroscopic phases (crumb cells and cell walls solids) should be considered. Digital image analysis studies of bread-crumb structure allow the determination of the characteristics of these macroscopic phases. 2D images are used in most studies, whereas 3D images (ie. using X-ray tomography) can bring more accurate information. The average size of the cell can rst be considered. The distributions of the size of the cells, and in turn the wall thickness are important parameters that affect the texture of cellular foam like bread crumb. The size of the cells and the thickness of the walls depend on several factors linked to the

Abbreviations: E, Young modulus of the degassed crumb MPa); E*, Young modulus of the bread crumb (cellular) (MPa); E0, Young modulus of the degassed crumb at initial time (MPa); EN, Young modulus of the degassed crumb at end of staling (MPa); DH0, melting enthalpy of amylopectin at initial time (J/gdm); DHN, melting enthalpy of amylopectin at end of staling (J/gdm); t, : time (days); s or s, time constant of rst order model (days); r or r*, density of the degassed crumb and of cellular crumb (kg/m3); 3 DL/L, strain of the sample during compression (dimensionless). * Corresponding author. Tel.: 33 2 51 78 54 54; fax: 33 2 51 78 5467. E-mail address: alain.lebail@enitiaa-nantes.fr (A. Le-Bail). 0733-5210/$ see front matter 2009 Published by Elsevier Ltd. doi:10.1016/j.jcs.2009.05.008

processing conditions and to the formulation. The duration of fermentation for example can have an impact on the cell distribution. During fermentation, bubble to bubble interaction starts to occur for dough porosity larger than around 50%. Therefore, a longer fermentation may result in a less uniform cell distribution. Baking at a lower temperature than for conventional baking, such as in the case of partial baking of bread usually results in a lower oven rise. The structure of the crumb of part baked bread is most of the time made of smaller cells (vs. conventional baking) with a narrow size distribution. The location of the breads in the oven can also induce an artifact. Indeed, the loaves that are installed close to the wall of the oven are exposed to different radiative heat transfer conditions than at the centre of the oven and may heat up at a different rate than the bread placed at the centre of the oven depending on the design of the oven. The bread close to the wall of the oven may also have a thicker crust. Baik and Chinachoti (2000) showed that the staling rate of bread with crust was faster than the staling of bread without crust. This was attributed to the equilibration of moisture between the crust and the crumb. During storage, the crust tends to trap moisture from the crumb, resulting in a dehydration of the crumb and in a faster staling in the case of bread with crust. Another point is that the duration of baking and the time-temperature history cannot be considered as strictly homogenous in all locations of the crumb in a given bread. The crumb located just below the crust undergoes a much longer baking than the crumb located at the very centre of the bread as shown by Marston and Wannan (1976). This difference in the

236

A. Le-Bail et al. / Journal of Cereal Science 50 (2009) 235240

degree of baking between the centre and the surface of the bread increases with the size of the bread Patel et al. (2005). One impact of this is that the degree of starch gelatinization and the degree of starch granule swelling and destructuration can be non-uniform in a same bread as shown by Marston and Wannan (1976). Indeed, the duration of baking has an inuence on the degree of destructuration of starch granules as shown by Borczak et al. (2008) and therefore a longer duration is supposed to have an impact on the staling rate as proposed by Boumali et al. (2008). The review presented above shows that the processing conditions, the crust crumb ratio, the location considered in a bread . interact with the structure and therefore with the texture of the crumb. If one wants to evaluate objectively the impact of formulation or of processing on the texture of the crumb, the size of the crumb sample, the spot considered in a given slice of bread to make the texture test can induce an artifact in the data that will be obtained. Moreover, a given batch of breads cannot be considered as strictly uniform, even though a strict procedure and protocol are used. The storage temperature also plays an important role in the kinetics of staling. Storing the bread at chilling temperature (i.e. 4  C) accelerates the staling in comparison to storage at room temperature Guinet and Godon (1994). Standardized methods are available to measure the texture of the crumb. For example, the American Association of Cereal Chemistry (AACC, 1995) proposes a standard method based on the compression of a slice of crumb of around 25 mm thickness. The texture of the bread crumb is often determined as the force in Newtons measured at a selected compression ratio (i.e. 25%). The percentage of deformation is always subject to discussion as the onset of compression is determined from the increase of the force when the plunger is contacting the crumb. There is thus a certain inaccuracy in the determination of the percentage of deformation. For example, if the slice of crumb is not perfectly at, then the force may rise before the compression is effectively applied to the slice of crumb; this may induce an error in the origin of the deformation and therefore in the force that will be determined after a given ratio of compression. Therefore, at least 36 repetitions must be done to obtain for the mean value one representative data of the texture of crumb. Another problem concerns the fact that small breads are often used for laboratory tests. In this case, the number of slices that can be obtained in a given bread is not very large. Sometimes, large cells can be present in the crumb (i.e. in the case of a baguette) with a size similar to the thickness of the slice. In this case, the compression test can result in wrong measurements. The objective of this work was to develop a new protocol to assess the staling of bread crumb that allows both a good control of the processing (baking) conditions and a good control of the conditions of the texture test. In other words, the objective of this new protocol was to overcome the difculties presented in the introduction of the manuscript. The protocol is based on the baking of a miniaturized sample of degassed dough with very well dened conditions. First results on the kinetics of staling are presented and are compared to the texture of real crumb from bread baked in similar conditions. 2. Materials and methods 2.1. Raw materials The white our (Moulins Soufet Pantin Pornic, France) used for the formulation had the following properties: 0.53 Ash content, 10.58% protein on dry matter, falling number 402 s, Zeleny 38, alveograph parameters W 183, Pm/L 0.62. The recipe was as follows: 100 g of our, 58 g of water, 5 g of compressed yeast (Michard SAS Theix, France), 2 g of salt (Esco, Levallois-Perret, France) and 1 g of improver (PURATOS Groot-Bijgaarden Belgium).

2.2. Bread-making procedure Ingredients were mixed in an SP10 spiral mixer (VMI, Montaigu, France) during 2 min at 100 rpm and 7 min at 200 rpm. At the end of the mixing, dough temperature was between 18 and 20  C. Dough was then rested for 15 min. It was then divided and moulded in a divider moulder (Bongard, Holtzheim, France) to obtain 30 dough pieces of 70 g each (initial batch of dough was 2100 g). Dough pieces were pierced with 5 holes (needle piercing) and were placed in a fermentation cabinet (Panimatic, Souppes s/Loing, France) for 60 min at 35  C, 95% relative humidity. The expansion ratio of the dough at the end of fermentation was 4.8. This relatively high expansion ratio is justied by the fact that no cuts were done on the surface of the breads before baking. Baking was done with two different protocols. In the case of normal bread making, baking was done in a static sole oven (MIWE Condo, Arnstein, Germany) during 20 min at selected temperatures with 0.5 L of steam at start of baking (0.4 m3 oven volume). Bread core temperature was measured with a K type thermocouple (0.3 mm diameter). Alternatively, the dough was baked as degassed dough. In this case, at the end of fermentation, the dough was gently squeezed manually. Then 20 g of dough was placed in a plastic pouch and was vacuum treated for 1 min to remove all porosity. The pouch was then sealed and was installed in a miniature baking system made of two Peltier elements (5 cm 5 cm) Jiang et al. (2008). The dough sample was attened between the two Peltier elements. A spacer (3 mm) located between the two Peltier elements was used to control the thickness and the parallelism of the dough sample. A time-temperature evolution was then programmed in a control unit and the baking was done, including temperature rise at a pre selected heating rate, a possible plateau at crumb baking temperature (98  C) and then a cooling step (40 min to reach 20  C). This procedure offered a perfect control of the time temperature history of the dough during baking. 2.3. Kinetics of baking Breads were baked in the oven (MIWE-Germany) with voute (upper heating) and sole (deck heating) temperatures of 180  C, 200  C and 220  C. Temperatures in the bread were logged with K type thermocouples (0.3 mm diameter) connected to a data logger (SA 32 AOIP France). The baking duration (corresponding to partial baking) was considered at the time needed to reach the temperature of 98  C at the centre of the bread (starting from 20  C). Breads were then cooled down in ambient air at around 20  C (40 min). The analysis of time temperature histories showed that it took 6, 8 and 10 min until a central temperature of 98  C is reached for oven temperatures of 220  C, 200  C and 180  C respectively, corresponding to heating rates of 13, 9.8 and 7.8  C/min respectively (initial temperature 20  C and baking temperature of 98  C). These three conditions were then considered for the baking of the degassed dough samples, considering that reaching this temperature corresponded to the case of partial baking of bread. A linear evolution of the temperature during heating was programmed on the temperature controller of the miniaturized baking system and samples were heated at the given heating rate until a temperature of 98  C was reached. Then the samples were cooled down to 20  C. 2.4. Analysis of the texture of bread and degassed crumb Measurements were performed after 1 h equilibration at room temperature. The texture of bread was determined at day 0 at 20  C by a compression test with a JJ LLOYD LR5K texturometer. For each measurement, 15 slices of 15 mm were used for compression tests.

A. Le-Bail et al. / Journal of Cereal Science 50 (2009) 235240

237

A cylindrical plunger of 20 mm diameter was used. The velocity of the plunger was 60 mm/min and compression was made up to 40% deformation. The displacement was logged when the force was higher than 0.1 N. The linear section of the forcedeformation plot was used to determine the Young modulus of the crumb. The Young modulus is dened as proposed in equation (1) from the slope of the stress (Force/surface of the plunger) as a function of relative displacement or strain (3 DL/L).

Ds D3

determine the amount of freezable water and the melting enthalpy of amylopectin. Besides, the two additional scans provided two other values of the melting enthalpy of amylopectin. The melting enthalpy of amylopectin and of freezable water DH was determined in Joules per gram of dry matter (J/gdm). The amount of freezable water was obtained by dividing the melting enthalpy of ice by the latent heat of melting of pure water. The results were expressed as grams of freezable water per gram of crumb on a dry matter basis (gwater/gdm).

(1)
2.7. Conditions of ageing tests and models used to describe the staling Ageing was done at 20  C. The samples (breads, slabs of degassed crumb and Eppendorf test tubes containing some pieces of degassed baked crumb for DSC tests) were packed in sealed plastic pouches and were installed in a temperature controlled cabinet. The storage was carried out for 14 days. Texture and DSC tests of baked crumb and of degassed crumb were done at day 0, 4, 6, 8, 12 and 14. First order kinetic models were used to model the evolution of the melting enthalpy of amylopectin DH(t) (equation (2)) with DH0 and DHN being the melting enthalpy at initial time and for long storage (time N). A similar equation was used to model the hardening (from the Young modulus E(t)) of the bread crumb and of the degassed bread crumb (equation (3)) with E0 and EN being the Young modulus at initial time and for long storage (time N). Characteristic times sand s of equation (2) and (3) were used to compare the kinetics of the two phenomena, namely amylopectin retrogradation and crumb hardening.

Similar tests were done using the degassed crumb slabs. Disks of 8 mm diameter and ca. 3 mm thick were obtained from the slab of baked crumb. The disks were installed in a Dynamical Mechanical Analyser (DMA Q800 TA Waters instruments 78-Guyancourt France) at 20  C. The DMA was used for compression tests. The interest of the DMA was that it allowed the measure with small samples. A parallel plate system was used. The compression was done with a compression rate of 2%/min until 10% strain. Three slabs were considered for each baking condition so that 6 to 10 disks were available for compression tests of degassed crumb. The Young modulus of the degassed crumb was determined as previously from the linear section of the stressstrain plot. The Young modulus of the crumb is quoted as E*, whereas the Young modulus of the degassed crumb is quoted as E. The relative Young modulus of the crumb is the ratio E*/E. 2.5. Analysis of the crumb density The density of the degassed crumb was determined from the thickness of the disks measured by the DMA at the beginning of each compression test. A helium picnometer (Accupyc 1330 picnometer retailed by Micrometrics, Creil, France produced by Norcross, USA) has also been used to check if any internal porosity was still present in the baked crumb. The effective density of the bread crumb was the density of the crumb containing gas cells (normal bread baked with the conventional baking process). It was measured by cutting a cylinder of crumb in a slice of crumb of known thickness. Knowing the weight and the dimensions of the sample, the crumb density was calculated. The effective crumb density is quoted as r*, whereas the density of the degassed crumb is quoted as r in the text. The relative density of the crumb is the ratio r*/r. 2.6. Calorimetry Samples were taken from the central section of the slabs of degassed crumb that were baked in the miniaturized baking system. They were placed in several Eppendorf test tubes just after the baking process. These samples were stored at 20  C. At day zero and then each 2 days, 4 DSC tests were carried out with a PYRIS 6 DSC (PERKIN ELMER). One Eppendorf test tube was open for each test. 4 DSC pans were prepared to run 4 different DSC tests with crumb coming from the same sample. The remaining pieces of samples were weighted and the moisture was checked by drying tests in a drying cabinet at 104  C for 24 h. Each DSC Pan contained ca. 5070 mg of crumb. A Mettler Toledo balance (accuracy 0.01 mg) was used. The two rst DSC runs were programmed as follow. The DSC temperature was equilibrated at 20  C for 5 min. Then the sample was cooled down to 60  C at 3  C/min and was maintained at 60  C for 10 min. Then a scan was performed from 60  C to 100  C with a heating rate of 3  C/min. The two other runs were similar to the previous ones except that the scan was skipping the subzero phase (no scan down to 60  C). The scan was performed directly from 20  C to 100  C. The two rst scans were used to

DHt DHN DH0 DHN e


 Et EN E0 EN e t 

 (2)

s0

(3)

3. Results and discussion The water content of degassed crumb was slightly higher than that of the crumb of the real bread (3% on a humid basis). The difference in moisture content between bread crumb and degassed crumb is due to the moisture lost during baking (ca. 2.2% on a wet basis for both bread crumb and degassed crumb). The average moisture contents of breads after baking were 35.37, 38.32 and 39.02% (wet basis) for baking at 220, 200 and 180  C respectively. During storage, both samples tend to lose water (2% on a wet basis during a 12 days storage), even though they were hermetically sealed in plastic pouches. Concerning the bread crumb samples (normal bread), the evolution of moisture content can be attributed to moisture transfer from the crumb to the crust (moisture equilibration). On the other hand, degassed crumb samples were stored in Eppendorf test tubes. The moisture lost for these samples may be explained by the small size of the sample and by some possible trapping of moisture by the polyethylene during storage or again by leakage toward the atmosphere. The evolution of the enthalpy of melting of amylopectin was measured during storage for the degassed crumb samples. Data were treated by calculating the logarithm of the difference of melting enthalpy at end of storage (DHN) when the enthalpy was not changing (this occurs at ca. day 12) and the melting enthalpy at a given time during storage as presented in equation (4). An example is given in Fig. 1.

238

A. Le-Bail et al. / Journal of Cereal Science 50 (2009) 235240

1.2

DAYS at 20 C
0 0 -1 H (J/g dm) -2 -3 -4 -5 -6 y = -0.3143x - 1.9219 R2 = 0.9873 y = -0.3088x - 1.0718 R2 = 0.9894 2 4 6 8 10 12 NORMAL BREAD CRUMB DEGASSED BREAD CRUMB

H (J/g dm)

1 0.8 0.6 0.4 0.2 0 0 2 4 6 8 10 12

DAYS at 20 C
Fig. 1. Evolution of the melting enthalpy of amylopectine (DH) during storage. Results obtained with degassed crumb baked with a heating rate of 9.8  C/min.

t ln DHN DHt ln DHN DH0 0

(4)

The time constant s was obtained from the slope of the linear regression of the logarithm of the enthalpy difference between nal value (long storage) and current value as a function of time. The slope of the linear regression plots is equal to 1/s as shown in equation (4) with s having the same unit as time (in days). Similar treatment of data obtained from compression tests was done with degassed crumb and with normal bread crumb. Experimental plots are shown in Fig. 2 for the Young modulus and in Fig. 3 for the data treatment that give the time constant s of equation (3) which was used to t the experimental results. The large number of repetitions for each measurement explains the small standard deviation as indicated by the error bars in the graphs. The evolution of the amount of freezable water is shown in Fig. 4. The values are given in grams of freezable water per gram of dry matter. The drift of the total moisture content of the sample during storage has been taken into account. The experimental values have been corrected with the moisture that was lost by the sample and has been accommodated in the results that are presented. In other words, it was considered that the amount of moisture that was lost by the sample would have been freezable water, so that the artifact induced by the slight evolution of the moisture content of the crumb is withdrawn in the results (the maximum correction is 0.02 at day 12). One can clearly see that the amount of freezable water is decreasing during staling as shown by Ribotta and Le-Bail (2007a). In parallel, one can observe that the melting enthalpy of amylopectin crystallites increases during storage as shown by Ribotta and Le-Bail (2007b). One can observe in Fig. 4 that the amount of freezable water was stabilized after ca. 6 days of storage. This period of time also corresponded to the time for which the melting enthalpy of amylopectin reached its

Fig. 3. Plot of the logarithm of the evolution of the Young modulus vs. time (from data in Fig. 2). The slope of the linear t gives the time constant s of model presented in equation (3).

maximum (see in Fig. 1). This observation conrms the fact that some moisture is trapped by crystallites of amylopectin during storage resulting in a reduction of the amount of freezable water. In turn, the transfer of moisture from the crumb components (gluten, starch) toward crystallites of amylopectin (as described by Gray and Bemiller (2003) resulted in a hardening of the crumb. In addition, it seems that for the slow heating rate (180  C10 min to reach 98  C), it took more time to reach the stabilization (around 8 days instead of 6 see Fig. 4). Even though the difference is not very large, it could be explained by the difference in baking kinetics. Indeed, during baking, the swelling of the starch granules result in a leaching of amylose outside the starch granules. The swelling and the destructuration of the starch granules is a progressive phenomenon that occurs during the rst 10 min of baking at 98  C as shown by Boumali et al. (2008) (using the same recipe and ingredients as in the present study). In the case of baking at 180  C (longer baking), the degree of gelatinization of the starch components is supposed to be higher than for short baking, resulting in a more jellied structure that delays the diffusion of water from the matrix (gluten network) towards amylopectin crystallites that are located mainly in the ghosts of the starch granules. The time constant of the rst order kinetics models are presented in Fig. 5 with conditions 1, 2 and 3 corresponding to baking at 180  C, 200  C and 220  C or 10, 8 and 6 min to reach 98  C). It appeared that the kinetics of hardening of the crumb, being the degassed crumb or the bread crumb was signicantly
0.55

FREEZABLE WATER (g/g dm)

LN(Eoo-E(t))

Freezable Water g/gdb - 180C 0.50 Freezable Water g/gdb - 200C Freezable Water g/gdb - 220C

0.50 0.45 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00 0

NORMAL BREAD CRUMB DEGASSED BREAD CRUMB

E (MPa)

0.45

0.40

0.35 4 6 10 0 1 2 3 4 5 6 7 8 9 10 11 12

DAYS at 20 C
Fig. 2. Evolution of the Young modulus of normal bread crumb and of degassed bread crumb during storage (Baking at 200  C).

DAYS at 20 C
Fig. 4. Evolution of the amount of freezable water during storage (g freezable water/g dry mater).

A. Le-Bail et al. / Journal of Cereal Science 50 (2009) 235240

239

E CRUMB / E DEGAZED CRUMB

BREAD CRUMB 5

DEGAZED CRUMB

AMYLOPECTINE

1.00
220C 220C 200C 200C 180C 180C

TIME CONSTANT (DAYS)

4.5 4 3.5 3 2.5 2 1.5 1 2 3

0.10

EXPERIMENTAL EXPERIMENTAL

ASHBY MODEL ASHBY MODEL

0.01 0.1 1.0

(DENSITY CRUMB / DENSITY DEGAZED CRUMB)


Fig. 6. Evolution of the relative Young modulus (E* crumb/E degassed crumb) in function of the relative crumb density (r CRUMB/rDEGASSED CRUMB). A comparison is done with the model proposed by (Gibson and Ashby, 1997) (see equation (5). The groups of points (circles) correspond to the 4 repetitions done for each baking condition.

Fig. 5. Time constant (s in days see eq. (2)&(3)) of the different phenomena tracked during staling : melting enthalpy of amylopectine, texture of bread crumb and texture of degassed crumb. Condition 1, 2, 3 corresponds to baking at 180, 200 and 220  C.

slower than the kinetics of formation of amylopectin crystallites. The time constant of the kinetic models obtained for hardening of both bread crumb and degassed bread crumb were around two times higher than for the amylopectin recrystallisation as shown in Fig. 5. Therefore, the faster the baking (from 6 min at 220  C10 min at 180  C to reach 98  C), the faster the staling. This trend is also conrmed by Fig. 4 in which it has been observed that in the case of baking at 180  C (slowest heating rate 10 min to reach 98  C), it took 8 days to reach a stabilized value of the amount of freezable water whereas for the other baking conditions, 6 days were sufcient. Concerning the texture of the crumb, the moisture transfer from the crumb to the crust has to be considered. Baking at 220  C resulted in a higher dehydration of the bread and therefore of the crumb; indeed, the average moisture content of the bread crumb was 35.37, 38.32 and 39.02% on a humid basis for baking at 220, 200 and 180  C respectively (normal baking conditions). One can assume that there is an interaction between the moisture content of the crumb and the staling rate. Tests done with the degassed crumb (DSC tests and texture of degassed crumb) showed the same trend than for the normal bread crumb. In the case of the degassed crumb, the moisture of all samples was very close from one sample to another. A difference appears in Fig. 5 between the normal bread crumb and the degassed crumb in the case of the long baking (condition 1, baking at 180  C or 10 min to reach 98  C). In this case, the staling was much slower (higher value of the time constant) for the normal bread than for the degassed crumb. The mechanical behavior of porous material has been studied by several researchers (Guessasma et al., 2008, Keetels et al., 1996). Mechanical models can be used to describe the relationship between effective mechanical properties of a cellular solid, the mechanical properties of the material constituting the cellular system and the structure of the cellular system. A macroscopic model has been proposed by Gibson and Ashby (1997) and is presented in equation (5). E* and E represent the Young modulus of the cellular solid and of the solid material constituting the cellular system respectively. r* and r represent the density of the cellular system and of the solid material respectively. It is important to point out that the model presented in equation (5) has been established for a foam with open cells.

the cellular crumb was determined during compression tests from the mass of the sample and from its geometry. The results obtained in our case are presented in Fig. 6. The relative Young modulus is higher than predicted by the model of (Gibson and Ashby, 1997) proposed in equation Error: Reference source not found. However, the data t quite well with the data shown in gure 9 of (Scanlon and Zghal, 2001). The difference between experimental data and the model of (Gibson and Ashby, 1997) could be explained by the presence of partially closed cells in the crumb; indeed, during compression test, the mechanical properties of the cellular matrix comes from the mechanical resistance of the material constituting the matrix plus the force needed to compress the gas contained in closed cells. The relative Young Modulus should have been lower to t the theoretical model. The ratio usually The baking temperature does not seem to have a signicant impact on the relative Young modulus which remains almost constant whatever the baking temperature was, baking at higher temperature (short baking) may result in a less uniform distribution of the cells and also in thicker wall of the cells, resulting in a higher Young modulus; this could be an explanation to the fact that the relative Young modulus remains high for the highest baking temperature in comparison with the lowest baking temperature. It is not possible to propose another analysis of the situation.

4. Conclusion In this paper, the impact of the kinetics of baking on the staling rate of bread has been investigated. Results showed that rapid baking (6 min to reach 98  C) resulted in a faster staling. Staling was evidenced by texture tests of both degassed bread crumb and of cellular bread crumb. The formation of amylopectin crystallites observed by calorimetry occurred before the hardening of the crumb. The time constant of the rst order kinetic model describing the hardening of the crumb was 40% longer than the time constant describing the formation of amylopectine crystallites. In addition, a reduction of the amount of freezable water was observed during storage. This conrms the hypothesis that free water was trapped by amylopectin crystallites during storage. A mechanical model based on the relative Young modulus and on the relative density of the crumb showed that the relative Young modulus (ratio of the modulus of the cellular crumb vs. the modulus of the degassed crumb) was proportional to the square of the relative density of the crumb. This paper showed that the control of the baking time and of the kinetics of baking is an important parameter to consider in

E* E

r* r

!2 (5)

The density of the solid crumb has been determined by helium pycnometry and was 1.23 kg m3. The effective density of

240

A. Le-Bail et al. / Journal of Cereal Science 50 (2009) 235240 Boumali, K., Jury, V., Chevallier, S., Queveau, D., Le-Bail, A., 2008. Impact du degre de cuisson sur la cinetique de rassissement. In: B. Industries des Cereales Cnam Case 306-292, rue Saint-Martin 75141 Paris Cedex 03-. 59es JOURNEES ` TECHNIQUES des INDUSTRIES CEREALIERES 2008, Paris, France. www. industriesdescereales.com. Gibson, L.J., Ashby, M.F., 1997. Cellular Solids- Structure and Properties. Cambridge University press, Cambridge. Gray, J.A., Bemiller, J.N., 2003. Bread staling: molecular basis and control. Comprehensive Reviews in Food Science and Food Safety 2, 121. Guessasma, S., Babin, P., Della Valle, G., Dendievel, R., 2008. Relating cellular structure of open solid food foams to their Youngs modulus: Finite element calculation. International Journal of Solids and Structures 45, 28812896. Guinet, R., Godon, B., 1994. La Panication Franaise, Techniques et Documentation. Lavoisier ed. Jiang, Z., Le-Bail, A., Wu, A., 2008. Effect of the thermostable xylanase B (XynB) from Thermotoga maritima on the quality of frozen partially baked bread. Journal of Cereal Science 47, 172179. Keetels, C., vanVliet, T., Walstra, P., 1996. Relationship between the sponge structure of starch bread and its mechanical properties. Journal of Cereal Science 24, 2731. Liu, Z., Scanlon, M.G., 2003. Predicting mechanical properties of bread crumb. Food and Bioproducts Processing 81, 224238. Marston, P.E., Wannan, T.L., 1976. Bread baking the transformation from dough to bread. Bakers Digest 50, 24. Patel, B.K., Waniska, R.D., Seetharaman, K., 2005. Impact of different baking processes on bread rmness and starch properties in breadcrumb. Journal of Cereal Science 42, 173184. Ribotta, P., Le-Bail, A., 2007a. Thermo-physical assessment of bread during staling. LWT 40, 879884. Ribotta, P.D., Le-Bail, A., 2007b. Thermo-physical assessment of bread during staling. LWT Food Science and Technology 40, 879884. Scanlon, M.G., Zghal, M.C., 2001. Bread properties and crumb structure. Food Research International 34, 841864.

the case of staling studies. Future investigations are envisaged to better understand the impact of the degree of baking on the mechanical behavior of starch biopolymers and on the destructuration of starch granules. Acknowledgements This study was carried out with nancial support from the European Commission, FP6, Thematic Area Food Quality and Safety, FOOD-2006-36302 EU-FRESH BAKE. It does not necessarily reect its views and in no way anticipates the Commissions future policy in this area. Luc GUIHARD, Christophe COUEDEL and Delphine QUEVEAU are gratefully acknowledged for their help. References
AACC, 1995. Method 74-09, Bread Firmness by Universal Testing Machine, vol. II. American Association of Cereal Chemists, Approved Methods AACC, St Paul, MN, USA. Baik, M.-Y., Chinachoti, P., 2000. Moisture redistribution and phase transitions during bread staling. Cereal Chemistry 77, 484488. Borczak, B., Pisulewski, P., Sikora, M., Krawontka, J., Perronnet, A., Roelens, G., Chevallier, S., Boumali, K., Le-Bail, A., 2008. Impact du procede de cuisson sur lindice glycemique dun pain de type franais. In: B. Industries des Cereales CnamCase 306-292, rue Saint-Martin 75141 Paris Cedex 03-. 59es JOURNEES ` TECHNIQUES des INDUSTRIES CEREALIERES 2008, Paris, France. www. industriesdescereales.com.

Journal of Cereal Science 50 (2009) 241247

Contents lists available at ScienceDirect

Journal of Cereal Science


journal homepage: www.elsevier.com/locate/jcs

Morphologies and microstructures of cornstarches with different amyloseamylopectin ratios studied by confocal laser scanning microscope
Pei Chen a, b, c, Long Yu a, c, *, George Simon b, **, Eustathios Petinakis c, Katherine Dean c, Ling Chen a
a

Centre for Polymers from Renewable Resources, ERCPSP South China University of Technology, Guangzhou, PR China Department of Materials Engineering, Monash University, Melbourne, Vic 3169, Australia c CSIRO, Materials Science and Engineering, Clayton South, Melbourne, Vic 3169, Australia
b

a r t i c l e i n f o
Article history: Received 11 March 2009 Received in revised form 1 June 2009 Accepted 2 June 2009 Keywords: Starch Amylose Microstructure Confocal image Gelatinization

a b s t r a c t
The morphologies and microstructures of cornstarches with different amyloseamylopectin ratios (waxy: 0/100; maize: 23/77; Gelose 50: 50/50; and Gelose 80: 80/20) were studied by a confocal laser scanning microscope (CLSM). The temperature-induced changes of the cornstarch granules in excess of water were also studied under CLSM. Acid hybridization of starch by HCl was used to enhance the difference between amorphous and crystalline ranges. It was found that the high-amylose starches (G50 and G80) were brighter than those of low-amylopectin starches (waxy and maize) under confocal laser light, and the average (decreasing) uorescence intensity sequence of the granules was G80 > G50 > maize > waxy. Waxy and maize starches showed clear internal cavities and channels, whilst G50 and G80 had bright cores. Sharp growth ring structures can be clearly observed for low-amylose starches (waxy and maize) after acid hydrolysis. Gelatinization of all starches starts at the hilum and the adjacent of the channels, and spreads rapidly to the periphery. This work is the rst time that three-dimensional images of partly gelatinized granules have been constructed and presented from different confocal images, which allows further exploration of the mechanisms of gelatinization. 2009 Elsevier Ltd. All rights reserved.

1. Introduction The starch granule is a heterogeneous material containing both amorphous and crystalline regions. In terms of their chemical structure, most native starches are a mixture of amylose (a linear structure of alpha-1,4-linked glucose units) and amylopectin (a highly branched structure of short alpha-1,4 chains linked by alpha-1,6 bonds) (Zobel, 1984, 1988). The physicochemical properties of starch and its suitability for various industrial applications depend on the proportion, composition, and structure of the amylose and amylopectin molecules, the amyloseamylopectin ratio, the length and pattern of the amylopectin branches, the substitution of the glucose monomers, and the size and modality of the granules (Boren et al., 2008). Determination of the molecular distribution of amylose and amylopectin molecules in the native starch granule, along with the analysis of starch granule internal and surface topography, is important in order to gain an understanding of the behavior of starch and starch-based products.

* Corresponding author. CSIRO, Materials Science and Engineering, Clayton South, Melbourne, Vic 3169, Australia. Tel.: 61 3 9545 2797; fax: 61 3 9544 1128. ** Corresponding author. Tel.: 61 3 9905 4936; fax: 61 3 9905 4934. E-mail addresses: long.yu@csiro.au (L. Yu), george.simon@monash.edu.au (G. Simon). 0733-5210/$ see front matter 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2009.06.001

Various native starches exhibit signicant differences in granular appearance and microstructure (Chen et al., 2006; Gregorov et al., 2006), and research involving the imaging of the morphology of starch granules and the subsequent changes which take place as they undergo digestion has been carried out for a long time. With the improvement and digitisation of analytical and microscopic techniques, a wider range of microscopic techniques can be usefully applied to this eld, giving rise to a large amount of new and structural information about starch. The relevant imaging technologies developed include scanning electron microscopy (SEM) (Chanzy et al., 1990; Evers et al., 1970; Yamaguchi et al., 1979), transmission electron microscopy (TEM) (Chanzy et al., 1990; Helbert and Chanzy, 1996; Oostergetel and Van Bruggen, 1989), atomic force microscopy (AFM) (Baldwin et al., 1998; Neethikajan et al., 2008), and confocal laser scanning microscopy (CLSM) (Boren et al., 2008; Chung and Lai, 2006; van de Velde et al., 2002), assisted by other equipments able to probe microstructure such as X-ray(Jenkins et al., 1993; Zobel and Illinois, 1988), small angle neutron scattering (Donald et al., 1997, 2001), and 13C NMR (Cheetham and Tao, 1998; Wang et al., 2008). Despite signicant efforts to characterize and differentiate various aspects of starch microstructure, the nature of their granular microstructure has not yet been fully elucidated (Chen et al., 2006; Kim and Huber, 2008; Liu et al., 2006, 2007, 2009; Xie et al., 2009).

242 Table 1 Materials used in the experimental work. Sample No. 1 2 3 4


a

P. Chen et al. / Journal of Cereal Science 50 (2009) 241247

Name Waxy Maize G50 G80

Ratio of amyloseamylopectin 0/100 23/77 50/50 80/20

Molecular weight (g/mol)a 20,787,000 13,000,000 5,115,000 673,000

Supplier Australia Australia Australia Australia

Molecular weight measured by GPC [provided by Penford (Australia)].

The development of novel microscopy techniques, in particular CLSM, has enabled the study of the morphology and internal structure of starch granules without the use of complicated or invasive sample preparation methods, thus increasing new possibilities (Boren et al., 2008; Chung and Lai, 2006; Huber and BeMiller, 2000; Mira et al., 2007; van de Velde, 2002). The key feature of confocal scanning laser microscopy is its ability to image a single focal plane of the sample. This enables researchers to visualize the internal structure of starch granules without the need for sectioning techniques, which can result in destruction of the microstructure of the sample (Durrenberger et al., 2001). Furthermore, CSLM imaging enables the visualization of a single starch granule in slices of less than 0.1 mm, whereas sectioning with a precision microtome system results in slices of 10 mm without the possibility to view several sections of the same granule. van de Velde et al. (2002) have studied granule characteristics by CSLM, as

well as the gelatinization behavior of starch granules from different botanical sources, and detected the cross-section of cornstarch granules. Blennow et al. (2003) studied the molecular deposition of transgenically modied starch in the potato starch granule using a combination of confocal laser scanning microscopy, light microscopy, and environmental SEM techniques. They found that starch extracted from tubers with low-amylose contents showed very weak 8-amino-1,3,6-pyrenetrisulfonic acid (APTS) uorescence. The starch granules with low molecular weight amylopectin and/or high-amylose granules showed a relatively even distribution in uorescene, whilst normal and low-amylose granules had an intense uorescence in the hilum, indicating a high concentration of amylose in the center of these granules. In our previous work (Chen et al., 2006), the morphologies and microstructures of cornstarches with different amyloseamylopectin ratios (waxy: 0/100; normal maize: 23/77; Gelose 50: 50/50; and Gelose 80: 80/20) were studied by microscopy with normal and polarized light, SEM, and XRD. It was found that birefringence of the granules under polarized light were waxy > maize > G50 > G80, which correspond with trends in molecular weight and crystallinity. In this work, the cornstarches were investigated by CLSM to systematically determine internal morphologies and microstructures of a single type of starch with different amyloseamylopectin ratios. The temperature-induced changes of the cornstarch granules in excess of water were also studied under CLSM.

Fig. 1. CLSM optical sections of different starches with same PMT: (a) waxy, (b) maize, (c) G50, and (d) G80.

P. Chen et al. / Journal of Cereal Science 50 (2009) 241247 Table 2 Observation results of different starches under CLSM. Sample Waxy Maize G50 G80
a

243

Fluorescence intensitya (arbitrary units) 40.27 61.61 83.64 90.73

Channel Many Some Rare Rare

Hilum Dark Main dark Highlight Highlight

Growth ring/crack Growth ring Growth ring Crack Crack

suspension was then ltered with Whatman No. 1 lter paper, with the acid-hydrolyzed starch dried at room temperature overnight under an air stream. 2.2.2. Partly gelatinized starch A quantity of 10 mg (dry basis) native starch was dispersed with 1 ml distilled water in a glass vial and then the glass vials were placed in the water bath for 2 min. Waxy and normal maize starch were heated to 70  C; and G50 and G80 were heated to 80  C. After thermal treatment, the samples were immediately cooled with tap water, the slurry found to separate into two layers after standing for 30 min. 2.2.3. Dying Native and partly gelatinized starch samples were prepared for CLCM as previously described (Blennow et al., 2003). Starch granules (10 mg) were dispersed in 15 ml of freshly made APTS solution (10 mM APTS dissolved in 15% acetic acid), and 15 ml of 1 M sodium cyanoborohydride was added. The reaction mixture was incubated at 30  C for 1518 h, with the granules washed ve times with 1 ml of distilled water and nally suspended in 20 ml of 1:1 (v/v) glycerolwater mixture. A drop of the mixture was then mounted on a glass plate for microscopy. 2.3. Confocal laser canning microscopy A CLSM equipped with an ArHg laser. (TCS SP2, Leica Microsystems, Wetzlar) with a stand for xed uorescent cell samples was used for the detection of the uorescene signal from

Parenthesis: Fluorescence intensity from ve independent measurements.

2. Experimental 2.1. Materials Cornstarches with different amyloseamylopectin ratios were used in the experimental work as model materials. All the starches are commercially available and were kindly supplied by Penford P/L (Australia). Table 1 lists the starches and their amyloseamylopectin ratios and other properties. APTS, sodium cyanoborohydride, and HCl acid were chemically pure and purchased as received from Sigma. 2.2. Sample preparation 2.2.1. Acid hydrolysis A quantity of 5 g (dry basis) of native starch was hydrolyzed by suspending it in 100 ml of 4 mol/L HCl solution at room temperature (about 21  C) for 5 days. After hydrolysis, the insoluble residue was washed several times with distilled water to neutrality. The

Fig. 2. CLSM optical sections of HCl-hydrolyzed different starches: (a) waxy, (b) maize, (c) G50, and (d) G80.

244

P. Chen et al. / Journal of Cereal Science 50 (2009) 241247

dye-stained starch granules. The details of the Leica objective lens used were: 100 plan apo/1.40 oil UV. The excitation wavelength was 488 nm with 52 capacity, and the format of the image was 512 512. During image acquisition, each line was scanned four times and averaged to reduce noise. For each starch sample, a stack of horizontal optical sections was obtained, encompassing the whole starch granule in three dimensions. Measurements of the uorescence intensity were made with Imaris 6.1.5 and the 3D-image analysis was performed using the software of the Leica TCS SP2. 3. Results and discussion There have been many reports over the years utilizing various methods to unravel the internal structures of starch. The key feature of CLSM is the ability to image a single focal plane in the sample. This enables researchers to visualize cross-sections of starch granules without the need for sectioning techniques which can result in destruction of the microstructure of the sample (Durrenberger et al., 2001). However, starch does not show autouorescence. It is necessary to choose an effective uorescent dye. APTS is one of the uorescence dyes with a small molecular structure that has previously been used successfully for specic labeling of the reducing ends of starch fragments (OShea et al., 1998). The intense and stable uorescence of APTS allows both surface and internal images of starch granules to be recorded at high resolution. Previous researches have used APTS to dye starches for CLSM observation (Blennow et al., 2003; Chung and Lai, 2006;

Glaring et al., 2006). In these studies, CLSM imaging with APTS as a probe for starch molecule distribution was used to study the internal structural features of starches, such as growth ring, channels, and cracks, as well as the general molecular distribution of amylose and amylopectin within the starch granule. Fig. 1 shows the CLSM optical sections of different cornstarches obtained at the same photomultiplier tube (PMT) after dying with APTS. It can be seen that most of the high-amylose starches (G50 and G80) showed an intensely stained center and were brighter than low-amylose starches (waxy and maize). It is well known that amylose has a lower molecular weight than amylopectin and contains a much higher molar ratio of reducing ends per anhydrous glucose residues, which results in a higher by-weight labeling of amylose (Blennow et al., 2003) and thus the granules with high-amylose content showed higher uorescence. It can be seen from Table 2 that the uorescence intensity sequence of the granules is G80 > G50 > maize > waxy, which corresponds to decreasing intensity with decreasing amylose content. The center section of CLSM also reveals the structures in the granule core, where it can be seen that the core of most of the waxy granules (low amylose) were dark, and that the dark core is connected with the channels in the granules. Some dark cores can also be found in maize starch. However, the cores of most of the G50 and G80 (high amylose) starch granules are bright, and the bright core in G80 is larger than that in G50, which indicates that high-amylose starch granules were more easily highlighted by uorescent dye in the hilum. One of the possible explanations for this phenomenon is that amylose content at the hilum is higher for the high-amylose

Fig. 3. CLSM optical sections of different starches after particularly gelatinized: (a) waxy, (b) maize, (c) G50, and (d) G80.

P. Chen et al. / Journal of Cereal Science 50 (2009) 241247

245

cornstarches G50 and G80, whilst another explanation is related to the different microstructures of the channel and hilum. The cavities in the granular core for waxy and maize starches have been clearly observed, but with different structures for G50 and G80, the higher density of hilum of G50 and G80 showing brighter cores. Further studies are warranted in this area due to the observation of such contradictory phenomena. It should also be noted that there is a dark line or channel in the center for G80, which separates the

bright core into two parts. This channel could relate to the granular growth and will be studied in detail in future work. From the cross-sections of the starch granule, the internal channels in starch granules can also be seen and are visible as dark lines running from the border of the granule toward the hilum in waxy and maize. The channels in starch granules have been observed by SEM and CLSM previously (Huber and BeMiller, 2000; Kim and Huber, 2008), and it has been noticed that the channels are

Fig. 4. 3D images of native (left) and partially gelatinized starches (right).

246

P. Chen et al. / Journal of Cereal Science 50 (2009) 241247

more frequently and clearly observed in waxy and maize starch (low amylose) than that in the G50 and G80 (high amylose). Such channels extend to the granule surface connected with cavities within the granules in waxy and maize starch, whilst the channel (if it is observed) is a ring along the bright core in G50 and G80. Recently Kim and Huber (2008) studied the channels within different wheat starch granules and found that A-type wheat starch has relatively larger channels than B-type wheat starch, which corresponds to our observations reported above since waxy and maize are A-type, while G50 and G80 are B-type starches. Acid hydrolysis can be used to remove the amorphous and less crystalline parts in the starch granule. The acid-etched granule resulted in better contrast between amorphous and semicrystalline layers within starch granules (Chung and Lai, 2006). The growth rings within high-amylopectin starches (waxy and maize) were clearly observed after acid hydrolysis, whilst sharp edge cracks from the hilum toward the border in the high-amylose starches (G50 and G80) were observed after such acid treatment (see Fig. 2). The growth rings showed spacings of about 300500 nm in waxy starches, whilst there are slightly ner rings with widths of about 200400 nm in maize material. Generally, the high-amylose starch (G50 and G80) granules have a more compact microstructure than that of low-amylose starches, which show more acid-resistant and thermally stable behavior (Chen et al., 2007; Liu et al., 2006). Moreover, the acid-hydrolyzed corn starches showed stronger uorescence intensity than native starches, indicating that there are more ends available to react with APTS groups after hydrolysis. The temperature-induced changes of the cornstarch granules in excess of water were also studied under CLSM. Phase transitions of starch granular during thermal processing are complicated and unique, especially the gelatinization process (Chen et al., 2007; Liu et al., 2006; Yeh and Li, 1996; Yu et al., 2006). The well-accepted denition of gelatinization for starch involves destruction of its order or crystalline regions. The starch suspensions were heated to just above their gelatinization temperatures (waxy and maize 70  C; G50 and G80 80  C) for 2 min, in order to avoid whole granules being completely destroyed, followed by rapid cooling by quenching with water. Following this they were then dyed with APTS and visualized using CSLM (see Fig. 3). The heterogeneity and complexity of such a starch structure within granules directly affects the gelatinization of starches. It can be seen from Fig. 3 that all the gelatinization starts at the hilum or botanical center of the granules and spreads rapidly to the periphery. Gelatinization thus begins in the intercellular and leastorganized region. As soluble components (primarily amylose) leach out of the granules during the process, they form a solution, whilst the center and adjacent components of the channels begin to gelatinize, and subsequently the other parts of the granules. The central area of the granule around the hilum is believed to be the least-organized region, since both acid hydrolysis and gelatinization start from there, which is in agreement with the previous observations (Chung and Lai, 2006). Fig. 4 shows 3D images of native and partly gelatinized starch granules constructed from different sections of CLSM images. The 3D images of native starch granules are similar to those observed from SEM (Chen et al., 2006). To the best of our knowledge, this is rst time the full image of partly gelatinized granules in water suspension has been observed since other techniques are not able to view it under these conditions. For example, SEM cannot be used to view a water suspension, whilst other optical microscope techniques are not able to slice the object and construct such a 3D image. Using confocal microscopy, it can be clearly seen that the granules of waxy and maize are broken through the cavity and channels when the granules swell, whilst the granules of G50 and G80 remained as granular after swelling. This technique clearly

shows promise to further explore the mechanisms of gelatinization, and we are currently carrying out this work. 4. Conclusion The morphology and microstructure of maize starches with different amyloseamylopectin ratios were systematically studied by confocal laser scanning microscopy (CLSM). Temperatureinduced changes of the cornstarch granules in excess water were also studied by this technique. The different microstructures of cornstarch granules with different amylose contents were evident, in particular the distribution of amylose and amylopectin materials and cavity, channel, and growth ringcrack locations within the starch granule. It can be seen that the high-amylose starches (G50 and G80) are brighter than low-amylose starches (waxy and normal maize), and the average uorescence intensity sequence of the granules is G80 > G50 > maize > waxy. The acid-hydrolyzed starches show stronger uorescence intensity than native starches due to the increased end groups as a result of chain scission being more amenable to dye attachment. Waxy and maize starches clearly show cavities and channels, whilst G50 and G80 present a bright core, indicative of dyed amylose. Waxy and maize starches show sharp growth ring structures after acid hydrolysis due to etching of the amorphous layer, whilst G50 and G80 granules showed different structures. The gelatinization starts at the hilum and adjacent to the channels of the granules and spreads rapidly to the periphery. Three-dimensional images of partly gelatinized granules composed of different sections of CLSM images were obtained, which were used to study the mechanisms of gelatinization. The granules of waxy and maize broke through the cavity and channels when the granules became swollen during gelatinization, while the granules of G50 and G80 remained as granular. Acknowledgements The authors from SCUT, China, would like to acknowledge the research funds NRDPHT (863) (2007AA10Z312, 2007AA100407), NFSC (50540420129), and GNSF (05200617). We acknowledge Monash Micro Imaging Laboratory for providing CLSM for this work. P. Chen would like to acknowledge the State Scholarship Fund provided by China Scholarship Council which supported her study in Australia. References
Baldwin, P.M., Adler, J., Davies, M.C., Melia, C.D., 1998. High resolution imaging of starch granule surfaces by atomic force microscopy. Journal of Cereal Science 27, 255265. Blennow, A., Hansen, M., Schulz, A., Jorgensen, K., Donald, A.M., Sanderson, J., 2003. The molecular deposition of transgenically modied starch in the starch granule as imaged by functional microscopy. Journal of Structural Biology 143, 229241. Boren, M., Glaring, M.A., Ghebremedhin, H., Olsson, H., Blennow, A., Jansson, C., 2008. Molecular and physicochemical characterization of the high-amylose barley mutant Amo1. Journal of Cereal Science 47, 7989. Chanzy, H., Vuong, R., Jesior, J.C., 1990. An electron diffraction study on whole granules of lintnerized potato starch. Starch/Starke 42, 377379. Cheetham, N.W.H., Tao, L., 1998. Solid state NMR studies on the structural and conformational properties of natural maize starches. Carbohydrate Polymers 36, 285292. Chen, P., Yu, L., Chen, L., Li, X., 2006. Morphology and microstructure of maize starches with different amylose/amylopectin content. Starch/Starke 58, 611615. Chen, P., Yu, L., Kealy, T., Chen, L., Li, L., 2007. Phase transition of starch granules observed by microscope under shearless and shear conditions. Carbohydrate Polymers 68, 495501. Chung, Y.-L., Lai, H.-M., 2006. Molecular and granular characteristics of corn starch modied by HClmethanol at different temperatures. Carbohydrate Polymers 63, 527534. Donald, A.M., Waigh, T.A., Jenkins, P.J., Gidley, M.J., Debet, M., Smith, A., 1997. Internal structure of starch granules revealed by scattering studies. In:

P. Chen et al. / Journal of Cereal Science 50 (2009) 241247 Frazier, P.J., Richmond, P., Donald, A.M. (Eds.), Starch: Starch Structure and Functionality. The Royal Society of Chemistry, London, pp. 172179. Donald, A.M., Kato, K.L., Perry, P.A., Waigh, T.A., 2001. Scattering studies of the internal structure of starch granules. Starch/Starke 53, 504512. Durrenberger, M.B., Handschin, S., Conde-Petit, B., Escher, F., 2001. Visualization of food structure by confocal laser scanning microscopy (CLSM). LebensmittelWissenschaft und-Technologie 34, 1117. Evers, A.D., McDermott, E.E., Albans, S., 1970. Scanning electron microscopy of wheat starch. II. Structure of granules modied by alpha-amylolysis Prelim inary report. Starch/Starke 22, 2326. Glaring, M.A., Koch, C.B., Blennow, A., 2006. Genotype-specic spatial distribution of starch molecules in the starch granule: A combined CLSM and SEM approach. Biomacromolecules 7, 23102320. Gregorov, E., Pabst, W., Bohacenko, I., 2006. Characterization of different starch types for their application in ceramic processing. Journal of the European Ceramic Society 26, 13011309. Helbert, W., Chanzy, H., 1996. The ultrastructure of starch from ultrathin sectioning in melamine resin. Starch/Starke 48, 185188. Huber, K.C., BeMiller, J.N., 2000. Channels of maize and sorghum starch granules. Carbohydrate Polymers 41, 269276. Jenkins, P.J., Cameron, R.E., Donald, A.M., 1993. A universal feature in the structure of starch granules from different botanical sources. Starch/Starke 45, 417420. Kim, H.-S., Huber, K.C., 2008. Channels within soft wheat starch A- and B-type granules. Journal of Cereal Science 48, 159172. Liu, H., Yu, L., Xie, F., Chen, L., 2006. Gelatinization of cornstarch with different amylose/amylopectin content. Carbohydrate Polymers 65, 357363. Liu, H., Yu, L., Chen, L., Li, L., 2007. Retrogradation of corn starch after thermal treatment at different temperatures. Carbohydrate Polymers 69, 756762. Liu, P., Yu, L., Liu, H., Chen, L., Li, L., 2009. Glass transition temperature of starch studied by a high-speed DSC. Carbohydrate Polymers 77, 250253. Mira, I., Kurtis Villwock, V., Persson, K., 2007. On the effect of surface active agents and their structure on the temperature-induced changes of normal and waxy

247

wheat starch in aqueous suspension. Part II: A confocal laser scanning microscopy study. Carbohydrate Polymers 68, 637646. Neethikajan, S., Thomson, D.J., Jayas, D.S., White, N.D.G., 2008. Characterization of the surface morphology of durum wheat starch granules using atomic force microscopy. Microscopy Research and Technique 71, 125132. OShea, M.G., Samuel, M.S., Konik, C.M., Morell, M.K., 1998. Fluorophore-assisted carbohydrate electrophoresis (FACE) of oligosaccharides: efciency of labelling and high-resolution separation. Carbohydrate Research 307, 112. Oostergetel, G.T., Van Bruggen, E.F.J., 1989. On the origin of a low angle spacing in starch. Starch/Starke 41, 331335. van de Velde, F., van Riel, J., Hans Tromp, R., 2002. Visualisation of starch granule morphologies using confocal scanning laser microscopy (CSLM). Journal of the Science of Food and Agriculture 82, 15281536. Wang, S., Yu, J., Yu, J., 2008. Conformation and location of amorphous and semicrystalline regions in C-type starch granules revealed by SEM, NMR and XRD. Food Chemistry 110, 3946. Xie, F., Yu, L., Su, B., Liu, P., Wang, J., Liu, H., Chen, L., 2009. Rheological properties of starches with different amylose/amylopectin ratios. Journal of Cereal Science 49, 371377. Yamaguchi, M., Kainuma, K., French, D., 1979. Electron microscopic observations of waxy maize starch. Journal of Ultrastructure Research 69, 249261. Yeh, A.-I., Li, J.-Y., 1996. A continuous measurement of swelling of rice starch during heating. Journal of Cereal Science 23, 277283. Yu, L., Kealy, T., Chen, P., 2006. Study of starch gelatinization in a ow eld using simultaneous rheometric data collection and microscopic observation. International Polymer Processing XXI, 283289. Zobel, H.F., 1984. Gelatinization of starch and mechanical properties of starch pastes. In: Whistler, R.L., BeMiller, J.N., Paschall, E.F. (Eds.), Starch: Chemistry and Technology, second ed. Academic Press, Orlando, pp. 285298. Zobel, H.F., 1988. Molecules to granules: A comprehensive starch review. Starch/ Starke 40, 4450. Zobel, H.F., Illinois, A., 1988. Starch crystal transformations and their industrial importance. Starch/Starke 40, 17.

Journal of Cereal Science 50 (2009) 248254

Contents lists available at ScienceDirect

Journal of Cereal Science


journal homepage: www.elsevier.com/locate/jcs

Atmospheric CO2 enrichment changes the wheat grain proteome


P. Hogy a, *, C. Zorb b,1, G. Langenkamper b, T. Betsche b, A. Fangmeier a
a

Universitat Hohenheim, Institute for Landscape and Plant Ecology (320), Plant Ecology and Ecotoxicology, Oekologiezentrum 2, August-von-Hartmann-Str. 3, 70599 Stuttgart, Germany b Max Rubner-Institut, Federal Research Institute of Nutrition and Food, Schuetzenberg 12, 32756 Detmold, Germany

a r t i c l e i n f o
Article history: Received 19 February 2009 Received in revised form 19 May 2009 Accepted 4 June 2009 Keywords: CO2 enrichment FACE Wheat Grain proteome

a b s t r a c t
Spring wheat (Triticum aestivum L. cv. Triso) was grown in a free-air CO2 enrichment (FACE) eld experiment in order to gain information on CO2-induced effects on grain composition and quality at maturity. A proteome analysis was performed using two-dimensional gel electrophoresis (2-DE) and protein identication was done with mass spectrometry (MALDI-TOF MS). In elevated CO2 (526 ml l1), an increase of 13.5% in grain yield was observed relative to 375 ml l1 at a low level of signicance (P 0.528). Total grain protein concentration was decreased by 3.5% at a high level of statistical significance. Most importantly, a number of statistically signicant changes within the grain proteome were observed, as the levels of 32 proteins were affected by elevated CO2: 16 proteins were up-regulated and 16 were down-regulated. Our experiment demonstrates that high-CO2 can markedly affect the proteome of mature wheat grain. The potential role of the proteins, changed in response to CO2 enrichment, is discussed as some may affect grain quality. For the task of selecting cultivars resistant to CO2-induced quality loss, we propose to consider the proteins affected by elevated CO2 identied in this work here. 2009 Elsevier Ltd. All rights reserved.

1. Introduction The atmospheric carbon dioxide (CO2) concentration will probably rise from the current 385 ml l1 to a level as high as 550 ml l1 by 2050 (Meehl et al., 2007). CO2 is an important greenhouse gas, but it is also a key substrate for plant growth. CO2 enrichment will directly affect important C3 crops such as wheat (Triticum aestivum L.) by stimulation of photosynthesis and inhibition of photorespiration (Ainsworth and Rogers, 2007). Increased accumulation of carbohydrates in vegetative parts, higher biomass and grain yield were observed (e.g. Ziska and Bunce, 2007). In spite of the benecial effects on water use efciency (WUE; Kimball et al., 2002), the higher canopy temperatures due to lower transpiration (Erbs et al., 2009; Kimball et al., 1999) might, in turn, lead to
Abbreviations: 2-DE, two-dimensional gel electrophoresis; FACE, free-air CO2 enrichment; GS, glutenin subunits; HMW, high molecular weight; IEF, isoelectric focusing; IPG, immobilised pH gradient; LMW, low molecular weight; MALDI-TOF MS, matrix-assisted laser desorption/ionisation time-of-ight mass spectrometry; NUE, nitrogen use efciency; pI, isoelectric point; SDS-PAGE, sodium dodecyl sulphate polyacrylamide gel electrophoresis; WUE, water use efciency. * Corresponding author. Tel.: 49 711 459 23819; fax: 49 711 459 23044. E-mail addresses: pethoegy@uni-hohenheim.de (P. Hogy), czoerb@ plantnutrition.uni-kiel.de (C. Zorb), georg.langenkaemper@mri.bund.de (G. Lan genkamper), mt.betsche@t-online.de (T. Betsche), afangm@uni-hohenheim.de (A. Fangmeier). 1 Present address: Christian-Albrechts-Universitat Kiel, Institute of Plant Nutrition and Soil Sciences, Hermann-Rodewald-Strasse 2, 24118 Kiel, Germany. 0733-5210/$ see front matter 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2009.06.002

oxidative stress. Concomitantly to increased nitrogen (N) use efciency (NUE), N and thus protein concentrations were often found to decrease in different plant parts (Cotrufo et al., 1998; Fangmeier et al., 1999), resulting in changed key quality traits for consumer nutrition and health, industrial processing and marketing. Wheat grain proteins are usually classied into three functional groups comprising storage proteins, structural and metabolic proteins as well as protective proteins. Metabolic and structural proteins have important functions in processes such as carbohydrate metabolism and protein synthesis/assembly (Vensel et al., 2005). Wheat storage proteins, characterised by high glutamine and proline contents, comprise monomeric gliadins and large disulphide linked polymeric glutenins made up of high molecular weight (HMW) and low molecular weight (LMW) subunits (Weegels et al., 1996). The HMW glutenin subunits (GS), which are minor components in terms of quantity, are particularly key factors in the process of breadmaking as they are positively correlated with our performance such as dough strength, elasticity and stability as well as bread volume (MacRitchie, 1999). Nevertheless, differences in the total amount of LMW-GS are also important for technological properties such as resistance and extensibility of dough (Metakovsky et al., 1990). Most, storage proteins have no known physiological function during grain maturation, but they constitute a major nutrient source during seed germination (Shewry and Halford, 2002). Since CO2 enrichment causes metabolic perturbations in vegetative plant parts, the N availability for the developing grain must

P. Hogy et al. / Journal of Cereal Science 50 (2009) 248254

249

be considered with regard to the grain proteome. The synthesis of the storage proteins is regulated by the availability of N, with glutamine-rich gliadins being more affected by availability of N than glutenins (Wieser and Seilmeier, 1998). In contrast, the accumulation of structural and metabolic proteins and of defence proteins is most likely constrained by N sink strength of the grain during lling. Thus, these proteins increase when the total protein per grain decreases. Consequently, CO2 enrichment has been seen to modify total protein concentration (Hogy and Fangmeier, 2008; Kimball et al., 2001; Taub et al., 2008) and protein composition due to the higher responsiveness of storage proteins in wheat grains (Hogy et al., 2009; Wieser et al., 2008). Indeed, storage proteinrelated our and dough properties and important end-use quality characteristics were reported to be impaired in response to elevated CO2 levels in chamber-based (Fangmeier et al., 1999) and free-air CO2 enrichment (FACE) eld experiments (Hogy et al., 2009; Kimball et al., 2001). The latter allow for the exposure of agroecosystems to predicted future atmospheric CO2 levels, avoiding alterations of the microclimatic conditions (chamber effects; Long et al., 2004) and restrictions of the rooting volume inevitable when using small pots. Using a proteomic approach, the objective of the present experiment was to identify proteins of mature wheat grains affected by CO2 enrichment. To our knowledge, studies on CO2-induced changes of the grain proteome of wheat under nearnatural eld conditions have not yet been performed. This experiment will serve as an initial report in this regard and may help us to understand the effects of high CO2 levels on the nutritional and baking qualities of wheat grains. 2. Experimental 2.1. Plant cultivation and CO2 exposure The experiments were conducted in 2005 at the Heidfeldhof farm, managed by the Plant Breeding Station of the Universitat Hohenheim (9110 E, 48 420 N, 395 m above sea level) south of Stuttgart, Germany. Spring wheat (T. aestivum L. cv. Triso) as main crop was grown together with 13 typical arable weed species on 15 circular plots 2 m in diameter in order to test wheat response under conditions of the agricultural practice to CO2 enrichment (Erbs et al., 2009; Hogy et al., 2009). The clayey to loamy soil was dug to a depth of 0.3 m and wheat (13 lines with 15 cm row spacing; 200 plants m2) was sown on 24 March 2005. CO2 enrichment was provided by a FACE exposure system for the whole growing season (04 April 2005 to 01 August 2005). Frames with four computer controlled solenoid valves and Teon tubes for the CO2 supply were established at each of ve FACE plots, while frames without additional CO2 fumigation were installed at ve AMBIENT plots. Five more plots without frames and CO2 enrichment were operated as CONTROL plots in order to identify potential effects of the frames. To achieve the target gas concentrations of 150 ml l1 above AMBIENT during daylight hours (average from 07:00 to 20:00), pure CO2 (Westfalen Gas, Germany) was added continuously and concentrations achieved at canopy height were monitored. To avoid contamination from gas released on other plots, 10 m distance was maintained between plots. The principle of operation and the performance of the FACE system have been described by Erbs and Fangmeier (2006). Plots were fertilised with 140 kg ha1 N, 30 kg ha1 phosphorous and 60 kg ha1 potassium, which was supplied in 50/25/25% portions at different development stages of the spring wheat (EC 1329, EC 3039 and EC 4059, respectively; Tottman and Broad, 1987). In addition, a trace element fertiliser (Microtop, Germany) was utilised at EC 3039 containing 0.03 kg ha1 boron,

0.45 kg ha1 magnesium, 0.36 kg ha1 sulphur and 0.03 kg ha1 manganese. 2.2. Climatic and soil conditions Wind speed and direction, air temperature, relative humidity (Rotronic, Switzerland) and solar radiation (Kipp & Zonen, The Netherlands) were measured at a reference height of 2 m at the centre of the eld site. Precipitation was recorded at the nearby Stuttgart airport weather station of the Deutscher Wetterdienst. Time domain reectometry sensors (IMKO, Germany) for soil moisture were horizontally installed at a depth of 15 cm at the centre of each plot. Irrigation was done manually when the volumetric soil moisture declined to 25% with the same amount of water applied to all treatments. Volumes of manual irrigation and natural rainfall were recorded and total water supply per plot was calculated. 2.3. Harvest and sample preparation Wheat plants of the central square metre of each plot were harvested at grain maturity (EC 89; Tottman and Broad, 1987). Plant dry weight and grain yield were determined and wheat grains were ground into wholemeal our using a laboratory mill (Retsch MM 301, Germany). Protein N concentration was analysed from ve AMBIENT and ve FACE plots by elemental analysis (Elementar, Germany), respectively, and multiplied by the conversion factor 5.7, which gives the total protein concentrations in grain samples. As grain protein concentrations were identical between the AMBIENT and CONTROL plots (data not shown), proteome analyses were conducted for three AMBIENT and three FACE plots. 2.4. Protein extraction and two-dimensional gel electrophoresis The wholemeal our was further ground with liquid N and proteins were extracted from a 200 mg aliquot using a modied DTT-TCA-acetone precipitation method (Zorb et al., 2009). A 2-DE clean-up kit (GE Healthcare, USA) was used for a second precipitation step and the protein concentration of the extract was determined with three replicates each using a 2-DE Quant protein determination kit (GE Healthcare, USA). Protein extracts (400 mg) were subjected to 2-DE with isoelectric focusing (IEF) in rst dimension and sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE) in second dimension by following a pub lished protocol (Zorb et al., 2009). Immobilised pH gradient (IPG) strips (18 cm, pH 310, linear) were from GE Healthcare, USA. For the second dimension, molecular weight standards (10150 kDa; Sigma, Germany) were applied onto 12.5% (v/v) acrylamide gels. The gels were stained with Coomassie R350 tablets (GE Health Care, USA; Zorb et al., 2009). 2.5. Image analyses Gels were scanned in the transillumination mode (HP Scan-Jet 4890, USA; 300 dpi and 16 bits per pixel). Computer-assisted 2-DE analysis was performed with the software Delta2D 3.4 (Decodon, Germany). Gels were warped using a group warping strategy to obtain a so-called average fusion gel. A virtual master fusion gel was created by the software using every matched protein spot from all gels. This master fusion gel was used to delete artefacts, specks and freckles on individual gels before further processing. Subsequently, normalised protein spot intensity was determined for all individual gels. Average relative protein spot intensities were calculated for AMBIENT and FACE treatments and compared.

250

P. Hogy et al. / Journal of Cereal Science 50 (2009) 248254 Table 1 Carbon dioxide (CO2) effects on wheat biomass productivity, grain yield and total grain protein.a Conditions Biomass (g m2) Grain yield (g m2) Protein concentration [% DW] Protein yield [g m2] AMBIENT 1041 191 350.9 84.0 15.71 0.29 54.99 12.39 FACE 1236 169 398.2 129.8 15.16 0.23 60.23 19.08 CO2 effect (%) 18.8 NS 13.5 NS 3.5** 9.5 NS

2.6. Data, replication, statistics For proteomics, the reliability of the results was ensured by the following measures: (i) two technical 2-DE replications from grain of each eld plot were used to create an average 2-DE gel. (ii) Average 2-DE gels including standard errors of the mean were calculated from the three independent eld plot replicates. (iii) Students t-tests were performed to evaluate the reproducibility of the spots over the corresponding eld trial replicates. Spots were disregarded when the t-test value was below 95%. (iv) A speckle lter with an intensity limit of 0.02 was applied. (v) Differences in levels of protein spots at ambient and elevated CO2 concentrations were disregarded, unless different by a factor 1.5 to <2 (trend) or factor !2. The statistical analyses for the CO2 effects on biomass productivity, grain yield and total protein concentration of wheat in the AMBIENT and FACE treatments were performed using SPSS V. 15.0 software (SPSS, USA); data from CONTROL plots were excluded since differences to AMBIENT treatments were not signicant. Data were subjected to the Levene test and ANOVA with CO2 as treatment factor was used to identify signicant differences between means of ambient (xa) and elevated (xe) CO2-grown grains (P 0.05). The response ratio (r xe/xa) was calculated according to Long et al. (2004) and the CO2 effect was computed as relative percentage change ([r 1] 100). 2.7. Protein identication Protein identication was performed as described (Zorb et al., 2009). Database searches were performed at the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov) using MASCOT 1.9 with one missed cleavage allowed. The molecular weight of the proteins was calculated on the basis of a standard 10 kDa ladder using 2-DE software. The pI (isoelectric point) of the protein spots was calculated from their position on the IEF-strips as mentioned by the manufacturers specication (GE Healthcare, USA). 3. Results 3.1. Climatic conditions and experimental CO2 concentrations The 130 days cultivation period of the year 2005 was rather moist with frequent thunderstorms and high precipitation in relatively isolated rain events. The crops received 425 mm of water, which comprised rainfall supplemented by irrigation from sowing to maturity. The growing degree days (baseline > 5  C) were 1120  C. The averages standard deviations of daily minimum, mean and maximum air temperatures averaged over the cropping season were 9.2 4.8  C, 14.8 5.3  C and 20.1 6.4  C, respectively. Daytime CO2 concentrations were 375 11 ml l1 (AMBIENT) and 526 5 ml l1 (FACE), which demonstrates that the exposure system worked satisfactorily. The average volumetric soil moisture to which the plants were exposed over the whole growing period was 26.4 3.4% (AMBIENT) and 27.9 3.8% (FACE). 3.2. Biomass, wheat grain yield and total protein concentration In the high-CO2 treatment, the total wheat biomass and grain yield were increased (Table 1). However, these CO2 effects were below statistical signicance (P 0.154 for biomass and P 0.528 for grain yield). In contrast, total grain protein concentration was signicantly decreased by CO2, whereas there was no signicant change in protein yield (i.e. protein harvested per ground area) at nal harvest.

**Signicant at P 0.01; a Wheat was grown in the eld at either ambient (AMBIENT) or elevated (FACE) CO2 concentrations. Presented are means and SD of ve replicates per treatment, the calculated CO2 effect (FACE vs. AMBIENT) and levels of statistical signicance of ANOVA. NS, not signicant at P 0.05.

3.3. Effects of CO2 enrichment on the wheat grain proteome Overall, 2-DE resolved a total of 770 proteins, of which 32 proteins (4.2% of total) were changed in relative concentrations due to elevated CO2 (Fig. 1a,b; Table 2). Our analyses revealed 28 proteins which were altered by a factor of !2 between AMBIENT and FACE, while for four proteins, only a trend (factor 1.5 to <2) was observed (Table 2). No proteins were found in FACE grains that did not occur also in AMBIENT, whereas one protein (D16) disappeared under elevated CO2. Sixteen proteins were up-regulated (2.1% of total; U1U16), while 16 proteins (2.1% of total) were downregulated (D1D16) in FACE. Of the 32 proteins with differential expression levels between AMBIENT and FACE treatments, 24 proteins (3.1% of total) were identied in database searches. Proteins were clustered according to their functional annotations into (i) storage proteins, (ii) stress- and defence-related proteins and (iii) structural and metabolic proteins. The eight unidentied proteins were denoted (iv) unknown proteins. 3.3.1. Up-regulation of wheat grain proteins under elevated CO2 Most of the up-regulated proteins showed sequence homologies with proteins involved in important structural functions and metabolic pathways (six proteins; U7, U9, U12U14, U16). Proteins U7 and U13 demonstrated resemblance to putative proteins of Arabidopsis thaliana. Protein U9 was identied as a starch synthase I enzyme. Proteins U12 and U16 were identied as glyceraldehyde3-phosphate dehydrogenase (GAPDH) of the Triticeae. Protein U14 is a glucose and ribitol dehydrogenase (GRDH) homolog of barley (Table 2). Stress- and defence-related proteins (four proteins; U6, U8, U10, U11) were also up-regulated by elevated CO2. Protein U6 was identied as serpin. Also protein U8 increased and was identied as peroxidase 1. Two other proteins U10 and U11 had high levels of similarity to a family 11 xylanase chain-A protein of wheat. Among the two up-regulated seed storage proteins, U4 was identied as triticin precursor of wheat. Protein U5 increased too and was identied as putative avenin-like b precursor. In addition, four unknown proteins U1, U2, U3 and U15 were up-regulated in the high-CO2 treatment. 3.3.2. Down-regulation of wheat grain proteins under elevated CO2 The 16 proteins down-regulated under elevated CO2 are presented in Table 2. Four proteins (D4, D11, D14, D16) decreased and could not be identied (unknown proteins). Moreover, one of these unidentied proteins (D16) was not detectable in FACE. Most of the identied proteins were seed storage proteins (66.7%), such as LMW-GS and HMW-GS. Among these eight proteins, D3 was identied as globulin and proteins D5, D6, D7 and D9 were revealed as HWM-GS 1By9 subunits. Proteins D8 and D15 were identied as LMW-GS and LMW-s KS2. In addition, D12 was matched to the HMW-GS 1Ax1 subunit. The other three wheat grain proteins (D1,

P. Hogy et al. / Journal of Cereal Science 50 (2009) 248254

251

Fig. 1. Example of Coomassie stained wheat grain proteins, separated by two-dimensional gel electrophoresis (2-DE). Horizontal, increasing isoelectric point from range pH 310; vertical, separation by molecular weight. (a) Ambient carbon dioxide (CO2; AMBIENT); (b) elevated CO2 (FACE). Only proteins which were differentially expressed under elevated CO2 concentration were labelled. Protein identities correspond to spot numbers in Table 2.

252

P. Hogy et al. / Journal of Cereal Science 50 (2009) 248254

Table 2 Effects of carbon dioxide (CO2) enrichment on wheat grain proteomea. Spot no. Protein name Mascot score Protein Mr (kDa)/pI Accession no. Species Factor CO2 effect

(i) Up-regulated proteins Storage proteins U4 Triticin precursor U5 Putative avenin-like b precursor Stress- and defence-related proteins U6 Serpin U8 Peroxidase 1 U10 A-chain family 11 xylanase U11 A-chain family 11 xylanase Structural and metabolic proteins U7 Putative protein: MuDR family transposase U9 Starch synthase I U12 Cytosolic glyceraldehyde-3-phosphate dehydrogenase (GAPDH) U13 Putative protein: Chromosome segregation ATPase U14 T06212 glucose and ribitol dehydrogenase homolog (GRDH) U16 Glyceraldehyde-3-phosphate dehydrogenase (GAPDH), cytosolic Unknown proteins U1 U2 U3 U15 (ii) Down-regulated proteins Storage proteins D3 Globulin (Gbl1) storage protein D5 HMW glutenin subunit 1By9 D6 HMW glutenin subunit 1By9 D7 HMW glutenin subunit 1By9 D8 LMW glutenin D9 HMW glutenin subunit 1By9 D12 HMW glutenin subunit 1Ax1 D15 LMW-s KS2 Stress- and defence-related proteins D2 Resistance protein-like protein Structural and metabolic proteins D1 ATP synthase subunit D10 Aldose reductase-related protein D13 Alpha-tubulin Unknown proteins D4 D11 D14 D16b

121 84.5 149 191 82.1 89.4 73.8 96.2 113 75.8 128 171

56.9/10.0 32.1/6.8 42.9/5.6 38.8/9.5 35.0/6.7 30.3/9.0 104.5/7.2 71.0/5.7 18.2/6.4 50.5/5.1 31.6/6.7 33.2/6.2

gij7548844 gij89143130 gij5734504 gij22001285 gij51247633 gij51247633 gij7269795 gij5880466 gij32478662 gij62320248 gij7431022 G3PC_HORVU

T. aestivum T. aestivum T. T. T. T. aestivum aestivum aestivum aestivum

5.0 4.3 3.6 3.2 2.9 2.8 3.5 2.9 2.6 2.3 2.2 2.1

A. thaliana T. aestivum T. aestivum A. thaliana H. vulgare H. vulgare

48.1/6.5 49.5/6.7 48.3/7.6 35.8/7.0

25.4 8.0 5.9 2.2

72.8 156 162 188 94.2 231 222 77.6 80.3 258 77.3 208

72.2/6.9 76.7/9.3 75.7/9.2 74.4/9.1 32.4/10.0 76.2/9.3 89.8/5.6 39.0/10.5 19.5/6.7 59.2/5.5 35.8/6.3 49.7/4.7 69.9/9.3 64.9/10.5 30.3/9.0 35.7/6.2

gij170696 gij22090 gij22090 gij22090 gij56480772 gij109452233 gij21743 gij51870702 gij53680936 gij525291 gij167113 gij4098272

T. aestivum T. aestivum T. aestivum T. aestivum T. aestivum T. aestivum T. aestivum T. aestivum Poncirus trifoliate L. T. aestivum Bromus inermis L. T. aestivum

3.3 2.8 2.6 2.5 2.3 2.3 1.9 1.5 4.1 4.7 2.2 1.8 2.8 2.2 1.7 >15

a Wheat was harvested at maturity and total grain protein extracts were subjected to two-dimensional gel electrophoresis (2-DE) and matrix-assisted laser desorption/ ionisation time-of-ight mass spectrometry (MALDI-TOF MS) analysis. Image analysis revealed proteins that were (i) up-regulated and (ii) down-regulated at elevated CO2 concentrations. Proteins identied are listed with the Mascot score as a measure of condence for MS-based identication. Experimental molecular weight and experimental isoelectric point (pI) were calculated from 2-DE. Accession numbers are given according the National Center for Biotechnology Information (NCBI) database (http://www.ncbi. nlm.nih.gov). Species names refer to the highest hit in the database: Triticum aestivum (T. aestivum L.), Arabidopsis thaliana (A. thaliana L.), Hordeum vulgare (H. vulgare L.). The CO2 effect factor was calculated as the ratio of average protein spot areas in AMBIENT and FACE treatment. For calculation see Section 2. b Protein was not detected at elevated CO2 concentrations.

D10, D13) involved in the general metabolism were reduced upon CO2 exposure. Protein D1 was identied as ATP synthase betasubunit and protein D10 as aldose reductase-related protein of Bromus inermis. Down-regulated protein D13 was identied as alpha-tubulin of wheat. With regard to plant stress and defence, one protein (D2) was down-regulated. Protein D2 was identied as a resistance protein-like protein homologue of Poncirus trifoliata. 4. Discussion 4.1. Biomass production, grain yield and grain proteome of wheat In our experiment, increase of biomass (18.8%) and grain yield (13.5%) under elevated CO2 corresponds to changes observed in other FACE studies (Kimball et al., 2001, 2002), although these effects were not statistically signicant. CO2 enrichment caused

a signicant decrease of total grain protein concentration by 3.5% (from 15.71 to 15.16%), which is well in agreement with other studies performed at ample water and N supply (e.g. recent meta analyses conducted on different wheat cultivars by Hogy and Fangmeier (2008) and Taub et al. (2008)), while Wieser et al. (2008) reported reductions in protein concentration of about 14% in comparable CO2 treatments. Moreover, our ndings demonstrate that elevated CO2 levels change the grain protein composition signicantly and may thus affect grain quality traits. By using 2-DE, a total of 770 proteins were detected in wheat grains grown in AMBIENT and FACE treatments. CO2 enrichment resulted in signicant changes in concentrations of 32 (4.2%) grain proteins. The up-regulated proteins predominantly belonged to the group of structural and metabolic proteins, while the down-regulated proteins mainly had storage functions. The implications of these ndings are discussed in Section 4.1.1.

P. Hogy et al. / Journal of Cereal Science 50 (2009) 248254

253

4.1.1. Seed storage proteins The CO2-induced decrease of storage proteins was mainly inferred from the down-regulation of ve HWM-GS and two LMW-GS. Also, Wieser et al. (2008) reported that the HMW-GS fraction was more affected than LMW subunits under CO2 enrichment. The HMW-GS are main determinants of visco-elastic properties of dough (MacRitchie, 1999). Particularly HMW 1Ax subunits contribute positively to the bread-making quality. Therefore, a reduction of the HMW 1Ax subunit concentration caused by elevated CO2 could lead to lower bread-making quality. Moreover, variations in the total amount of LMW-GS also affect other technological properties such as dough resistance and extensibility (Metakovsky et al., 1990). In our work, gliadins were not signicantly affected by CO2 enrichment, although linear relationships between protein levels and gliadin levels were reported before (Wieser and Seilmeier, 1998). However, Wieser et al. (2008) reported perturbations of storage proteins more by total gliadins (20%) and their types than glutenins (15%), resulting in decreases of the gliadin to glutenin ratio under CO2 enrichment. Thus, variations in the gliadin/glutenin ratio are one explanation for the CO2-induced impacts on the visco-elastic dough properties such as lower dough resistance found in wheat grain samples in FACE studies (Hogy et al., 2009). CO2 enrichment also resulted in an increase in concentration of a triticin precursor that is related to our quality aspects, but also suggested to function in plant stress responses. Moreover, a putative avenin-like b precursor that comprises a cereal-type alphaamylase inhibitor was increased by elevated CO2. These storage proteins are thought to protect the starch reserves in the endosperm from degradation. From a food technological perspective amylase inhibitors may also be useful markers to distinguish between aspects of grain quality (Skylas et al., 2005). Thus, the CO2-induced impact on the avenin-like b precursor might indicate changes of grain quality. Concomitantly, CO2-induced changes in protein composition of grains may also explain vari ations of the amino acid composition in FACE experiments (Hogy et al., 2009), resulting in alterations of nutritious properties of wheat grains. 4.1.2. Structural and metabolic proteins The expressions of six structural and metabolic proteins such as putative proteins, starch synthase I, GAPDH and a homologue of barley GRDH were increased, while ATP synthase betasubunit, aldose reductase-related protein and cytoskeletal alphatubulin were found to be decreased under elevated CO2. One of the up-regulated putative proteins is associated with the chromosome segregation ATPases that are involved in cell division and chromosome partitioning, while another putative protein is a transposase for mutator transposable elements. A higher level of the glycolytic enzyme GAPDH, involved in glycolysis and glyconeogenesis in grains, may be associated with drought stress (Velasco et al., 1994). However, the link between CO2 enrichment and up-regulation of the former proteins is unclear. The higher amount of starch synthase I, responsible for the synthesis of amylase and involved in building the nal structure of amylopectin, is in agreement with earlier ndings of higher starch contents in wheat grains under CO2 enrichment (Hogy and Fangmeier, 2008). Concerning down-regulated proteins, the ATP synthase beta-subunit is involved in energy metabolism, playing an important role in the synthesis and transformation of ATP, while the aldose reductase-related protein has a function in proteolysis. Moreover, reactive aldehydes produced by imbalances in the redox metabolism can be removed by the action of aldose reductases. Currently, the background behind the CO2induced down-regulation of these proteins is unclear. The

observed relationship between alpha-tubulin, a main protein of the cytoskeleton involved in cell division, which is lowered under CO2 enrichment, remains obscure. Our ndings further demonstrate up- and down-regulation of single grain proteins revealed in the albumin and globulin families, which is not in contrast to the unaffected total amounts of albumins/globulins reported under CO2 enrichment (Hogy et al., 2009; Wieser et al., 2008). 4.1.3. Stress- and defence-related proteins There were substantially more stress/defence proteins upregulated than down-regulated in wheat grain under CO2 enrichment. As soil moisture content was slightly increased over long periods and carbon isotope analysis revealed that wheat plants had signicantly improved WUE in the high-CO2 treatment in 2005 (Erbs et al., 2009) we suggest that most likely more energy and biochemical precursors were available for inputs offering protection against stressors. The group of CO2-induced proteins consisted of serpin, peroxidase 1 and two isoforms of family 11 xylanase. The latter can inhibit glycohydrolases secreted by fungal pathogens during infection or insect attacks (Payan et al., 2003). Serpins are chymotrypsin-like serine protease inhibitors of the albumin globulin fraction and may therefore prevent premature proteolysis of the storage proteins of wheat grains. In addition, peroxidases, catalysing the oxidation of phenolic compounds and protecting cells from peroxide attack, have been recognised to play a role in pathogen defence and dehydration stress. Reduced transpiration in wheat under CO2 enrichment (Erbs et al., 2009) and thus increased canopy temperatures (Kimball et al., 1999) would cause oxidative stress which might be a reason for higher peroxidase levels in grains. However, elevated CO2 and stress- and defence-related proteins in wheat grains interact in several ways that are currently not well understood. Overall, variations in total and individual grain protein concentration did not only originate from changes of the grain biomass via the so-called dilution effect under elevated CO2 conditions (Pleijel et al., 1999). Rather, CO2 enrichment impacts on the vegetative plant metabolism (e.g. Riviere-Rolland et al., 1996), most likely leading to differences in biosynthesis of proteins and of protein partitioning during grain lling. The factors that inuence relationships between N sources and sinks interact in ways that are currently not well understood. The impact on grain quality traits, including nutritional and technological properties, is likely to become more pronounced considering future CO2 enrichment and cannot be compensated by relatively minor adjustments to crop management. For instance, increased N fertilisation may translate into higher biomass and yield rather than into enhanced redistribution of N to the grains (Fangmeier et al., 1999). Alterations in grain proteins may also be relevant for the developing plant during germination as they provide a major nutrient source. In addition, the proteins with unknown identity may possibly reveal further clues providing new insights into grain quality attributes of wheat. Since many of the up-regulated proteins identied correspond to allergens involved in bakers asthma, including GAPDH, peroxidase and the serpins (Hurkman et al., 2008), the risk for human health may increase under elevated CO2. Interactive effects of CO2 enrichment and alterations of climatic variables such as temperature and nutrient and water availability may additionally affect the end-product quality and food safety with regard to nutritional quality, allergenicity and industrial processing in the future. To alleviate some of these problems, CO2affected proteins can probably serve as markers to select useful quality traits in wheat breeding, quality testing and product analysis. However, an experimental database for this purpose is largely missing as yet.

254

P. Hogy et al. / Journal of Cereal Science 50 (2009) 248254 Long, S.P., Ainsworth, E.A., Rogers, A., Ort, D.R., 2004. Rising atmospheric carbon dioxide: plants FACE the future. Annual Review in Plant Biology 55, 591628. MacRitchie, F., 1999. Wheat proteins: characterization and role in our functionality. Cereal Food World 44, 188193. Meehl, G.A., Stocker, T.F., Collins, W.D., Friedlingstein, P., Gaye, A.T., Gregory, J.M., Kitoh, A., Knutti, R., Murphy, J.M., Noda, A., Raper, S.C.B., Watterson, I.G., Weaver, A.J., Zhao, Z.C., 2007. Global climate projections. In: Solomon, S., Qin, D., Manning, M., Chen, Z., Marquis, M., Averyt, K.B., Tignor, M., Miller, H.L. (Eds.), Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, UK and New York, pp. 748846. Metakovsky, E.V., Wrigley, C.W., Bekes, F., Gupta, R.B., 1990. Gluten polypeptides as useful genetic markers of dough quality in Australian wheats. Australian Journal of Agricultural Research 41, 289306. Payan, F., Flatman, R., Porciero, S., Williamson, G., Juge, N., Roussel, A., 2003. Structural analysis of XIP-I, a xylanase proteinaceous inhibitor from wheat. Biochemical Journal 372, 395405. Pleijel, H., Mortensen, L., Fuhrer, J., Ojanpera, K., Danielsson, H., 1999. Grain protein accumulation in relation to grain yield of spring wheat (Triticum aestivum L.) grown in open-top chambers with different concentrations of ozone, carbon dioxide and water availability. Agriculture. Ecosystems & Environment 72, 265 270. Riviere-Rolland, H., Contard, P., Betsche, T., 1996. Adaptation of pea to elevated atmospheric CO2: rubisco, phosphoenolpyruvate carboxylase and chloroplast phosphate translocator at different levels of nitrogen and phosphorus nutrition. Plant. Cell & Environment 19, 109117. Shewry, P.R., Halford, N.G., 2002. Cereal seed storage proteins: structures, properties and role in grain utilization. Journal of Experimental Botany 53, 947958. Skylas, D.J., Cordwell, S.J., Craft, G., McInerney, B., Wu, M.J., Chin, J., Wrigley, C.W., Sharp, P.J., 2005. Proteome analysis of two soft-grained wheats of different processing quality: cultivar-specic proteins. Australian Journal of Agricultural Research 56, 145155. Taub, D.R., Miller, B., Allen, H., 2008. Effects of elevated CO2 on the protein concentration of food crops: a meta-analysis. Global Change Biology 14, 565 575. Tottman, D.R., Broad, H., 1987. The decimal code for the growth stages of cereals, with illustrations. Annals of Applied Biology 110, 441454. Velasco, R., Salamini, F., Bartels, D., 1994. Dehydration and ABA increase mRNA levels and enzyme activity of cytosolic GAPDH in the resurrection plant Craterostigma plantagineum. Plant Molecular Biology 26, 541546. Vensel, W.H., Tanaka, C.K., Cai, N., Wong, J.H., Buchanan, B.B., Hurkman, W.J., 2005. Developmental changes in the metabolic protein proles of wheat endosperm. Proteomics 5, 15941611. Weegels, P.L., Hamer, R.J., Schoeld, J.D., 1996. Critical review: functional properties of wheat glutenin. Journal of Cereal Science 23, 118. Wieser, H., Seilmeier, W., 1998. The inuence of nitrogen fertilization on quantities and proportions of different protein types in wheat our. Journal of the Science of Food and Agriculture 76, 4955. Wieser, H., Manderscheid, R., Erbs, M., Weigel, H.J., 2008. Effects of elevated atmospheric CO2 concentrations on the quantitative protein composition of wheat grain. Journal of Agricultural and Food Chemistry 56, 65316535. Ziska, L.H., Bunce, J.A., 2007. Predicting the impact of changing CO2 on crop yields: some thoughts on food. New Phytologist 175, 607618. Zorb, C., Betsche, L., Langenkamper, G., 2009. Search for diagnostic proteins to prove authenticity of organic wheat grains (Triticum aestivum L.). Journal of Agricultural and Food Chemistry 57, 29322937.

Acknowledgements The authors appreciate Dr M. Erbs, G. Gensheimer, Dr S. Weber and all student assistants for their valued participation in the Hohenheim Mini-FACE experiment. We are grateful to Dr J. Breuer (Universitat Hohenheim, Landesanstalt fur Landwirtschaftliche Chemie) for the analyses of total protein concentrations. The authors thank M. Null-Greulich (Max Rubner-Institut, Federal Research Institute of Nutrition and Food, Detmold) for her excellent technical assistance with the 2-DE. Grateful acknowledgements to J. Hommer (Center for Molecular Medicine, University of Cologne (CMMC), Bioanalytical Laboratory) for processing the MALDI-TOF MS analysis. The authors are grateful to Dr J. Franzaring, G. Gensheimer and Dr A. Klumpp who read the manuscript carefully and suggested improvements. References
Ainsworth, E.A., Rogers, A., 2007. The response of photosynthesis and stomatal conductance to rising [CO2]: mechanisms and environmental interactions. Plant, Cell & Environment 30, 258270. Cotrufo, M.F., Ineson, P., Scott, A., 1998. Elevated CO2 reduces the nitrogen concentration of plant tissues. Global Change Biology 4, 4354. Erbs, M., Fangmeier, A., 2006. Atmospheric CO2 enrichment effects on ecosystems experiments and real world. Progress in Botany 67, 441459. Erbs, M., Franzaring, J., Hogy, P., Fangmeier, A., 2009. Free-air CO2 enrichment in a wheat-weed assembly effects on water relations. Basic and Applied Ecology 10, 358367. Fangmeier, A., De Temmerman, L., Mortensen, L., Kemp, K., Burke, J., Mitchell, R., Van Oijen, M., Weigel, H.J., 1999. Effects on nutrients and on grain quality in spring wheat crops grown under elevated CO2 concentrations and stress conditions in the European, multiple-site experiment ESPACE-wheat. European Journal of Agronomy 10, 215229. Hogy, P., Fangmeier, A., 2008. Effects of elevated atmospheric CO2 on grain quality of wheat. Journal of Cereal Science 48, 580591. Hogy, P., Wieser, H., Kohler, P., Schwadorf, K., Breuer, J., Erbs, M., Weber, S., Fangmeier, A., 2009. Does elevated atmospheric CO2 allow for sufcient wheat grain quality in the future? Journal of Applied Botany and Food Quality 82,114121. Hurkman, W.J., Vensel, W.H., DuPont, F.M., Altenbach, S.B., Buchanan, B.B., 2008. Endosperm and amyloplast proteomes of wheat grain. In: Agrawal, G.K., Rakwal, R. (Eds.), Plant Proteomics: Technologies, Strategies, and Applications. John Wiley & Sons, pp. 207222. Kimball, B.A., LaMorte, R.L., Pinter Jr., P.J., Wall, G.W., Hunsaker, D.J., Adamsen, F.J., Leavitt, S.W., Thompson, T.L., Matthias, A.D., Brooks, T.J., 1999. Free-air CO2 enrichment and soil nitrogen effects on energy balance and evapotranspiration of wheat. Water Resources Research 35, 11791190. Kimball, B.A., Morris, C.F., Pinter, P.J., Wall, G.W., Hunsaker, D.J., Adamsen, F.J., LaMorte, R.L., Leavitt, S.W., Thompson, T.L., Matthias, A.D., Brooks, T.J., 2001. Elevated CO2, drought and soil nitrogen effects on wheat grain quality. New Phytologist 150, 295303. Kimball, B.A., Kobayashi, K., Bindi, M., 2002. Responses of agricultural crops to freeair CO2 enrichment. Advances in Agronomy 77, 293368.

Journal of Cereal Science 50 (2009) 255261

Contents lists available at ScienceDirect

Journal of Cereal Science


journal homepage: www.elsevier.com/locate/jcs

Rheological characterization of refrigerated and frozen non-fermented gluten-free dough: Effect of hydrocolloids and lipid phase
Gabriel Lorenzo a, b, Noem E. Zaritzky a, b, Alicia N. Califano a, *
a b

Centro de Investigacion y Desarrollo en Criotecnologa de Alimentos (CIDCA-CCT La Plata), Facultad de Cs. Exactas, UNLP CONICET, 47 y 116, La Plata (1900), Argentina Area Departamental Ingeniera Qumica, Facultad de Ingeniera, UNLP, Argentina

a r t i c l e i n f o
Article history: Received 5 December 2008 Received in revised form 9 June 2009 Accepted 12 June 2009 Keywords: Gluten-free Dough Texture Rheology Hydrocolloids

a b s t r a c t
This work intended to study the rheological and textural behavior of gluten-free dough for empanadas and pie-crusts production. Traditionally, these products are made with non-fermented wheat-based dough. They are highly consumed in Latin America, but unsuitable for celiac people. Gluten matrix is a major determinant of the properties of dough, thus it must be replaced with other network forming components, such as hydrocolloids. Different hydocolloids were tested: hydroxypropylmethylcellulose (HPMC) and mixtures of xanthan/guar, and xathan/HPMC gums. Three different kinds of lipid phase were also studied; i.e. sunower oil and low and high solid content margarine at two different levels (2030%). Changes in rheological and textural properties during refrigerated storage were evaluated by dynamic oscillatory measurements and puncture and elongation tests on the unbaked dough. Best results were obtained using either of the hydrocolloid mixtures. Besides, the textural characteristics of cooked empanadas were also studied. Freezing before baking did not alter the quality of the crust in the products. ESEM micrographs revealed a continuous matrix formed by hydrocolloid entanglements. Starch granules were homogenously distributed in the dough and acted as inactive llers. An untrained panel accepted the xanthan/HPMC dough with a 74/90 score and it was signicantly preferred over a commercial gluten-free dough. 2009 Elsevier Ltd. All rights reserved.

1. Introduction Theoretical and experimental progress in polymer and biomaterial science has contributed to increase the knowledge of the processing structure rheology relationships of starch and other carbohydrates in food. Many authors have studied the effect of hydrocolloid and protein addition to several bakery products studying their inuence for different food processing conditions (Barcenas and Rosell, 2005; Collar et al., 1999; Conde-Petitt, 2003). Development of gluten-free bakery products involves the application of hydrocolloids to produce a high quality gluten-free food considering the important fact that they must replace the gluten matrix network. Population-based studies, using serological

Abbreviations: D, deformation at break in elongation test (mm); FE, maximum breaking force in elongation test (mN); FP, maximum breaking force in puncture test (mN); XG, xanthan and guar gum mixture; XH, xanthan and hidroxipropylmethyl cellulose mixture; SO, Sunower Oil; RM, Low solid fat content margarine (retail margarine); IM, High solid fat content margarine (margarine for industrial uses); G0 , storage modulus (Pa); G00 , loss modulus (Pa); h*, complex dynamic viscosity (Pa.s). * Corresponding author. Tel.: 54 221 4254853, fax: 54 221 4249287. E-mail address: anc@quimica.unlp.edu.ar (A.N. Califano). 0733-5210/$ see front matter 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2009.06.003

screening, have indicated that the true prevalence of celiac disease is higher than previously thought, at up to 1 in 150 individuals (Dewar et al., 2004). Celiac disease is an autoimmune enteropathy triggered by the ingestion of gluten-containing grains in susceptible individuals. The gliadin fraction of wheat gluten and similar alcohol-soluble proteins in other grains (mainly barley, rye and oat) are the environmental factors responsible for the development of the intestinal damage (Fasano and Catassi, 2001). The pathological changes and symptoms generally resolve with withdrawal of gluten from the diet and a strict adherence to a gluten-free diet throughout the patients lifetime. There is a traditional meal in Latin America called empanadas, which is quite similar to Cornish pasties. The non-fermented dough used for empanadas normally includes wheat our, fat, and water (Lupano, 2003). Small circles (11 cm diam.) of pastry dough are topped with different llings and folded over into a half-moon. The pastry edges are rmly pressed together to seal the lling and uted. Empanadas are baked or fried before eating. This product is commercialized in two distinct ways, either as refrigerated small disks of dough or ready-to-bake frozen empanadas. Low quality of the gluten-free products currently in the market has led to important research in order to improve their poor structure,

256

G. Lorenzo et al. / Journal of Cereal Science 50 (2009) 255261

mouthfeel and avor (Arendt et al., 2002; Gallagher et al., 2002, 2004; Lazaridou et al., 2007; Toufeili et al., 1994). Production of gluten-free pastries involves research to nd an adequate combination of the key components to produce a dough with a good elasticity, resistance to puncture, and stretching. Type of hydrocolloid and lipid phase are two of the most important components, which affect the handling characteristics of the uncooked dough as well as the overall quality of the baked product. Storage of bakery products often leads to deterioration of the overall quality of the product. Upon freezing, moisture in the food transforms into ice, often resulting in physical stress to the food matrix. When a frozen food is thawed for consumption, the moisture is readily separated from the matrix causing a softening of the texture. Hydrocolloids were found to be able to give stability to food products during freezing-thawing cycles, and help to minimize negative effects of the freezing and frozen storage of starch based products (Barcenas et al., 2004; Sanderson, 1996; Selomulyo and Zhou, 2007). Textural properties of the baked dough and sensorial perception of the consumers are crucial aspects to be considered to assess the quality of the nal product. The evaluation of gluten-free dough for empanada production was analyzed in a previous work (Lorenzo et al., 2008) and the contribution of gum, proteins, and water content to the rheological properties of the dough were described. It was found that those formulations containing higher percentages of gums (3%) and lower water content (51%) led to an adequate composition for industrial production. However, higher protein contents (whey protein concentrate and dry egg) interfered with the formation of the three-dimensional gum network making the doughs more fragile. The aim of the present work was to examine changes in the rheological attributes of non-fermented gluten-free dough as affected by composition: i) type of hydrocolloid (hydroxy propylmethylcellulose (HPMC), xanthan/guar, and xathan/HPMC gums), ii) lipid phase (sunower oil, low and high solid content margarines); and to consider the inuence of iii) the storage time (refrigerated and frozen storage), and iv) the baking process on the nal product. 2. Experimental 2.1. Materials Corn starch, (12.5% moisture, 0.3% protein) was obtained from Droguera Saporiti (Argentina); cassava starch (14% moisture and 0.2% protein) from Santa Mara (Argentina). Retail margarine (Flora Danica S.A., Argentina) and 100% sunower oil (Molinos Ro de La Plata SACIFI, Buenos Aires) were purchased from a local supermarket and used without further treatment. High solid fat content margarine (for industrial uses) was from CALSA, Argentina. Whey protein concentrate (WPC) containing 80% protein (Arla Food Ingredients S.A., Argentina), food-grade commercial xanthan and guar gums (Sigma Chemical Co., St. Louis, MO), hydroxypropylmethylcellulose (HPMC), Methocell E4M (The Dow Chemical Company, Michigan), dry egg (6% moisture, 38% lipids, Tecnovo S.A., Argentina), and analytical grade NaCl were used. Distilled water was used in all formulations and ethyl alcohol was added as a preserving agent. 2.2. Dough sample preparation Basic dough formula consisted in a mixture of starches, hydrocolloids (3%), dry egg (3.5%), whey protein concentrate (WPC, 6.5%), NaCl (2%), water, and a lipid phase. Percentage concentrations of the formula are given on a g/100 g total starch basis. Concentration of total hydrocolloids, dry egg, WPC, and NaCl were maintained

constant in all the formulations. Dry ingredients were premixed for 1 min in a commercial food processor (Universo, Rowenta, Germany) at 400 rpm (setting #2). With the processor still running, the lipid phase was added and mixed for one more min. Finally, water was added and the dough was mixed for 5 more min to combine the ingredients. The dough was briey kneaded by hand, wrapped in a lm, put in a tightly sealed container, and kept refrigerated (4  C) for 24 h to let the starches hydrate and to let the dough consistency stabilize (Manley, 2001). Dough was divided in three pieces, and each piece was rolled out with a rolling pin (2025 passes) over a at board. A spacer was placed on each side of the dough to maintain a consistent level while the dough was rolled until a nal thickness of 2 mm. Each dough formulation was prepared twice and both replicates were analyzed. In the description of each assay the number of subsamples taken from each dough batch is recorded. 2.2.1. Effect of hydrocolloids Different hydrocolloids were tested in dough preparation: HPMC (H), a mixture of xanthan/guar gums (XG, ratio 2:1), and xathan/HPMC (XH, ratio 3:2). In all cases 3% total hydrocolloids content, 20% sunower oil and 51% distilled water were used. 2.2.2. Effect of lipid phase Sunower oil, margarine with high solid fat content (for industrial uses), and retail margarine (with low solid fat content) were studied in dough preparation at two different levels (20% and 30%; g/100 g total starch) for the two hydrocolloid mixtures that presented the best textural results in 2.2.1. Studied formulations are shown in Table 1. The fatty acid compositions of the three lipid phases were provided by the producers: Sunower oil: Lauric (0.5), Myristic (0.20), Palmitic (6.80), Palmitoleic (0.10), Stearic (4.70), Oleic (18.60), Linoleic (68.20), Linolenic (0.50), Arachidic (0.40), Gadoleic (1.00); Retail Margarine: Myristic (0.12), Palmitic (11.05), Palmitoleic (0.70), Margaric (0.13), Margaroleic (0.03), Stearic (8.87), Elaidic (36.27), Oleic (24.59), Linoleic (15.85), Linolenic (2.05), Arachidic (0.40), Gadoleic (0.09), Behenic (0.38), Lignoceric (0.11); Industrial Margarine: Myristic (3.20), Myristoleic (0.90), Pentadecanoic (0.80), Pentadecenoic (0.30), Palmitic (26.50), Palmitoleic (3.30), Margaric (1.60), Margaroleic (0.50), Stearic (26.20), Elaidic (4.10), Oleic (29.30), Linoleic (2.20), Linolenic (0.50), Arachidic (0.20), Behenic (0.40). The concentration of water was modied according to the lipid content in order to keep constant the sum of these two ingredients; otherwise consistency of the dough was not adequate to allow its manipulation. 2.3. Solid fat content determination Solid fat content of the margarines (SFC) was measured by pulsed nuclear magnetic resonance (p-NMR) in a Minispec PC/120 series NMR analyzer (Bruker, Karlsruhe, Germany), following the AOCS 16b-93. Prior to NMR analysis, samples were melted at 80  C for 30 min, the NMR tubes were then lled at this temperature and nally, maintained for 30 min at 20  C (temperature test) and 0  C (Martini and Herrera, 2000; Puppo et al., 2002). 2.4. Storage 2.4.1. Refrigerated storage of the dough Disks of dough were packaged under vacuum conditions in bags CN510 (PO2: 35 cm3/(m2day.bar) at 23  C, Sealed Air, Argentina) and kept refrigerated at 4  C for 20 days to perform textural and dynamic rheological analysis approximately every four days.

G. Lorenzo et al. / Journal of Cereal Science 50 (2009) 255261

257

Table 1 Compositions of the tested formulations with different type and concentration of lipids; results obtained by textural analyses: deformation at breaking (D), puncture (FP) and elongation forces (FE), and average values of material strength parameter (Aa) as a function of dough composition. Formulation Gums Lipid phase Type XGSO20 XGSO30 XGRM20 XGRM30 XGIM20 XGIM30 XHSO20 XHSO30 XHRM20 XHRM30 XHIM20 XHIM30 Xanthan gum/guar gum Oil Retail Margarine Industrial Margarine Xanthan gum/HPMC Oil Retail Margarine Industrial Margarine % (wt.) 20 30 20 30 20 30 20 30 20 30 20 30 357agh 104b 285cde 315cdh 262ce 413fg 376afgh 117b 275cde 343adgh 279cde 399fg 294aef 77b 211cd 227cde 284aef 255ade 301af 80b 212cde 202cd 278adef 280aef 14.13aef 7.30bc 9.49bcd 7.72bc 10.82cdf 16.33ae 14.62aef 7.82bc 7.87bc 9.21bcd 12.9adf 16.69ae 81.5a 45.5b 107.7c 136.6d 202.4e 212.8e 55.6ab 36.1b 126.4cd 125.9cd 165.1f 174.9f FP(mN) FE(mN) D(mm) Aa (kPas)

Different superscripts within the same column indicate signicant differences (P < 0.05).

2.4.2. Frozen storage of empanadas Empanadas were prepared using a traditional ground beef lling (Lupano, 2003), frozen, stored at 20 C during 15 days, and then baked (200  C, 20 min) without previously thawing them. Freshly prepared pastries with the same dough formulations were baked and used as controls. 2.5. Texture analyses Textural analyses were performed in a TAXT2i Texture Analyzer (Stable Micro Systems,UK), using the Texture Expert Exceed software supplied by Texture Technologies Corp. For every formulation six repeated measurements (three specimens for each batch) were done and mean values were reported. 2.5.1. Puncture tests On raw specimens, a cylindrical probe 2 mm in diameter at a constant rate of 1 mm/s was used. Tests were performed on disks of 40 mm diameter and 2 mm thick from each dough formulation (Lorenzo et al., 2008). The Volodkevich bite jaws (HDP/VB) probe, which simulates the action of an incisor tooth (Wen-Ching et al., 2007), was used in measuring the texture of the baked nal product, because it has been observed that the rst bite of the product is done with the fore teeth. Maximum breaking force (FP, mN) was determined in both procedures. 2.5.2. Elongation tests Elongation tests were performed on dog-bone shaped specimens of unbaked dough (Lorenzo et al., 2008). Maximum breaking force (FE, mN) and deformation at break (extension at the moment of rupture, D) were obtained from force vs. deformation curves. 2.6. Oscillatory shear tests Small amplitude oscillatory shear tests (storage modulus (G0 ), loss modulus (G00 ) vs. frequency, (u)) were performed in a Controlled Stress Rheometer RS600 (Haake, Germany) using a serrated plate-and-plate geometry (35 mm diameter, 1.6 mm gap). Frequency was swept from 0.01 to 22 Hz. To determine the limit of the linear viscoelastic region, dynamics tests were performed at a xed frequency (1 Hz) and amplitude of the stress was stepwise increased from 0.5 Pa to 100 Pa. Temperature was maintained at 20  C throughout the experiment. Two replicates of each test were performed on the unbaked samples.

2.7. Environmental scanning electron microscopy (ESEM) An environmental scanning electron microscope (Phillip-Electroscan 2010, Netherland) was used to examine the dough surface of different samples. Micrographs of raw and baked doughs were taken without any previous treatment of the samples. 2.8. Sensory assessment Sensory analyses were conducted by forty panellists from graduate students and faculty members in our Institute who were experienced in sensory evaluation of foods, but received no specic training relevant to these products. Panellists evaluated ground beef lled empanadas prepared with two dough formulations, one of the formulations developed in our laboratory and a glutenfree dough purchased from a local market. The samples were simultaneously baked at 200  C during 20 min. Panellists were asked to indicate how much they liked or disliked each product on a 9-point hedonic scale (9 like extremely; 1 dislike extremely) according to avor and texture characteristics. They were also asked to mark which formulation they preferred according to global evaluation. Warm specimens were distributed in white polystyrene plates and presented to the panellist with three-digit codes and in random order for evaluation. Tap water was supplied to the panellist for rinsing between samples. Experiments were conducted in an appropriately designed and lighted room. 2.9. Statistical analysis All statistical analyses were accomplished using the SYSTAT software (SYSTAT, Inc., Evanston, IL). Analyses of Variance (ANOVA) were performed separately for the puncture, elongation, and oscillatory shear tests measurements. Type of hydrocolloid or type and content of lipids were the factors analyzed. Bonferronis test was chosen for simultaneous pairwise comparisons. Differences in means and F-tests were considered signicant when P < 0.05 (95% of condence). The standard error of the mean (SEM) was calculated as the standard deviation divided by the square root of the number of observations. Each mean value was reported as mean standard error of the mean. 3. Results and discussion The basic formulation of the dough was taken from previous work where percentages of dry egg, WPC, water, and gums were optimized (Lorenzo et al., 2008). Initial formulation contained 3%

258

G. Lorenzo et al. / Journal of Cereal Science 50 (2009) 255261

gums, 3.5% dry egg, 6.5% WPC, and 51% water. Percentages are given in g/100 g total starch basis. 3.1. Effect of hydrocolloids on dough rheology

450 400 350

Force (mN)

300 250 200 150 100 50 0 -50

10

15

20

25

Distance (mm)

G'; G'' (Pa)

1.E+05

1.E+04 0.01

0.1

10

100

Frequency (Hz)
Fig. 1. (a) Environmental scanning electron micrograph of a dough containing 20% sunower oil and xanthan/HPMC gums (bars 50 mm) C cassava starch; M corn starch; H hydrocolloid network. (b) Effect of hydrocolloids on dough texture: puncture (P, lled symbols) and elongation (E, open symbols). Xanthan/guar mixture (-; ,); HPMC (C; B), and xanthan/HPMC mixture (:; 6). (c) Effect of hydrocolloids on dough rheology: storage, (G0 , lled symbols) and loss moduli (G00 , open symbols) vs. frequency (u). Xanthan/guar mixture (-; ,); HPMC (C; B), and xanthan/HPMC mixture (:; 6). The values are expressed as mean standard error of the mean.

Fig. 1a presents the dough microstructure for a formulation containing a mixture of xanthan and HPMC gums as an example. All the tested formulations presented qualitatively the same appearance. The ESEM micrograph shows a continuous hydrocolloid network and a random organization of both starch granules, where clusters of either corn or cassava starches cannot be distinguished. Fig. 1b shows the results of textural analyses (puncture and elongation) for the formulations containing different hydrocolloids (HPMC and combinations of xanthan/guar, xanthan/HPMC). Each curve corresponds to the mean of six replicates. It could be observed that those formulations containing a mixture of xanthan/guar gums exhibited the same mean values of FP (366 mN) and FE (297 mN) as those formulated with xanthan/ HPMC (Fig. 1b). When the hydrocolloid matrix was only formed by HPMC the samples presented extremely low values for FP, FE, and D (104 mN, 77 mN, 7 mm, respectively). This fragile behavior turns the formulation into an unsuitable dough for industrial handling since it would not resist the large stretching forces leading to cracks and tears in the dough disks. Moreover, this lack of resistance in the structure would cause spills during baking, producing an important quality loss in the nal product. Commercial dough disks formulated with wheat our were measured for comparison. They presented FE 344 mN (SEM: 9.9) and FP 325 mN (SEM: 16.6), which were similar to those formulated with the tested hydrocolloid mixtures and much larger than the HPMC formulation. The results of dynamic rheological measurements in the linear viscoelastic range were expressed in terms of the storage modulus (G0 ) and loss modulus (G00 ). Results of the dynamic oscillatory tests are presented in Fig. 1c for the three formulations; the curves were qualitatively similar for all the formulations assayed. G0 was always greater than G00 in the frequency range measured and the increase of the two moduli with frequency was small. Two characteristic regions may be distinguished: a pseudo-terminal region at low frequencies that shows a tendency to a crossover of both viscoelastic functions, and the plateau region. The plateau region is an intermediate zone between the terminal and the transition zones (Ferry, 1980). It is characterized by a decrease in the slope of both moduli (lower than 1) and a possible minimum in the loss modulus (G00 ). In the present work the frequency range in Fig. 1b entirely corresponds to what is known as the plateau zone. The power law exponent of the storage modulus is lower than 0.2 in the whole range, indicating a slight dependence with frequency. The plateau region in G0 is related to the formation of physical entanglements among polymeric chains that form a three-dimensional network of interacting molecules (Ferry, 1980). Several authors have reported a similar trend for our dough with G0 and G00 increasing with frequency (Agyare et al., 2004; Dreese et al., 1988; Kenny et al., 2001; Lefebvre et al., 2003; Lorenzo et al., 2008; Ribotta et al., 2004). Doughs formulated with xanthan/guar gums presented the highest values of elastic modulus, while the lowest values corresponded to the HPMC doughs (with or without xanthan gum) in the whole frequency range analyzed (P < 0.05). Guarda et al. (2004) detected higher water absorption on wheat bread when adding HPMC instead of xanthan gum. This effect has been attributed to the hydroxyl groups in the hydrocolloid structure which allow more water interactions through hydrogen bonding (Rosell et al., 2001). The lower G0 and G00 curves for both formulations containing HPMC (P < 0.05), as well as the observed FP and FE values, could be attributed to this softening phenomenon.

G. Lorenzo et al. / Journal of Cereal Science 50 (2009) 255261

259

3.2. Effect of lipid phase on dough rheology The effect of type and content of the lipid phase was evaluated in those formulations that presented the best textural behavior in Section 3.1, i.e. those with xanthan/guar and xanthan/HPMC mixtures. Table 1 shows the maximum forces in puncture (FP) and elongation (FE) tests for all the assayed formulations and also the extensibility or deformation at the breaking point (D). Textural analyses did not detected any signicant differences (P < 0.05) between both hydrocolloid combinations used, regardless of the lipid type and concentration. When margarine was used, the increase in fat content produced doughs with higher resistance to puncture, while the opposite effect was found using sunower oil. Both type and lipid content also affected FE and D signicantly (P < 0.05). Margarine provided ductility to the dough; more markedly in the case of the industrial margarine which contained higher solid fat content (46.8%) than the retail (all purpose) one (24.7%). The formulation with high sunower oil content showed the lowest FP FE, and D, values, making it the least suitable for industrial handling (Fig. 2 and Table 1). The effect of lipid phase on the viscoelastic characteristics of the doughs was analyzed using the Friedrich and Heymann theory (Friedrich and Heymann, 1988; Lorenzo et al., 2008). According to this theory, complex viscosity (h*) could be expressed as

hydrocolloid mixture, the decrease in Aa value was more noticeable in the doughs containing industrial margarine. The order of the relaxation function (a) was lower than 0.2 for all formulations, indicating the pronounced elastic character of the doughs, typically observed in gel-like samples (Doublier et al., 1992; Steffe, 1996). However, a was signicantly higher for those formulations containing oil (a 0.19) than those values corresponding to doughs made with margarine, either retail or industrial (a 0.15), since a liquid lipid phase decreased the solid characteristics of the matrix.

3.3. Storage 3.3.1. Refrigerated storage of dough disks The 20% sunower oil or 30% industrial margarine showed the highest values for maximum breaking forces (FP and FE) as well as the highest deformation at break but did not differ signicantly among themselves; thus, formulations with the lower lipid content (healthier formulations), i.e. XGSO20 and XHSO20 (Table 1) , were chosen to analyze the effect of refrigerated storage (4  C). Their rheological and textural behavior was evaluated in dough disks at 1, 5, 8, 12, 16, and 20 days. FP (364 mN), FE (290 mN), remained constant during refrigerated storage for both formulations and signicant differences were not detected between them (P < 0.05). G0 , and G00 did not change either during the storage time. However, the mixture of xanthan/ SEM 7.0$103 Pa; G00 2.30$104 Pa, guar (G0 1.41$105 Pa, 2 SEM 7.7$10 Pa at 1 Hz) produced a more elastic dough than the xanthan/HPMC formulation (G0 9.88$104 Pa, SEM 3.1$103 Pa; G00 1.92$104 Pa, SEM 7.7$102 Pa at 1 Hz). The deformation at breaking point in the elongation tests (D) remained constant for the formulation containing HPMC. On the other hand, XGSO20 presented a decrease of the extensibility as is shown in Fig. 3. The capability to retain water of hydroxyl groups present in HPMC could probably be responsible for the longer keepability of the dough. A similar trend was observed in wheat bread when HPMC was added (Collar et al., 1998). 3.3.2. Frozen storage of ready-to-bake empanadas The effect of frozen storage of ready-to-bake empanadas on the microstructure and textural behavior of the baked nal product was evaluated. Three formulations: XGSO20, XHSO20, and XHIM30

G*

p G02 G002 zAa ua1

(1)

where a is the order of the relaxation function and Aa the material strength parameter, related to the rigidity of the network. Equation (1) was tted to the frequency sweep curves and average values of the material strength parameter obtained for each formulation are listed in the last column of Table 1. A marked increment of Aa with the solid fat content of the lipid phase is observed. When the matrix contained higher solid content (industrial margarine) it showed more elastic behavior, which was reected in an increase in Aa (Aa(IM) > Aa(RM)). Additionally, type of hydrocolloids signicantly affected the material strength of the dough (P < 0.05). When HPMC replaced guar gum in the

300 250

RM20 IM20 SO20

RM30 IM30 SO30

20 18

Deformation at break (mm)

200

16 14 12 10 8 6 4 2 D XH D XG 0 5 10 15 20 25

FE (mN)

150 100 50 0 0 5 10 15 20 25

D (mm)
Fig. 2. Effect of lipid phase on dough elongation tests (E). Results obtained for the formulations with xanthan/HPMC mixture. Sunower oil (C; B), industrial margarine (high solid fat content) (:; 6), and retail margarine (low solid fat content) (-; ,). Filled symbols correspond to 20% and open symbols to 30% lipid content.

Storage time (days)


Fig. 3. Effect of refrigerated storage on deformation at breaking point (D) for formulations containing 20% sunower oil with: xanthan/guar (:) and xanthan/HPMC (-). The values are expressed as mean standard error of the mean.

260

G. Lorenzo et al. / Journal of Cereal Science 50 (2009) 255261

(Table 1) were tested. The formulation with industrial margarine was included to examine whether a signicant difference would be detected after the baking process. Baking irreversibly alters the structural nature of dough constituents through a series of physical, chemical and biochemical reactions (Pyler, 1988). Temperature, humidity, and duration of baking inuence the nal product. Oven heat is responsible for the formation of an enveloping crust, coagulation of proteins, gelatinization of starch, and the stabilization of the colloidal dough system (Freeman and Shelton, 1991). Microscopic observations revealed that the crust appeared as a continuous sheet of proteins and hydrocolloids with embedded starch granules, which were not gelatinized because of the rapid dehydration of the surface at the oven temperature (Fig. 4a). The inner structure of the dough showed both intact and gelatinized granules (Fig. 4b and c). When going from the crust to the interior of the dough, less intact starch granules were observed (Fig. 4b). Near the lling, gelatinization is more complete and a distinction cannot be made between starch and the hydrocolloid matrix (Fig. 4c), while in the middle region of the dough (between crust and lling), starch granules are still recognizable. The limited water content of the dough may account for these differences. Fig. 4d presents a photograph of the nal baked product. The maximum puncture force of the baked specimens did not present signicant differences (P < 0.05) between the control (unfrozen) and frozen samples. Besides, FP was not affected by the

storage time in any of the three tested compositions. This effect is in agreement with Sanderson (1981) who has stated that xanthan gum induced cooking and cooling stability of wheat our-based products and improved the freezethaw stability of starch-thickened frozen foods. This result implies an important technological advantage since the product could be frozen without changing its textural quality. 3.4. Sensory assessment The forty experienced panelists evaluated the appearance, texture, avor and overall acceptability of the formulations that have been stored for ten days at 4  C. A formulation developed with xanthan gum/HPMC mixture and 20% sunower oil (XHSO20) was compared with a commercial gluten-free dough. There were no signicant differences in avor, texture and overall acceptability between both formulations (P < 0.05). The appearance of the commercial dough was perceived as signicantly worse by the panel, which was related with the notorious cracks that appeared in these products during baking. The observations were classied into three perception sensorial groups, the rst one corresponded to those that disliked the product (scores 1 to 4, dislike extremely to slightly), the second one was indifferent (scores 5), and the third group expressed that they liked the samples (scores 6 to 9, like slightly to like extremely). More than 70% of the panelists liked all the attributes of the

Fig. 4. (a) Photograph of the nal baked product (bars 5 cm); (bd) Environmental scanning electron micrograph of a dough containing 20% sunower oil and xanthan/HPMC gums (b) baked dough, external surface, crust (bars 200 mm); (c) baked dough, middle region between the crust and the internal surface (bars 200 mm); (d) baked dough, internal surface (bars 200 mm).

G. Lorenzo et al. / Journal of Cereal Science 50 (2009) 255261 Table 2 Sensory scores of gluten-free empanadas dough. Percentage of panelists that scored each tested property between 6 and 9 is given between parentheses. Appearance XHSO20 Commercial dough 7.4a (88.9%) 6.5b (75%) Texture 6.8a (70.1%) 6.9a (83.1%) Flavor 7.1a (80.6%) 7.0a (83.3%) Global Acceptability 7.5a (94.4%) 7.0a (80.6%) Preferences 67.5%a 32.5%b

261

Different superscripts within the same column indicate signicant differences (P < 0.05).

products. In particular, over 94% of the panelists liked the formulation containing xanthan/HPMC mixture and sunower oil. Two thirds of the panelists preferred formulation XHSO20 over the commercial dough (Table 2) (P < 0.05). It is important to remark that none of the formulations developed in the present work showed signs of rupture during baking. Meanwhile, more than 90% of the commercial gluten-free disks presented this problem, which is an important parameter to consider since crust rupture produces a low quality product. 4. Conclusion Textural and rheological behavior of non-fermented gluten-free doughs, a product intended for celiac people, was studied. The study of different hydrocolloid mixtures revealed that formulations containing xanthan gum exhibited the best elasticity and resistance to puncture regardless of the other hydrocolloids present in the dough (guar or HPMC). Formulations containing only HPMC were the least suitable for industrial production of empanadas or piecrusts. When margarine was used, the increase in fat content produced doughs with higher resistance to puncture, while the opposite effect was found using sunower oil. Margarine for industrial uses (with higher solids content) provided greater elasticity to the dough than the retail margarine rendering it more ductile. Doughs were stored at 4  C for twenty days without any signicant changes in rheological properties. Frozen storage did not affect textural characteristics of baked dough either. The panel accepted the xanthan/HPMC dough with a 75/90 score and it was signicantly preferred over a commercial gluten-free dough. Acknowledgments The authors are grateful to Tecnovo, S.A., Argentina and to Arla Food Ingredients S.A., Argentina, who provided the dried egg and milk protein concentrate for this study, respectively. The nancial support of the Consejo Nacional de Investigaciones Cientcas y Tecnologicas (CONICET), Agencia Nacional de Promocion Cientca y Tecnologica and Universidad Nacional de La Plata are also acknowledged. References
Agyare, K.K., Xiong, Y.L., Addo, K., Akoh, C.C., 2004. Dynamic rheological and thermal properties of soft wheat our dough containing structured lipid. Journal of Food Science 69, 297302. Arendt, E.K., Obrien, C.M., Schober, T., Gormley, T.R., Gallagher, E., 2002. Development of gluten-free cereal products. Farm and Food 12, 2127. Barcenas, M.E., Benedito, C., Rosell, C.M., 2004. Use of hydrocolloids as bread improvers in interrupted baking process with frozen storage. Food Hydrocolloids 18, 769774.

Barcenas, M.E., Rosell, C.M., 2005. Effect of HPMC addition on the microstructure, quality and aging of wheat bread. Food Hydrocolloids 19, 10371043. Collar, C., Andreu, P., Martnez, J.C., Armero, E., 1999. Optimization of hydrocolloid addition to improve wheat bread dough functionality: a response surface methodology study. Food Hydrocolloids 13, 467475. Collar, C., Armero, E., Martnez, J., 1998. Lipid binding of formula bread doughs. Zeitschrift fur Lebensmitteluntersuchung und -Forschung A 207, 110121. Conde-Petitt, B., 2003. The structure and texture of starch-based foods. In: Mackena, B.M. (Ed.), Texture in Food, Vol 1: Semi-solid Foods. CRC Press, Boca Raton, pp. 87110. Dewar, D., Pereira, S.P., Ciclitira, P.J., 2004. The pathogenesis of coeliac disease. The International Journal of Biochemistry & Cell Biology 36, 1724. Doublier, J.L., Launay, B., Cuvelier, G., 1992. Viscoelastic properties of food gels. In: Rao, M.A., Steffe, J.F. (Eds.), Viscoelastic Properties of Foods. Elsevier Science, Barking, pp. 371434. Dreese, P.C., Faubion, J.M., Hoseney, R.C., 1988. Dynamic rheological properties of our, gluten, and gluten-starch doughs. I. Temperature-dependent changes during heating. Cereal Chemistry 65, 348353. Fasano, A., Catassi, C., 2001. Current approaches to diagnosis and treatment of celiac disease: an evolving spectrum. Gastroenterology 120, 636651. Ferry, J.D., 1980. Viscoelastic Properties of Polymers. John Wiley & Sons Inc., New York. Freeman, T.P., Shelton, D.R., 1991. Microstructure of wheat starch: from kernel to bread. Food Technology 45, 162168. Friedrich, C.H.R., Heymann, L., 1988. Extension of a model for crosslinking polymer at the gel point. Journal of Rheology 32, 235241. Gallagher, E., McCarthy, D., Gormley, R., Arendt, E., 2004. Improving the Quality of Gluten-Free Products. Final Report: Project RMIS No. 4881. Agriculture and Food Development Authority, Dublin, Ireland. Gallagher, E., Polenghi, O., Gormley, T.R., 2002. Improving the quality of gluten-free breads. Farm and Food 12, 813. Guarda, A., Rosell, C.M., Benedito, C., Galotto, M.J., 2004. Different hydrocolloids as bread improvers and antistaling agents. Food Hydrocolloids 18, 241247. Kenny, S., Wehrle, K., Auty, M., Arendt, E.K., 2001. Inuence of sodium caseinate and whey protein on baking properties and rheology of frozen dough. Cereal Chemistry 78, 458463. Lazaridou, A., Duta, D., Papageorgiou, M., Belc, N., Biliaderis, C.G., 2007. Effects of hydrocolloids on dough rheology and bread quality parameters in gluten-free formulations. Journal of Food Engineering 79, 10331047. Lefebvre, J., Pruska-Kedzior, A., Kedzior, Z., Lavenant, L., 2003. A phenomenological analysis of wheat gluten viscoelastic response in retardation and dynamic experiments over a large time scale. Journal of Cereal Science 38, 257267. Lorenzo, G., Zaritzky, N., Califano, A., 2008. Optimization of non-fermented gluten-free dough composition based on rheological behavior for industrial production of empanadas and pie-crusts. Journal of Cereal Science 48, 224231. Lupano, C.E., 2003. Discs for "Empanadas" with whey protein concentrate. Journal of Food Technology 1, 182186. Manley, D., 2001. Biscuit, Cracker and Cookie Recipes for the Food Industry. Woodhead Publishing Ltd., Cambridge. Martini, S., Herrera, L., 2000. Crystallization behavior of milk fat model systems: kinetics and polymorphism. In: Mohan, R.M. (Ed.), Research Advances in Oil Chemistry. Global Research Network, Kerala, pp. 123. Puppo, M.C., Martini, S., Hartel, R.W., Herrera, M.L., 2002. Effects of sucrose esters on isothermal crystallization and rheological behavior of blends of milk-fat fraction sunower oil. Journal of Food Science 67, 34193426. Pyler, E.J., 1988. In: Baking Science and Technology, third ed., vol. II. Sosland Pub Co, Kansas City. Ribotta, P.D., Perez, G.T., Leon, A.E., Anon, M.C., 2004. Effect of emulsier and guar gum on microstructural, rheological and baking performance of frozen bread dough. Food Hydrocolloids 18, 305313. Rosell, C.M., Rojas, J.A., Benedito de Barber, C., 2001. Inuence of hydrocolloids on dough rheology and bread quality. Food Hydrocolloids 15, 7581. Sanderson, G.R., 1981. Polysaccharides in food. Food Technology 35, 5057. 83. Sanderson, G.R., 1996. Gums and their use in food systems. Food Technology 50, 8184. Selomulyo, V.O., Zhou, W., 2007. Frozen bread dough: effects of freezing storage and dough improvers. Journal of Cereal Science 45, 117. Steffe, J.F., 1996. Rheological Methods in Food Process Engineering. Freeman Press, East Lansig. Toufeili, I., Dagher, S., Shadarevian, S., Noureddine, A., Sarakbi, M., Farran, M.T., 1994. Formulation of gluten-free pocket-type at breads: optimization of methylcellulose, gum arabic, and egg albumen levels by response surface methodology. Cereal Chemistry 71, 594601. Wen-Ching, K., Wen-Chian, L., Yu-Ting, T., Chang-Wei, H., 2007. Kinetics of winter mushrooms (Flammulina velutipes) microstructure and quality changes during thermal processing. Journal of Food Engineering 81, 587598.

Journal of Cereal Science 50 (2009) 262265

Contents lists available at ScienceDirect

Journal of Cereal Science


journal homepage: www.elsevier.com/locate/jcs

Comparison of color techniques to measure the color of parboiled rice


Bin Lv, Bin Li*, Sha Chen, Jian Chen, Bo Zhu
College of Food Science and Technology, Huazhong Agricultural University, Wuhan, 430070, China

a r t i c l e i n f o
Article history: Received 11 November 2008 Received in revised form 25 April 2009 Accepted 12 June 2009 Keywords: Parboiled Rice Color Vision Builder

a b s t r a c t
A combination of digital camera, computer and graphic software can provide a less expensive and more versatile technique to determine the surface color of parboiled rice compared to instrumental color measurement. The instrument was used to measure rice powder and whole rice. Pearson correlation coefcients and sample paired t-test on total color difference (DE), L and b values were calculated. The value of DE of samples from the instrumental technique was 0.694.61 (powder), 4.710.2 (whole rice) with a coefcient of variance (CV) ranging from 3.5 to 25.3% and from 15.4 to 46.6%. Meanwhile, the digital image technique gave a DE value ranging from 4.2 to 13.77 with a CV from 6.3 to 21.2%, respectively. A highly signicant correlation (R2 0.7451, R2 0.8074, R2 0.7518,) was obtained for DE between instrumental (powder and whole rice), Vision Builder and instrumental (powder), and instrumental (whole rice) and Vision Builder. The chromatic b value of instrumental for powder had a signicant correlation with the Vision Builder data (R2 0.7741). The results suggest that although the digital image provided the surface color of parboiled rice, it was less accurate than the instrumental for powder. Therefore results from the digital image should be interpreted with caution. 2009 Elsevier Ltd. All rights reserved.

1. Introduction Rice is one of the leading food crops of the world and is the staple food of over half of the worlds population (Ali and Pandya, 1974). Different methods of production for rice lead to different results. It was reported that about one-fth of the worlds rice is parboiled (Bhattacharya, 1985), which is a hydrothermal treatment applied to raw paddy or brown rice (Soponronnarit et al., 2006), including soaking, steaming and drying. During parboiling, the physico-chemical qualities of rice change a lot, such as starch gelatinization (Kaddus Miah, 2002) and starch polymorphism (Mei et al., 1995). Parboiling also improves the yields of head rice (Elbert et al., 2001) and the nutritional value of rice (Rao, 1966). The latest research shows that a higher head rice yield of 93.3% was obtained with rice parboiled with the higher process steam pressure of 5.5 104 N/m2 (Agidi et al., 2008). However, a problem which concerns most consumers is that parboiling affects the color of milled rice. The milled rice becomes darker and more yellow after parboiling. The degree of color change during parboiling is affected by different factors including soaking

Abbreviations: AI, automated inspection; CV, coefcient of variance; DOM, degrees of milling; DE, total color difference; IMAQ, image acquisition. * Corresponding author. Tel.: 86 27 63730040; fax: 86 27 87282966. E-mail address: libinfood@mail.hzau.edu.cn (B. Li). 0733-5210/$ see front matter 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2009.06.004

temperature and duration, steaming temperature and duration and drying methods (Elbert et al., 2001; Pillaiyar and Mohandoss, 1981; Jayanarayanan, 1965; Johnson, 1965). It has been hypothesized that the color changes during parboiling are caused by (i) migration of husk and/or bran pigments, (ii) non-enzymic browning of the Maillard type and (iii) enzymic color changes that occur during soaking (Lamberts et al., 2006a, 2008). Several researchers have studied the color of parboiled rice with instrumental color measurement (Lamberts et al., 2006a, 2008; Elbert, 2001; Bhattacharya, 1995). It is necessary to mention here that the tested samples were powders (Lamberts et al., 2006a). They could not reect the surface color of parboiled rice which truly affects the rice value in the market place. Lamberts et al. (2006b, 2008) studied the surface color of whole parboiled rice. The coefcient of variation (CV) for the determination of the color parameters was relatively high. However, visual examination could be employed to estimate the surface color of parboiled rice with the limitation of dependence on the person judging the color and could not be used for quality control in rice processing industries. Thus, a new method should be established urgently to determine the color of the parboiled rice surface rapidly and precisely. National Instruments Vision Builder for Automated Inspection (AI) is a congurable machine vision development environment that requires no programming. IMAQ (image acquisition) Vision Builder has been applied in many elds including food quality,

B. Lv et al. / Journal of Cereal Science 50 (2009) 262265

263

material science, and environmental engineering (Nopens et al., 2008; Nian et al., 2006; Thalmann et al., 2003). However, there was no study on the color of parboiled rice with Vision Builder, which could determine the surface color of parboiled rice rapidly, precisely and at low cost. The advantage of this tool is that it can measure the color proles on the whole surface of the parboiled rice at one time. However, the use of Vision Builder to measure the color of food is still limited because the method is relatively new compared to the instrumental techniques. In this study, parboiled rice was prepared by various processes. The color of parboiled rice was studied by two methods, instrument (rice powder and whole rice) and Vision Builder. The objective of this study was (i) to develop a method to evaluate the surface color of parboiled rice, which affects the acceptability to consumers; (ii) to evaluate the accuracy of Vision Builder compared with the instrumental method in measuring the color of parboiled rice.

2.2. Moisture content Moisture contents were determined according to AACC-method 44-15A (American Association of Cereal Chemists, 2000).

2.3. Color measurement with the instrument Color measurements of the milled parboiled rice were obtained using a Hunterlab Labscan spectro colorimeter (Hunter Associates Laboratory, Reston, VA, USA). Prior to color measurement, the instrument was calibrated with a white and black calibration tile. Color measurements were made at least 10 fold on powder samples (80100 mesh) and whole rice placed in a clear Petri dish. The color was measured in CIE 1976 L, a, b color space. L is a measure of the brightness from black (0) to white (100). Parameter a describes redgreen color with positive a values indicating redness and negative a values indicating greenness. Parameter b describes yellowblue color with positive b values indicating yellowness and negative b values indicating blueness (Good, 2002). The total color difference (DE) was calculated (Rhim et al., 1989) from the Hunter L, a and b values according to Eq. (1).

2. Materials and methods 2.1. Parboiling One local long grain rough rice named Huanghuazhan, grown industrially, was selected for the present study. The mean length/ breadth ratio of 10 randomly selected grains was 3.4. The apparent amylose content, determined by iodine colorimetric method, was 17.38% (Knutson, 1986), and thousand kernel weights were 22 0.8 g. 2.1.1. Soaking Samples (100 g) of paddy rice were soaked in 150 mL boiled water in a 500-mL beaker, and the water temperature was decreased to 6070  C immediately. Then the beaker was placed in a water bath shaker (180 rpm) maintained at 60 0.5  C for 4 h as suggested by Adhikaritanayake and Noomhorm (1997) with little modication. In order to minimize splitting of the kernels (Luh and Mikus, 1980) and avoid discoloration (Saif et al., 2003), it is important that the soaking temperature be below the gelatinization temperature. 2.1.2. Tempering The tempering was performed after the soaking step. The soaked grains were quickly withdrawn and transferred to an adiabatic container which was sealed and maintained at 25  C for 24 h. In these conditions the mean moisture content of the grain remained constant. 2.1.3. Steaming After tempering, the grains were steamed in a laboratory cooking chamber. Then samples were steamed for 10, 20, 30, 40, 50, 60 min at temperatures of 80 and 90  C. Considering the organoleptic properties of rice, the longest steaming times were 40 and 30 min at 95  C and 100  C, respectively (Islam et al., 2002). 2.1.4. Drying The grains were then air-dried at room temperature in the absence of direct sunlight to avoid any stress on the kernels until the water content was 1214% (WB). 2.1.5. Milling Samples were dehusked with a rice husker. Brown rice was polished with a test rice polisher. The milling time was 40 s. Degrees of milling (DOM) were 7.5% (non-parboiled rice) and 6.2% (parboiled rice).

DE L L0 2 a a0 2 b b0 2

i0:5

(1)

2.4. Color measurement with Vision Builder Before the experiment, an image analysis set-up was built (Fig. 1) in order to standardize the procedure of taking pictures. The set-up consisted of a digital camera (FUJIFILM FinePix S5800/S800 10 optical zoom) and two TL-D light sources (Philips, China). The samples were placed on a support at a height of 21 cm from the table. On this support, a black piece of cloth was placed to form the background of the images. The resulting pictures from the digital camera are in TIFF format and have a resolution of 8 megapixels (3264 2448). The pictures are in 24-bit RGB-color format (8 bits for redgreenblue) resulting in a le size of 1.51.6 Mb per picture. The collected pictures were analyzed using an image analysis procedure developed in the IMAQ Vision Builder AI 2.6.1 (National instruments, USA). Image analysis procedures were as follows:

Fig. 1. Image analysis set-up.

264

B. Lv et al. / Journal of Cereal Science 50 (2009) 262265

1. Use the Simulate Acquisition step as the rst step in your inspection sequence when you want to explore Vision Builder AI steps without connecting a camera to your computer. In the Step Name control, enter a descriptive name for the step. 2. Click Measure Features and select Measure Colors. The step determines the histogram of the pixel color values for the selected color space (CIE L*a*b*) in the region of interest. Use an image with a white area in it, and then select the CIE XYZ color space. Create a region of interest that only includes the white part of the image. Record the average X, Y, and Z values. Then switch to the CIE L*a*b* color space and use the previously recorded X, Y, and Z values as the White Reference. 3. After that, choose the broken line tool, then draw a broken line along the edge of a kernel of rice in the image. Record the average CIE L*, a* and b* values. All tests were carried out 10 times. In this software, L* is given in pixels (256 levels, ranging from values of 0 to 255) and for its conversion into L, the following equation was applied (Papadakis et al., 2000):

15

E instrument (whole rice)

13 11 9 7 5 3 0

y = 0.9776x + 4.8361 R2 = 0.7451

E instrument (powder)
Fig. 2. Relationship of DE between instrument (powder) and instrument (whole rice) of parboiled rice.

L L*=250 100
2.5. Statistical analyses

(2)
E instrument (powder)

6 5 4 3 2 1 0 3 5 7 9 y = 0.4074x - 1.211 R2 = 0.8074

All data were analyzed with SPSS (version 16.0; SPSS Inc, Chicago, IL, USA). The paired samples t-test was conducted for comparison purposes between instrumental and Vision Builder with a 95% condence interval. 3. Results and discussion The color change of parboiled rice has existed for a hundred years, and many researchers have investigated the mechanism of color change during parboiling, the effect of processing on the color change, and also measured the L, a, b, and DE by instruments. The means of DE attained from the instrument and Vision Builder for each sample with various parboilings ranged from 0.69 to 4.61 (powder), 4.7 to 10.2 (whole rice) and 4.2 to 13.77, respectively. Standard deviation varied from 0.102 to 0.308 (powder), 0.901 to 4.015 (whole rice) for instrument and from 0.341 to 1.621 for Vision Builder, respectively. DE for powder with instrument and Vision Builder had coefcients of variation (CV) of 3.525.3% and 6.321.2%. However, the measurement using the instrument for whole rice showed a relatively high range of CV (15.446.6%), which may be due to the void space among the whole rice. Although there were different levels of accuracy, a highly signicant correlation (R2 0.7451, Fig. 2; R2 0.8074, Fig. 3; R2 0.7518, Fig. 4) was obtained for DE between instrumental (powder and whole rice), Vision Builder and instrumental (powder), and instrumental (whole rice) and Vision Builder for color determination of parboiled rice. Generally, values of L and b were used to evaluate the color of parboiled rice. The means of L values and standard deviation of the instrument for powder and whole rice varied from 89.30 to 92.23 and 0.15 to 0.38, 55.15 to 60.74 and 0.91 to 4.22, whereas those of Vision Builder varied from 90.84 to 93.43 and 0.35 to 1.58. The CV of the instrument for powder and whole rice ranged from 0.1 to 0.4% and 1.5 to 7.4% compared with 0.1 to 1.7% for Vision Builder. Obviously, the measurement of the instrument for whole rice was less accurate for the L value. The chromatic b value of the instrument for powder had a signicant correlation with the Vision Builder data (R2 0.7741, Fig. 5), not found with the instrument for whole rice. The means of b values and standard deviation of the instrument for powder and

11

13

15

E vision builder
Fig. 3. Relationship of DE between Vision Builder and instrument (powder) of parboiled rice.

whole rice varied from 6.24 to 10.24 and 0.09 to 0.22, 8.59 to 11.09 and 0.42 to 1.71, whereas those of Vision Builder varied from 6.22 to 15.82 and 0.29 to 1.73. The measurement with the instrument for powder had lower CV (0.92.6%) compared with the instrument for whole rice (4.419.3%) and Vision Builder (4.721.5%). These ndings imply that both instruments for whole rice and Vision Builder cannot be used as indicators for chromatic b value (from blue to yellow) in parboiled rice. This may be because the surface color of parboiled rice was not uniform.
18 16 14 12 10 8 6 4 2 0 4 6 8 10 12 y = 1.6886x - 3.6393 R2 = 0.7518

E vision builder

E instrument (whole rice)


Fig. 4. Relationship of DE between instrument (whole rice) and Vision Builder of parboiled rice.

B. Lv et al. / Journal of Cereal Science 50 (2009) 262265

265

20

b value of vision builder

18 16 14 12 10 8 6 4 2 0 5 6 7 8 y = 1.7278x - 3.6568 R2 = 0.7741 9 10 11

Vision Builder in the color measurement of parboiled rice. However, the two methods evaluate the color of parboiled rice in two different ways. The results suggest that data from the instrument for powder could estimate the surface color of parboiled rice, which concerns most consumers. The detailed relationship between the two methods should be explored in the future.

Reference
Adhikaritanayake, T.B., Noomhorm, A., 1997. Effect of continuous steaming on parboiled rice quality. Journal of Food Engineering 36, 135143. Agidi, G., et al., 2008. Effect of variety, pressure and specic volume of steam on the head rice yield of milled parboiled rice. Journal of Food Science and Technology Mysore 45 (3), 282283. Ali, N., Pandya, A.C., 1974. Basic concept of parboiling of paddy. Journal of Agricultural Engineering Research 19, 111115. American Association of Cereal Chemists, 2000. Approved methods of the AACC, 10th ed. AACC-method. 4415A. Bhattacharya, K.R., 1985. Parboiling of rice. In: Juliano, B.O. (Ed.), Rice Chemistry and Technology, second ed. American Association of Cereal Chemists, Inc., St Paul, Minnesota, pp. 289348. Bhattacharya, S., 1995. Kinetics on colour changes in rice due to parboiling. Journal of Food Engineering 29, 99106. Elbert, G., Tolaba, M., Suarez, C., 2001. Effects of drying conditions on head rice yield and browning index of parboiled rice. Journal of Food Engineering 47, 3741. Good, H., 2002. Measurement of color in cereal products. Cereal Foods World 4, 56. Islam, R., Roy, P., Shimizu, N., Kimura, T., 2002. Effect of processing conditions on physical properties of parboiled rice. Food Science and Technology Research 8 (2), 106112. Jayanarayanan, E.K., 1965. Inuence of processing conditions on the browning of parboiled rice. Rice Journal 68, 1617. Johnson, R.M., 1965. Light reectance meter measures degree of milling and parboiling of parboiled rice. Cereal Chemistry 42, 167174. Kaddus Miah, M.A., 2002. Parboiling of rice. Part II: effect of hot soaking time on the degree of starch gelatinization. International Journal of Food Science and Technology 37, 539545. Knutson, C.A., 1986. A simple colorimetric procedure for determination of amylose in maize starch. Cereal Chemistry 63, 8992. Lamberts, L., et al., 2006a. Impact of browning reactions and bran pigments on color of parboiled rice. Journal of Agricultural and Food Chemistry 54, 99249929. Lamberts, L., et al., 2006b. Effect of processing condition on color change of brown and milled parboiled rice. Cereal Chemistry 83, 8085. Lamberts, L., et al., 2008. Impact of parboiling conditions on Maillard precursors and indicators in long-grain rice cultivars. Food Chemistry 110 (4), 916922. Luh, B.S., Mikus, R.R., 1980. Parboiled rice. In: Luh, B.S. (Ed.), Rice Production and Utilization. AVI, Westport, CT, pp. 501542. Mei, H., Ong, J., Blanshard, M.V., 1995. The signicance of starch polymorphism in commercially produced parboiled rice. Starch/Starke 47 (1), 713. Nian, C.Y., et al., 2006. A new algorithm for a three-axis auto-alignment system using vision inspection. Journal of Materials Processing Technology 171, 319329. Nopens, I., Foubert, I., De Graef, V., van Laere, D., Dewettinck, K., Vanrolleghem, P., 2008. Automated image analysis tool for migration fat bloom evaluation of chocolate coated food products. Food Science and Technology 41, 18841891. Papadakis, S.E., et al., 2000. A versatile and inexpensive technique for measuring colour of foods. Food Technology 54 (12), 4851. Pillaiyar, P., Mohandoss, R., 1981. Hardness and color in parboiled rices produced at low and high temperatures. Journal of Food Science and Technology 18, 79. Rhim, et al., 1989. Kinetics of color change of grape juice generated using linearly increasing temperature. Journal of Food Science 54, 776777. Saif, S.M.H., et al., 2003. Gelatinisation properties of rice and our. International Journal of Food Properties 6 (3), 531542. Soponronnarit, S., Nathakaranakule, A., Irajindalert, A., Taechapairoj, C., 2006. Parboiling brown rice using super heated steam uidization technique. Journal of Food Engineering 75, 423432. Subba Rao, P.B., Bhattacharya, K.R., 1966. Effect of parboiling on thiamine content of rice. Journal of Agricultural and Food Chemistry 14 (5), 479482. Thalmann, C., Freise, J., Heitland, W., Bacher, S., 2003. Effects of defoliation by horse chestnut leafminer (Cameraria ohridella) on reproduction in Aesculus hippocastanum. Trees Structure and Function 17, 383388.

b value of instrument (powder)


Fig. 5. Relationship of b value between instrument (powder) and Vision Builder of parboiled rice.

Paired sample t-test showed signicant differences in DE of all samples (t 47.85, p < 0.001,) between the instrument for powder and Vision Builder with a high correlation coefcient (r 0.863). Also, there were signicant differences (t 37.10, p < 0.001; t 6.003, p < 0.001) between the instrument for powder and whole rice, instrument for whole rice and Vision Builder, respectively. For the L value, there were signicant differences (t 224.792, p < 0.001; t 13.906, p < 0.001; t 206.936, p < 0.001), and all had poor correlation coefcients. Although the b value had a good correlation coefcient (r 0.844), there was signicant difference (t 15.606, p < 0.001) between instrument for powder and Vision Builder. However, there was no signicant difference (t 0.322, p > 0.05) between the instrument for whole rice and Vision Builder with a poor correlation coefcient. As expected, there was a signicant difference (t 20.02, p < 0.001) between the instrument for powder and whole rice with a poor correlation coefcient. It was observed that the values of DE, L, and b obtained from the instrument for whole rice were signicantly different from those of the instrument for powder and Vision Builder. This was mainly due to the difference between the color of the surface and the interior of parboiled rice. The instrument for whole rice was less accurate than the other two methods because of the void among the whole rice and the properties of the instrument. To some extent there was a relationship between the instrument for powder and Vision Builder. The instrument for powder had a good correlation coefcient for DE (R2 0.8074, Fig. 3) and the b value (R2 0.7741, Fig. 5) with Vision Builder, but this was not found for the L value. This may be explained by the light conditions, which was an open system. 4. Conclusion The results suggest that the combination of digital image from digital camera, computer and Vision Builder have good potential to be used for color measurement and this technique can provide color information such as for parboiled rice compared to a colormeasuring instrument. There was a high signicant correlation but signicant difference for DE and the b value between the instrument for powder and

Journal of Cereal Science 50 (2009) 266271

Contents lists available at ScienceDirect

Journal of Cereal Science


journal homepage: www.elsevier.com/locate/jcs

Optimization of ultrasound-assisted extraction of defatted wheat germ proteins by reverse micelles


Ke-Xue Zhu a, *, Xiao-Hong Sun a, b, Hui-Ming Zhou a
a b

State Key Laboratory of Food Science and Technology, School of Food Science and Technology, Jiangnan University, 1800 Lihu Avenue, Wuxi-214122, Jiangsu Province, PR China College of Life Science and Engineering, Qiqihar University, Qiqihar-161006, Heilongjiang Province, PR China

a r t i c l e i n f o
Article history: Received 29 March 2009 Received in revised form 10 June 2009 Accepted 12 June 2009 Keywords: Ultrasound-assisted extraction Defatted wheat germ proteins Reverse micelles Forward extraction efciency

a b s t r a c t
In this work, an ultrasound-assisted procedure for the extraction of defatted wheat germ proteins by reverse micelles was established. Response surface methodology (RSM) was used to optimize the ultrasound-assisted extraction (UAE) parameters (ultrasonic output power, ultrasonic time and pulse mode) for enhancing the forward extraction efciency of defatted wheat germ proteins by implementing a three-level, three-variable BoxBehnken experimental design. The independent variable with the largest effect on response was X2 (ultrasonic time), followed by X1 (ultrasonic output power) and X3 (pulse mode). The optimum extraction conditions were found to be ultrasonic output power 363 W, ultrasonic time 24 min and pulse mode 2.4 s on and 2 s off (2.4 s:2 s). Under these conditions, the forward extraction efciency of defatted wheat germ proteins can increase from 37% to 57%, and the nal protein extraction efciency by reverse micelles can reach 45.6%, making the advantage of reverse micelles much more obvious than alkaline extraction and isoelectric precipitation on the separation of protein. 2009 Elsevier Ltd. All rights reserved.

1. Introduction Defatted wheat germ (DWG), after extraction of valuable wheat germ oil, contains a relatively high content of protein (more than 30%). Defatted wheat germ protein has been classied with effectively superior animal protein, and it is rich in amino acids, especially the essential amino acids (lysine, methionine and threonine) (Zhu et al., 2006), in which many cereals are decient. Moreover, it also has good functional properties, such as emulsifying activity, emulsifying stability, foaming capacity, water retention and solubility (Ge et al., 2000). Therefore, defatted wheat germ proteins can be regarded as one of the most potentially valuable vegetable proteins. Traditionally, defatted wheat germ proteins have been extracted by alkaline extraction and isoelectric precipitation (Hettlarachchy et al., 1996). However, this traditional approach has some serious defects: a great deal of acerbic or alkaline wastewater is produced which can cause serious environmental pollution. Moreover, it is easy to cause protein denaturation because of extreme pH, and it is
Abbreviations: DWG, defatted wheat germ; AOT, sulphosuccinic acid bis (2-ethylhexyl) ester sodium salt; RSM, response surface methodology; UAE, ultrasound-assisted extraction; RWG, raw wheat germ; ANOVA, the results of analysis of variance; CV, coefcient of variation; R2, the coefcient of determination. * Corresponding author. Tel./fax: 86 510 85329037. E-mail addresses: kxzhu@jiangnan.edu.cn, zkx1978@hotmail.com (K.-X. Zhu). 0733-5210/$ see front matter 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2009.06.006

also high in consumption of acid, alkali and water. Thus, it is imperative to explore an alternative extraction method for defatted wheat germ proteins (Sun et al., 2008). The reverse micelles are nanometer sized aggregates of surfactant molecules containing inner cores of water molecules in nonpolar solvents. The biotechnological relevance of reverse micelles arises from their ability to solubilize water and hydrophilic molecules, such as proteins, in their polar cores (Lye et al., 1994). The extraction of proteins by reverse micelles has many advantages compared to the traditional alkaline extraction and isoelectric precipitation. The method of reverse micelles is not only low cost (the surfactant and organic solvents could be used repeatedly after recovery), convenient and easy to scale-up, but also could solubilize proteins without damaging their native conformation and activity (Marcozzi et al., 1991; Noh and Imm, 2005; Liu et al., 2004). Thus, reverse micelles have emerged as one of the most attractive and promising extraction methods for proteins (Andrews and Haywood, 1994; Matzke et al., 1992; Leser et al., 1989). Defatted wheat germ proteins have already been extracted by reverse micelles (Sun et al., 2008, 2009), and the study demonstrated that the forward extraction of defatted wheat germ proteins by reverse micelles was feasible. However, the forward extraction efciency of defatted wheat germ proteins was only 37% by reverse micelles using stirring, and it is necessary to combine other methods to improve the forward extraction efciency.

K.-X. Zhu et al. / Journal of Cereal Science 50 (2009) 266271

267

Recently, ultrasound-assisted extraction (UAE) has emerged as a potential approach to extract substances from plant raw materials. UAE has been proven to be a very effective tool for improving the extraction efciency and shortening the extraction time (Kim et al., 2007; Stanisavljevic et al., 2007; Rodrigues and Pinto, 2007; Rostagno et al., 2007; Alissandrakis et al., 2003; Jerkovic et al., 2007). The extraction efciency enhancement by ultrasound is attributed to the phenomenon of cavitation produced in the solvent by the passage of an ultrasonic wave. Ultrasound also exerts a mechanical effect, breaking up the matrix and causing smaller particles to be produced, thereby exposing more surface area to the extraction solvent (Rostagno et al., 2003). Thus, the objective of the present study was to combine UAE and reverse micelles in order to increase the forward extraction efciency of defatted wheat germ proteins. Response surface methodology (RSM) was further employed to optimize the UAE parameters for enhancing the forward extraction efciency of defatted wheat germ proteins by implementing a three-level, three-variable BoxBehnken experimental design (Box and Behnken, 1960).

2.4. Protein determination The protein content in the reverse micellar system was determined using the micro-Kjeldahl method and a conversion factor of 5.45 was used to quantify the protein content (AOAC, 1984; Tkachuk, 1969). The extracted defatted wheat germ proteins were expressed as:

Extracted protein% total protein in the reverse micelle system 100 total protein in DWG

2.5. Experimental design In this study, the BoxBehnken experimental design with three variables and three levels was chosen to investigate the inuence of ultrasonic output power (X1), ultrasonic time (X2) and pulse mode (X3) on the UAE efciency of defatted wheat germ proteins. The complete design consisted of 15 combinations including three replicates of the center point. The independent variables Xi were coded as xi, which were dened as dimensionless, according to Eq. (1):

2. Materials and methods

xi
2.1. Materials and chemicals Raw wheat germ (RWG) was a gift from Huaian Xinfeng Flour Mill (Jiangsu, China). Sulphosuccinic acid bis (2-ethylhexyl) ester sodium salt (AOT) was purchased from Sigma Chemical Co. (St. Louis, MO) and used without further purication. Spectrophotometric-grade isooctane was obtained from Aldrich (Milwaukee, WI). All other chemicals used in the experiments were of analytical grade.

Xi X0 DXi

(1)

2.2. Preparation of reverse micellar systems The reverse micellar systems were formed by AOT, isooctane and KCl solution. The water content of reverse micellar solution was determined by the KarlFischer method and expressed as W0, the molar ratio of water to surfactant, i.e. W0 [H2O]/[AOT]. Firstly, AOT was mixed with isooctane (3:50, w/v) and stirred at room temperature. When AOT dissolved completely, 0.1 mol/l KCl solution (pH 8) was added and W0 was adjusted to 25 (Sun et al., 2008).

where xi is the coded value of an independent variable, Xi is the real value of an independent variable, X0 is the real value of an independent variable at the center point, and DXi is the step change value (Guan and Yao, 2008). Coded levels and actual values of the independent variables are presented in Table 1. Experiments were randomized in order to minimize bias. All the experiments were done in triplicate and the average UAE efciency of defatted wheat germ proteins was taken as the response, Y. The experimental results were tted to the following predictive quadratic polynomial equation as the correlation between the response and the independent variables:

Y A0

Ai Xi

Aii Xi2

Aij Xi Xj

(2)

where Y is the response variable, A0, Ai, Aii and Aij are the regression coefcients of variables for intercept, linear, quadratic and interaction terms, respectively, and Xi and Xj are independent variables (Cai et al., 2008). 2.6. Statistical analysis All data were analyzed using the SAS statistical package (version 8.1, SAS Institute, Cary, NC, USA). Analysis of variance (ANOVA) was performed to evaluate signicant differences between independent variables. This analysis included the sum of squares, Fishers F-test, its associated probability p-value and determination coefcient R2. To visualize the relationships between the responses and the
Table 1 Independent variable and their levels employed in BoxBehnken experimental design. Independent variable Ultrasonic output power (W) Symbol X1 Level 1 0 1 1 0 1 1 0 1 Real value 250 350 450 10 20 30 1:2 2:2 3:2

2.3. The forward extraction of defatted wheat germ proteins combined UAE and reverse micelles The forward extraction of defatted wheat germ proteins applied in this work is based on the previous study (Sun et al., 2008) with a modication (UAE instead of stirring). Defatted wheat germ our (0.5 g) was mixed with 50 ml of reverse micellar solution in a 100 ml beaker. This suspension was ultrasonicated at various levels of ultrasonic output power (250450 W) for different ultrasonic times (1030 min) using a 20 kHz ultrasonic generator (VCX 500, 100500 W; Sonics and Materials, USA) with a 12.7 mm probe that was directly inserted in the suspension. The pulse mode was set at 1 s:2 s, 2 s:2 s and 3 s:2 s, respectively. During the extraction period, the suspension was controlled by an ice bath in order to prevent proteins from denaturation by high temperature. After extraction, the undissolved residue was separated by centrifugation at 3000 rpm for 20 min, and the volume of supernatant was measured.

Ultrasonic time (min)

X2

Pulse mode (s)

X3

268

K.-X. Zhu et al. / Journal of Cereal Science 50 (2009) 266271 Table 3 Signicance of regression coefcients for response (Y). Model term A0 A1 A2 A3 A11 A22 A33 A12 A13 A23 Estimate 56.473333 1.588750 2.878750 1.245000 4.302917 3.292917 1.160417 0.182500 1.440000 0.515000 Standard error 0.649543 0.397762 0.397762 0.397762 0.585490 0.585490 0.585490 0.562521 0.562521 0.562521 t-value 86.94 3.99 7.24 3.13 7.35 5.62 1.98 0.32 2.56 0.92 p-value <0.001 0.0104 0.0008 0.0260 0.0007 0.0025 0.1043 0.7587 0.0507 0.4019

independent variables and also to deduce the optimum conditions, the tted quadratic polynomial equation was expressed as both response surface and contour plots. 3. Results and discussion 3.1. Fitting the models The protein content of DWG was on a wet basis 31.48% (Sun et al., 2008). The experimental conditions and the corresponding response values from the experimental design are given in Table 2. As observed in Table 2, treatment runs 4, 7, 8, 10, 13, 14 and 15 showed high levels of the UAE efciency of defatted wheat germ proteins at 54%, 53%, 56%, 54%, 56%, 57% and 56%, respectively. The maximum UAE efciency was observed in run number 14, while the minimum UAE efciency was achieved in run number 1. The values of the coefcients in the second-order polynomial equation are presented in Table 3. The regression equation for the UAE efciency of defatted wheat germ proteins (Y) was as follows:

Y 56:473333 1:588750X1 2:878750X2 1:245000X3


2 2 2 4:302917X1 3:292917X2 1:160417X3

0:182500X1 X2 1:440000X1 X3 0:515000X2 X3

(3)

The statistical signicance of Eq. (3) was checked by F-test, and the results of analysis of variance (ANOVA) are shown in Table 4. Because of the model F-value and a low p-value ((p > F) 0.0025), it can be easily deduced that the model was highly signicant. Usually, the lower the value of the coefcient of variation (CV), the higher the reliability of experiment. Here, a lower value of CV (2.1718) indicated a better precision and reliability of the experiments was carried out (Li et al., 2007). For the model tted, the coefcient of determination (R2), which can check the goodness of a model, was 0.9708. This implied that the sample variation of 97.08% for the UAE efciency of defatted wheat germ proteins was attributed to the independent variables, and only 2.92% of the total variation cannot be explained by the model. From the above, it can be concluded that the developed model could adequately represent the real relationship among the parameters chosen. 3.2. Effects of independent variable on responses At the early stage of extraction (24 min), the UAE efciency of defatted wheat germ proteins was remarkably enhanced with increasing ultrasonic time, because the major action of ultrasound
Table 2 BoxBehnken experiments design matrix with experimental values of the UAE efciency of defatted wheat germ proteins. Run 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 X1 1 1 1 1 0 0 0 0 1 1 1 1 0 0 0 X2 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 X3 0 0 0 0 1 1 1 1 1 1 1 1 0 0 0 Y (%) 44 51 46 54 47 52 53 56 47 54 51 52 56 57 56

was acoustic cavitation during this period, which disrupted the cell walls of DWG and accelerated the washing out of the proteins. As the time increased, more and more cells were broken, and defatted wheat germ proteins were released gradually. However, there was no obvious observed yield change in the prolonged time periods (Hemwimol et al., 2006; Zhao et al., 2007; Wei et al., 2008). The UAE efciency of defatted wheat germ proteins also increased with increasing ultrasonic output power. An explanation for this is that the increased ultrasonic output power leads to an enhanced mass transfer procedure. However, when the ultrasonic output power was set above 363 W, no signicant increase in the UAE efciency was observed which can be attributed to that the balance of the extraction system having been reached (Wei et al., 2008). When the pulse mode was set at 2.4 s on and 2 s off, the UAE efciency of defatted wheat germ proteins reached the highest value. However, the electrical energy was wasted and the UAE efciency was not signicantly increased for further increase in pulse mode (Sivakumar et al., 2007). The signicance of each coefcient can be seen from the Student t-test and p-value in Table 3. The corresponding variables will be more signicant if the absolute t-value becomes larger and the pvalue becomes smaller (Amin and Anggoro, 2004). Thus, the independent variable with the largest effect was X2 (ultrasonic time), followed by X1 (ultrasonic output power) and X3 (pulse mode). The factor t-value (7.24) and p-value (0.0008) corresponded to X2, while the t-values for X1 and X3 were smaller at 3.99 and 3.13, respectively, but the p-values were still signicant at p 0.0104 and 0.0260, respectively. Positive coefcients for X1, X2 and X3 2 revealed a linear effect of increasing Y. The quadratic terms X1 and 2 X2 also had a highly signicant effect on Y (p < 0.01), whereas all the cross-product terms were not found to contribute to the 2 response at a signicant level (p > 0.05). In this work, X1, X2, X3, X1 2 and X2 were signicant model terms. 3.3. Comparison of actual and predicted UAE efciencies of defatted wheat germ proteins A regression model can be used to predict the future response Y (the UAE efciency of defatted wheat germ proteins) corresponding to particular values of the regressor variables. Fig. 1(a) showed the actual UAE efciency of defatted wheat germ proteins versus the predicted one under the optimum extraction conditions. The gure
Table 4 Analysis of variance (ANOVA) for the response surface model. Regression Linear Quadratic Cross product Total model Degree of freedom 3 3 3 9 Sum of squares 98.890825 102.373992 9.488525 210.753342 R2 0.4555 0.4716 0.0437 0.9708 F-value 26.04 26.96 2.50 18.50 p-value 0.0018 0.0016 0.1740 0.0025

K.-X. Zhu et al. / Journal of Cereal Science 50 (2009) 266271

269

proved that the predicted value of the response from the tted second-order polynomial model was in agreement with the actual one in the range of the operating variables. 3.4. Localization of optimum condition In order to gain a better understanding of the interaction effects of the independent variables on the response, the response surface

plot and their corresponding counter plot were formed based on the tted second-order polynomial model (Fig. 1(b)(d)). The effects of ultrasonic output power (X1) and ultrasonic time (X2) on the UAE efciency of defatted wheat germ proteins (Y) are depicted in Fig. 1(b), while pulse mode (X3) was xed at its middle level 2 s on and 2 s off (2 s:2 s). Fig. 1(b) shows that the UAE efciency of defatted wheat germ proteins increased slightly as the ultrasonic time increased at a low ultrasonic output power. With

Fig. 1. (a) Actual UAE efciency of defatted wheat germ proteins (DWGP) versus the predicted one under the optimum extraction conditions. (b) Response surface and contour plots for the effects of ultrasonic output power (X1) and ultrasonic time (X2) on Y, when pulse mode (X3) was 2 s on and 2 s off. (c) Response surface and contour plots for the effects of ultrasonic output power (X1) and pulse mode (X3) on Y, when ultrasonic time (X2) was 20 min. (d) Response surface and contour plots for the effects of ultrasonic time (X2) and pulse mode (X3) on Y, when ultrasonic output power (X1) was 350 W.

270

K.-X. Zhu et al. / Journal of Cereal Science 50 (2009) 266271 Table 5 Optimum UAE conditions of defatted wheat germ proteins, predicted and experimental UAE efciencies of defatted wheat germ proteins from RSM. Optimum UAE conditions Ultrasonic output power (W) 363 Ultrasonic time (min) 24 Pulse mode (s) 2.4:2 The UAE efciencies (%) Predicted value 58 Experimental value 57

the increase in ultrasonic output power, the UAE efciency of defatted wheat germ proteins increased from 50% to 56% at a high level of ultrasonic time, and increased only from 45% to 50% at a low level of ultrasonic time. This suggested that the increase of both the ultrasonic output power and ultrasonic time within the tested range was favorable to the accumulation of the UAE efciency of defatted wheat germ proteins. The data also showed that, under optimal ultrasonic output power and ultrasonic time, an increase in ultrasonic output power and ultrasonic time would not further increase the UAE efciency of defatted wheat germ proteins. Similar results have been reported in the literature (Sivakumar et al., 2007). These facts were important in making the whole extraction process economically more feasible and possible to shorten the ultrasonic time and economize power in the potential industry application. Fig. 1(c) presents the response surface plot and its corresponding counter plot, showing the effects of ultrasonic output power (X1) and pulse mode (X3) on the UAE efciency of defatted wheat germ proteins (Y), while holding ultrasonic time (X2) at its middle level (20 min). It was evident that the UAE efciency of defatted wheat germ proteins considerably increased with prolonged pulse time when the ultrasonic output power was near its middle level and the maximum UAE efciency could be achieved in the range of ultrasonic output power 350375 W and pulse time 2.22.7 s. The effects of ultrasonic time (X2) and pulse mode (X3) on the UAE efciency of defatted wheat germ proteins (Y) can be seen in Fig. 1(d), while ultrasonic output power (X1) was xed at its middle level (350 W). With increase of pulse time, the UAE efciency of defatted wheat germ proteins drastically increased at a high level of ultrasonic time. However, no signicant effect of pulse time on the UAE efciency of defatted wheat germ proteins was observed at a low level of ultrasonic time. The predicted maximum UAE efciency of defatted wheat germ proteins (57%) was obtained in this situation when the ultrasonic time and pulse mode were 23 min and 2.5 s:2 s, respectively. The forward extraction efciency of defatted wheat germ protein enhancement by ultrasound may be attributed to the fact that ultrasound exerted a mechanical effect, disrupting the cell walls of DWG and accelerating the washing out of the proteins. Moreover, ultrasound could also break up the defatted wheat germ proteins and cause smaller protein particles to be produced, and the smaller particles were less likely to be excluded from reverse micelles because of size limitations (Woll et al., 1989; Wei et al., 2008).

4. Conclusions Combining UAE and reverse micelles to increase the forward extraction efciency of defatted wheat germ proteins has been proposed. Our work provided evidence that UAE can not only shorten extraction time, but also increase the forward extraction efciency of defatted wheat germ proteins from 37% to 57%, making the advantage of reverse micelles much more obvious than alkaline extraction and isoelectric precipitation on the separation of protein. The backward extraction efciency was 80% (Sun et al., 2009), so the nal extraction efciency by reverse micelles was 45.6% (57 80% 45.6%). While the nal protein extraction efciency using the traditional alkaline extraction and isoelectric precipitation was only 24.037.0% (Zhu et al., 2006). Therefore, the method of reverse micelles combined with UAE may have great potential for being scaled-up to a commercial process. Acknowledgements This research was supported by Jiangsu Provincial Natural Science Foundation of China (BK2008102), Doctoral Program Foundation of Institutions of Higher Education of China (20070295034) and Program for Changjiang Scholars and Innovative Research Team in University (IRT0627). References
Amin, N.A.S., Anggoro, D.D., 2004. Optimization of direct conversion of methane to liquid fuels over Cu loaded W/ZSM-5 catalyst. Fuel 83, 487494. Alissandrakis, E., Daferera, D., Tarantilis, P.A., Polissiou, M., Harizanis, P.C., 2003. Ultrasound-assisted extraction of volatile compounds from citrus owers and citrus honey. Food Chemistry 82, 575582. Andrews, B.A., Haywood, K., 1994. Effect of pH, ion type and ionic strength on partitioning of proteins in reverse micelles systems. Journal of Chromatography A 668, 5560. AOAC, 1984. Ofcial Methods of Analysis, fourteenth ed. Association of Ofcial Analytical Chemists, Washington, DC. Box, G.E.P., Behnken, D.W., 1960. Some new three level designs for the study of quantitative variables. Technometrics 2, 455475. Cai, W.R., Gu, X.H., Tang, J., 2008. Extraction, purication, and characterization of the polysaccharides from Opuntia Milpa Alta. Carbohydrate Polymers 71, 403410. Guan, X., Yao, H.Y., 2008. Optimization of Viscozyme L-assisted extraction of oat bran protein using response surface methodology. Food Chemistry 106, 345351. Ge, Y.Q., Sun, A.D., Ni, Y.Y., Cai, T.Y., 2000. Some nutritional and functional properties of defatted wheat germ protein. Journal of Agriculture and Food Chemistry 48, 62156218. Hemwimol, S., Pavasant, P., Shotipruk, A., 2006. Ultrasound-assisted extraction of anthraquinones from roots of Morinda citrifolia. Ultrasonics Sonochemistry 13, 543548. Hettlarachchy, N.S., Grifn, V.K., Gnanasambandam, R., 1996. Preparation and functional properties of a protein isolate from defatted wheat germ. Cereal Chemistry 73, 364367. Jerkovic, I., Mastelic, J., Marijanovic, Z., Klein, Z., Jelic, M., 2007. Comparison of hydrodistillation and ultrasonic solvent extraction for the isolation of volatile compounds from two unioral honeys of Robinia pseudoacacia L. and Castanea sativa L. Ultrasonics Sonochemistry 14, 750756. Kim, S.J., Murthy, H.N., Hahn, E.J., Lee, H.L., Pack, K.Y., 2007. Parameters affecting the extraction of ginsenosides from the adventitious roots of ginseng. Separation and Purication Technology 56, 401406. Li, Y., Liu, Z.Q., Zhao, H., Xu, Y.Y., Cui, F.J., 2007. Statistical optimization of xylanase production from new isolated Penicillium oxalicum ZH-30 in submerged fermentation. Biochemical Engineering Journal 34, 8286.

3.5. Validation of the model The optimum UAE conditions depended on ultrasonic output power (X1), ultrasonic time (X2) and pulse mode (X3), and were obtained by RSM. For validation of the model, the defatted wheat germ proteins were extracted under optimal UAE conditions and the extracted protein was determined. The experimental and predicted values were compared in order to determine the validity of the model. From the model, optimized conditions were obtained as presented in Table 5. The stationary point giving a maximum UAE efciency of defatted wheat germ proteins had the following critical values: ultrasonic output power 363 W, ultrasonic time 24 min and pulse mode 2.4 s:2 s. The predicted UAE efciency of defatted wheat germ proteins for these conditions was 58%, while the experimental value for these conditions was 57%. The results demonstrated that the experimental value (57%) was reasonably close to the predicted one (58%) conrming the validity and adequacy of the predicted model.

K.-X. Zhu et al. / Journal of Cereal Science 50 (2009) 266271 Liu, J.G., Xing, J.M., Shen, R., Yang, C.L., Liu, H.Z., 2004. Reverse micelles extraction of nattokinase from fermentation broth. Biochemical Engineering Journal 21, 273278. Lye, G.J., Asenjo, J.A., Pyle, D.L., 1994. Protein extraction using reverse micelles: kinetics of protein partitioning. Chemical Engineering Science 49, 31953204. Leser, M.E., Luisi, P.L., Palmieri, S., 1989. The use of reverse micelles for the simultaneous extraction of oil and proteins from vegetable meal. Biotechnology and Bioengineering 34, 11401146. Matzke, S.F., Creagh, A.L., Haynes, C.A., Prausnitz, J.M., Blanch, H.W., 1992. Mechanisms of protein solubilization in reverse micelles. Biotechnology and Bioengineering 40, 91102. Marcozzi, G., Correa, N., Luisi, P.L., Caselli, M.,1991. Protein extraction by reverse micelles: a study of the factors affecting the forward and backward transfer of a-chymotrotrypsin and its activity. Biotechnology and Bioengineering 38, 12391246. Noh, K.H., Imm, J.Y., 2005. One-step separation of lysozyme by reverse micelles formed by the cationic surfactant, cetyldimethlammonium bromide. Food Chemistry 93, 95101. Rodrigues, S., Pinto, G.A.S., 2007. Ultrasound extraction of phenolic compounds from coconut shell powder. Journal of Food Engineering 80, 869872. Rostagno, M.A., Palma, M., Barroso, C.G., 2007. Ultrasound-assisted extraction of isoavones from soy beverages blended with fruit juices. Analytica Chimica Acta 597, 265272. Rostagno, M.A., Palma, M., Barroso, C.G., 2003. Ultrasound-assisted extraction of soy isoavones. Journal of Chromatography A 1012, 119128.

271

Sun, X.H., Zhu, K.X., Zhou, H.M., 2009. Optimization of a novel backward extraction of defatted wheat germ protein from reverse micelles. Innovative Food Science and Emerging Technologies 10, 328333. Sun, X.H., Zhu, K.X., Zhou, H.M., 2008. Protein extraction from defatted wheat germ by reverse micelles: optimization of the forward extraction. Journal of Cereal Science 48, 829835. Stanisavljevic, I.T., Lazic, M.L., Veljkovic, V.B., 2007. Ultrasonic extraction of oil from tobacco seeds. Ultrasonics Sonochemistry 14, 646652. Sivakumar, V., Verma, V.R., Rao, P.G., Swaminathan, G., 2007. Studies on the use of power ultrasound in solid-liquid myrobalan extraction process. Journal of Cleaner Production 15, 18131818. Tkachuk, R., 1969. Nitrogen-to-protein conversion factors for cereals and oilseed meals. Cereal Chemistry 46, 419424. Wei, F., Gao, G.Z., Wang, X.F., Dong, X.Y., Li, P.P., Hua, W., Wang, X., Wu, X.M., Chen, H., 2008. Quantitative determination of oil content in small quantity of oilseed rape by ultrasound-assisted extraction combined with gas chromatography. Ultrasonics Sonochemistry 15, 938942. Woll, J.M., Hatton, T.A., Yarmush, M.L., 1989. Bioafnity separations using reverse micellar extraction. Biotechnology Progress 5, 5762. Zhao, S., Kwok, K.C., Liang, H.H., 2007. Investigation on ultrasound assisted extraction of saikosaponins from Radix Bupleuri. Separation and Purication Technology 55, 307312. Zhu, K.X., Zhou, H.M., Qian, H.F., 2006. Proteins extracted from defatted wheat germ: nutritional and structural properties. Cereal Chemistry 83, 6975.

Journal of Cereal Science 50 (2009) 272277

Contents lists available at ScienceDirect

Journal of Cereal Science


journal homepage: www.elsevier.com/locate/jcs

Wholemeal wheat bread: A comparison of different breadmaking processes and fungal phytase addition
Cristina M. Rosell*, Eva Santos, Juan M. Sanz Penella, Monica Haros
Food Science Department, Institute of Agrochemistry and Food Technology (IATA-CSIC), PO Box 73, 46100 Burjassot, Valencia, Spain

a r t i c l e i n f o
Article history: Received 13 January 2009 Received in revised form 29 May 2009 Accepted 23 June 2009 Keywords: Wholemeal Breadmaking Quality Phytates

a b s t r a c t
The effect of different breadmaking processes (conventional, frozen dough, frozen partially baked bread) and the effect of the storage period on the technological quality of fresh wholemeal wheat breads are investigated. In addition, the impact of the exogenous fungal phytase on the phytate content was also determined. Results showed that breadmaking technology signicantly affected the quality parameters of wholemeal breads (specic volume, moisture content, crumb and crust colour, crumb texture prole analysis and crust aking) and frozen storage affected to a different extent the quality of the loaves obtained from partially baked breads and those obtained from frozen dough, particularly crust aking. Freezing and frozen storage of wholemeal bread in the presence of fungal phytase decreased signicantly the phytate content in whole wheat breads. The combination of fungal phytase addition, breadmaking process and frozen storage could be advisable for overcoming the detrimental effect of bran on the mineral bioavailability. 2009 Elsevier Ltd. All rights reserved.

1. Introduction Until recently, the nutrient value of wheat bran, although long understood, has been ignored and the bran discarded and used as animal feed. However, its rich nutrient composition (Antoine et al., 2004), and its dietary bre content has motivated numerous campaigns for increasing the consumption of whole wheat products. Whole grain products are perceived as more nutritionally balanced, healthy and natural, bread being the most consumed product (Claupein et al., 2007). Nevertheless, recommendations for the increase of wholemeal cereal consumption have raised questions about the increased intake of antinutritive compounds such as phytates (Sanz Penella et al., 2008). Whole wheat ours contain phytate or myo-inositol hexaphosphate that decreases the bioavailability of multivalent cations due to the formation of insoluble complexes in the gastrointestinal tract (Lopez et al., 2002). Signicant reduction in the content of phytates in whole wheat bread have been obtained by adding exogenous phytase (Haros et al., 2001a,b), and also using different lactic acid bacteria with phytase activity (Palacios et al., 2008). Despite the benecial effect of consuming whole wheat bread, the publics acceptance of this product is limited due to its lower

* Corresponding author. Tel.: 34 963900022; fax: 34 963636301. E-mail address: crosell@iata.csic.es (C.M. Rosell). 0733-5210/$ see front matter 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2009.06.007

volume, coarser texture and faster staling compared to rened wheat bread (Gan et al., 1992; Zhang and Moore, 1999). Therefore, some technological efforts are needed in order that the performance of whole wheat breads meet consumers needs and demands. Convenience bakery products are solving the constraints on food preparation and shopping imposed by the accelerated consumer lifestyles. Concerning bakery products, freezing dough, partially baked or fully baked bread becomes in many cases necessary to face the present demands (Rosell and Gomez, 2007). The use of frozen dough is very attractive because the low volume of the unfermented dough is very convenient when frozen storage is involved in the process, and it has satisfying quality characteristics even after 9 months of storage (Giannou and Tzia, 2007). The partially baked bread (part-baked, par-baked bread or pre-baked bread), also called bake off technology (BOT), consists in partial baking till the dough structure is xed, giving a product with aerated crumb but without a crunchy crust that is formed during the second baking (Barcenas and Rosell, 2006a,b). Numerous studies have focussed on the sensory and technological quality of rened wheat loaves obtained from frozen dough or partially baked breads (Barcenas and Rosell, 2006a,b; Fik and Surowka, 2002; Poinot et al., 2008). Those revealed that breads with qualities close to the ones obtained from conventional breadmaking processes are obtained. Nevertheless, little information exists about the impact of those breadmaking processes on whole wheat bread loaves, where studies have addressed the improvement of formulation for

C.M. Rosell et al. / Journal of Cereal Science 50 (2009) 272277

273

counteracting the negative effects of the bran particle size on the breadmaking performance and bread quality (Collar et al., 2006; Rosell et al., 2006; Shogren et al., 1981; Zhang and Moore, 1999). The aim of this research was to determine the effect of different breadmaking processes (conventional, frozen dough, frozen partially baked bread) and diverse storage time on the technological quality (specic volume, texture, crumb structure and crust aking) of the resulting fresh wholemeal wheat breads and to assess the impact of exogenous fungal phytase on the phytate prole of the different breads. 2. Materials and methods Commercial wholemeal wheat our for breadmaking was used in this study. The characteristics of the our were: 14.20% moisture content (ICC 110/1), 12.61% protein content (ICC 105/2), 1.82% fat content (ICC 136) and 1.46% ash content (ICC 104/1). Commercial phytase (3.13.8) from Aspergillus niger (11.4 U/ml, Ronozyme Phytase) provided by Novozymes (Madrid, Spain) was used. One unit of phytase activity was dened as 1.0 mg of Pi liberated per minute at pH 5.0 and 30  C (Haros et al., 2001a). The bread improver was provided by Puracor (Groot-Bijgaarden, Belgium). The rest of the ingredients were acquired in the food market. 2.1. Breadmaking process Three different breadmaking processes were followed: conventional, frozen dough (FD) and partially baked frozen (PBF). Basic wholemeal wheat dough formula on 100 g our basis consisted of 3% (w/w) compressed yeast, 1.8% (w/w) salt and 1% (w/w) bread improver. When required, fungal phytase (200 mL/100 g our) was added. In the conventional and partially baked process, the amount of water necessary to give 500 Brabender Units (BU) of dough consistency was used, whereas dough consistency of 600 BU was used in the case of frozen dough. Bread doughs were prepared by mixing ingredients in a spiral mixer (AV18/2, Vimar Industries 1900, S.L., Spain) for 6 min. After 10 min resting, dough was divided into 70 g pieces and hand moulded. Fermentation was carried out in a proong cabinet at 35  C and 95% relative humidity for 45 min. In the conventional process, complete baking was carried out in an electric oven at 185  C for 15 min with steam injection at the beginning of the baking. Partial baking was carried out at 170  C for 16 min, and then loaves were cooled down to room temperature. Partially baked bread and doughs were placed directly in a blaster freezer at 30  C till the bread core reached 18  C. Loaves and frozen doughs were taken out of the freezer and thawed at room temperature for 60 min. Full baking of partially baked breads was carried out at 185  C for 8 min. Frozenthawed doughs were fermented in a proong cabinet at 35  C and 95% relative humidity for 1 h and then baked in an electric oven at 185  C for 15 min, as has been previously described for the conventional process. For storage studies, partially baked frozen loaves and frozen dough were packed in polypropylene bags and stored at 18  C for 3 months. Technological characteristics were evaluated during the storage period, taking samples after 1, 2, and 3 months. Two different sets of breads were made for each breadmaking process, comprising at least 20 loaves for performing further analysis. 2.2. Bread quality parameters Technological parameters of bread quality included: volume, specic volume (rapeseed displacement, AACC Standard 10-05), moisture content (AACC Standard 44-15A), crumb texture prole analysis (TPA), crust and crumb colour, crust aking, and crumb

image analysis. TPA was measured in a Texture Analyzer TA-XT2i (Stable Micro Systems, Surrey, UK) using two bread slices of 1-cm thickness, which underwent two double compression tests up to 50% penetration of its original height at a crosshead speed of 1 mm/s and a 30 s gap between compressions, with a cylindrical stainless steel probe (diameter 25 mm) (Collar and Bollan, 2005). Bread crust and crumb coloration were measured in four different locations by using a Minolta colorimeter (Chroma Meter CR-400/410, Konica Minolta, Japan) after standardization with a white calibration plate (L 96.9, a 0.04, b 1.84). The colour was recorded as L*, C and h colour parameters, where L* is the lightness or clearness, C the chroma and h the shade or hue angle. For crumb to crust ratio determination, crust was separated from the crumb using a razor blade. Crumb to crust ratio was expressed as weight ratio and as volume ratio on a wet basis. The crust aking test was carried out in a specic crushing system developed by Le Bail et al. (2005). Bread was crushed on its anks and on its base by 30% of its diameter and height in the crushing system. Crust pieces were collected and weighed and then a digital picture of crust pieces was taken. Using UTHSCSA Image Tool 3.0 Software, average crust akes size was measured. Crumb cell analysis was performed by scanning longitudinal and cross sections of bread samples, 10 mm thick. Images were analyzed by Image J software according to Gonzales-Barron and Butler (2006). Number of cells, average cell area, average diameter and cell circularity were calculated. Values were the means of four replicates. 2.3. Determination of myo-inositol phosphates myo-Inositol hexaphosphate (InsP6) concentration in our, the remaining concentration of InsP6 and the lower inositol phosphates contained in bread were measured by an HPLC method following the method described by Turk and Sandberg (1992) and lately modied by Sanz Penella et al. (2008). 2.4. Statistical analysis Experimental data were subjected to analysis of variance (ANOVA) using the Statgraphics V.7.1 program (Bitstream, Cambridge, MN), to determine signicant differences among the factors combination. When ANOVA indicated signicant F values, multiple sample comparison was also performed and Fishers least signicant difference (LSD) procedure was used to discriminate between the means. 3. Results and discussion 3.1. Effect of breadmaking processes and phytase addition on bread technological quality The type of process signicantly affected the specic volume of the bread (P < 0.001), and also the moisture content (P < 0.01) (Table 1). Phytase addition did not promote any signicant effect on those parameters. The combination process phytase affected the specic volume of the loaves (P < 0.05). In the absence of phytase (), bread from partially baked showed the lowest specic volume (Table 2), whereas no signicant differences were observed between the ones obtained from the conventional process or frozen dough. The same trend was observed in the moisture content, although in this case, bread from PB had the highest value. Taking into account that FD and PB, in the absence of phytase, were subjected to freezing during breadmaking, it seems that freezing itself did not affect the specic volume or it could affect the volume in a different manner to frozen dough and

274

C.M. Rosell et al. / Journal of Cereal Science 50 (2009) 272277

Table 1 Signicant effects of the breadmaking process and the presence of fungal phytase on the technological quality parameters of wholemeal breads. Parameters Specic volume Crust colour L* C h Crumb colour L* C h Crumb texture, TPA Hardness Springiness Cohesiveness Chewiness Resilience Moisture content Crust aking CFM CFS Crust section Crumb to crust ratio Volume Weight Crumb cell analysis Number of alveoli Average area Average diameter Circularity Process *** *** * ** * ** ns *** *** *** *** *** ** ns * ns ns * ns ns ns ns Enzyme ns ns ns ns *** ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns Process enzyme * ns ns ns * ns ns ** ns ns ** * ns ns ns ns ns ns ns ns ns ns

CFM, g crust/100 g bread; CFS, average crust aking size. ns, no signicant effect; signicant effect at *P < 0.05; **P < 0.01; ***P < 0.001.

partially baked bread. Poinot et al. (2008) found that breadmaking processes (conventional, frozen dough and partially baked bread), when running with the same formulation, does not produce any effect on the density of white wheat breads. The presence of phytase decreased the specic volume of the bread when it was obtained following the conventional process. This result disagrees with previous ndings of Haros et al. (2001a). Likely wholemeal our composition might be responsible for this divergence, since the action of phytase affects the endogenous alpha-amylase activity through the inhibitory role of phytates. No signicant effect on the bread specic volume was observed when phytase was added in the other breadmaking processes. Haros et al. (2001b) did not nd any signicant effect when fungal phytase was added to wholemeal bread. Neither breadmaking process, nor the presence of phytase induced any signicant effect on the bread shape (width/height ratio), in agreement with previous results when a similar amount of phytase was added to wholemeal conventional breadmaking (Haros et al., 2001a). Freezing and thawing processes exert some stress on the rened wheat dough

that cause a deterioration in the quality of the baked product, mainly affecting the protein fraction of the wheat our and the shelf-life of the bakers yeasts. In consequence, extended proong times are needed and reduced loaf volumes are obtained from frozen dough (Phimolsiripol et al., 2008). However, the freezing process without frozen storage seems to have less detrimental effect on the whole wheat dough as revealed by results of the present study. The process signicantly (P < 0.001) affected texture prole parameters of the crumb, whereas phytase addition did not show any signicant effect on those parameters (Table 1). Regarding the breadmaking process, a similar effect has been observed for white wheat breads obtained by different breadmaking processes (Poinot et al., 2008). The combined effect of process and enzyme had a signicant (P < 0.01) effect on hardness, chewiness and (P < 0.05) resilience. Regarding the individual effect induced by each breadmaking process (Table 2), partially baked bread gave the softest crumb followed by the conventional process and frozen dough. Only in the frozen dough process, a signicant softening effect was observed from phytase addition, likely derived from the relationship existing between phytase and alpha-amylase activities, previously described. In the present study, the presence of bran could likely modify that behaviour in whole meal bread, since the disruption of the structure induced by the bran particles will be enhanced with the formation of ice crystals. However, it seems that in the case of PB, where crumb was already formed when freezing, ice crystal formation could induce breaking of the gluten brils that form the skeletal framework of coarse pore walls, as has been previously described for frozen dough (Naito et al., 2004). As a consequence, a disrupted crumb structure might be obtained, which is manifested as softer crumb. Breads from PB also showed signicantly lower springiness, cohesiveness, chewiness and resilience, which again could be attributed to crumb rupture. When crumb cells were analyzed deeply (number of alveoli, average area, average diameter and circularity), no signicant difference could be attributed to the breadmaking process, or to the presence of phytase (Table 1). However, although they were not signicant (Table 3), some differences were induced by phytase in the number of alveoli, which could explain differences observed in the crumb texture. Rapid freezing and the absence of frozen storage might be responsible for those results, since the main effect on crumb microstructure occurs during prolonged frozen storage (Barcenas et al., 2004; Ribotta et al., 2001). Colour of crumb and crust changes due to process and phytase addition were estimated by L*C*h colour space (Table 3). Crust colour was signicantly affected by type of process but not by the presence of enzyme, and no process enzyme interaction was observed (Table 1). Luminosity and hue angle were signicantly (P < 0.05) higher in the case of bread from PB, whereas crust of

Table 2 Effect of different breadmaking processes and fungal phytase addition on technological quality parameters of the fresh loaves. Process Phytase Specic volume (g/cm3) Moisture content (%) Width/height ratio Crumb texture fresh bread Hardness, g Conventional FD PB + + + 3.6 b 3.3 a 3.7 b 3.9 b 3.2 a 3.2 a 35.4 a 34.9 a 35.1 a 35.3 a 38.5 b 38.4 b 1.43 a 1.43 a 1.65 a 1.69 a 1.66 a 1.60 a 340 b 337 b 418 c 344 b 256 a 278 a Springiness 0.990 d 0.982 cd 0.960 bc 0.981 cd 0.922 a 0.947 b Cohesiveness 0.839 b 0.829 b 0.834 b 0.837 b 0.794 a 0.805 a Chewiness, g 281 b 274 b 334 c 282 b 188 a 212 a Resilience 0.432 c 0.423 c 0.420 c 0.427 c 0.345 a 0.366 b

Means followed by the same letter within a column were not signicantly different (P < 0.05) (n 4). Process: FD, bread obtained from frozen dough; PB, bread obtained from frozen partially baked bread. Phytase: , in the absence of phytase; +, in the presence of phytase.

C.M. Rosell et al. / Journal of Cereal Science 50 (2009) 272277 Table 3 Effect of different breadmaking processes and fungal phytase addition on the characteristics of crust and crumb of the crust and crumb of fresh loaves. Process Phytase Crust aking CFM Conventional FD PB + + + 0.10 a 0.10 a 0.08 a 0.07 a 0.13 a 0.13 a CFS (mm2) 0.77 b 0.58 ab 0.37 a 0.53 ab 0.67 ab 0.83 b 2.7 a 2.5 a 2.9 a 2.7 a 2.9 a 2.9 a Crust section (mm) Crumb to crust ratio Volume 1.19 ab 0.94 a 1.48 ab 1.73 b 1.36 ab 1.31 ab Weight 1.38 abc 1.17 abc 1.60 c 1.54 bc 1.07 a 1.11 ab Crust colour L* 56.5 ab 58.4 b 56.2 ab 55.7 a 63.5 c 61.5 c C* 34.0 b 34.3 b 31.4 a 32.5 ab 31.4 a 33.1 ab h 70.6 a 71.2 a 71.9 a 71.6 a 75.8 b 74.7 b Crumb colour L* 61.7 ab 64.3 c 59.9 a 62.2 b 66.4 d 65.3 cd C* 18.1 a 18.6 abc 19.3 c 19.4 c 19.1 bc 18.2 ab h

275

78.7 b 78.6 ab 77.4 a 78.6 ab 78.7 ab 78.9 b

Phytase: , in the absence of phytase; +, in the presence of phytase. CFM, g crust/100 g bread; CFS, average crust akes size.

conventional breads showed signicantly higher chroma values (C) than FD() and PB(), indicating more vivid coloration (Table 3). The two-stage baking that occurs in the bread obtained from partially baked signicantly modied the crust luminosity. The lower baking temperatures or shorter baking times are responsible for the reduced Maillard reactions, yielding increased lightness. In fact, the aromatic prole observed in loaves from partially baked breads shows a signicantly lower amount of volatile compounds compared to frozen dough, mainly due to a reduced amount of volatile compounds from the fermentative process (Poinot et al., 2008). Despite differences observed in the crust colour, no signicant differences were observed in the crust section or crust wideness (Table 3). Crumb colour parameters L* and C were signicantly affected by process and only L* was signicantly (P < 0.001) modied by the enzyme. Also the combined action of process enzyme did modify signicantly this parameter (Table 1). Again, crumb of bread from PB had signicantly higher L* than the ones obtained from the other processes. Phytase presence signicantly affected L* in bread from conventional and FD, but no change was observed in breads from PB (Table 3). Crust properties have been considered an important factor for bread quality assessment. Crust aking resulting from the detachment of some part of the crust constitutes an important drawback, which has been related to excessive drying of the bread surface at the end of the post-baking chilling and freezing process (Hamdami et al., 2007; Lucas et al., 2005). Breadmaking processes only affected signicantly (P < 0.05) the crust aking size and the crumb to crust weight ratio, whereas the enzyme did not inuence crust properties (Tables 1 and 3). In the absence of phytase, bread from FD gave signicantly smaller crust akes than bread obtained from the conventional process, whereas breads from PB showed an intermediate behaviour (Table 3). Crust aking has been mainly investigated for partially baked bread stored under frozen conditions (Le Bail et al., 2005). This phenomenon has been ascribed to two different processes. The rst is the concentration of water as ice under the crust due to the presence of the freezing front. The second process is the interfacial differences between the crust and crumb associated with the tensile forces and stresses induced by the thermo-mechanical shock (Lucas et al., 2005). However, the similar values of crust ake amount (CFM) obtained with the different breadmaking processes stressed the role of the frozen storage on this phenomenon, because freezing itself did not induce signicant differences. In addition, temperature uctuations between crust and crumb produced during baking and cooling could be responsible for the observed effect on the size of the crust akes. Le Bail et al. (2005) reported that chilling conditions after partial baking are the most determinant parameter for crust aking followed by the proong conditions. It is advisable to have high air humidity during those processes for reducing crust aking.

3.2. Effect of frozen storage and phytase addition on bread technological quality Frozen dough and partially baked bread were stored at subzero temperatures during 3 months to determine the effect of frozen storage on the technological quality of wholemeal breads. Storage and breadmaking processes induced signicant effects on the specic volume (P < 0.001), crumb hardness (P < 0.001) and crust aking (P < 0.001) (Fig. 1), whereas phytase addition did not modify those parameters. Breads obtained from FD showed increasing specic volume when extending the storage period (Fig. 1). Taking into consideration that the bran particles contained in the wholemeal dough cause disruption of the dough structure, structural changes derived from frozen storage and the redistribution of water associated with them might affect positively further wholemeal dough expansion leading to a slight increase during storage. No trend was observed with the wholemeal breads obtained from PB. Barcenas et al. (2004) observed that although the specic volume of white partially baked bread was not signicantly affected by the duration of the frozen storage after 42 days, breads obtained from those PB showed a slight hardness increase. Conversely, previous ndings regarding the quality of white breads obtained from stored frozen dough revealed that dough weight loss and bread crumb rmness increase with increasing storage time, although temperature uctuations during storage could explain this divergence (Phimolsiripol et al., 2008; Ribotta et al., 2001). The breadmaking process promoted a signicant effect (P < 0.001) on the crumb hardness of bread; loaves obtained from PB were signicantly (P < 0.001) softer than those obtained from FD during all the frozen storage period tested (Fig. 1). Results obtained during frozen storage conrmed that ice crystal formation and growing do not affect FD and PB in the same way. Crumb hardness of the breads obtained from PB was kept almost constant during the whole storage (3 months) and the presence of phytase did not induce any effect on this parameter. Breads from FD showed irregular values of crumb hardness, although an increasing trend could be envisaged, and no effect was clearly observed due to the presence of phytase. Ice crystals initially formed during freezing at the gaspore interface (Esselink et al., 2003) grow during frozen storage since the amount of freezable water in frozen doughs increases with frozen storage (Lu and Grant, 1999), yielding baked breads with harder crumbs and coarse texture (Sharadanant and Khan, 2003). The amount of crust akes underwent the greatest variation related to frozen storage (Fig. 1). During frozen storage, the breadmaking process signicantly (P < 0.001) affected the amount of crust akes (CFM). Although freezing did not signicantly affect the CFM, frozen storage dramatically augmented this parameter. Principally, in breads obtained from FD, crust aking signicantly (P < 0.001) increased after 1 month of storage, and further storage only induced a smooth increase in CFM. In breads obtained from PB,

276

C.M. Rosell et al. / Journal of Cereal Science 50 (2009) 272277

5.0 4.0 3.0 2.0 1.0 0.0 FD FD + PB PB + 0 1 2 3

Time (months)

600

Hardness (g)

500 400 300 200 100 0 FD FD + PB PB + 0 3 2 1 Time (months)

CFm (g/100g bread)

0.30 0.25 0.20 0.15 0.10 0.05 0.00 FD FD + PB PB + 0 1 2 3

Time (months)

Fig. 1. Effect of breadmaking process and frozen storage on some technological quality parameters of breads containing phytase () or in the absence of phytase (). FD, frozen dough; PB, partially baked.

the effect of frozen storage on CFM was only signicant (P < 0.05) after prolonged storage (3 months). 3.3. Effect of different breadmaking technologies on the phytate content During fermentation the cereal phytate-degrading enzyme degraded the total InsP6 initially present in the wheat our and generated lower myo-inositol phosphates as released hydrolysis
Table 4 Residual amount of myo-inositol phosphates in whole wheat bread.a,b Process Phytase Hydrolysis (%)

products (Table 4). With the exception of breads obtained from PB, the residual content of phytates was signicantly and positively affected by the addition of fungal phytase, which increased the percentage of InsP6 hydrolysis from 22.934.3% to 54.960.9%. The myo-inositol phosphate prole was also signicantly affected by fungal phytase addition, particularly the lower phosphate compounds (InsP4 and InsP3) and that effect was independent of the breadmaking process followed (Table 4). In the case of higher phosphate compounds, fungal phytase signicantly reduced the amount of InsP6 and InsP5 in breads obtained from the conventional process and frozen dough. The breadmaking process and the addition of fungal phytase signicantly (P < 0.05) affected the phytate hydrolysis during frozen storage (Fig. 2). Initially, the InsP6 amount showed a slight increase during frozen storage compared to the samples without storage, which was signicant (P < 0.05) in the case of PB samples. Presumably, the effect of freezing and frozen storage on dough microstructure could favor both the substrate liberation and the accessibility of the fungal phytase to the phytate compounds, as has been observed at dough level (Sanz Penella et al., 2007), the overall result being an increase in the level of InsP6. A reverse tendency was observed with longer storage, the InsP6 amount signicantly (P < 0.05) decreased, thus enzymes were not inactivated. On the other hand, in frozen systems, although the low temperatures decrease reactions rate, the increment of solute concentration in the unfrozen phase could increase the rates. Another factor that may be involved is a possible catalytic effect of ice crystals, greater proton mobility in ice than in water, a favorable substrate catalyst orientation caused by freezing or a greater dielectric constant for water than ice (Fennema et al., 1973). Regarding the type of process, in general the InsP6 level was signicantly (P < 0.05) higher in breads from PB samples than those from FD samples (Fig. 2). Partial baking could induce partial inactivation of the fungal phytases, whereas in FD samples, the enzyme might remain active during the storage. Therefore, although phytase did not induce a signicant effect on bread specic volume and crust aking, signicantly softer crumbs were obtained for breads from FD containing phytase. A comparison of different breadmaking processes for obtaining wholemeal breads showed that the type of process signicantly affects technological quality parameters like specic volume, crust and crumb colour, crumb texture and moisture content, whereas the addition of fungal phytase only induced signicant effect on the crumb lightness. However, there was signicant interaction between the breadmaking process and enzyme addition concerning specic volume, crumb lightness and crumb texture. Freezing and frozen storage inuenced in a diverse way the quality of wholemeal breads obtained from frozen dough or partially baked breads. Freezing and frozen storage of wholemeal bread in the presence of fungal phytase decreased signicantly the phytate content, independently of the breadmaking process followed. Thus, the combination of both variables could be a good approach to increase the mineral bioavailability in whole wheat breads.

Specific Volume (cm3/g)

Myo-Inositol phosphates (mmol/ g dm) InsP6 InsP5 a b a b a a 1.00 0.27 0.42 0.06 0.98 0.32 0.53 0.10 0.88 0.12 0.58 0.09 a b a b a b InsP4 0.47 0.08 0.15 0.04 0.58 0.26 0.20 0.04 0.45 0.05 0.13 0.07 a b a b a b InsP3 0.13 0.06 a n.d. 0.18 0.01 a n.d. 0.11 0.03 a n.d.

Conventional FD PB
a b

+ + +

22.9 13.5 a 60.9 7.8 b 34.3 14.7 a 54.9 5.4 b 21.7 11.1 a 35.0 14.4 a

5.88 1.03 2.99 0.60 5.01 0.71 3.44 0.49 5.97 0.91 4.96 0.55

Means standard deviation followed by the same letter in the same column are not signicantly different at 95% condence level. d.m., dry matter; n.d., not detected; InsP3InsP6, myo-inositol phosphate containing 36 phosphates per inositol residue.

C.M. Rosell et al. / Journal of Cereal Science 50 (2009) 272277

277

InsP6, mol/g d.m.

6 5 4 3 2 1 0 1 FD FD + PB PB + 0 2 3

Time (months)

Fig. 2. Effect of fungal phytase addition, breadmaking process and frozen storage on residual InsP6 content in whole wheat bread. Breads containing phytase () or in the absence of phytase (). FD, frozen dough; PB, partially baked.

Acknowledgements This work was nancially supported by the Commission of the European Communities, FP6, Thematic Area Food quality and safety, FOOD-2006-36302 EU-FRESH BAKE. It does not necessarily reect its views and in no way anticipates the Commissions future policy in this area.

References
AACC Methods. American Association of Cereal Chemistry International. St Paul, MN. Antoine, C., Lullien-Pellerin, V., Abecassis, J., Rouau, X., 2004. Effect of bran ballmilling on fragmentation and marker extractability of the aleurone layer. Journal of Cereal Science 40, 275282. Barcenas, M.E., Rosell, C.M., 2006a. Different approaches for improving the quality and extending the shelf life of the partially baked bread: low temperatures and HPMC addition. Journal Food Engineering 72, 9299. Barcenas, M.E., Rosell, C.M., 2006b. Effect of frozen storage time on the quality and aging of par-baked bread. Food Chemistry 95, 438445. Barcenas, M.E., Benedito, C., Rosell, C.M., 2004. Use of hydrocolloids as bread improvers in interrupted baking process with frozen storage. Food Hydrocolloids 18, 769774. Claupein, E., Saba, A., Shepherd, R., 2007. Consumers beliefs about whole and rened grain products in the UK, Italy and Finland. Journal of Cereal Science 46, 197206. Collar, C., Bollan, C., 2005. Relationships between dough functional indicators during breadmaking steps in formulated samples. European Food Research and Technology 220, 372379. Collar, C., Santos, E., Rosell, C.M., 2006. Signicance of dietary ber on the viscometric pattern of pasted and gelled our ber blends. Cereal Chemistry 83, 370376. Esselink, E.F.J., Van Aalst, H., Maliepaard, M., Van Duynhoven, J.P.M., 2003. Longterm storage effect in frozen dough spectroscopy and microscopy. Cereal Chemistry 80, 396403. Fennema, O.R., Powrie, W.D., Marth, E.H., 1973. Low Temperature Preservation of Foods and Living Matter. Marcel Dekker, New York, pp. 3-207.

Fik, M., Surowka, K., 2002. Effect of prebaking and frozen storage on the sensory quality and instrumental texture of bread. Journal of the Science of Food and Agriculture 82, 12681275. Gan, Z., Gallieard, T., Ellis, P.R., Angold, R.E., Vaughan, J.G., 1992. Effect of the outer bran layers on the loaf volume of wheat bread. Journal of Cereal Science 15, 151163. Giannou, V., Tzia, C., 2007. Frozen dough bread: quality and textural behavior during prolonged storage prediction of nal product characteristics. Journal of Food Engineering 79, 929934. Gonzales-Barron, U., Butler, F., 2006. A comparison of seven thresholding techniques with the k-means clustering algorithm for measurement of bread-crumb features by digital image analysis. Journal of Food Engineering 74, 268278. Hamdami, N., Tuan Pham, Q., Le-Bail, A., Monteau, J.Y., 2007. Two stage freezing of part baked breads: application and optimization. Journal of Food Engineering 82, 418426. Haros, M., Rosell, C.M., Benedito, C., 2001a. Use of fungal phytase to improve breadmaking performance of whole wheat bread. Journal Agricultural of Food Chemistry 49, 54505454. Haros, M., Rosell, C.M., Benedito, C., 2001b. Fungal phytase as a potential breadmaking additive. European Food Research and Technology 213, 317322. ICC-Standard Methods. International Cereal Science and Technology Association. Vienna. Austria. Le Bail, A., Monteau, J.Y., Margerie, F., Lucas, T., Chargelegue, A., Reverdy, Y., 2005. Impact of selected process parameters on crust aking of frozen partly baked dough. Journal of Food Engineering 69, 503509. Lopez, H.W., Leenhardt, F., Coudray, C., Remesy, C., 2002. Minerals and phytic acid interactions: is it a real problem for human nutrition? International Journal of Food Science & Technology 37, 727739. Lu, W., Grant, L.A., 1999. Effects of prolonged storage at freezing temperatures on starch and baking quality of frozen doughs. Cereal Chemistry 76, 656662. Lucas, T., Quellec, S., Le Bail, A., Davenel, A., 2005. Chilling and freezing of partbaked breads. II. Experimental assessment of water phase changes and of structure collapse. Journal of Food Engineering 70, 151164. Naito, S., Fukami, S., Mizokami, Y., Ishida, N., Takano, H., Koizumi, M., Kano, H., 2004. Effect of freeze-thaw cycles on the gluten brils and crumb grain structures of breads made from frozen doughs. Cereal Chemistry 81, 8086. Palacios, C., Haros, M., Sanz, Y., Rosell, C.M., 2008. Selection of lactic acid bacteria with high phytate degrading activity for application in whole wheat breadmaking. LWT - Food Science and Technology 41, 8292. Phimolsiripol, Y., Siripatrawan, U., Tulyathan, V., Cleland, D.J., 2008. Effects of freezing and temperature uctuations during frozen storage on frozen dough and bread quality. Journal of Food Engineering 84, 4856. Poinot, P., Arvisent, G., Grua-Priol, J., Colas, D., Filllonneau, C., Le Bail, A., Prost, C., 2008. Inuence of formulation and process on the aromatic prole and physical characteristics of bread. Journal of Cereal Science 48, 686697. Ribotta, P.D., Leon, A.E., Anon, M.C., 2001. Effect of freezing and frozen storage of doughs on bread quality. Journal of Agricultural and Food Chemistry 49, 913918. Rosell, C.M., Santos, E., Collar, C., 2006. Mixing properties of bre enriched wheat bread doughs: a response surface methodology study. European Food Research and Technology 223, 333340. Rosell, C.M., Gomez, M., 2007. Freezing in breadmaking performance: frozen dough and part-baked bread. Food Reviews International 23, 303319. Sanz Penella, J.M., Rosell, C.M., Haros, M., 2007. Effect of fungal phytase addition, fermentation and freezing on the phytate degradation of whole wheat dough, in: Proceedings of the 1st Latin American ICC Conference, Rosario, Argentina. Sanz Penella, J.M., Collar, C., Haros, M., 2008. Effect of wheat bran and enzyme addition on dough rheological performance and phytic acid levels in bread. Journal of Cereal Science 48, 715721. Sharadanant, R., Khan, K., 2003. Effect of hydrophilic gums on the quality of frozen dough. II. Bread characteristics. Cereal Chemistry 80, 773780. Shogren, M.D., Pomeranz, Y., Finney, K.F., 1981. Counteracting the deleterious effects of ber in breadmaking. Cereal Chemistry 58, 142144. Turk, M., Sandberg, A.S., 1992. Phytate degradation during breadmaking: effect of phytase addition. Journal of Cereal Science 15, 281294. Zhang, D., Moore, W.R., 1999. Wheat bran particle size effects on bread baking performance and quality. Journal of the Science of Food and Agriculture 79, 805809.

Journal of Cereal Science 50 (2009) 278282

Contents lists available at ScienceDirect

Journal of Cereal Science


journal homepage: www.elsevier.com/locate/jcs

Yield dilution of grain Zn in wheat grown in open-top chamber experiments with elevated CO2 and O3 exposure
Hkan Pleijel a, *, Helena Danielsson b
a b

University of Gothenburg, Plant and Environmental Sciences, P.O. Box 461, 40530 Goteborg, Sweden Swedish Environmental Research Institute, P.O. Box 5302, 40014 Goteborg, Sweden

a r t i c l e i n f o
Article history: Received 26 March 2009 Received in revised form 22 June 2009 Accepted 23 June 2009 Keywords: Carbon dioxide Ozone Zinc Wheat

a b s t r a c t
Wheat (Triticum aestivum L.) grain Zn data from six open-top chamber experiments performed in southwest Sweden were combined to study the relationship between Zn accumulation and grain yield, grain protein, and yield components. Treatments included, in addition to open-top chamber controls, elevated CO2, elevated O3, combined CO2 and O3 exposure, combined elevated CO2 and supplemental irrigation, supplemental irrigation, and ambient air comparison plots. The grain Zn concentration was strongly correlated with grain protein (R2 0.90) over the range of the experimental treatments, representing non-soil factors. A signicant yield dilution effect was found for Zn. For a 10% increase in grain yield, Zn yield was increased by 6.8% on average. Effects on Zn yield correlated strongly with effects on grain protein yield, with a slope close to unity, showing that yield dilution effects for grain Zn and grain protein were similar. Treatment effects on grain Zn concentration were related to effects on grain weight (P < 0.01) and grain number (P < 0.05), but not to harvest index. It was concluded that yield stimulation caused by rising CO2 concentrations is likely to lead to reduced Zn concentrations of wheat grain, thus reducing the nutritional quality. 2009 Elsevier Ltd. All rights reserved.

1. Introduction Global environmental change and its effects on agriculture have a multitude of aspects. Rising CO2 has the potential to stimulate yield in crops, including wheat, Triticum aestivum L. (Amthor, 2001). Moderately elevated ozone can reduce wheat yield (Feng et al., 2008). The net result of these opposing trends may vary from site to site and from time to time. For human nutrition, not only yield magnitude but also nutritional quality is of interest. Several studies have indicated that yield stimulation by elevated CO2 is associated with a reduction in grain protein concentration (Cotrufo et al., 1998; Hogy and Fangmeier, 2008; Taub et al., 2008), while yield reduction by ozone typically leads to a higher grain protein concentration (Pleijel et al., 1999). Although the data are more limited than for grain protein, there are indications that the concentrations of other nutrients, such as Zn, may be diluted by increasing yield (Oury et al., 2006), including yield stimulation by elevated CO2 (Hogy and Fangmeier, 2008). Loladze (2002) suggested a simple denition of yield dilution. This

* Corresponding author. Tel.: 46 31 7862532; fax: 46 31 7862984. E-mail address: hakan.pleijel@dpes.gu.se (H. Pleijel). 0733-5210/$ see front matter 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2009.06.009

was, as a rst approximation, based on the assumption that a constant amount of mineral is taken up regardless of yield. Dilution can then be calculated for a certain level of yield stimulation by CO2 (or yield reduction by O3). It is, however, obvious that the stimulation of growth by CO2 is paralleled by a smaller but signicant stimulation of the accumulation of grain protein per unit area (Pleijel et al., 1999), and that an increased grain yield can be associated with a larger grain protein accumulation per unit area (Pleijel et al., 1999) in wheat. This is likely to be the case also for other mineral nutrient components. In triticale ( Triticosecale Wittm.) grains, for example, the concentrations of protein and micronutrients, including Zn, have been shown to be strongly correlated (Feil and Fossati, 1995). Thus, there is a need for a growth dilution concept which simultaneously takes into account the parallel effects on both grain yield and mineral accumulation. The importance of such an approach for plant breeding purposes was stressed by McDonald et al. (2008), describing a method to evaluate Zn grain concentration under different degrees of yield dilution. They concluded that measuring grain Zn concentration, without considering grain yield, may result in incorrect statements about the true level of genetic variation in grain Zn. Generally decreasing Zn, Fe, Cu, and Mg concentrations in wheat grain in the last decades has been established by Fan et al. (2008).

H. Pleijel, H. Danielsson / Journal of Cereal Science 50 (2009) 278282

279

They found that in the period from the mid 19th century until the present, the Zn concentrations of wheat in south England remained largely constant until the 1960s, when high-yielding cultivars with high harvest index were introduced. These authors concluded that this effect was due to the growth pattern of the plant, where less Zn and other micronutrients are distributed to the grain in relation to the enhanced grain yield, and that soil factors were probably not important for the result. A similar historical shift in wheat grain micronutrient concentrations has been found in the US (Garvin et al., 2006). Since human nutrition with respect to Zn is already inadequate in many areas of the world (FAOSTAT, 2007; Loladze, 2002; Miraglia et al., 2009), it is important to understand if yield stimulation by elevated CO2 may lead to an aggravation of the problem. The undesirable association of grain protein concentration and grain yield has been found to be strongly linked to non-genetic factors (Kibite and Evans, 1984). CO2, O3, and open-top chamber (OTC) enclosure with its change in microclimate all have the potential to inuence grain yield and yield components, without altering soil factors. Ozone is known to negatively affect harvest index (Pleijel et al., 1995) and to reduce grain weight (Feng et al., 2008; Gelang et al., 2000). In many cases elevated CO2 has been shown to increase grain number (Amthor, 2001; Piikki et al., 2008). These effects may be relevant for the Zn accumulation in wheat grain. For example, yield stimulation by increased grain set leads to a large proportion of distal grains, which tend to have lower nutrient concentrations (Calderini and Ortiz-Monasterio, 2003). The aim of the present investigation was to suggest a theoretical framework for the evaluation of yield dilution effects and to apply this framework to grain Zn data from six OTC experiments with wheat, using different levels of O3 and CO2, CO2 in combination with different irrigation, and unchambered ambient air comparison plots. Thus, the environmental treatments were related to atmospheric or meteorological conditions and not to soil factors. The hypotheses of the present investigation were:  There exists a yield dilution effect for Zn in response to effects on yield by elevated CO2, O3, and OTC enclosure.  The yield dilution effect can partly be explained by effects on grain weight, grain number, and harvest index.  The grain concentration of Zn is linked to the grain concentration of protein.

are presented in Table 1. Experiment 1, with a grain yield range of 4.67.1 ton ha1, used ambient and elevated CO2 in a factorial design with ambient and three levels of elevated ozone (number of experimental replicates: n 3). Experiment 2 used ambient and two levels of elevated CO2 (n 5; grain yield range: 4.76.2 ton ha1). Experiment 3 had a factorial design with ambient and elevated CO2 and two levels of irrigation (n 6; grain yield range: 4.55.7 ton ha1). Experiment 4 compared ambient O3 with elevated O3, either before or during and after anthesis (grain yield range: 5.36.1 ton ha1). Experiments 2 and 4 had the same AA control treatment. In the present study, the AA control was only included in the analysis of Experiment 2, in order not to use the same treatment twice. The cultivar used was Dragon in all the experiments except in one of the experiments described in Pleijel et al. (2006), which used a w100-year-old landrace (Lantvete). The treatments of the two different cultivars (exposed in different OTCs) used in that study were considered to represent two separate data sets in the present study. Grain yields were in the range of 4.35.2 and 5.67.9 ton ha1 for Dragon (Experiment 5) and Lantvete (Experiment 6), respectively. Statistical signicance of experimental treatments on Zn concentration, Zn yield, protein concentration, protein yield, and grain yield as well as relevant references for the different experiments are provided in Table 2. The concentration of Zn was analysed by induced coupled plasma atomic emission spectroscopy (ICP-AES). Prior to the analysis, the samples were digested at a temperature of 550  C and then elements were dissolved using HCl prior to dilution with HNO3 (NKL No. 139 1991 mod.). Nitrogen was determined using the Kjeldahl method, and protein concentration was obtained by multiplying the nitrogen concentration by 5.7.

Table 1 Experiments and treatments used in the study including information on CO2 and ozone concentrations (average 08.0020.00) and Zn concentration of the wheat grain. Experiment 1 (1994) Treatment OTC control CO2 O3 1 CO2 O3 1 O3 2 CO2 O3 2 O3 3 CO2 O3 3 AA OTC control CO2 1 CO2 2 AA OTC control CO2 H CO2 H AA OTC control O3 1 O3 2 OTC control O3 AA OTC control O3 AA [CO2] (ppm) 355 660 355 646 355 663 355 654 355 347 515 667 349 355 675 358 696 362 347 [O3] (ppb) 33 33 39 38 46 47 54 54 23 25 26 32 23 41 46 9 57 32 9 52 32 Zn conc (mg kg1) 33 30 34 32 38 31 37 31 35 24 21 23 25 33 31 31 30 28 24 25 27 35 40 31 49 58 54

2. Experimental
2 (1995)

2.1. Data and analysis methods The data for the present study were obtained from six OTC experiments performed at Ostads sateri, south-west Sweden (57 540 N and 12 240 E). All the experiments were performed with eld-grown wheat crops (soil type: arenosol with loamy sand texture) following common agricultural practice. Experimental treatments of the different experiments included, in addition to the control chamber treatments (OTC control), elevated CO2 (CO2), elevated O3 (O3), elevated CO2 in combination with elevated O3 (CO2 O3), elevated CO2 in combination with supplementary irrigation (CO2 H), supplementary irrigation (H) and ambient air comparison treatments without OTCs (AA). The CO2 H and H treatments received double the amount of water compared to the OTC control and CO2 treatments in that experiment. The experiments, as well as yield and growth effects and effects on grain protein, but not Zn data, have been described in the scientic literature. The treatments used in the six experiments, CO2 concentrations, ozone concentrations, and grain Zn concentrations,
3 (1996)

4 (1995)

5 (1999)

6 (1999)

Experiments 46 were ozone experiments where the CO2 concentration was not recorded. Experiment 3 was a CO2 experiment where ozone was not monitored. OTC, open-top chamber; CO2, elevated CO2 treatment; O3, elevated ozone treatment; AA, ambient air (no chamber); H, supplemental irrigation. In experiments where more than one level of elevated CO2 or ozone was used, these are denoted by adding 1, 2, or 3 after the type of treatment.

280

H. Pleijel, H. Danielsson / Journal of Cereal Science 50 (2009) 278282

Table 2 Results of the statistical analysis of treatment effects in the experiments included in the study for Zn concentration, Zn yield, protein concentration, protein yield, and grain yield. Experiment 1 Treatment CO2 O3 CO2 O3 AA CO2 AA CO2 H CO2 H AA O3 O3 AA O3 AA Zn concentration ** ns ns ns ns ns ns ns ns * ns (*) ns (*) ns Zn yield * ns ns ns ns *** ns ns ns * ns (*) ns ns ns Protein concentration *** ns ns ns ** (*) ns * ns ** * * (*) ns ns Protein yield * ns ns ns ns *** ns ns ns ** ns * ns (*) (*) Grain yield *** ns ns ns ns * (*) ns ns ns * *** ns * ns Reference Pleijel et al. (2000)

2 3

Pleijel et al. (2000) Pleijel et al. (2000)

4 5 6

Pleijel et al. (1998) Pleijel et al. (2006) Pleijel et al. (2006)

References for the different experiments are also given. AA, ambient air (no chamber); CO2, elevated CO2; O3, elevated ozone; H, supplemental irrigation; ***, P < 0.001; ** 0.001 P < 0.01; * 0.01 P < 0.05; (*), 0.05 P < 0.1.

2.2. Yield dilution concept and statistical analysis To enhance the comparison between different experiments and treatments, relative scales for the experimental effects on biomass yield, yield components (grain weight, grain number, and harvest index), and on the concentration and yield (offtake) of protein and Zn were used. The relative effect of a certain treatment on a certain variable can be expressed as:

Ai;j;k 1 ai;j;k Aref ;j;k

(1)

where Ai,j,k is the value of variable j of treatment i in experiment k, Aref,j,k is the value of variable j of the control OTC treatment of experiment k, and ai,j,k represents the relative effect on the variable j of treatment i in relation to the OTC control of experiment k. Now, the relative effects for the yield of Zn can be plotted against the relative effect for grain yield, for the whole population of treatments and experiments. The OTC control treatments are only used to calculate experimental effects and not used in further analysis (by denition a would always get the value of zero for all variables). If the data thus plotted result in a linear regression with a slope signicantly smaller than unity, this can be interpreted as a significant yield dilution effect. Similarly, the magnitude of effects of other variables can be compared using the relative scale dened by Eq. (1). The signicance of linear regressions, as well as the test of signicant deviation from 1 of the coefcient of determination of the linear regression, was estimated according to Underwood (1997). The difference of linear regressions was tested using ANCOVA (analysis of covariance). For the data presented in Table 2, Students t-test was used for the test of the chamber effect (AA vs. chamber control) and the ozone effects in Experiments 5 and 6. For Experiments 1 and 3, involving factorial design with two independent variables, two-way ANOVA was used including test for an interaction effect. For the CO2 effects in Experiment 2, and the ozone effects in Experiments 4, one-way ANOVA was used. 3. Results In Fig. 1, the relationship between grain Zn concentration and grain protein concentration for all the included treatments is presented. A strong and highly signicant linear relationship was obtained. There was no indication of the relationship between Zn and protein concentration to be different for different types of

experimental treatments. The three points with the highest values for both Zn concentration and protein concentration represent the landrace (Lantvete). The average relationship obtained in Fig. 1 changes rather little if these values are removed (y 3.3x 13.6; R2 0.73). This relationship is not signicantly different from that presented in Fig. 1. Fig. 2 shows the relative treatment effect on Zn yield plotted against the relative treatment effect on grain yield. A strong linear relationship was found with a slope of 0.68, which was signicantly different from 1 (P < 0.01). Thus, there was a strong indication of yield dilution for Zn over the range of experimental treatments. The linear relationship had no statistically signicant intercept. This indicates that the yield dilution of Zn associated with grain yield stimulation is on average similar to the yield concentration effect on Zn caused by the yield reduction in some of the experimental treatments.

70 60 50 y = 3.58x - 16.7 R2 = 0.90 P < 0.001

grain Zn, mg kg

-1

40 30 20 10 0 AA CO2+ CO2+O3+ CO2+H O3+ H OTC control

10

15

20

25

grain protein, %
Fig. 1. Relationship between grain Zn concentration and grain protein concentrations for all experimental treatments included. AA, ambient air; CO2, elevated CO2 concentration; CO2 O3, elevated CO2 concentration in combination elevated ozone concentration; CO2 H, elevated CO2 concentration in combination with supplemental irrigation; O3, elevated ozone concentration; H, supplemental irrigation; and OTC control, open-top chamber control treatment.

H. Pleijel, H. Danielsson / Journal of Cereal Science 50 (2009) 278282

281

0.5 0.4 y = 0.68x + 0.006 R2 = 0.70 P < 0.001

0.4 0.3 y = 1.06x + 0.05 R2 = 0.85 P < 0.001

relative effect on Zn yield

0.3

relative effect on Zn yield

0.2 0.1 0.0 -0.1 -0.2 -0.3 -0.4 -0.4 -0.3 -0.2 -0.1 0.0 0.1 0.2 0.3
AA CO2+ CO2+O3+ CO2+H O3+ H

0.2 0.1 0.0


AA

-0.1 -0.2

CO2+ CO2+O3+ CO2+H O3+ H

0.4

0.5

relative effect on grain yield


Fig. 2. Relative treatment effects on grain Zn yield plotted against the relative treatment effects on grain yield. For abbreviations see Fig. 1.

-0.3 -0.3

-0.2

-0.1

0.0

0.1

0.2

0.3

0.4

relative effect on grain protein yield


Fig. 3. Relative treatment effects on grain Zn yield plotted against the relative treatment effects on grain protein yield. For abbreviations see Fig. 1.

relative effect on Zn concentration

The data set is too limited to make a detailed comparison of the effects of ozone and carbon dioxide. However, when separate linear regressions for the carbon dioxide treatments and ozone treatments (interaction treatments excluded) were made, the following relationships were derived: y 0.81 * x 0.05 (CO2) and y 0.76 * x 0.08 (O3). Thus, the slopes are similar but the intercepts are somewhat different. The two relationships are not signicantly different. In Fig. 3, the relationship between relative treatment effects on the yield of Zn and the relative treatment effects on grain protein yield is presented. A strong linear relationship was obtained with a slope of 1.06, which was not signicantly different from 1. Thus, the yield dilution was indicated to be similar for Zn and protein. The relative treatment effects on grain Zn concentration were compared with the relative treatment effects on grain weight, grain number, and harvest index. Signicant negative relationships were obtained for grain weight (R2 0.32; P < 0.01) and grain number (R2 0.24; P < 0.05), while only a very weak and non-signicant relationship was obtained for harvest index (R2 0.04) (data not shown). A multiple linear regression model, using the effects on grain weight and grain number as independent variables, explained 40% of the variation in the effect on grain Zn concentration, as presented in Fig. 4. This gure shows that the elevated CO2 treatments typically resulted in a reduction in Zn concentration, while ozone treatments were associated with increased Zn concentrations. The effect of chamber enclosure on Zn concentration was variable. 4. Discussion In the present investigation a signicant yield dilution effect was indicated for Zn in wheat grain over the range of non-soil experimental conditions included, supporting that grain mineral concentration may vary considerably in response to growth pattern, regardless of soil factors (Fan et al., 2008). Furthermore, the Zn concentration was strongly correlated with grain protein concentration, as found earlier by Feil and Fossati (1995) in triticale. Zn yield was inuenced by the different experimental treatments in a way similar to grain protein, further emphasising the strong link between grain Zn and grain protein. Thus, at least for conditions similar to those in the experiments used, the yield dilution effects on grain protein found in many studies (e.g. Kibite and Evans, 1984; Pleijel et al., 1999; Taub et al., 2008) are likely to be followed by

a similar effect on Zn. A very clear example of CO2-induced yield dilution of Zn was presented by Wu et al. (2004). These authors received unusually strong yield stimulation (166%) in well-water plants by doubling the CO2 concentration, which was followed by a reduction in grain Zn concentration from 56 to 38 mg kg1. Similar but smaller effects by CO2 on both grain yield and grain Zn concentration have earlier been observed by Fangmeier et al. (1997) and Manderscheid et al. (1995). In the present study, effects on both grain weight and grain number contributed signicantly to yield dilution. Since both grain nitrogen and grain Zn are present to a large extent in the outer layers of the kernels, a higher proportion of endosperm to total grain weight can explain the yield dilution effect of grain weight (Petersen et al., 1983). The contribution to yield dilution by grain number is likely to be associated with the fact that a larger number of grains depend on the development of a larger proportion of

0.20 0.15 0.10 0.05 0.00 -0.05 -0.10 -0.15 -0.20 -0.20
AA CO2+ CO2+O3+ CO2+H O3+ H

R2 = 0.40

-0.15

-0.10

-0.05

0.00

0.05

0.10

0.15

0.67*(effect on grain weight) + 0.21 *(effect on grain number) - 0.010


Fig. 4. Relative treatment effects on grain Zn concentration plotted against the linear combination of treatment effects on grain weight and grain number from multiple regression analysis. For abbreviations see Fig. 1.

282

H. Pleijel, H. Danielsson / Journal of Cereal Science 50 (2009) 278282 Feil, B., Fossati, D., 1995. Mineral composition of triticale grains as related to grain yield and grain protein. Crop Science 35, 14261431. Feng, Z., Kobayashi, K., Ainsworth, E.A., 2008. Impact of elevated ozone concentration on growth, physiology, and yield of wheat (Triticum aestivum L.): a metaanalysis. Global Change Biology 14, 113. Garvin, D.F., Welch, R.M., Finley, J.W., 2006. Historical shifts in the seed mineral micronutrient concentration of US hard red winter wheat germplasm. Journal of the Science of Food and Agriculture 86, 22132220. Gelang, J., Pleijel, H., Sild, E., Danielsson, H., Younis, S., Sellden, G., 2000. Rate and duration of grain lling in relation to ag leaf senescence and grain yield in spring wheat (Triticum aestivum) exposed to different concentrations of ozone. Physiologia Plantarum 110, 366375. Hogy, P., Fangmeier, A., 2008. Effects of elevated atmospheric CO2 on grain quality of wheat. Journal of Cereal Science 48, 580591. Kibite, S., Evans, L.E., 1984. Causes of negative correlations between grain yield and grain protein concentration in common wheat. Euphytica 33, 801810. Loladze, I., 2002. Rising atmospheric CO2 and human nutrition: toward globally imbalanced plant stoichiometry? Trends in Ecology and Evolution 17, 457461. Manderscheid, R., Bender, J., Jager, H.J., Weigel, H.J., 1995. Effects of season long CO2 enrichment on cereals. II. Nutrient concentrations and grain quality. Agriculture, Ecosystems and Environment 54, 175185. McDonald, G.K., Genc, Y., Graham, R.D., 2008. A simple method to evaluate genetic variation in grain zinc concentration by correcting for differences in grain yield. Plant Soil 306, 4955. Miraglia, M., Marvin, H.J.P., Kleter, G.A., Battilani, P., Brera, C., Coni, E., Cubadda, F., Croci, L., De Santis, B., Dekkers, S., Filippi, L., Hutjes, R.W.A., Noordam, M.Y., Pisante, M., Piva, G., Pradini, A., Toti, L., van den Born, G.J., Vesperman, A., 2009. Climate change and food safety an emerging issue with special focus on Europe. Food and Chemical Toxicology. doi:10.1016/j.fct.2009.02.005. Oury, F.X., Leenhardt, F., Remesy, C., Chanliaud, E., Duperrier, B., Balfourier, F., Charmet, G., 2006. Genetic variability and stability of grain magnesium, zinc and iron concentration in bread wheat. European Journal of Agronomy 25, 177185. Petersen, C.J., Johnson, V.A., Mattern, P.J., 1983. Evaluation of variation in mineral element concentration in wheat our and bran of different cultivars. Cereal Chemistry 60, 450455. Piikki, K., De Temmerman, L., Ojanpera, K., Danielsson, H., Pleijel, H., 2008. The grain quality of spring wheat (Triticum aestivum L.) in relation to ozone uptake and carbon dioxide exposure. European Journal of Agronomy 28, 245254. Pleijel, H., Skarby, L., Wallin, G., Sellden, G., 1995. A process-oriented explanation of the nonlinear relationship between grain yield in wheat and ozone exposure. New Phytologist 131, 241246. Pleijel, H., Danielsson, H., Gelang, J., Sild, E., Sellden, G., 1998. Growth stage dependence of the grain yield response to ozone in spring wheat (Triticum aestivum L.). Agriculture. Ecosystems and Environment 70, 6168. Pleijel, H., Mortensen, L., Fuhrer, J., Ojanpera, K., Danielsson, H., 1999. Grain protein accumulation in relation to grain yield of spring wheat (Triticum aestivum L.) grown in open-top chambers with different concentrations of ozone, carbon dioxide and water availability. Agriculture. Ecosystems and Environment 72, 265270. Pleijel, H., Gelang, J., Sild, E., Danielsson, H., Younis, S., Karlsson, P.E., Wallin, G., Skarby, L., Sellden, G., 2000. Effects of elevated carbon dioxide, ozone and water availability on spring wheat growth and yield. Physiologia Plantarum 108, 6170. Pleijel, H., Berglen Eriksen, A., Danielsson, H., Bondesson, N., Sellden, G., 2006. Differential ozone sensitivity in an old and a modern Swedish wheat cultivar grain yield and quality, leaf chlorophyll and stomatal conductance. Environmental and Experimental Botany 56, 6371. Taub, D., Miller, B., Allen, H., 2008. Effects of elevated CO2 on the protein concentration of food crops: a meta-analysis. Global Change Biology 14, 565575. Underwood, A.J., 1997. Experiments in Ecology. Cambridge University Press, Cambridge. Wu, D.X., Wang, G.X., Bai, Y.F., Liao, J.X., 2004. Effects of elevated CO2 concentration on growth, water use, yield and grain quality of wheat under two soil water levels. Agriculture. Ecosystems and Environment 104, 493507.

distal grains. These typically have lower levels of minerals such as Zn (Calderini and Ortiz-Monasterio, 2003). It was obvious that positive treatment effects on grain yield were associated with more Zn being accumulated in the grain, while yield reductions were followed by a lower Zn yield, although the variation in Zn yield was smaller than in grain yield. It is, thus, obvious that the assumption (Loladze, 2002) that the accumulation of grain mineral components like Zn is constant under grain yield variation caused by, e.g. elevated CO2 is untenable. A yield dilution concept that simultaneously takes into account variation in grain yield and mineral (e.g. Zn) yield is required to understand the effect of environmental change (e.g. CO2, O3, and climate) as well as the genetic variation in grain mineral used in plant breeding considerations (McDonald et al., 2008). The present data set is not sufciently large to discern differences in the relationship between Zn accumulation and grain biomass accumulation between carbon dioxide and ozone. Such differences cannot be excluded. The pattern of response was not identical, although not signicantly different, when the CO2 and ozone effects on yield dilution of the present experiment were compared. The main conclusions of the present study were that atmosphericclimatic conditions can signicantly inuence the Zn concentration of wheat grain that this effect was associated with a signicant degree of yield dilution, that both grain weight and grain number contributed signicantly to yield dilution, and that grain Zn was strongly linked with grain protein. Yield stimulation by further elevated CO2 in the future is likely to lead to lower concentrations of minerals like Zn in wheat grain, with potentially large consequences for human nutrition. Acknowledgement Thanks are due to Patrik Alstromer, Ostads sateri, for making the investigation area available and for support related to agricultural practices and preparation of the experimental area. References
Amthor, J.S., 2001. Effects of atmospheric CO2 concentration on wheat yield: review of results from experiments using various approaches to control CO2 concentration. Field Crops Research 73, 134. Calderini, D.F., Ortiz-Monasterio, I., 2003. Grain position affects grain macronutrient and micronutrient concentrations in wheat. Crop Science 43, 141151. Cotrufo, M.F., Ineson, P., Scott, A., 1998. Elevated CO2 reduces the nitrogen concentration of plant tissues. Global Change Biology 4, 4354. Fan, M.-S., Zhao, F.-J., Fairweather-Tait, S., Poulton, P.R., Dunham, S.J., McGrath, S.P., 2008. Evidence of decreasing mineral density in wheat grain over the last 160 years. Journal of Trace Elements in Medicine and Biology 22, 315324. Fangmeier, A., Gruters, U., Hogy, P., Vermehren, B., Jager, H.J., 1997. Effects of elevated CO2, nitrogen supply and tropospheric ozone on spring wheat II. Nutrients (N, P, K, S, Ca, Mg, Fe, Mn, and Zn). Environmental Pollution 96, 4359. FAOSTAT, 2007. Agricultural Data. Food and Agricultural Organisation of the United Nations, Rome. <http://faostat.fao.org>.

Journal of Cereal Science 50 (2009) 283290

Contents lists available at ScienceDirect

Journal of Cereal Science


journal homepage: www.elsevier.com/locate/jcs

Association analysis reveals effects of wheat glutenin alleles and rye translocations on dough-mixing properties
Shusong Zheng a, Patrick F. Byrne a, *, Guihua Bai b, Xueyan Shan a,1, Scott D. Reid a, Scott D. Haley a, Bradford W. Seabourn c
a b c

Department of Soil and Crop Sciences, Colorado State University, Fort Collins, CO 80523-1170, USA United States Department of Agriculture, Agricultural Research Service, Plant Science and Entomology Research Unit, 4008 Throckmorton Hall, Manhattan, KS 66506, USA United States Department of Agriculture, Agricultural Research Service, Hard Winter Wheat Quality Laboratory, 1515 College Avenue, Manhattan, KS 66502, USA

a r t i c l e i n f o
Article history: Received 5 March 2009 Received in revised form 28 June 2009 Accepted 29 June 2009 Keywords: Glutenin subunits End-use quality Wheat Association mapping

a b s t r a c t
The glutenin loci of wheat (Triticum aestivum L.) are important determinants of bread-making quality, although the effects of alleles at those loci are incompletely understood. We applied an association analysis method to assess the effects of glutenin alleles and 1RS wheatrye (Secale cereale L.) translocations on dough-mixing properties in 96 wheat cultivars and advanced lines grown at three Colorado locations while accounting for population structure and relatedness of individuals in the population. The results indicated that (1) in the majority of cases, controlling relatedness of individuals reduced the signicance of associations between glutenin loci and Mixograph traits; (2) the Glu-D1 and Glu-B3 loci and 1RS translocations had greater impacts on dough-mixing properties compared to other glutenin loci; (3) Glu-B1w, Glu-D1d, and Glu-B3b were consistently associated with greater (more favorable) Mixograph peak time (MPT) than other alleles at the respective loci, whereas Glu-B1e, Glu-D1a, and Glu-B3c were associated with reduced MPT; (4) the 1BL.1RS translocation was associated with a decrease in Mixograph properties. Our results indicate that taking multiple-level relatedness of individuals into account can improve the results of association analysis for wheat-quality traits. 2009 Elsevier Ltd. All rights reserved.

1. Introduction Wheat (Triticum aestivum L.) bread-making quality is a complex collection of traits inuenced by a number of environmental, genetic, and biochemical factors. Among them, grain protein concentration and its composition are of primary importance (reviewed by Singh and Khatkar, 2005). Glutenin, a major component of wheat protein, is formed of polymers linked by disulphide bonds. Glutenin proteins are further classied into two groups, the

Abbreviations: HMW-GS, high-molecular-weight glutenin subunits; LMW-GS, low-molecular-weight glutenin subunits; MPT, Mixograph midline peak time; MPV, Mixograph midline peak value (height); MPW, Mixograph band width at peak; MRS, Mixograph midline right slope, calculated as (MRV-MPV)/2; MRV, Mixograph midline right value (height) at two minutes after peak; MRW, Mixograph band width at two minutes after peak; PCR, polymerase chain reaction; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; SSR, simple-sequence repeat; 1RS, the short arm of rye chromosome. * Corresponding author. Tel.: 1 970 491 6985; fax: 1 970 491 0564. E-mail address: Patrick.Byrne@colostate.edu (P.F. Byrne). 1 Present address: Department of Plant and Soil Sciences, Mississippi State University, Mississippi State, MS 39762, USA. 0733-5210/$ see front matter 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2009.06.008

high-molecular-weight glutenin subunits (HMW-GS) and lowmolecular-weight glutenin subunits (LMW-GS), based on their mobility in sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) under reduced conditions (Singh and Khatkar, 2005). The HMW-GS have been studied in detail because of their important roles in determining bread-making quality, as well as their ease of separation on SDS-PAGE. The LWM-GS have not been studied as thoroughly because they are relatively difcult to separate in gels (Singh and Khatkar, 2005). A number of genes of agronomic importance, including disease and pest resistance genes, have been transferred from rye (Secale cereale L.) to wheat in the form of 1AL.1RS, 1BL.1RS, and 1DL.1RS wheatrye chromosomal translocations (1AL, 1BL, and 1DL are the long arms of group 1 wheat chromosomes; 1RS signies the short arm of rye chromosome 1) (Graybosch, 2001). Among these, the 1AL.1RS and 1BL.1RS translocations are most common in wheat cultivars. Unfortunately, the 1RS translocation has often resulted in signicant reduction in wheat end-use quality (Graybosch, 2001). Severe quality problems are recognized when 1BL.1RS is present in hard wheat backgrounds, including low SDS sedimentation volumes, the production of sticky doughs, a lack of

284

S. Zheng et al. / Journal of Cereal Science 50 (2009) 283290

tolerance of dough to overmixing, and low loaf volumes (Graybosch, 2001). Wheats cultivars with the 1AL.1RS translocation have better end-product qualities than those with the 1BL.1RS translocation (Graybosch et al., 1993; Kumlay et al., 2003), although 1AL.1RS reduces mixing strength and SDS sedimentation volumes compared to non-1RS sibling lines (Espitia-Rangel et al., 1999; Graybosch et al., 1993). Both irrigated and rainfed (dryland) environments are important for wheat production in semi-arid areas of the west central Great Plains of the US. For example, irrigated wheat accounted for one-sixth of the total wheat production in Colorado in 2007, with rainfed wheat making up the rest (www.nass.usda.gov/Statistics_ by_State/Colorado/Publications/County_Estimates/alwhtce07.pdf, veried on Dec. 15, 2008). Breeders would benet from knowing whether the same genetic factors affect bread-making quality under both sets of moisture conditions. The Mixograph has traditionally been used to determine the rheological characteristics and bread-making potential of dough systems by the wheat milling and baking industry (Kunerth and DAppolonia, 1985). US breeding programs typically prefer the Mixograph over other methods for assessing the functional properties of our because the method is fast (typically no more than 8 min), can accommodate a relatively small amount of our (10 g), and imparts a comparatively large amount of work input to the dough that provides greater discrimination of the end-use performance of breeding lines. The quality effects of glutenin alleles and 1RS have been studied in two types of populations: structured populations, typically derived from biparental crosses and investigated via quantitative trait locus (QTL) analysis, and non-structured populations, usually collections of cultivars and breeding lines that are analyzed by association analysis methods (Yu and Buckler, 2006). Non-structured populations have several advantages, including the ability to simultaneously evaluate many alleles at each locus, assess allele function in genetic backgrounds of current breeding materials, and use multi-location and multi-year quality data at no additional costs because these data are collected routinely in many wheat-breeding programs (Breseghello and Sorrells, 2006; Cane et al., 2008; Eagles et al., 2002). Due to these advantages, the largest number of associations between glutenin alleles and quality characteristics have been established in collections of wheat cultivars and advanced lines (Bekes et al., 2006; Branlard et al., 2001; Cane et al., 2008; Eagles et al., 2002; He et al., 2005). Nevertheless, there are certain precautions that must be heeded with association analysis of non-structured populations to avoid reaching incorrect conclusions. Due to geographic isolation, natural selection, or articial selection, population stratication is common for plant germplasm (Yu et al., 2006), and has been observed in collections of wheat germplasm (Breseghello and Sorrells, 2006; Ravel et al., 2006). Population stratication is the presence of a systematic difference in allele frequencies between subpopulations within a population (Yu and Buckler, 2006). This population stratication and an unequal distribution of alleles within a population can result in spurious associations (Knowler et al., 1988). Therefore, genome-wide distributed molecular markers have been used to estimate population structure and relatedness of individuals, and incorporating descriptions of these relationship into the statistical model has helped to reduce false associations (Pritchard et al., 2000; Yu and Buckler, 2006). A unied mixed model (QK) was proposed to account for both population structure and relatedness among individuals (Yu et al., 2006). Therefore, the objective of this study was to evaluate the effects of alleles at the Glu-1 and Glu-3 loci and the presence of the 1RS translocation on dough-mixing properties in a collection of 96

wheat cultivars and lines in irrigated and rainfed Colorado environments while accounting for multiple-level relatedness of individuals. 2. Experimental 2.1. Plant materials A population of 96 hard red and hard white winter wheat cultivars and advanced lines was evaluated in this study. These included 87 entries chosen from a larger set of wheat germplasm released or developed since 1991 in the Hard Winter Wheat Region of the US Great Plains and which were mostly included in the glutenin survey by Shan et al. (2007). Based on pedigree information, we attempted to avoid including sister lines or other closely related materials. In addition, eight entries from Eastern Europe and Central Asia were included (Supplementary Table S1). 2.2. Field experiments The 96 genotypes were grown at three locations in the northeast quadrant of Colorado in the 20052006 growing season: the Agricultural Research, Development and Education Center of Colorado State University (CSU), Fort Collins; the Central Great Plains Research Station, USDA Agricultural Research Service, Akron; and a farmers eld near Dailey. At all locations, the elds were arranged in an alpha-lattice incomplete block design (Patterson et al., 1978) with two replications, and each plot consisted of two rows with 25 cm spacing between rows. Plots at Fort Collins were 3.4 m long, and plots at Akron and Dailey were 3.7 m long. Ten grams of seed per plot were planted in each trial. The trial at Fort Collins was irrigated with a linear overhead sprinkler system as needed to maintain optimum moisture conditions until three weeks before harvest. Akron and Dailey were rainfed locations with no additional water applied to those trials. All plots were hand harvested and threshed separately. There was very little difference in growing season temperature among the three locations according to the CoAgMet weather database (http://ccc. atmos.colostate.edu/wcoagmet/, veried on June 25, 2009). Of the rainfed locations, Dailey received twice as much precipitation as Akron in May (10.5 and 4.5 cm, respectively), coinciding with anthesis and early grain lling, a critical time for wheat grain development. 2.3. Phenotypic evaluation 2.3.1. Agronomic traits Phenotypic data were recorded or calculated for each plot for grain yield, kernel weight, days to heading, and grain ll duration. The single kernel weight was estimated using a 200-kernel subsample from each plot. Days to heading was the number of days from January 1, 2006 to the day on which 50% of the spikes were fully visible above the ag leaf collar. Grain ll duration was determined by subtracting days to heading from the number of days from January 1, 2006 to the date when 50% of peduncles showed complete loss of green color. 2.3.2. Mixograph properties Fifty grams of seed from each plot were milled to our with a Brabender Quadrumat Jr. Mill. (C. W. Brabender Instruments, Inc., South Hackensack, NJ) following AACC Method 26-50 (AACC, 2004). Dough-mixing properties were analyzed for each plot with a 10 g-Mixograph at room temperature, according to the approved AACC method 54-40A (AACC, 2004). Mixograph characteristics were determined with the software program Mixsmart version

S. Zheng et al. / Journal of Cereal Science 50 (2009) 283290

285

3.80 (AEW Consulting, Lincoln, NE, commercially available through National Manufacturing Division of TMCO, Lincoln NE, USA). An explanation of the Mixogram output was presented by Bordes et al. (2008). Briey, all mixing characteristics were measured from the center line (Mixograph midline) of the mixogram, and all characteristics except mixing time were reported on a 100Mixograph unit scale (% height). Mixograph midline peak time (MPT) was visually determined. Other characteristics were automatically estimated by Mixsmart, including midline peak (height) value (MPV), midline peak width (MPW), midline right value (height) at two minutes after peak (MRV), right width at two minutes after peak (MRW), and midline right slope (MRS) calculated as (MRV-MPV)/2, as shown in the mixogram (Fig. 1, Bordes et al., 2008). The optimum MPT for hard red winter wheat is 36 min, and the target for mixing tolerance is at least 3 on a 7-point scale (06), as suggested in the report by Chen and Seabourn (2006). For this population, higher values of MPT, MPW, MRW, and MRS and lower value of MPV are considered desirable for bread-making quality. 2.4. Genotypic evaluation 2.4.1. Glutenin analysis The glutenin proteins from each entry (cultivar or advanced line) were extracted and analyzed by SDS-PAGE according to the protocols described by Shan et al. (2007). Ten random seeds from each genotype were bulked for each extraction. For each genotype, loci with protein bands corresponding to more than one allele were scored as the allele with the higher intensity. The only exception was for Glu-A1a and b whose intensities were very similar; these were scored as having both alleles present and denoted as Glu-A1a/b. Assignment of alleles at Glu-A3 and Glu-B3 was sometimes problematic in cases where the cultivar contained a 1AL.1RS or 1BL.1RS translocation, respectively. If a translocation replaces the whole short arm of the respective wheat chromosome, then no Glu-A3 or Glu-B3 locus would be present. However, we detected bands for Glu-B3 alleles in 13 1BL.1RS cultivars, indicating most likely that these lines were heterogeneous for the 1RS translocation. For Glu-A3, only seven of 17 lines scored as having the Glu-A3e (null) allele also carried the 1AL.1RS translocation and we were unable to resolve that potential discrepancy. 2.4.2. Simple-sequence repeats (SSR) marker analysis DNA was extracted from bulked leaves of 10 seedlings of each entry using the method described by J.A. Anderson, Department of Agronomy and Plant Genetics, University of Minnesota (http:// maswheat.ucdavis.edu/PDF/DNA0002.pdf, veried on June 25, 2009). SSR primers were selected based on the map of Somers et al. (2004). For each chromosome, we chose at least three SSR markers (one in each telomere region and one in the centromere region) for testing, and had the primers synthesized according to information available in the GrainGenes database (http://wheat.pw.usda.gov/ GG2/index.shtml, veried on June 25, 2009). A total of 63 SSR markers were applied to the 96 entries. The presence of the 1AL.1RS and 1BL.1RS translocations was determined with the rye-specic SSR marker SCM9 (Saal and Wricke, 1999). All SSR genotyping was carried out in the USDA Small Grain Genotyping Laboratory in Manhattan, KS, following the method of Chen et al. (2008) with some modications. In brief, a 13ml polymerase chain reaction (PCR) mixture contained 50150 ng/mL DNA template, 200 mM of dNTP, 1 ASB buffer (Bioline USA Inc. Taunton, MA), 2.5 mM MgCl2, 1 U Taq polymerase, 100 nM forward primer, 300 nM of reverse primer, and 200 nM uorescence dye-labeled M13 primer compatible with an ABI 3730 DNA Analyzer (Applied Biosystems, Foster City, CA). The PCR were conducted in an MJ-PTC200 Cycler (Biorad, Hercules, CA) and detected by an ABI

3730 DNA analyzer. The marker data were scored using GeneMarker version 1.5 (SoftGenetics, LLC, State College, PA). Markers producing a single band were preferred. If markers produced two well-separated bands, those markers were scored as independent loci. A total of 69 SSR loci were scored from these SSR markers, including one locus each from 57 markers and two loci each from the other six markers. 2.5. Data analysis 2.5.1. Multiple levels of relatedness of 96 entries To minimize spurious associations, the mixed model (QK) (Yu et al., 2006) was used to account for population structure and cousin-level relatedness of individuals among 96 entries. The genetic structure among the 96 entries was inferred with the 60 SSR loci that had a maximum of 16 missing values, using the admixture and correlated allele frequency options in the program STRUCTURE version 2.2 (Pritchard et al., 2000). All the alleles of these SSR markers were used to estimate the population structure. The a priori subpopulation number (k) was estimated with the program StructuRAMA (Huelsenbeck and Andolfatto, 2007) with the same set of 60 SSR loci used for the STRUCTURE analysis. The estimated k was eight. The Q-matrix was estimated in STRUCTURE with 60 SSR markers, eight subpopulations, and one million burnin periods, followed by two million Markov Chain Monte Carlo iterations. The Q-matrix was visualized with the program Distruct (Rosenberg, 2004). The relative kinship (K) matrix (K being distinct from the a priori subpopulation number k) was calculated on the basis of 69 SSR loci using the method proposed by Ritland (1996), which is built into the program SPAGeDi (Hardy and Vekemans, 2002). 2.5.2. Least squares means, correlation, and association analysis Data were analyzed with SAS software version 9.2 (SAS Institute, Cary, NC). First, exploratory association analysis was performed on Mixograph properties combining data from the three locations. This analysis revealed highly signicant (P <0.0001) genotype environment interactions between most of the glutenin loci and growing locations (data not shown). Thus, our data were analyzed and are presented separately by location. Second, least squares means of all traits at each location were estimated with the MIXED procedure of SAS. In the model, entry was treated as a xed effect and incomplete block as a random effect. Third, homogeneity of error variance of Mixograph properties was evaluated by the KolmogorovSmirnov test and visually with residual plots. The expected and residual values used in residual plots were obtained from the MIXED procedure for calculating least square means. MPW and MRW were not consistent with an acceptable level of homogeneity of error variance across the three locations, so they were loge (ln) transformed to achieve homogeneous variance in all the three environments. The results were back transformed to original units for presentation purposes. Fourth, the association of glutenin loci and 1RS translocations with Mixograph properties was tested with the QK model (Yu et al., 2006). The SAS code for the QK model was adapted from E. Buckler (USDA-ARS, Plant, Soil and Nutrition Research Unit, Ithaca, NY, http://www2. maizegenetics.net/images/stories/interests/qk.txt, veried on June 25, 2009). The subpopulations together with one glutenin locus or 1RS translocation were treated as xed effects, incomplete block as a random effect, and entry as a random effect with the K-matrix dening the degree of genetic covariance between a pair of individuals. The LSMEAN statement in the MIXED procedure was used to estimate the least squares means for each allele of the glutenin loci and 1RS status. Henceforth, we will use mean to refer to the least squares mean. The means for different allelic classes at

286

S. Zheng et al. / Journal of Cereal Science 50 (2009) 283290 Table 2 Least squares means of Mixograph properties for glutenin allelic classes and 1RS translocations of 96 wheat genotypes at Akron, CO, in the 20052006 growing season. Loci AlleleA n Mixograph propertiesB MPWC MRW 10.3a 11.0a 9.8a 9.3a 10.2a 10.2a 8.7a 10.1a 10.1a 11.6a 11.5a 10.0a 9.7a 9.7a 9.9a 11.1a 9.0a 13.1a 11.1a 5.7b 10.9a 10.3a 9.9a 10.0a 9.5a 9.5a 10.6a 10.3a 8.8a 10.9a 9.1b 7.6c 10.6a MPT 3.2b 3.9ab 3.4ab 4.0a 3.5a 3.6a 2.7b 3.1ab 3.7a 2.7b 3.2ab 3.4a 3.2ab 3.2b 3.8ab 3.9a 2.9ab 3.8ab 3.9a 2.8b 3.1ab 3.2ab 3.3ab 2.8b 3.0b 3.4ab 3.7a 3.3ab 2.8b 3.3ab 3.2a 3.1a 3.4a MPV 48.3ab 50.2a 48.7a 43.9b 48.2a 48.8a 51.0a 48.8a 48.2a 50.9a 50.2a 48.2a 47.8a 49.3a 49.0a 47.7a 47.0a 43.3bc 48.8ab 41.9c 49.3ab 51.2a 48.8ab 51.0a 48.2ab 48.2a 48.3a 49.2a 50.3a 50.0a 48.4a 43.2b 49.8a MRS 3.0a 3.1a 3.2a 2.4a 2.9b 3.1b 4.1a 3.5ab 3.0ab 3.6a 3.0a 3.1a 3.6a 3.3a 3.1ab 2.5b 3.1ab 1.6b 2.8ab 3.4ab 3.4ab 3.3ab 3.3ab 4.3a 3.2ab 3.3a 3.0a 3.1a 3.6a 3.0a 3.3a 3.0a 3.3a

glutenin loci or presenceabsence of 1RS were compared with TukeyKramer multiple-range tests.

3. Results 3.1. Trait means Wheat from the irrigated location, Fort Collins, had higher values (P < 0.05) of most agronomic traits (grain yield, single kernel weight, and grain ll duration) compared to those from the rainfed locations, Dailey and Akron (Table 1). Wheat from Akron had the highest grain protein concentration (170 g/kg), followed by those from Fort Collins (165 g/kg), and Dailey (142.7 g/kg). The MPT for the three locations were in the order of Dailey > Akron > Fort Collins (P < 0.001). Wheat from Fort Collins had wider MPW than those from Dailey and Akron (P < 0.001), whereas wheat from Dailey had a wider MRW than those from Fort Collins and Akron (P < 0.001). Wheat from Dailey had lower MPV and higher MRS than Fort Collins and Akron (P < 0.001), indicating that ours from wheat grown at Dailey were more tolerant to mixing. Signicant (P < 0.05) differences among the entries were found for each trait at each location (data not shown) except single kernel weight at Fort Collins and grain ll duration at Akron.

Glu-A1

a a/b b c b c e i w a b d a c d e g a b c e f g h i a b c d e 1AL.1RS 1BL.1RS Non-1RS

14 8 71 3 33 38 10 7 7 10 6 79 14 51 7 18 4 3 14 5 8 11 40 4 10 30 27 28 6 5 17 13 65

19.1a 18.8a 18.6a 16.0a 19.0a 18.4a 18.9a 19.6a 17.3a 19.5ab 22.4a 18.2b 18.2a 19.1a 19.2a 18.2a 17.6a 15.8a 17.8a 16.9a 19.4a 20.4a 18.9a 18.6a 18.4a 18.6a 18.2a 18.5a 18.6a 21.9a 18.0ab 16.8b 19.2a

Glu-B1

Glu-D1

Glu-A3

Glu-B3

3.2. Glutenin allelic and 1RS diversity A high level of diversity was observed for most glutenin loci in this set of materials (Tables 24). The most common alleles at different glutenin loci were Glu-A1b, Glu-B1b and c, Glu-D1d, GluA3c, Glu-B3g, and Glu-D3a, b, and c. Some alleles were present in less than three entries, including Glu-B1a (one entry), Glu-D1e (one entry), Glu-A3b (two entries), and Glu-B3d (one entry). These rare alleles were not included in the analysis to reduce the level of multiple comparisons. There were four classes of entries with regard to the presence of 1RS translocations (Tables 24). Sixty-ve entries had no rye translocation (non-1RS), 17 entries had a 1AL.1RS translocation, 13 entries had a 1BL.1RS translocation, and one entry (Innity) was found to have both 1AL.1RS and 1BL.1RS translocations. Because Innity was the only entry in this class, it was excluded from the analysis for 1RS effects. When the genotypic data were analyzed with the program STRUCTURE, population stratication of the 96 wheat cultivars was obvious (Supplementary Fig. S1). The entries formed eight hypothetical ancestral groups. Most entries were not assigned to a single group, but to different proportions of multiple groups.

Glu-D3

1RS

A Subunit designations for the HMW-GS loci are Glu-A1: a, 1; b, 2*; c, null; Glu-B1: b, 78; b, 79; e, 20x20y; i, 1718; w, 6*8*; Glu-D1: a, 212; b, 312; d, 510. B Means followed by different letters indicate a signicant difference at the P < 0.05 level. C MPT, Mixograph peak time; MPW, Mixograph band width at MPT; MRW, Mixograph band width at two minutes after MPT; MPV, Mixograph midline peak value; MRS, Mixograph midline right slope, calculated by (Mixograph right value at two minutes after peak-MPV)/2.

Table 1 Least squares means and ranges of Mixograph properties and agronomic traits of 96 wheat genotypes grown in three Colorado locations in the 20052006 growing season (n 96 except KW and SPH from Dailey, for which n 95). Variablea Fort Collins Mean SD Grain yield (kg/ha) Kernel weight (mg) Days to heading (days) Grain lling duration (days) Grain protein (g/kg) MPT (min) MPW (% heightc) MRW (% height) MPV (% height) MRS (% height/min) 2867 28.5 141.2 26.3 165.1 2.9 21.6 10.8 50.6 3.8 655 2.4 1.5 1.4 8.9 0.6 3.7 3.4 4.2 1.1 Range 11514365 20.234.1 138.5145.4 23.230.2 135.4183.6 1.54.7 13.333.8 5.322.5 41.360.7 6.1 to 1.3 Akron Mean SD 1313 22.7 142.7 23.2 170.1 3.3 19.1 10.5 48.6 3.2 308 2.5 1.5 1.1 5.7 0.8 3.6 3.4 4.1 1.1 Range 7262046 16.928.9 140.1146.8 20.626.4 157.51872 1.85.1 12.630.7 5.122.2 38.161.6 5.6 to 1 Dailey Mean SD 1440 25.0 142.3 23.4 142.7 4.6 19.1 15 43.5 1.4 270 2.5 1.7 1.2 8 0.9 3.1 3.5 3.9 0.8 Range 3792082 1931.3 138.7147.4 20.426.2 118.0159.4 2.67.1 10.527.5 5.423.7 34.557.1 4.1 to 0.5 127 0.7 0.4 0.3 2.2 0.2 1.0 1.0 1.2 0.3 LSD0.05b

a SD, standard deviation; MPT, Mixograph peak time; MPW, Mixograph band width at MPT; MRW, Mixograph band width at two minutes after MPT; MPV, Mixograph midline peak value; MRS, Mixograph midline right slope, calculated by (Mixograph right value at two minutes after MPT-MPV)/2. b LSD0.05, Fishers least signicant difference among environments at P < 0.05, calculated based on the SAS PROC MIXED analysis of variance table. c % height, Mixograph units in 100-unit scale.

S. Zheng et al. / Journal of Cereal Science 50 (2009) 283290 Table 3 Least squares means of Mixograph properties for glutenin allelic classes and 1RS translocations of 96 wheat genotypes at Dailey, CO, in the 20052006 growing season. Loci Allele n Mixograph propertiesA MPWB Glu-A1 a a/b b c b c e i w a b d a c d e g a b c e f g h i a b c d e 1AL.1RS 1BL.1RS Non-1RS 14 8 71 3 33 38 10 7 7 10 6 79 14 51 7 18 4 3 14 5 8 11 40 4 10 30 27 28 6 5 17 13 65 18.6a 18.7a 19.0a 16.0a 19.1a 18.8a 19.1a 18.4a 17.7a 21.0a 21.4a 18.4b 18.8a 19.0a 20.3a 18.2a 18.1a 16.5a 20.1a 18.0a 19.0a 19.3a 18.8a 18.2a 18.8a 18.6a 19.2a 18.6a 18.7a 21.2a 18.3a 16.4b 19.5a MRW 13.4a 15.3a 14.7a 14.0a 15.5a 14.7a 13.2a 13.4a 13.9a 14.8a 14.0a 14.4a 13.5a 14.4a 14.8a 15.5a 13.7a 16.3a 16.2a 8.7b 14.2a 15.3a 14.6a 12.6ab 13.6a 14.1a 14.9a 15.4a 12.3a 15.2a 14.0a 10.6b 15.3a MPT 4.8a 4.5a 4.5a 5.0a 4.8a 4.6a 3.9a 4.2a 5.0a 3.7b 3.8b 4.7a 4.3a 4.5a 4.6a 5.0a 4.7a 5.2a 5.0a 4.1a 4.6a 4.5a 4.5a 4.1a 4.1a 4.7a 4.7a 4.4a 4.3a 4.8a 4.3a 4.3a 4.6a MPV 42.4a 41.7a 44.1a 40.9a 43.4a 43.7a 44.8a 44.6a 41.1a 46.2a 46.2a 42.9b 44.2a 44.0a 42.6a 42.7a 43.4a 38.1b 43.8ab 41.4ab 43.6ab 45.7a 43.5ab 46.7a 43.8ab 43.5a 43.6a 44.3a 41.7a 44.4a 44.3a 40.5b 44.1a MRS 1.4a 0.7a 1.6a 0.7a Glu-B1 Glu-B1 1.2a 1.3a 1.8a 1.8a 0.8a Glu-D1 Glu-D1 2.1a 1.9ab 1.2b Glu-A3 Glu-A3 1.6a 1.4a 1.2a 0.9a 2.0a Glu-B3 Glu-B3 0.6a 1.2a 2.0a 1.5a 1.6a 1.3a 2.0a 1.6a Glu-D3 Glu-D3 1.6ab 1.3a 1.3a 3.0b 1.3ab 1RS 1RS 1.3a 1.5a 1.4a Glu-A1 a a/b b c b c e i w a b d a c d e g a b c e f g h i a b c d e 1AL.1RS 1BL.1RS Non-1RS 14 8 71 3 33 38 10 7 7 10 6 79 14 51 7 18 4 3 14 5 8 11 40 4 10 30 27 28 6 5 17 13 65

287

Table 4 Least squares means of Mixograph properties for glutenin allelic classes of 96 wheat genotypes at Fort Collins, CO, in the 20052006 growing season. Loci Allele n Mixograph propertiesA MPWB 21.3a 20.2a 21.2a 20.6a 21.6a 20.8a 19.5a 21.5a 21.9a 21.1a 22.3a 21.0a 21.0a 21.1a 21.8a 20.8a 22.4a 21.7a 22.3ab 15.9c 22.1ab 23.3a 20.8ab 24.7a 19.3bc 21.0a 21.0a 21.8a 20.7a 19.2a 19.8b 17.6c 22.1a MRW 9.7a 12.4a 10.2a 9.3a 11.0a 10.0a 8.6a 10.6a 11.0a 9.3a 9.6a 10.5a 9.7a 9.9a 11.0a 11.3a 10.0a 14.0a 11.7a 6.7b 11.5a 10.7a 9.9a 9.8ab 9.1ab 10.0a 10.7a 10.7a 9.3a 9.3a 8.9b 8.2b 10.9a MPT 3.0a 3.5a 2.9a 3.2a 3.0a 3.0a 2.2b 2.8ab 3.4a 2.4b 2.6ab 3.0a 2.8a 2.9a 3.0a 3.2a 2.4a 3.3ab 3.2a 2.3b 2.9ab 2.9ab 2.9ab 2.8ab 2.5b 3.0a 3.1a 2.9a 2.4a 2.8a 2.8a 2.8a 2.9a MPV 50.5a 50.4a 50.8a 48.6a 50.1a 51.1a 50.7a 49.5a 52.2a 52.1a 51.6a 50.1a 50.3a 51.6a 48.8a 49.7a 48.0a 47.8ab 50.2a 42.8b 52.4a 53.8a 50.5a 53.3a 50.8a 50.8a 50.1a 51.3a 49.9a 50.7a 49.5b 45.8c 51.8a MRS 3.5a 2.9a 3.8a 3.5a 3.4b 3.8ab 4.8a 3.9ab 3.0ab 4.6a 4.3ab 3.5b 4.0a 3.9a 3.3a 3.3a 3.9a 2.1b 3.2ab 4.1ab 3.9ab 4.1a 3.7ab 4.6a 4.3a 3.9a 3.4a 3.7a 4.0a 4.1a 3.9a 3.6a 3.8a

Means followed by different letters indicate a signicant difference at the P<0.05 level. B Trait abbreviations are the same as for Table 2.

A Means followed by different letters indicate a signicant difference at the P < 0.05 level. B Trait abbreviations are the same as for Table 2.

3.3. Relationship of glutenin loci and 1RS translocations with dough-mixing properties We tested the association of the six glutenin loci and the 1RS translocations with Mixograph properties by using the QK model (Yu et al., 2006) and a basic model (not accounting for QK). The QK model gave better model t as indicated by lower Akaike Information Criterion (AIC) values (data not shown). The QK model reduced the signicance levels of most associations in the basic association model (Table 5). For example, Glu-D3 was significantly (P < 0.05) associated with MPW, MRW, and MPV at Fort Collins in the basic model, whereas none of these associations was signicant in the QK model. On the other hand, the QK model detected additional signicant (P < 0.05) associations among glutenin loci and Mixograph properties which were not signicant with the basic model. For example, in the QK model, there were signicant (P < 0.05) associations between Glu-D1 and MPW and Glu-A3 and MPT at Dailey and between Glu-A1, Glu-A3, and Glu-D3 and MPT at Akron. These associations were not signicant in the basic model. Hereafter, only the results from the QK model are presented. All six glutenin loci and translocation status were associated with Mixograph properties, with Glu-D1, Glu-B3, and 1RS status

affecting most traits at all the three locations (Table 5). The associations among genetic factors and Mixograph properties were highly inuenced by wheat-growing environment (Table 5). The associations detected at the irrigated location, Fort Collins, were very similar to those of the rainfed location Akron, but different from the rainfed location Dailey. Results for specic loci are described in the following paragraphs. 3.3.1. Glu-A1 Glu-A1c was associated with higher (P < 0.05) MPT than GluA1a, and lower (P < 0.05) MPV than Glu-A1b and a/b at Akron, but not at other locations (Tables 24). Glu-A1a, b, and a/b were not signicantly different from each other for any of the Mixograph properties. 3.3.2. Glu-B1 Variation at the Glu-B1 locus was signicantly associated with MPT (P < 0.01) at Fort Collins and Akron (Tables 2 and 4). This was mainly because Glu-B1e produced lower MPT than Glu-B1b, c, and w. In addition, Glu-B1e was associated with a more negative MRS than Glu-B1b at Akron and Fort Collins, and therefore is an undesirable allele at the Glu-B1 locus. In this rst study of the effects of Glu-B1w, we observed that Glu-B1w had similar values for

288

S. Zheng et al. / Journal of Cereal Science 50 (2009) 283290

Table 5 The effects of glutenin loci and the presence of 1RS translocations on Mixograph properties of 96 wheat genotypes at Fort Collins, Dailey, and Akron, CO, in the 20052006 growing season. Location Loci Basic model (not adjusted for QK) MPWa Akron Glu-A1 Glu-B1 Glu-D1 Glu-A3 Glu-B3 Glu-D3 1RS Glu-A1 Glu-B1 Glu-D1 Glu-A3 Glu-B3 Glu-D3 1RS Glu-A1 Glu-B1 Glu-D1 Glu-A3 Glu-B3 Glu-D3 1RS MRW MPT *** *** *** *** ** ** *** ** * * * *** *** ** *** * *** *** MPV *,b * * ** * ** *** *** * * * * * *** *** * *** ** * * *** * *** MRS QK model MPW MRW MPT ** *** * ** ** * MPV * * * * MRS

*** ***

Dailey

** * * *** * *** ** * *** *** ***

** *

** *

** * *

***

*** * *** **,b **

***

Fort Collins

*** ** *** *** *** *** * *** ***

* ** **

*** * ***

*** ** ***

Results from the basic model are to the left and those from the QK model are to the right. *, **, ***, F test signicant at 0.05, 0.01, and 0.001 levels, respectively; empty cells indicate not signicant (P > 0.05) in the F test. a Trait abbreviations are the same as for Table 2. b Cells shaded gray in the basic model indicate that the strength of associations was reduced in the QK model. Cells shaded black in the QK model indicate the strength of associations was increased in the QK model.

all Mixograph traits compared to other Glu-B1 alleles, except for MPT, which tended to be higher. Other alleles were not signicantly different from each other for any of the Mixograph properties. 3.3.3. Glu-D1 Glu-D1d was associated with a higher MPT (P < 0.01) than GluD1a at all the three locations and higher (P < 0.01) MPT than GluD1b at Dailey. Glu-D1d was associated with lower MPW (P < 0.05) than Glu-D1b at Akron and Dailey, and higher MRS than Glu-D1a at Fort Collins and Dailey (Tables 24). 3.3.4. Glu-A3 Variation at Glu-A3 was signicantly (P < 0.05) associated with MPT and MRS at Akron (Tables 2 and 5). The allele e was associated with higher (P < 0.05) MPT and MRS than allele a or c at Akron. The means of allelic classes for MPT were in the order of e > d > a > c > g at Akron (Tables 24). 3.3.5. Glu-B3 Variation at Glu-B3 was signicantly (P < 0.05) associated with MRW, MPT, MPV, and MRS at Akron, MRW and MPV at Dailey, and all the properties at Fort Collins (Table 5). Glu-B3c was associated with lower (P < 0.05) MRW than Glu-B3a, b, e, f, and g at all the three locations, lower MPV (P < 0.01) than most other Glu-B3 alleles at Akron and Fort Collins, and lower MPW (P < 0.01) than most other Glu-B3 alleles at Fort Collins except Glu-B3i (Tables 24). Glu-B3i produced lower MPT (P < 0.05) than Glu-B3b at Akron and Fort Collins (Tables 2 and 4). Glu-B3h was associated with the most negative MRS among Glu-B3 alleles at Akron and Fort Collins (Tables 2 and 4). 3.3.6. Glu-D3 The Glu-D3d allelic class was signicantly (P < 0.05) associated with shorter MPT than Glu-D3b at Akron and more negative

(P < 0.05) MRS than Glu-D3b and c at Dailey (Tables 24). Other alleles were associated with similar values for Mixograph properties across the three locations (Tables 24). 3.3.7. 1RS translocation 1BL.1RS was associated with the lowest values of MPW, MRW, and MPV at all the three locations (Tables 24). Non-1RS was associated with higher MPW, MRW, and MPV than 1AL.1RS at Fort Collins and higher MRW than 1AL.1RS at Akron. 4. Discussion We used a candidate gene association analysis to study the HMW-GS, LMW-GS, and 1RS translocation effects on wheat quality as measured by the Mixograph. The analysis accounted for multiple-level relatedness with the unied mixed model suggested by Yu et al. (2006). Of 64 traitlocus associations that were signicant (P< 0.05) in the basic model, 52 showed a reduced strength of association when the QK model was applied (Table 5). Many of these were presumably spurious associations or showed inated signicance levels due to the relatedness among some of the entries. In a smaller number of cases (11), the signicance of the association was increased with the QK model compared to the basic model, suggesting that false negative as well as false positive associations can be detected in the absence of adjustment for population structure (Table 5). Population stratication in wheat has been observed in the previous studies (Breseghello and Sorrells, 2006; Ravel et al., 2006). Cane et al. (2008) were aware that the relatedness of cultivars might inuence the association between alleles and quality traits, so they included a relationship matrix generated from pedigree information in the analysis to minimize bias. However, when they compared the effects of glutenin allelic classes with the results published earlier on a similar dataset (Eagles et al., 2006), the

S. Zheng et al. / Journal of Cereal Science 50 (2009) 283290

289

results were nearly the same. They concluded that a large dataset (n 899) minimized the inuence of relationship among entries for the association between glutenin alleles and quality traits. In our smaller study, the 96 wheats were from eight hypothetical ancestral groups, with most entries showing a high level of admixture (Supplementary Fig. S1). This study demonstrated the importance of accounting for multiple-level relatedness of individuals when studying the quality effects of the glutenin loci in collections of breeding lines and cultivars. Our results conrmed that the HMW-GS, LMW-GS, and 1RS translocations play important roles in determining dough-mixing properties. Among them, 1RS translocations, and the Glu-B3 and Glu-D1 loci were of major importance in this set of genotypes, as discussed in detail below. 4.1. Glu-A1 Variation at Glu-A1 played a minor role in accounting for variation in dough-mixing properties. In this population, most entries carried Glu-A1a, b, or both, and those alleles had equivalent effects. This is in agreement with the previous studies on materials from Australia (Bekes et al., 2001; Eagles et al., 2002; Ma et al., 2005) and France (Branlard et al., 2001). However, Moonen et al. (1983) found that Glu-A1b was associated with higher loaf volume and SDS sedimentation volumes than Glu-A1a in wheat grown in the Netherlands. There was some indication that Glu-A1c, the null allele, was inferior in dough strength compared to Glu-A1a and Glu-A1b as shown by lower MPW and MRW. However, only three of the 96 entries in our study carried Glu-A1c. 4.2. Glu-B1 To the best of our knowledge, the quality effects of Glu-B1w have not been previously reported. In our study, the seven entries with Glu-B1w also carried Glu-D1d. To minimize the inuence of Glu-D1d on the results for Glu-B1w, we compared Glu-B1w to other Glu-B1 alleles within the set of 79 lines carrying Glu-D1d. The results showed that Glu-B1w was associated with higher MPV, and lower MRW, MRS, and MPT than Glu-B1b and c at all the three locations (Supplementary Table S2). Compared to Glu-B1e, Glu-B1w was associated with higher MPW, MRW, MPT, and MRS at all the three locations (Supplementary Table S2). However, none of these associations was signicant at a 0.05. In general, Glu-B1w could be considered slightly inferior to Glu-B1b and c, but better than GluB1e for dough-mixing properties. The frequency of Glu-B1w in US wheat is low, only 5% in 111 US wheat cultivars released after 1991 in the US Great Plains (Shan et al., 2007). Glu-B1e appeared to be an inferior allele compared to other alleles at the Glu-B1 locus for Mixograph properties, especially for MPT (Tables 24). This is in agreement with the score of Payne et al. (1987) and previous reports for bread wheat (Cane et al., 2008; Eagles et al., 2004; Ma et al., 2005) and durum wheat (Brites and Carrillo, 2001; Martinez et al., 2005). 4.3. Glu-D1 We conrmed the benecial effects of Glu-D1d over Glu-D1a (Tables 24). This agreed with the gluten-quality score (Payne et al., 1987) and many other studies (Bekes et al., 2001; Eagles et al., 2002; He et al., 2005; Huang et al., 2006). Selection for improved bread-making quality in breeding programs might explain why the ratio between Glu-D1d and Glu-D1a changed from 3:2 in US hard winter wheat cultivars released between 1970 and 1990 compared to 4:1 in those released from 1991 to 1996 (Graybosch, 1992; Shan et al., 2007).

4.4. Glu-A3 Glu-A3e, the null allele, was a favorable allele for bread-making quality because there was a trend for Glu-A3e to be associated with higher MRS, MRW, and MPT than other Glu-A3 alleles at all the three locations, though not all differences reached a signicance level of P < 0.05 (Tables 24). However, this differs from a previous report in which Glu-A3e was associated with a lower extensibility and Rmax than Glu-A3d, b, and c (Cane et al., 2008; Eagles et al., 2002). 4.5. Glu-B3 Our study conrmed the important role of the Glu-B3 locus for dough-mixing properties. Glu-B3c was associated with low Mixograph properties at all the locations. However, Glu-B3c was present in only ve of the 96 entries (Wendy, SD97538, Cougar, Rowdy, and Postrock) and four of these entries carried 1BL.1RS. To minimize the inuence of 1BL.1RS, we compared lines with Glu-B3c to lines with other alleles at the Glu-B3 locus within a subgroup of 13 lines having 1BL.1RS. We observed that Glu-B3c was consistently associated with low values of Mixograph properties in all the three environments; however, the t-tests were not all signicant at a 0.05 due to the limited number of entries (Supplementary Table S3). Inferior quality effects of Glu-B3c are in agreement with many previous studies (Bekes et al., 2001; Branlard et al., 2001; Cane et al., 2008; Eagles et al., 2002; Ma et al., 2005). 4.6. Glu-D3 The ve alleles at the Glu-D3 locus produced similar values for Mixograph properties in our study. The only signicant differences observed were for higher MPT for Glu-D3b compared to Glu-D3d at Akron, and more negative MRS for Glu-D3d compared to Glu-D3b and c at Dailey. Other studies have shown that the Glu-D3 alleles had similar Rmax and extensibility in a large collection of wheats (Branlard et al., 2001; Cane et al., 2008; Eagles et al., 2002). These results indicate that Glu-D3 would have the lowest priority among glutenin loci in which to invest efforts to improve wheat quality because of its minor role in determining quality. Lack of independence between the six glutenin loci has been observed in other studies (Eagles et al., 2002; He et al., 2005). However, due to the low number of entries in our study, we were not comfortable in inferring any interaction effects, although it is likely there will be signicant interactions among glutenin loci and among glutenin loci and other non-glutenin loci as previously reported (Eagles et al., 2002; Huang et al., 2006). 4.7. 1RS translocation In this study, 1BL.1RS was clearly the more detrimental form of the 1RS translocation in all the three locations. Wheat with 1AL.1RS was comparable to non-1RS wheat in most cases, but was inferior for three Mixograph properties at Fort Collins and one Mixograph property in Akron (Tables 24). The quality defects associated with the 1RS translocation may be the consequence of a change in the protein composition of the recipient wheats because incoming rye genes, such as secalin genes, are detrimental to wheat end-use quality. In addition, 1RS replaces the short arm of at least one wheat chromosome pair, leading to a permanent loss of some potentially important wheat genes, such as those encoding the LMW glutenins (Graybosch, 2001; Kumlay et al., 2003). In conclusion, we have completed one of the few association analysis studies in wheat that has taken multiple-level relatedness into account. Controlling for population structure did indeed

290

S. Zheng et al. / Journal of Cereal Science 50 (2009) 283290 1A vs. 1AL.1RS genotypes derived from winter wheat Nekota. Crop Science 39, 649654. Graybosch, R.A., 1992. High-molecular-weight glutenin subunit composition of cultivars, germplasm, and parents of United States red winter wheat. Crop Science 32, 11511155. Graybosch, R.A., 2001. Uneasy unions: quality effects of rye chromatin transfers to wheat. Journal of Cereal Science 33, 316. Graybosch, R.A., Peterson, C.J., Hansen, L.E., Worrall, D., Shelton, D.R., Lukaszewski, A., 1993. Comparative our quality and protein-characteristics of 1BL.1RS and 1AL.1RS wheatrye translocation lines. Journal of Cereal Science 17, 95106. Hardy, O.J., Vekemans, X., 2002. SPAGEDi: a versatile computer program to analyse spatial genetic structure at the individual or population levels. Molecular Ecology Notes 2, 618620. He, Z.H., Liu, L., Xia, X.C., Liu, J.J., Pena, R.J., 2005. Composition of HMW and LMW glutenin subunits and their effects on dough properties, pan bread, and noodle quality of Chinese bread wheats. Cereal Chemistry 82, 345350. Huang, X.Q., Cloutier, S., Lycar, L., Radovanovic, N., Humphreys, D.G., Noll, J.S., Somers, D.J., Brown, P.D., 2006. Molecular detection of QTLs for agronomic and quality traits in a doubled haploid population derived from two Canadian wheats (Triticum aestivum L.). Theoretical and Applied Genetics 113, 753766. Huelsenbeck, J.P., Andolfatto, P., 2007. Inference of population structure under a Dirichlet process model. Genetics 175, 17871802. Knowler, W.C., Williams, R.C., Pettitt, D.J., Steinberg, A.G., 1988. Gm3; 5, 13, 14 and type 2 diabetes mellitus: an association in American Indians with genetic admixture. American Journal of Human Genetics 43, 520526. Kumlay, A.M., Baenziger, P.S., Gill, K.S., Shelton, D.R., Graybosch, R.A., Lukaszewski, A.J., Wesenberg, D.M., 2003. Understanding the effect of rye chromatin in bread wheat. Crop Science 43, 16431651. Kunerth, W.H., DAppolonia, B.L., 1985. Use of the mixograph and farinograph in wheat quality evaluation. In: Faridi, H. (Ed.), Rheology of Wheat Products. AACCI, St. Paul, MN, pp. 2749. Ma, W., Appels, R., Bekes, F., Larroque, O., Morell, M.K., Gale, K.R., 2005. Genetic characterization of dough rheological properties in a wheat doubled haploid population: additive genetic effects and epistatic interactions. Theoretical and Applied Genetics 111, 410422. Martinez, M.D., Ruiz, M., Carrillo, J.M., 2005. Effects of different prolamin alleles on durum wheat quality properties. Journal of Cereal Science 41, 123131. Moonen, J.H.E., Scheepstra, A., Graveland, A., 1983. The positive effects of the high molecular-weight subunits 3 10 and 2-star of glutenin on the breadmaking quality of wheat cultivars. Euphytica 32, 735742. Patterson, H.D., Williams, E.R., Hunter, E.A., 1978. Block designs for variety trials. Journal of Agricultural Science 90, 395400. Payne, P.I., Nightingale, M.A., Krattiger, A.F., Holt, L.M., 1987. The relationship between HMW glutenin subunit composition and the breadmaking quality of British-grown wheat-varieties. Journal of the Science of Food and Agriculture 40, 5165. Pritchard, J.K., Stephens, M., Donnelly, P., 2000. Inference of population structure using multilocus genotype data. Genetics 155, 945959. Ravel, C., Praud, S., Murigneux, A., Linossier, L., Dardevet, M., Balfourier, F., Dufour, P., Brunel, D., Charmet, G., 2006. Identication of Glu-B1-1 as a candidate gene for the quantity of high-molecular-weight glutenin in bread wheat (Triticum aestivum L.) by means of an association study. Theoretical and Applied Genetics 112, 738743. Rosenberg, N.A., 2004. DISTRUCT: a program for the graphical display of population structure. Molecular Ecology Notes 4, 137138. Ritland, K., 1996. Marker-based method for inferences about quantitative inheritance in natural populations. Evolution 50, 10621073. Saal, B., Wricke, G., 1999. Development of simple sequence repeat markers in rye (Secale cereale L.). Genome 42, 964972. Shan, X., Clayshulte, S.R., Haley, S.D., Byrne, P.F., 2007. Variation for glutenin and waxy alleles in the US hard winter wheat germplasm. Journal of Cereal Science 45, 199208. Singh, M., Khatkar, B.S., 2005. Structural and functional properties of wheat storage proteins: a review. Journal of Food Science and Technology 42, 455471. Somers, D.J., Isaac, P., Edwards, K., 2004. A high-density microsatellite consensus map for bread wheat (Triticum aestivum L.). Theoretical and Applied Genetics 109, 11051114. Yu, J.M., Buckler, E.S., 2006. Genetic association mapping and genome organization of maize. Current Opinion in Biotechnology 17, 155160. Yu, J.M., Pressoir, G., Briggs, W.H., Bi, I.V., Yamasaki, M., Doebley, J.F., McMullen, M.D., Gaut, B.S., Nielsen, D.M., Holland, J.B., Kresovich, S., Buckler, E.S., 2006. A unied mixed-model method for association mapping that accounts for multiple levels of relatedness.