International Journal of Food Microbiology 124 (2008) 126-134

Contents lists available at ScienceDirect

International Journal of Food Microbiology

j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / i j f o o d m i c r o

Analysis of bacterial community during the fermentation of pulque, a traditional Mexican alcoholic beverage, using a polyphasic approach
Adelfo Escalante a,, Martha Giles-Gmez b, Georgina Hernndez a, Mara Soledad Crdova-Aguilar a, Agustn Lpez-Mungua a, Guillermo Gosset a, Francisco Bolvar a
a Departamento de Ingeniera Celular y Biocatlisis, Instituto de Biotecnologa, Universidad Nacional Autnoma de Mxico (UNAM), Av. Universidad 2001, Col Chamilpa, 62210, Cuernavaca, Morelos, Mxico b Departamento de Biologa, Facultad de Qumica, UNAM, Ciudad Universitaria, Coyoacn, 04510, Mxico D.F. Mxico

A R T I C L E

I N F O

A B S T R A C T
In this study, the characterization of the bacterial community present during the fermentation of pulque, a traditional Mexican alcoholic beverage from maguey (Agave), was determined for the rst time by a polyphasic approach in which both culture and non-culture dependent methods were utilized. The work included the isolation of lactic acid bacteria (LAB), aerobic mesophiles, and 16S rDNA clone libraries from total DNA extracted from the maguey sap (aguamiel) used as substrate, after inoculation with a sample of previously produced pulque and followed by 6-h fermentation. Microbiological diversity results were correlated with fermentation process parameters such as sucrose, glucose, fructose and fermentation product concentrations. In addition, medium rheological behavior analysis and scanning electron microscopy in aguamiel and during pulque fermentation were also performed. Our results showed that both culture and non-culture dependent approaches allowed the detection of several new and previously reported species within the -, -Proteobacteria and Firmicutes. Bacteria diversity in aguamiel was composed by the heterofermentative Leuconostoc citreum, L. mesenteroides, L. kimchi, the -Proteobacteria Erwinia rhapontici, Enterobacter spp. and Acinetobacter radioresistens. Inoculation with previously fermented pulque incorporated to the system microbiota, homofermentative lactobacilli related to Lactobacillus acidophilus, several Proteobacteria such as Zymomonas mobilis and Acetobacter malorum, other -Proteobacteria and an important amount of yeasts, creating a starting metabolic diversity composed by homofermentative and heterofermentative LAB, acetic and ethanol producing microorganisms. At the end of the fermentation process, the bacterial diversity was mainly composed by the homofermentative Lactobacillus acidophilus, the heterofermentative L. mesenteroides, Lactococcus lactis subsp. lactis and the -Proteobacteria A. malorum. After a 6-h fermentation, 83.27% of total sugars detected after inoculation were consumed (228.4 mM hexose equivalents) and a carbon (C) recovery of 66.18% in fermentation products was estimated. They were produced 284.4 mM C as ethanol, 71.5 mM C as acetic acid and 19 mM C as lactic acid, demonstrating the presence of homo- and heterofermentative, acetic and alcoholic metabolisms in the nal product. It was also found, after hydrolysis, that the exopolysaccharide produced during the fermentation was mainly composed by fructose residues, probably inulin or levan. 2008 Elsevier B.V. All rights reserved.

Article history: Received 2 June 2007 Received in revised form 14 November 2007 Accepted 3 March 2008 Keywords: Pulque Traditional alcoholic fermentation Microbial diversity Lactic acid bacteria

1. Introduction Pulque is a non-distilled traditional alcoholic beverage produced by the fermentation of the sap known as aguamiel (AM), which is extracted from several species of maguey such as Agave atrovensis and A. americana. This beverage is currently produced and consumed mainly in the central states of Mexico. For its production, freshly collected AM is transported to large open barrels where fermentation takes place. The process is accelerated by the addition of a portion of

Corresponding author. Tel./fax: +52 55+5622 7648. E-mail address: adelfo@ibt.unam.mx (A. Escalante). 0168-1605/$ see front matter 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.ijfoodmicro.2008.03.003

previously produced pulque and the fermentation time varies from few hours to overnight, depending on whether the sap is collected at daybreak or dusk. The development of viscosity due to exopolysaccharide (EPS) synthesis and the alcohol content are the main parameters used to determine the degree of fermentation. The process is static and performed under non-aseptic conditions; therefore, the populations of microorganisms involved in the fermentation are those naturally occurring during AM accumulation in the maguey and those incorporated during collection, transport, inoculation and manipulation (Snchez-Marroqun and Hope, 1953; Godoy, 1987; Garca-Garibay and Lpez-Mungua, 1993; Escalante et al., 2004). Several studies have been performed to characterize the microbial diversity of AM and pulque samples using traditional culture dependent

A. Escalante et al. / International Journal of Food Microbiology 124 (2008) 126-134

127

and identication methods. Important industrial microorganisms isolated from various pulque samples, comprise several yeasts, such as the ethanol producing Saccharomyces cerevisiae and species of Kluyveromyces producers of inulinase (Snchez-Marroqun and Hope, 1953; Estrada-Godina et al., 2001; Cruz-Guerrero et al., 2006). Isolated bacterial species include the alcohol producing Zymomonas mobilis (Snchez-Marroqun and Hope, 1953; Godoy, 1987) and the dextran producing lactic acid bacteria (LAB) Leuconostoc mesenteroides (Chellapandian et al., 1988). Among them, S. cerevisiae, L. mesenteroides and Z. mobilis have been proposed as essential microorganisms for the pulque fermentation process (Snchez-Marroqun and Hope, 1953; Godoy, 1987; Favela, 1993; Garca-Garibay and Lpez-Mungua, 1993). We recently reported a study of the bacterial diversity present in three samples of pulque from different geographical origins, based on 16S rDNA sequence analysis amplied from total bacterial DNA. Our results indicated the presence of a microbial community composed of previously reported microorganisms such as L. mesenteroides and Z. mobilis, but several Proteobacteria, homo- and heterofermentative LAB were detected and reported for the rst time in pulque samples. This study revealed the presence of common microbial species in this traditional fermented beverage, independently of the sample origin (Escalante et al., 2004). However, to date, no information regarding the microbial dynamics during the fermentation process of pulque is available. Microbiological characterization of traditional fermented foods and beverages produced domestically and at semi-industrial level represent the rst stage in the process for industrialization and scaleup of this type of products. Results from this initial characterization should allow the identication of microorganisms that are essential in these processes (Blandino et al., 2003). This strategy is not only restricted to local fermented foods and beverages; several studies have been performed to characterize the microbial communities involved in fermentations in order to improve important alcoholic distilled beverages, such as Scotch malt whisky (Simpson et al., 2001; van Beek and Priest, 2002). In Mexico, interest in alcoholic beverages obtained from Agave has increased in recent years due to their acceptance both in local and international markets. Among these beverages, tequila and in a minor proportion mezcal are of particular importance due to their economical impact and the complex microbial process involved in their production; therefore, it is possible to foresee that some of these traditional fermentation processes could become industrialized. In the present study, we report the characterization of the bacterial community present in AM, and its dynamics during the fermentation of pulque using cultured and non-cultured dependent methods. The microbiological study was correlated with the sucrose, glucose, fructose and fermentation products concentration, as well as medium rheological behavior and scanning electron microscopy analyses in AM and during pulque fermentation. 2. Materials and methods 2.1. Pulque sampling and fermentation analysis Samples of freshly collected AM and overnight fermented pulque to be used as inoculum were obtained from Huitzilac, a town in Morelos State, Mexico (altitude of 2550 m in a cold weather mountainous region), placed in sterile plastic recipients and transported to the laboratory. One fermentation was carried out in a 4 L sterile glass container and was started by addition of previously fermented pulque and fresh AM (2:3 v/v) (as recommended by local supplier). Temperature and pH were monitored in AM, after inoculation (T0), 3 (T3) and 6 (T6) h of fermentation. Sucrose, glucose, fructose, ethanol, lactic- and acetic acids concentrations were determined in all samples using a Waters HPLC system, equipped with an Aminex column for fermentation

analysis. One ml of each sample was centrifuged at 10,000 g/5 min; supernatants were then ltered through a 0.45 m membrane and a 20 l aliquot was injected to the column. 2.2. Microbiological determinations Ten ml of each sample were diluted in 90 ml of sterile 0.1% Peptone (DIFCO) water and serial dilutions (101 to 106) were made. Growing aerobic mesophilic bacteria and LAB were determined by plating 1 ml of each serial dilution on Nutritive-agar (DIFCO) and on APT-agar (DIFCO), respectively. In order to obtain a general view of the interactions of the different microbial groups involved in the process, we decided to include the yeasts counts in our analysis. One ml of each serial dilution was plated by duplicate on Malt extract-agar (DIFCO) supplemented with 10% w/v tartaric acid as antibacterial agent. Microbial determinations were performed by triplicate. All plates were incubated at 30 C and monitored for colony appearance from 1 to 3 days. Representative plates of Nutritive-agar and APT-agar were used to isolate and purify an average of 100 CFU, from AM and pulque samples, respectively. Gram staining and light microscopy observations were used as criteria to determine the purity of each isolate. 2.3. Amplied rDNA restriction analysis (ARDRA) of isolated cultures and microbial identication ARDRA was performed to Gram-positive and Gram-negative isolated bacteria. Total DNA was released from each cultured isolate by a previously used method that ensures an effective lysis of diverse bacterial groups (Smit et al., 2001). One l of each lysate was used as template for 16S rDNA fragment amplication as previously reported (Escalante et al., 2004) using the primers Eu530F (5'-TGACTGGAGTGCCAGCAGCCGCGG-3') and Eu1449R (5'-TGACTGACTGAGGCTACCTTGTTACGACTT-3'), which target to Eubacteria. These primers had already been applied to determine the bacterial diversity present in soil (Borneman et al., 1996) and pulque samples (Escalante et al., 2004). Amplied 16S rDNA fragments were digested with Hae III, for 2 h at 37 C. Restriction proles were visualized using 2.5% agarose gel electrophoresis and analyzed with the molecular weight analysis mode of the Eagle Eye Image Capture and Analysis Software (Stratagene). Cultured isolates were grouped based on their ARDRA patterns after identication of unique common restriction proles. 16S rDNA fragments corresponding to unique restriction types were sequenced by the Taq FS Dye Terminator Cycle Fluorescence-Based Sequencing method, with a Perkin Elmer/Applied Biosystems Model 37718 sequencer. The resulting DNA sequences were submitted to the nonredundant nucleotide database at GeneBank using the BLAST program to determine their identity. 2.4. Total bacterial DNA extraction from AM and pulque samples, PCR amplication and cloning of 16S rDNAs Total bacterial DNA was extracted from AM and pulque samples using the microbial genomic DNA isolation kit of MoBio (12224-250). This kit was selected due to its lytic efciency reported for several Gram-positive and Gram-negative bacteria and EPS producing species (www.mobio.com). Five independent extractions were performed for each sample and pooled. One l of each pooled sample was used as template for 16S rDNA fragment amplication using the Eu530F/ Eu1449R primer set. Three independent PCR reactions were performed and pooled for AM and pulque samples. A 16S rDNA clone library was created for each sample by cloning 4 l of amplied 16S rDNA into TOPO Blunt vector (Invitrogen) and used to transform E. coli TOP10 electrocompetent cells (Invitrogen). Positive clones were screened by digestion of 2 l alkaline plasmid minipreps from each 16S rDNA clone library with EcoRI and visualized using 1% agarose gel electrophoresis in order to identify those clones with the 16S rDNA insert. An average of 100 positive TOPO Blunt 16S rDNA clones from

128

A. Escalante et al. / International Journal of Food Microbiology 124 (2008) 126-134

each AM, T0, T3 and T6 samples were selected, ARDRA was performed to the positive ones and data analyzed as described in Section 2.3. 2.5. Non-culture dependent diversity analysis in 16S rDNA clone libraries from AM and pulque samples Unique ARDRA types identied from AM and pulque 16S rDNA libraries were sequenced and identied as described in Section 2.3. The EstimateS program (version 8.0.0; Colwell, 2006) was used to calculate the nonparametric richness estimator Chao 1 (Hughes et al., 2001; Buckley et al., 2006), which species the extent of diversity richness based on 100 clones sampled. In order to determine whether the size of the libraries was statistically different from each other or whether one or more was a subset of others, a LIBSHUFF analysis was performed (Schloss et al., 2004). Multiple alignment of AM and pulque 16S rDNA clones and reference 16S rDNA sequences retrieved from GenBank database was performed using the Clustal W program. Distance matrix calculation of nucleotide substitution rates and a phylogenetic tree was constructed with the Jukes and Cantor algorithm and the Neighbor-Joining (NJ) method, respectively, using the Phylip (v 1.3b) program. Bootstrap methods were utilized to provide condence estimated for tree topology in the NJ method (1000 replicates). 2.6. Scanning electron microscopy (SEM) analysis One ml of AM and pulque samples were xed with 1 volume of 4% v/v glutaraldehyde and 0.01% w/v OsO4. Fixed samples were successively dehydrated in 30%, 50%, 70%, 100% ethanol, super critical CO2 dried in a Sandri-780 apparatus (Tousimis) and coated with gold particles in a JFC-110 ion sputter device (JEOL). Samples were observed in a JEOL scanning electron microscope model JSM 5410LV. 2.7. Hydrolysis of EPS produced in pulque fermentation and rheological analysis Acid and enzymatic hydrolysis were performed to determine EPS components produced at the end of the fermentation. For the acid hydrolysis, 1 ml aliquot of pulque sample T6 was centrifuged for 10 min at 12,000 g and the supernatant sterilized under UV exposition during 4 h. The sample was then hydrolyzed by addition of 0.3 N HCl, incubated for 1 h at 105 C and neutralized with 0.1 N NaOH. Reducing sugars were determined by the DNS method before and after hydrolysis. For the enzymatic hydrolysis, to a 1 ml aliquot of UV sterilized sample adjusted to pH 4.6 was added 1 ml of the enzyme Fructozyme (1% v/v nal concentration). Reaction was heated to 60 C and incubated 2 h. Finally, the resulting concentrations of reducing sugars were measured (DNS method) and analyzed by HPLC with a Waters HPLC system. Rheological determinations were performed using a rotational AR 1000 TA Instruments rheometer (New Castle, DE, USA) tted with a stainless steel cone/plate geometry (60 mm of diameter, 1 cone angle). AM and pulque samples were transferred to the peltier plate, removing the excess material with lter paper. All experiments were conducted at room temperature. Data in the laminar ow regime (Re b 10) was tted to the power law model (t = Kgn). From this model, polymer consistency index, K (Pa sn) and ow index, n (dimensionless) were determined. 3. Results 3.1. Culture-dependent analysis of bacterial diversity in AM and during pulque fermentation The log number of CFU/ml of aerobic mesophiles and LAB present in AM was 1.3 107 and 3.2 108 CFU/ml, respectively. The log number

for both microbial groups after inoculation was 1.2 107 and 1.5 108 CFU/ml, respectively, and remained relatively constant until the end of the process (3.5 107 and 1.5 108 CFU/ml at the end of the fermentation, respectively). Yeasts concentration detected in AM was 3.1 10 4 CFU/ml. Inoculation with previous fermented pulque increased the log number of yeasts to 8.8 106 CFU/ml and increased further to reach 1.4 107 CFU/ml after 3-h culture to remain constant for the rest of the fermentation process. An average of 100 colonies of aerobic mesophiles and LAB from representative plates of AM and each pulque sample were isolated. DNA from each colony from the different samples was isolated and subjected to ARDRA by digesting the 16S rDNA fragment with Hae III. Using restriction proles generated with this enzyme, microbial

Table 1 Identity of cultivated and non-cultivated microorganisms by 16S rDNA analysis from AM and during pulque fermentation Lineagea Identityb Isolate and/or 16S rDNAd clone detected by Cultivated 16S rDNA library analysis -Proteobacteria/ Rhodospirillales -Proteobacteria/ Sphingomonadales -Proteobacteria/ Enterobacteriales Acetobacterium malorum Acetobacter orientales Zymomonas mobilis subsp. pomaceae Citrobacter spp. Enterobacter agglomerans Enterobacter spp. Erwinia rhapontici Kluyvera ascorbata Kluyvera cochleae Providencia spp. Serratia grimensii Acinetobacter radioresistens T3332 T638 T321 AM116 T0100 AM119 AM230 T0330 AM143 T6126 AM14

T3344

-Proteobacteria/ Psedomonadales -Proteobacteria/ Sterotrophomonas spp. Xhantomonadales Flavobacterias Chryseobacterium spp. Firmicutes/Bacillales Bacillus sp. Bacillus licheniformis Firmicutes/ Lactobacillus spp. AC07c Lactobacillales Lactobacillus spp. ASF360c Lactobacillus spp. Y10c Lactobacillus acidophilus Lactobacillus hilgardii Lactobacillus paracollinoides Lactobacillus sanfranciscencis Lactococcus spp. Lactococcus lactis Lactococcus lactis subsp. cremoris Leuconostoc kimchi Leuconostoc citreum Leuconostoc gasicomitatum Leuconostoc mesenteroides Leuconostoc pseudomesenteroides Pediococcus urinaeequi Streptococcus devriesei Uncultured bacterial Uncultured bacterial clone clones BANW647c Uncultured bacterial clone P-2988-9F2c Uncultured bacterial clone rRNA342c Eukaryotic Saccharomyces cerevisiae
a

AM83 T040 T6342 T621 T023 T33 T01 T3239 T3160 T0225 T3150 T0218 AM96

AM123 T3239 T011 AM146 AM140 AM212 T3245 T325

AM235 AM132 T0121 AM118 T358 AM446 AM312 T3135 T365 T0323

Based on information retrieved from the NCBI taxonomy homepage. b16S rDNA or 16S rRNA sequences of organisms which showed the highest percent of identity in the output result from analysis in the non-redundant nucleotide database from NCBI using the BLAST program. cStrain or clone designation of closest 16S rDNA sequences in database are indicated only for Lactobacuillus spp. and uncultured bacterial clones. Previously reported microorganisms or 16S rDNA clone sequences for AM and pulque are highlighted in bold letters (Snchez-Marroqun and Hope, 1953; Godoy, 1987; Garca-Garibay and Lpez Mungua, 1993; Escalante et al., 2004). dAguamiel (AM), after inoculation (T0), 3 (T3) and 6 (T6) h indicate the sample source of isolates or 16S rDNA clones detected.

A. Escalante et al. / International Journal of Food Microbiology 124 (2008) 126-134

129

isolates with similar patterns were grouped and the 16S rDNA of one member of each group was sequenced and identied. From total isolates, 21 different species and 14 different genera were found, 51.44% were identied as Firmicutes (Bacillales and Lactobacillales) and 48.55% distributed among -Proteobacteria, -Proteobacteria and Flavobacteria. Identities of isolated bacteria in AM and pulque samples are presented in Table 1 and its distribution during the fermentation is given in Fig. 1. 3.2. Non-culture dependent diversity analysis in 16S rDNA clone libraries from AM and pulque samples Total microbial DNA was extracted from AM and pulque samples. Amplication of the 16S rDNA fragment was performed independently from each DNA extraction and cloned into the TOPOBlunt system. An average of 100 clones from each 16S rDNA library was then digested with Hae III and unique restriction types from each library were detected and selected for sequence analysis. Retrieved results from GenBank analysis revealed that 79.1% of analyzed clones were identied as cultivated Firmicutes and 8.2% as members of the cultivated -Proteobacteria, -Proteobacteria and Enterobacteriales. These included some of the identied microorganisms by the culturedependent approach (Table 1). Interestingly, 7.2% of the analyzed sequences were identied as uncultured bacterial clones. Finally, 5.4%

Fig. 2. Rank abundance curves of unique 16S rDNA clones from AM and pulque samples. A. AM, B. T0, C. T3, D. T6. Nick names for identied isolates are the same as in Fig. 1. Additional abbreviations for AM: Lki, Leuconostoc kimchi; Ara, Acinetobacter radioresistens; Unb, Uncultured bacterial clones; Sde, Streptocccus devriesei. For T0: Zmo, Zymomonas mobilis; Lhi, Lactobacillus hilgardii, Lga, Leuconostoc gasicomitatum. For T3: Lac, Lactobacillus acidophilus; Lsa, Lactobacillus sanfranciscencis. For T6: Aor, Acetobacter orientalis. Sequences identied as Saccharomyces cerevisiae (abbreviated Sce) are also included.

Fig. 1. Rank abundance curves of both total aerobic mesophiles and LAB isolated from AM and pulque samples. A. AM, B. T0, C. T3, D. T6. Abbreviations for identied isolates in AM are: Erh, Erwinia rhapontici; Lci, Leuconostoc citreum; Lme, L. mesenteroides; Ent, Enterobacter spp.; Cit, Citrobacter sp.; Lla, Lactococcus lactis; Sgr, Serratia grimensii; Kch, Kluyvera cochleae. For T0: Eag, Enterobacter aglomerans; Kas, Klyuvera ascorbata; Ama, Acetobacter malorum; Ste, Sterotrophomonas spp.; Lps, Leuconostoc pseudomesenteroides; Bli, Bacillus licheniformis; Llc, Lactococcus lactis spp. cremoris. For T3: Lbs, Lactobacillus spp.; Lcs, Lactococcus spp., Pur, Pediococcus urinaeequi. For T6: Pro, Providencia spp.; Chr, Chryseobacterium sp; Bac, Bacillus spp.

of analyzed clones were identied as S. cerevisiae, the most abundant yeast present in pulque fermentation. As it was observed for the culture-dependent approach, several clones were identied as microorganisms non-previously reported in pulque (Table 1). LIBSHUFF analysis was used to compare the phylogenetic composition of 16S rDNA clone libraries of all samples in order to nd out whether the average of the analyzed clones was statistically different. Results revealed that the AM library was statistically different to that of sample T0, whereas samples T3 and T6 were not statistically different as a consequence of the presence of common species in T0 and T3, and T3 and T6 libraries, that disappeared, were maintained without relative changes or were enriched during the fermentation (Fig. 2). Calculation of the Chao 1 estimator revealed a low difference with respect to the expected number of unique 16S rDNA clones for AM, T0 and T3 samples. These results indicated that the number of analyzed clones in these libraries adequately described the extent of diversity richness present, except for T6 library, in which the estimator showed a higher difference with respect to the expected diversity. Phylogenetic analysis of 16S rDNA sequences demonstrated a high similarity to isolates or environmental clones previously obtained from diverse sources and deposited in GenBank database (Fig. 3). However, other clones showed less than 95% relatedness to NCBI database sequences, which may indicate the presence of new species.

130

A. Escalante et al. / International Journal of Food Microbiology 124 (2008) 126-134

Fig. 3. Phylogenetic tree of unique 16S rDNA sequences detected in AM and pulque samples and sequences of closest neighbor 16S rDNA from identied bacteria or environmental clones in the NCBI database. GenBank accession number of closest 16S rDNA sequences in database are indicated. The 16S rDNA sequence of Sulfolobus acidocaldarius served as outgroup. The percentage of 1000 bootstrap samplings supporting each topological element in the neighbor-joining analysis is indicated. No values are given for groups with bootstrap values less than 80%. Identities of unique 16S rDNA sequences presented in the tree are shown in Table 1. Identity percentage with closest reference to 16S rDNA clones in the database is indicated in parenthesis.

3.3. pH, temperature, sugar content and fermentation products during AM and pulque fermentation AM had an initial pH of 6.0, which after inoculation with previous fermented pulque (pH = 4.0) decreased to 4.5 as a consequence of the addition of fermentation products. The nal pH of pulque after 6-h fermentation was 4.3. The traditional process took place at room temperature: The AM sample had an in situ temperature of 17 C, which after inoculation increased to 20 C and then gradually up to 22 C after 3 h and 24 C after 6 h. This was the result of important microbial activity during fermentation, in spite of the relatively constant total bacterial counts observed all through the process. To asses the overall metabolic activity of the microbial community in AM and pulque samples, the concentrations of sucrose, glucose, fructose and fermentation products were determined (Table 2). Total sugars consumed after 6 h of culture were 228.4 mM hexose equivalents; this is equivalent to 1370.5 mM C, which was found as CO2. It was assumed that in molar terms, the amount of CO2 produced is equal to the amount of acetic acid and ethanol. When 176.5 mM C corresponding to 5.3 g/L of fructose in EPS were also included (see Section 3.4), then

66.2% of total C in products were recovered. The concentration of ethanol reached a nal value of 142.17 mM; meanwhile nal concentrations of acetic and lactic acids were 35.73 mM and 6.32 mM, respectively (Table 2).

Table 2 Sugar content, fermentation product concentrations and physicochemical parameters in aguamiel (AM) and pulque samples Sample Sugar content (mM hexose equivalents) Product concentrations (mM C) Acetic acid 30.48 36.58 94.84 108.05 Viscosity

Sucrose Glucose Fructose Ethanol Lactic acid AM T0 T3 T6 410.16 334.79 210.69 180.19 83.53 30.86 28.19 3.82 533.35 188.05 71.21 12.43 4.34 655.10 749.31 939.46 46.95 87.25 95.91 106.24

n ()a K (Pa sn) 1.060 1.030 0.842 0.754 0.0012 0.0017 0.0071 0.0142

Changes during the fermentation process: after inoculation (T0), 3 (T3) and 6 (T6) h of fermentation. Values are the mean of triplicate determinations. aDimensionless.

A. Escalante et al. / International Journal of Food Microbiology 124 (2008) 126-134

131

Fig. 4. Scanning electron microscopy micrographs of AM and pulque samples. A. AM; B. T0; C. T3; D. T6.

132

A. Escalante et al. / International Journal of Food Microbiology 124 (2008) 126-134

3.4. Analysis of EPS produced in pulque fermentation and rheology EPS production is an important distinctive characteristic of fermented pulque. HPLC analysis of sugars generated from EPS by acid hydrolysis revealed no changes in glucose concentration while a signicant increase in fructose residues was observed. Calculations derived from HPLC and reducing sugars analysis indicated that 5.3 g/L of fructose were present in the polymeric structure. Based on these observations, the EPS produced after 6-h fermentation must be a polymer composed by fructose residues that could be either inulin or levan. Two rheological properties were determined, the ow index (n) and the consistence index (K). Results in Table 2 indicate that AM and pulque polymer behaved as non-Newtonians uids. As fermentation developed, n decreased indicating the change from a uid lightly dilatant (AM) to a pseudoplastic product after 6 h. Based on the evolution of K values, which increased from AM to T6, it is possible to propose the presence in AM of one type of EPS which is degraded during fermentation and the synthesis of another type of polymer with higher viscosity (probably inulin or levan). 3.5. Scanning electron microscopy of AM and pulque samples SEM observations of AM revealed the presence of bacteria with several morphologies including short bold and large thin rods, cocci organized in pairs and short chains with smooth surface, cocci in large chains with rough surface, and few yeasts, probably Saccahromyces cerevisiae (Fig. 4A). After inoculation (Fig. 4B), rough cocci in large chains, probably EPS producing Leuconcostoc species, were observed as the most abundant cell morphology followed by budding yeasts. Samples corresponding to T3 (Fig. 4C) and T6 (Fig. 4D) showed cocci in short chains and rods, an important increase in yeast counts, as well as the enrichment of cocci in large chains, which was dened as the most abundant bacterial morphology at the end of fermentation. 4. Discussion The bacterial diversity present in AM and during the fermentation of pulque, a Mexican traditional alcoholic beverage, was studied for the rst time to determine microbial community changes involved in the process. In addition, we determined several physicochemical parameters during the fermentation in order to obtain a more detailed description of the process. Bacterial diversity was explored using both culture and non-culture dependent methods in order to minimize biases currently associated with either of these methodologies. Although both approaches allowed the detection of common species among the -, -Proteobacteria and Firmicutes, it was observed that each method revealed a poor diversity for some species, highlighting the limitations of using only one approach to study bacterial diversity. Rank abundance curves were elaborated to integrate the results obtained with both approaches. Besides, important microbial diversity changes were determined during the fermentation from AM to the nal fermented product. Identities obtained from 16S rDNA clone sequence analysis and strain isolation revealed that species of Leuconostoc mesenteroides, L. kimchi and L. citreum were the most abundant LAB in AM (36.6% of total cultured organisms and 51.9% of total clones analyzed for AM). Among them, L. mesenteroides had been previously reported in pulque (Snchez-Marroqun and Hope, 1953; Godoy, 1987; Garca-Garibay and Lpez Mungua, 1993), whereas L. kimchi, L. citreum and Lactococcus lactis (in minor proportions) were detected for the rst time in both AM and pulque. Furthermore, it was observed that the -Proteobacteria identied by culturing as Erwinia rhapontici and Enterobacter spp. (47.9% of total isolates) and 16S rDNA clones identied as Acinetobacter radioresistens (18.3% of total clones analyzed) were the second dominant group in sap (Figs. 1A and 2A). Members of the genus Leuconostoc are heterofermentative LAB occuring in pairs and chain aggregates; they are commonly found in

fruits and vegetables. They play important roles in the fermentation of dairy products, vegetables and meats given their ability to ferment a wide diversity of sugars as those present in pulque (sucrose, glucose and fructose), producing CO2, D() lactic acid, glucans and fructans from sucrose (Carr et al., 2002). SEM of AM sample revealed the presence of abundant cocci organized in pairs and short chains with smooth surface and cocci in large chains, probably Leuconcostoc species. Although it has been reported that some -Proteobacteria detected in AM are naturally distributed in several environments including fresh water, soil, plant and vegetable surfaces, some of them are also considered opportunistic human pathogens (Ashraf et al., 1997; Janssen et al., 1997; Holt et al., 2000; Waleron et al., 2002). It is possible that some of the -Proteobacteria found in AM are members of the natural bacterial diversity of the sap that were incorporated from the environment during its accumulation in the maguey, whereas others could be incorporated during its extraction, handling and storage under non-aseptic conditions. Inoculation of AM with previously fermented pulque added new bacterial species to the system and an important content of yeasts; however, abundant clones in sap such as those identied as L. kimchi and A. radioresistens, were diluted and therefore detected in low proportions (5.6% of total clones), whereas culturable L. mesenteroides, L. citreum (33% of total isolates in T0) stayed relatively constant with respect to AM. After inoculation, the most abundant microorganisms detected by clone analysis were Lactobacillus spp. (75.3% of total clones), a LAB closely related to homofermentative L. acidophilus, and the -Proteobacteria Enterobacter agglomerans (21.2% of total isolates); both species were not detected in AM. Furthermore, -Proteobacteria, other -Proteobacteria and Gram-positive bacteria not detected in AM were observed after inoculation (Figs. 1B and 2B). The -Proteobacteria Zymomonas mobilis and Acetobacter malorum were found after inoculation. Z. mobilis, previously reported in pulque, has been traditionally considered in association with yeasts responsible for ethanol production (Snchez-Marroqun and Hope, 1953; Godoy, 1987, Favela, 1993; Garca-Garibay and Lpez-Munga, 1993). Several strains of Z. mobilis have been previously isolated from pulque samples and reported as high ethanol yield producers (Favela, 1993). Although in low proportions, A. malorum was also detected for the rst time in pulque. Species of this microorganism occur naturally in sugar rich environments such as alcoholic fermentation musts of sake, tequila (another fermented beverage from Agave), palm wine, grape wine, among others (Holt et al., 2000, Du Toit and Lambrechts, 2002). SEM after inoculation revealed the addition of yeasts to the system, while previously observed cocci in pairs and chains with smooth and rough surface remained relatively abundant. Important physicochemical changes were observed in T0, pH detected in AM changed from 6.0 to 4.5 as a consequence of the addition of fermented pulque, whereas total C in fermentation products increased from 116.6 mM to 1470.6 mM. At T0, starting bacterial diversity for pulque fermentation was composed by those species present in AM and those incorporated by inoculation; integrating a metabolic diversity of homofermentative and heterofermentative LAB, acetic and ethanol producing bacteria and yeasts. After 3-h fermentation important changes in bacterial diversity were observed. Lactobacillus spp. (49.5% of total clones), L. mesenteroides and Enterococcus agglomerans (61.8% of total isolates) became the dominant bacterial species. Other bacteria such as Z. mobilis, A. malorum and A. radioresistens were found relatively constant since inoculation; L. citreum was detected in minor proportion and some others disappeared (Figs. 1C and 2C). 16S rDNA clones identied for the rst time the homofermentative LAB L. acidophilus found in the fermentation as a minor group; also clones identied as Lactococcus lactis subsp. cremoris increased after inoculation. During the rst 3-h fermentation, yeasts increased

A. Escalante et al. / International Journal of Food Microbiology 124 (2008) 126-134

133

from 8.8 106 CFU/ml to 1.4 107 CFU/ml. Fields of this sample analyzed by SEM demonstrated abundant presence of yeasts, smooth cocci in pairs, rough cocci aggregated in chains and some rods. At the end of the fermentation process, the bacterial diversity observed by clone analysis was mainly composed of the homofermentative Lactobacillus acidophilus (88.1% of total clones), meanwhile major culture diversity consisted of the heterofermentative LAB L. mesentreoides (50%), L. lactis subsp. lactis (12.5%), a homofermentative Lactobacillus spp. (14.8%) and the -Proteobacteria A. malorum (13.6% of total isolates). Yeasts counts were maintained relatively constant with respect to T3. Analyzed elds by SEM showed abundant yeast and the same bacterial morphologies observed previously. As a consequence of the microbial activity, after 6-h fermentation 83.27% of total sugars detected after inoculation were consumed (228.4 mM hexose equivalents). It was estimated that 18% of C was incorporated into fermentation products in the following manner: 284.4 mM C as ethanol, 71.5 mM C as acetic acid and 19 mM C as lactic acid. Thus, the presence of homo- and heterofermentative, acetic and alcoholic metabolisms was demonstrated during the whole process. Traditionally, studies on the microbiology of pulque have reported that L. mesenteroides and Z. mobilis are the most important bacteria involved in the fermentation process. However, although Lactobacillus species have been previously detected as well, no detailed information about the identity and the role of this LAB group is available (Godoy, 1987; Escalante et al., 2004). In this work, it was shown that homofermentative lactobacilli species could probably play an important role during the fermentation process, given the high number of L. acidophilus identied clones and closely related Lactobacillus spp. in all pulque samples. Lactobacillus acidophilus has been reported to ferment glucose, fructose, sucrose and produce DL lactic acid (Carr et al., 2002). Other lactobacilli detected in minor proportions were L. hilgardii, Lactobacillus spp. ACO7 (related to L. hilgardii) and L. sanfranciscensis (heterofermentative LAB). They also ferment glucose and fructose but not sucrose, and produce DL lactic acid. It has been reported that some species are able to grow in environments with 20% ethanol (Carr et al., 2002.). Therefore, lactic acid bacteria resistance to high ethanol concentrations and their capacity to grow in acid environments (pHb 4.5) could explain their relative abundance during the fermentation process. It is possible that bacterial species that apparently disappeared during fermentation were not able to resist the acidication of the environment, increased ethanol concentration, possible antimicrobial effect of LAB such as H2O2 production, or other antibacterial activities. The liquid consistence of pulque avoids the formation of microenvironments or gradients observed in solid traditional fermentations that allow survival of large populations of Enterobacteriaceae despite the high acid production (Ampe et al., 1999). The use of universal primers Eu530F and Eu1449R for amplication of 16S rDNA fragments demonstrated the amplication of yeast S. cerevisiae and uncultured bacterial species ribosomal genes. Detection of eukaryotic genes using universal prokaryotic primers has been previously observed (Barns et al., 1994; Rheims et al., 1996; Escalante et al., 2001). These results could be explained as a bias associated with the following circumstances: the methodology used for extraction of total bacterial DNA could have lyzed non-bacterial cells, low primer annealing temperature used during amplication, and the relative abundance of S. cerevisiae in pulque. Uncultured bacterial clones detected in this study were related to known sequences of Acinetobacter spp. and Lactobacillus spp. (Fig. 3). One of the main distinctive properties of pulque is its viscous consistence, which is dependent on EPS production. The main microorganism associated with this property has been L. mesenteroides, widely studied for its dextran producing capacity (Snchez-Marroqun and Hope, 1953; Garca-Garibay and Lpez-Mungua, 1993; Chellapandian et al., 1998). However, in this study we reported for the rst time in pulque the presence of an EPS composed by fructose (levan or inulin

type) observed after 6-h fermentation. It has been published that L. citreum CW28 and L. mesenteroides (NRRL B512F and ATCC8293 strains) produce an EPS composed by fructose as a consequence of fructosyltranferase activitiy which transforms fructose from sucrose to a growing fructan polymer (Olivares-Illana et al., 2003; Ortiz-Soto et al., 2004). Characterization of this fructose-EPS is currently in progress. The bacterial diversity in AM and during the fermentation of traditional pulque was studied by using a polyphasic approach. The microbial diversity detected in AM and T0 was composed by homo-, heterofermentative, acetic, ethanol and EPS producer species. The process started in an acid and sugar rich environment, with a complex bacterial diversity composed by both Gram-positive and Gramnegative bacteria from which the heterofermentative L. citreum, L. mesenteroides and L. kimchi were the most abundant species during the rst hours. Homofermentative L. acidophilus was the most abundant species at the end of fermentation. Z. mobilis subsp. pomaceae and Citrobacter species contributed to the production of lactic acid (DL), CO2, ethanol and small amounts of acetic acid which are important for the sensorial characteristics of fermented pulque. All these microorganisms, their proportions and properties must be considered in order to dene a well known bacterial starter to perform the fermentation of pulque under controlled conditions. However, it is still necessary to organize additional studies to gain further information on the interactions between the bacterial diversity present in pulque with yeasts, the most important microbial group responsible for alcohol production. Acknowledgements We thank Fernando Gonzlez, Mercedes Enzaldo and Rodrigo Conca for technical assistance; Eugenio Lpez, Paul Gaytan, Jorge Yaz and Soledad Jurez for primer synthesis and sequencing support; Araceli Patrn and Rosa Mara Picasso for SEM support, and Mr. Alfonso Dvila for pulque sampling. This work was supported by grants CONACYT P46052-Z, D43243-Z, PAPIIT/UNAM IN-205005-2 and PAPIME/UNAM PE205606. References
Ampe, F., ben Omar, N., Moizan, C., Wacher, C., Guyot, J.P., 1999. Polyphasic study of the spatial distribution of microorganisms in Mexican pozol, a fermented maize dough, demonstrates the need for cultivation-independent methods to investigate traditional fermentations. Applied and Environmental Microbiology 65, 54645473. Ashraf, I., Gerner-Smidt, P., Liesack, W., 1997. Phylogenetic relationship of the twenty-one DNA groups of the genus Acinetobacter as revealed by 16S ribosomal DNA sequence analysis. International Journal of Food Microbiology 47, 837841. Barns, S.M., Fundyga, R.E., Jeffries, M.W., Pace, N., 1994. Remarkable archaeal diversity detected in a Yellowstone National Park hot spring environment. Proceedings of the National Academy of Sciences of the United States of America 91, 16011613. Blandino, A., Al-Aseeri, M.E., Pandiella, S.S., Cantero, D., Webb, C., 2003. Cereal-based fermented foods and beverages. Food Research International 36, 527543. Borneman, J., Skroch, P.W., O'Sullivan, K.M., Palus, J.A., Rumjanek, N.G., Jansen, J.L., Nienhhius, J., 1996. Molecular microbial diversity in an agricultural soil in Wisconsin. Applied and Environmental Microbiology 62, 19351943. Buckley, D.H., Huangyutitham, V., Nelson, T.A., Rumberger, A., Thies, J.E., 2006. Diversity of Planctomycetes in soil in relation to soil history and environmental heterogeneity. Applied and Environmental Microbiology 72, 45224531. Carr, F.J., Chill, D., Maidas, N., 2002. The lactic acid bacteria: a literature survey. Critical Reviews in Microbiology 28, 281370. Chellapandian, M., Larios, C., Snchez-Gonzlez, M., Lpez-Mungua, A., 1988. Production and properties of a dextransucrase from Leuconostoc mesenteroides IBT-PQ isolated from pulque, a traditional Aztec alcoholic beverage. Journal of Industrial Microbiology and Biotechnology 21, 5156. Colwell, R.K., 2006. EstimateS: Statistical estimation of species richness and shared species from samples. Version 8.0. Persistent URL bpurl.oclc.org/estimatesN. Cruz-Guerrero, A.E., Olvera, J.L., Garca-Garibay, M., Gmez-Ruiz, L., 2006. Inulinasehyperproducing strains of Kluyveromyces sp. isolated from aguamiel (agave sap) and pulque. World Journal of Microbiology and Biotechnology 22, 115117. Du Toit, W.J., Lambrechts, M.G., 2002. The enumeration and identication of acetic acid bacteria from South African red wine fermentations. International Journal of Food Microbiology 74, 5764. Escalante, A., Wacher, C., Farrs, A., 2001. Lactic acid bacterial diversity in the traditional Mexican fermented dough pozol as determined by 16S rDNA sequence analysis. International Journal of Food Microbiology 64, 2131.

134

A. Escalante et al. / International Journal of Food Microbiology 124 (2008) 126-134 Ortiz-Soto, M.E., Olivares-Illana, V., Lopez-Munguia, A., 2004. Biochemical properties of inulosucrase from Leuconostoc citreum CW28 used for inulin synthesis. Biocatalysis and Biotransformation 22, 275281. Rheims, H.C., Sprer, F., Rainey, A., Stackebrandt, E., 1996. Ribosome molecular biological evidence for the occurrence of uncultured members of the actinomycete line of descent in different environments and geographical locations. Microbiology 142, 28632870. Snchez-Marroqun, A., Hope, P.H., 1953. Agave juice. Fermentation and chemical composition of some species. Agricultural and Food Chemistry 1, 246249. Schloss, P.D., Larget, B.R., Handelsman, J., 2004. Integration of microbial ecology and statistics: a test to compare gene libraries. Applied and Environmental Microbiology 70, 54855492. Simpson, K.L., Petterson, P., Priest, F.G., 2001. Characterization of lactobacilli from Scotch malt whisky distilleries and description of Lactobacillus ferintoshensis sp. nov., a new species isolated from malt whisky fermentations. Microbiology 147, 10071016. Smit, E., Leeang, P., Gommans, S., Van Den Broek, J., Van Mil, S., Wernars, K., 2001. Diversity and seasonal uctuations of the dominant members of the bacterial soil community in a wheat eld as determined by cultivation and molecular methods. Applied and Environmental Microbiology 67, 22842291. van Beek, S., Priest, F.G., 2002. Evolution of the lactic acid bacterial community during malt whisky fermentations: a polyphasic study. Applied and Environmental Microbiology 68, 297305. Waleron, M., Waleron, K., Podhajska, A.J., ojkowska, E., 2002. Genotyping of bacteria belonging to the former Erwinia genus by PCR-RFLP analysis of a recA gene fragment. Microbiology 148, 583595.

Escalante, A., Flores, M.E., Martnez, A., Lpez-Mungua, A., Bolvar, F., Gosset, G., 2004. Characterization of bacterial diversity in pulque, a traditional Mexican alcoholic fermented beverage, as determined by 16S rDNA analysis. FEMS Microbiology Letters 235, 273279. Estrada-Godina, A.R., Cruz-Guerrero, A.F., Lappe, P., Ulloa, M., Garca-Garibay, M., GmezRuz, L., 2001. Isolation and characterization of killer-yeasts from agave sap (aguamiel) and pulque. World Journal of Microbiology and Biotechnology 17, 557560. Favela, E., 1993. Produccin de alcohol de Zymomonas mobilis. In: Garca-Garibay, M., Quintero-Ramrez, R., Lpez-Mungua, A. (Eds.), Biotecnologa Alimentaria. LIMUSA, Mxico, DF, pp. 617636. Garca-Garibay, M., Lpez Mungua, A., 1993. Bebidas alcohlicas no destiladas. In: Garca-Garibay, M., Quintero-Ramrez, R., Lpez-Munguia, A. (Eds.), Biotecnologa Alimentaria. LIMUSA, Mxico, DF, pp. 263311. Godoy, A., 1987. Recopilacin bibliogrca sobre aspectos: histrico, etnobiolgico, microbiolgico y qumico de bebidas alcohlicas no destiladas, indgenas de Mxico.Thesis. Universidad Nacional Autnoma de Mxico. Holt, J.G., Krieg, N.R., Sneath, P.H.A., Staley, J.T., Williams, S.T., 2000. Bergey's manual of determinative bacteriology, 9th edition. Lippincott Williams & Wilkins, Philadephia. Hughes, J.B., Hellmann, J.J., Ricketts, T.H., Bohannan, B.J.M., 2001. Counting the uncountable: statistical approach to estimating microbial diversity. Applied and Environmental Microbiology 67, 43994406. Janssen, P., Mzquelin, K., Coopman, R., Tjernberg, I., Bouvet, P., Kerster, K., Dijkshoorn, L., 1997. Discrimination of Acinetobacter genomic species by AFLP ngerprinting. International Journal of Systematic Bacteriology 47, 11791187. Olivares-Illana, V., Lopez-Munguia, A., Olvera, C., 2003. Molecular characterization of inulosucrase from Leuconostoc citreum: a fructosyltransferase within a glucosyltransferase. Journal of Bacteriology. 185, 36063612.