-

A C S S Y M P 0 S I U M S E R I E S 660
Spices
Flavor Chemistry and Antioxidant
Properties
Sara J. Risch, EDITOR
Science By Design
Chi-Tang Ho, EDITOR
Rlltgers, 71l e Stale University of New Jersey
from symposium
the DIVIsIOn of Agncultural and Food .
American Chemicol Society, WOshington, DC
Uhnry or Congr''!! Calalos;ng.!n. I'ubUu !lon Ila!a
Sl'tc,,: Il.wor che,niWy and pmpcrtie . I S,,.n I, Ri""h, rJilOI,
110. wiU)",
p. cm.-{ACS . ympo<ium series, ISSN 0097--61 660)
n.:",lopa l from a , ymposi um l ponsored by tile of
al ti,e 2 111h National Muting oflhe Anoerio.:an Cloemic"1
S<lci.,ly. New Much 24_28, 1\I'}6.'"
ISHN O M412_3495_7
.. SI'I<:<:.,-Congre. .. e... 2.
I. Ri:h. Sara 1., 1958- n. HO,o.iTang. 19044- III. American
C1Icmk.,1 SOCiely. Division of Agrieultulal und F"od O'CII,iSlry. IV. AUlerican
SOCtcty. Meeting (2 1 Ith: 1996: New OrIeIlllS, La .. ) V. SC'.!eS.
1996
664'S--<:Ic21
This book; < IJf\ Kid free, recycled paper.
Copy.iht C 1997 Amer;can l.'hr:micaJ Society
All Reproaraphic beyond that pcnnitted by Section5 107 or 108 of Ihe U.s.
CopyriGht Act;s allowed fOf ,nlern.1 use only. provided that a pc1-chap1er fee of plus $0.25 pcr
i5 paid to the CopyriGht Oearance Center. Inc., 222 Rmewood Dri'e. Dan'crs. M,\ 01923,
Republic-.ahon or re l"oducliOft for sale of in book i, permined only un<kr licen<e f.om ACS.
Di=I lhQC and othet jlCm1issions requesls 10 ACS Copyright Office, Publiutions Divi sion. 161h
SlIcet. N.W" Washington. OC 20036.
The ci talion of trade nanoe. andlor namc. of in thi s pubfication not to be con5\nH:d IS
an or as ap>"oval by ACS of the CQlnmefeial producu 01' ",."iets .de.ent ed hcn:in; 00f
sbovld the meTC .... ein to any drawing. speciflC"ioo. che ....ical process. 0, other data be
regarded as" Ikense or "5 con"ey"nc, of any riMll t or permission to the holdc:r. reade or any odler
or corporation. to m3l1ufx,u.e, .(produce. use, or sen any palcnled inventiort ()( copyrighted
work in be related thereto. names, tra<lemarh. etc .. used in
. "en wil llOUt indication thereof, arc not 10 be considered unprotected by law.
PRll\'TED];O; Tim UNIll; D OF AMERICA
Advisory Board
ACS Symposium SerI es
Robert J. Alaimo
l'roeler & Gamble PhJ'lI1l1<:euticals
Mark Arnold
University of Iowa
David Baker
Uoivenity ofTenne5$<'e
Arindam Bose
I'filtr Rcsearcl,
Robert F. Brady. Jr.
Re.o;car(h t abomtory
Mary E. Castellion
ChemEdil Company
Margaret A . Cavanaugh
NallQn.1 Science Foundatioo
Arthur B. Ellis
Univer,;ity of Wiscon!in at Madison
Gunda I. Georg
or Kansas
Mnddcine M. Joullie
Uni "cl"$ity of l'cnnsylvania
Lawrence P. Kl emann
1'00<15 GIOUp
Douglas R. Ll oyd
'Ille University of Teus at
Cynthia A. Maryanoff
R. w, JOII05011 J'llarmaccut ;cal
dt Instit ute
Roger A. Minear
University of 1I1ioois
M Urbnna- Chalnp"ign
Omkaram Nalarnasu
AT&T Ikll
Vincent Pecoraro
of
George W. Roberts
Non" Carolina StAte UniVeriiiity
John R. Shapley
University of Illinois
I! Urba.na-013mpalgn
Douglas A . Smith
Concurrent Teeh!lolvgies Corporation
L. Somasundaram
DuPont
Midlad D. Taylor
l'hilftnaCeutical Reseuch
William C. Walker
DuPont
Peter Willett
University or Sheffield (England)
Foreword
THE ACS SYMPOSIUM SERIES W3S first publi shed in 197410 provide
a mechanism for publishing symposia qui ckly in book form. The
purpose of Ihis series is to publish comprehensive books developed
from symposia, which are usually "snapshots in time" of the current
research bei ng done on a topic. plus some review material on the
lopic. For this reason, it is nccess:uy that Ihe papers be publ ished as
qui ckly as possible.
Before a symposium-based book is put under contract, the
proposed tllble of contents is reviewed for appropriateness 10 the topic
and for comprehensiveness of the collecti on. Some papers are
excl uded al this point, and others are added to round Qut the scope of
the volume. In addi ti on, a draft of each paper is peer-reviewed prior (0
fi nal acceptance or rejection. This anonymous review process is
supervi sed by the organizer(s) of the symposium, who become the
editor(s) of the book. TIlt": authors then revi se thei r papers according to
the recommendations of both the reviewers and the edi tors. prepare
camera-ready copy, and submit the final papers to the editors, who
check that all necessury revisions have been made.
As a rule, onl y original research papers and original review
papers are included in the volumes. Verbatim reproductions of
previously published papers are not :! ccepled.
ACS BOOKS DEPARTMENT
Contents
Preface ............................................................................................................. ix
GENERAL OVERVIEW Al"' D METHODS
I . Sllices: Sources, Processing, and Chemistry..... ......................... ..... 2
Sara J. Risch
2. Methods of Dactel"ial Reduction in Spices ................... .................... 7
W. Leistritz
FLAVOR CIl EMIS1"RY
3. The Principal Flal'or Components of Fenugreek ( l'rigo1lelfa
foe1lum-graecum L) ............................................................................... 12
Imre Blank. Jianming Li n, Stephani e Dev3ud,
R e n ~ Fumeaux, and L:l.Urent n. Fay
4. Vanil\ll............................... ........................................................ .............. 29
Daphna Havkin-Frcnkel and Ruth Dorn
5. Onion Flavor Chemistry and Factors Influencing Flavur
Intensity.................................................................................................. 41
Will iam M. Randle
6. Cuntribuli on IIf Nonvolati le SuUur-Containing Flavor
PI'ecursors of the Genus Allium to the Flavor of Thermally
I'rocessed Allium Vegetables ................................ "............................ 53
Tung"lsi Yu
ANALYTICAL 1):CHNIQUES
7. Characterization of Saffron Flavor by Aroma Extract
I)ilution Analysis................................................................................... 66
Keith R. Cadwall ader, Hyung !-Icc 8aek. and Min COli
,
H. Tile (]wracteri 7.atinn of Volatil e :lntl Se mi vulatile
in I'uwdered Turmeric hy Dired Therm:l l Extractiun
( ;a5 ChruIIllIlugr:lphy- M:lss Sllectrumetry ..................................... 1:10
Rich:l rd D. Iliscrodt , a li-Tang 1-10, and Robert T. Rosell
'J. !,unge nt Flavor I'rnfil es and Component s of Spices
by Chrolll., tugraphy and Chemiluminescent Nit r ngen
Ilc tectinn ...................................... " .............................. ..... ".................... 98
E. M. Fujinari
10. Supercritical "' Iuid Extracti on nf AlliulII Species ............................ 11 3
Eli;r.abcth M. C.,lvey and Erie Block
11. Determination of Glucosinolates in Mustard
by High. Performance: Liquid Chromatography-
Electrospray M:lss Spect rometry ....................................................... 125
Carol L. Zrybko and Robert 1'. Rosen
12. Reasons fur the Vari ation in Compos ition of Some
COInnler cial Essenti al Oil s ........................................... ..................... .. 138
Chi-Kuen Shu and Brian M. L1wrence
13. Cumponent of Mixed Spices .............................................. 160
C. K. Cheng, C. C. Chen, W. Y. Shu, L. L Shih,
and H. H. Feng
ANll0XIDA."lT PROPERTIES
14. Antioxidati ve Activity of Spices and Spice Extracts ....................... 176
Helle Lindberg Madsen, Grete Bertelsen,
and Leif H. Skibsted
15. Ant ioxidat ive Effect and Kineti cs Study of Capsanthin
on the Chlorophyll -Sens itized I)hotooxidation of Soybean Oil
and Selected Flavor Compounds ....................................................... 188
Chung-Wen Chen, Tung Ching Lee, and Chi-Tang Ho
16. Cureumin: An Ingredient that Reduces Plate let Aggregation
and Hyperlipidemia, and Enhances Antioxidant and Inunune
Functions ................................................................................................ 199
Yaguang Li u
17. Antioxidant Activity of L:l vandin (LtlVam/lIia x illfermedill ) Cell
Culture s in Relati ll n to Their R"s marinic Acid Content._ ............ 206
T. L6pez-Arnaldos, J. M. Zapata, A A. Calder6n,
and A. Ros Barccl6
18. Anti -innanunatory Antioxidants from Trupic:!1 Zingiberaceae
Plant s: Isolation and Synthes is of New Curcuminoids ................. 219
Toshiya Masuda
19. Curcumin: A Pulse Radiolys is Investigll tinn of the Radical
In Micellar Systems: A Model for 8eha\' ior 3S a Biological
Antioxidant in Both Hydrnphobic and Hydrophilic
EnvironJllents ......................................................................................... 234
A A Gorman, I. Hamblett, T. J. Hi ll . H. Jones,
V. S. Sri nivasan, and P. O. Wood
INDEXES
Author Index .................................................................................................. 246
Affil iati on I ndex ............................................................................................. 246
Subject Index .................................................................................................. 247
Preface
S PI CES ARE WIDELY USED IN FOOD PRODUCTS to create the di stinctive flavor
and I:haruClcr that is representative or di fferent cui sines. The symposium on the
flavor and antioxi dant propenies of spices was organized \0 look al new
development s in the area o[ spicc chemi stry.
111C nature of the volatile components in different spic..:s is important in
understanding the fl avors they impart 10 food products. Research is being
conducted \ 0 determine the significant volut il c and nonvolatile compounds thnl
create the di stinct flavor of various spices. Vari ations in spi ces C;l1l occur
dependi ng on the region in which Ihey arc grown and the climatic condit ion ...
which can dramatically infl uence their composition. This information j-.;
imponant in formulating food products and maintaining their (I\,...:r
time_ As spices vary in their flavor profile, adjust ments may need to be made in
the fonnulution of a product or spice blend to maintain the desired !lavOL
Comput er modcls are being developed to eVlllullte spi ces [lnd spice blends to
determine thcir composi tion and other rekvllnt variables_
Another major arcu of interest today is the antioxidant properties of spices.
Although spices have long been used to help preserve food, it has not been
known what components give the preservative eITect. Research in this area has
expanded beyond the use of spices as preservati ves 10 the potcnti ol hcalt h
benefits they confer as antioxi dants in the body_ There is considerable evidence
tha t specific components in spi ces may provide these benefi cial cl1ccts.
Research is being carried oul to determine the aclive components and \0 explai n
the mechanism of action. With increasing interest in the use of food product s to
help maintain health and prevent disease, spices may playa significtlnt role.
-Ille symposium on whi ch this book is based was sponsort..-d by the Division
of Agricult ural and Food Chemistry at the 21lth National Meeting of the
Americtlll Chemicill Socicty, which took place in New Orleans, Louisiana, from
March 24- 28, 1996. n lis symposi um provided a forum for researchers from all
over the world to prcsent information on possible specific roks .)J" spices ill
di sease prevcntion.
We thank all of the panicipants in thc symposium as well as ad,litioll;11
aut hors who have contributed papers \0 thi s hook_ All or 'lIlT wi",
gtlve of their time to Hssist in the review of the an: r,-n':'I !i/l-.I , ; lIltl
Ihei r work is ~ i n c c r d y appreciatcd. Thc American Spice Trauc Associati un was
helpful in providing general information about spices and their usage.
SARA J. Ri sen
Scicnce Oy Design
50S :-Jorth Lake Shore Drive, 113209
Chicaso. IL 606 11
CIIITANG 1-10
Dcp:_u1ment ofFoocl Science
Cook College
Rut gers, '111C Siale University orNcl\' Jersey
New Brunswick. NJ 08903- 023 (
Novcmber 22, 1996
GENERAL OVERVIEW AND M ETHODS
Chapt er 1
Spices
Sources, ProceSSing, and Chemist rY
Sara J. Risch
Science By DeSig n, 505 North L'lke Shore Drive, #3209,
Chicago, IL 60611
Spices are used throughout the world to season food products and
create the unique characteri st ic fl avors of different cuisi nes,
Spices are grown primarily in tropical count ries although the
United States has recently started growing a number of different
spices to supply the domestic needs. The use of spices has
increased significantly over the past few years, due in pan to the
high level of interest in different types of foods that use a wide
variety of spices. Interest is also developing in the ability of spices
to act as antioxidants in addition to seasoning a product
Spices have long been important for food products. It was found that small amounts
or various plants could be used to enhance the flavor of a food and also served to
help preserve that food. In somc cases spices wefe even used to mask spoilage or
off-flavors in products. This use continued for centuries wi thout any real
understanding of how the spices were being effective. People simply understood 1hat
spices helped create a more desi rable taste in the foods that were being prepared.
Different cuisines are noted for using types of spices to create their
characterist ic fl a\ors.
People also viewed spices as being imponant because in early history the
entire ecollomy of many regions was based solely on spice trade. Spices were the
major item of trade and the region that could control spice trade dominated as a
world power. That situation has changed dramat ically with spices now accounting
for less than 0. 1 percent of world trade. Other raw materials as well us processed
foods account for a larger percent agc of world- wide trade than spices. There are
still several countries that rely heavily on the trade of a specific spice. In
the production of clove account s for a large percentage of that country's economy.
Tanzania grows about 213 of tile world requirement for cloves. About half Granada's
revenue is derived from the sales of nutmeg and mace. !fyou also consider vanilla,
(tI l991 American Chemicat Society
1. RI SCH Sp i"fls: S(//m:u, l'rocb'l';ng, allll CJII'IIris(ry
which is noIteclmicaJly /I spice but is a Vf%'J 'Nidcly used flavoring material that comes
from a plant grawn in tropical regions, it is the second largest prodllct in the
Malagasy Republic.
There are 36 different herbs and spices that are generally recosnized and
commonly used, White and black pepper account for the largest amount of spices
used in terms of doll ar value The ne.'l:t highest value spices consumed are cloves,
nUlmeg, cardamom, ci nnamon ginger, mace and all spice It should be noted that
accurate figures for the value of spice imports and exports are difficult to track. ,
Countries typically do nOt repon spices separately from other food products
Most spices that are imported into the United States come in whole or
unprocessed, There are requi rements that they meet minimum standards for purity
and cleanliness. The Food and Drug Administration performs inspections on the
incoming spices mainly to conlirm that they are free from undesirable filth and that
they are safe for use in food. The American Spice Trade Association (I) and AOAC
(2) have defined specific tests for spices. A complete list of for the
quali ty of herbs and spices was developed a number of years ago by Heath (3) This
provi des a comprehensive listing of tesls that can give infor mation on Ihe volatile oil
cont ent as well as other quality related factors in spices.
Spices and herbs come from a number of di fferent pans of plants. This cnn
incl ude the bark, seeds, berries Dnd leavcs of a plant. In the United States, spices
have a legal definition that is spelled out in the Code of Federal Regulations (4). This
states that spices are "any arommie vegetable substance in whole, broken or ground
fonn, except those which have been traditionally regarded as foods, such
as onions, garlic and celery; whose significant function in food is seasoning rather
than nutritional; that is tAre to name; and from which no ponion of any vola:ile oil or
other na\'ori ng principle has been removed". nle form that the spice is use<! in ollen
depends on the application in the food product In some cases it is desirable to have
large, visible pieces of spice while in others the spice may be finely ground to be more
easily incorporated into the product. Some of the appl ications for whole spices
include peppercorns in sausages, bay leaves in soups and various seeds thllt are used
on or in breads and other baked products includi ng caraway seeds, sesame seeds and
poppy seeds.
For the vast majority of applicat ions, the spice is processed before using in a
food product. General processing of spices usually involves some means of panicle
size reduction. Processors use different to achieve the desired panicle
size. The process of grinding spices breaks down some of the cell structure to make
the volatile oils in the plant more readily released when used in a product than if the
spice were left in the whole form. Care must be laken during the gri nding process
to insure that the desired volatile oil level is maintained in the spice. One process Ihal
has been investigated to help preserve the qual ity of the spice is cryogenic grinding.
In addit ion to grinding or comminution of the plant part , a mcans or bacterial
reduction is also needed. The different methods empl oyed to reduce the bacterial
load on spices. including irradiation, steam trcatment and ethylene oxi{!c, will be
covered in a subsequent chapter in Ihis book. A comprehensive review ofindividllill
,
SI' II.:F.S: flAVOR CIII': MJ STIl V fi NO ANTlOXJUflNT l' IIOI'lmTJK"
spi..:;cs intludinij thcir flavor characters and various applicati ons is included in the
Source Book of Flavors by Reineccius (5). This rererence al so includes inrormation
aboUI processing ofspice5 as well as the specific volatile oils and the procedures that
arc lI sed to extract the oils from spices. The flavor of herbs and spices is derived
from the volatile viis that arc present in the plant. The only exceplion to this is the
which give The flavor to peppen. The essential oil content of herbs is
Ollcn less Ihan 1%. The conlenl of other spices ranges up to 18% essential oil in
clove buds Essential oils arc recovered from herbs and spices by steam distillati on.
'111i5 recovers The vollllile oils but not the non-volat ile components which contribute
to the pungency of some spices such as ginger and pepper. Oleoresins are oblained
by solvent extraction which yields both volatile and non-volatile components.
While spices have long been used in food products, their usage continues to
gro w al a significant rate. A recent estimate by Ihe American Spice Trade
Association indicates that the usage in the U. S. alone could reach one billion pounds
by the year 2000 (AST A personal communication, 1996). Interest in spices started
growing aner World War II and saw a sharp rise beginning in the 1980's. This
likely be attributed to first, the exposure that many people had to other spices in the
1940's and. more recently. to increased interest in different ethnic cuisines. 11 should
be noted that the consumption figures that are reported include dehydrated onion and
garlic even though by legal definition they are not spices. These items may be used
as spices in a food in the U. S. but must be listed separately on an ingredient
statement. The average consumption today is about 50"10 greater than the average
a decade AgO with the annual per capita consumption of spices in the United States
ovcr three pounds (ASTA, personal communication, 1996). The spices that have
shown the largest increase in use include sesame seeds., oregano and paprika. II is
likely that the growth in usc of sesame seeds parallels the growth in fast food
restaurants using sesame seeds on sandwich buns.
The other trends in spice consumption in thc U.S. follow thc growth of
cenai n food categories. One area is in hot foods where the AInerican public is
becomi ng much more adventuresome in the foods that they will eat . Consumption
of rcd peppcrwas up 105% from 1988 to 1992. Our bland foods arc becoming much
spicier One company rc:ccnlly introduced a fudge sauce for icc cream that contains
blnck pepper, red pepper and cinnamon ill addition to dark chocolate. At the 1996
Institute of Food Tcdlnologists convenlion, one demonstration showed thc use of red
pepper sauce in browni es. In cont rast, another area of growth is in the mild herbs
including basil and oregano. This growth follows the increase in consumplion or
piua and a proliferation of commercially available spaghcui sauces. As the desire for
ncw and unique foods cont inues, it is expected the use of spices will continue to
grow.
Spices are commonly grown in tropiCI'll regions ofthc world. This can present
chall enges to companies that are trying to fmd a consistent source of spices both from
a qlJal ity and supply perspective. Fanners in the U. S. have Slarted growing a number
of di fferent spices to meet our growing demand. When dehydrated onion and garlic
ar c includcd in the figures, the U.S. in 199] was able to supply 38% of its needs
I. JUSCI! Spices: SOllrces, Pmce.I'.I'i/'I:. and Chemistry
(ASTA, personal communication, 1996). The other countries that are impoJ' lant
sources of spices include India, Indonesia, Mexico, Guatemala, China and Canada.
Canada has gained prominence: for its production of mustard, caraway and coriander
seeds.
With the growt h in the use of spices., there has been continued research into
the active components of spices not only from a flavor standpoint but also from
functi onal perspective 10 explore the antioxidant properties of spices. Differcnt
spices are known to contain hundreds of acli ....e compounds. One area of interest is
to identify Ihe specific compounds in a spice that make the most significant
contribution to the flavor of that product. A technique that has been applied to flavor
compounds to detennine thclr relative importance to Oavor of a product is gas
chromatography coupled with effluent sniffing, referred to as GC-O (gas
chromatography-olfactometry). Several variations on this te<:hnique, including aroma
extract dilution analysis, can be used to determine the compounds that are most
important to the aroma. This uses the human nose to determine characteristic and
importance ofindividual compounds to compliment gas chromatography which can
only give an indication of relative amounts of different compounds but cannot give
any information on importance to the flavor profile. It is known that flavor
compounds can have widcly varying sensory thresholds so having only the analylical
dala does not teUthe entire story regarding importance of the individual compounds.
The two specific spices that were evaluated and are presented in later chapters were
salll'on (Cadwallader and Back) fenugreek (Blank et el). This data be used
in developing flavor systems that may be able to replace or supplcment spices in food
products.
There is ongoing de\-dopmcnl of new analytical techniques to evaluate spices
to dcvclop specific information about their composi tion. These techniques may nol
be applicable only to spices but are to expand our knowledge of spices and
their active components. One method is supercrit ical extraction which is used to
more effectively removc Ihe volatile compounds from a spice and eliminate Ihe
interference of olher plant malerials. Once the compounds are isolated, one detector
that is being employed is specific to nilTogen and can aid in determining Ihe amounts
of capsaicinoids which contribute the heat or pungency to different types of peppers.
Spice companies and food manufacturers arc also concerned about
maintaining the consistency of the spices and spice blends that they li re using. It is
know that a number offactors can influence the abundance of the aelive components
in spices. The climatic conditioJlll can have a significant impact on spi ce quality from
one year to the ne.'!:t . One chapter later in Ihis hook will address the influence or
different growing conditions on the volatile compounds that are present (Randle).
The same species of plant that is grown in different regions of the country can have
a diffcrent volatile profile from that grown in another region. Understanding these
dilTerences"is important for manufacturers to be able to maintain consistent flavor of
a product when the spice being used may be sourced from different areas. A variety
of plant species have been investigated to develop fingerprints of spices from different
growing regions. This information can be used to det ermine authenti city or a spice
,
6
. ,
SPICES: H.AVOM CIlt-:l\II STMY ANIl ANTlUXILIANT j' IWI'EMT1KIi
and determine optimum blends that should be used to maint!!.in the desired flavor
profile. Statistical programs arc also bting developed \0 be used 10 detemune the
most likely constituents in a spice blend and aid in replication oCa slJCCi lic blend.
This is presented in the chapter by Chen el al.
An area of considerable interest and the focus of many research projects is the
antioxidant propenies of spices. As was mentioned earlier, one of the earliest uses
of spices was to help preserve foods. Liule was understood III that time about the
mode of aaion of the spices. It was simply known that they would help maintain the
quality of the food to be stored, The area of antioxidant propenies of spices is
imponant for preserving the quality of foods but may also provide beneficial health
effects for people consumi ng the spices. This area will be covered in much more
detail in an ollCl"View of the antioxidant properties of spices that is presented later ill
this book in the chapter by Madsen et a!. One spice component in particular that has
received atlemion is curcumin. Several chapters address the effectiveness of
rurcumin and possible ways to synthesize new curruminoids. The investigators have
looked at the various ways in which the compounds may be effective as an
antioxidant in different systems. This is an area that will continue to receive attention
in the research conununity as people cominue [ 0 look for Ihe specific health benefilS
Ihal certain foods may offer. Spices will remain important for Ihe flavor lhat Ihey
impart to foods and may al so gain significance for other benefits they have to offer.
Lit er at ure Cited
I . Analytical Methods orthe American Spice Trade Association, )rd ed. ASTA..,
Englewood Cliffs, NJ.,1978.
2. Official Methods of Analysi s. 14th ed. AGAC, Washington, D.C. 1984.
3. li eat h, t-I . B. Call. lnSI. Food 'f"e,h. J. 1968, 1(1), pp. 936.
4. Code of Fcderal Regulations, Title 21, U.S. Government Printing Office.
1995.
5. Source Book oj Flavors; Reineccius, G. A., Ed., Second ed.; Chapman and
Hall, One Penn Plaza, New York, NY, 1994, pp234 255.
Chapter 2
Methods of Bacterial Reduction ill Spices
w. L..e is tr itz
limit ed, 185 Alexll ndr a Way, Carul Stream, IL 60 188
There IIfC three major Incthods for b.lclerial reduction currently being us.:d
in the spicc industry. These arc ethyknc oxide, irnlliiation and Sh::.m.
Irradiation has received a gl"l!lll deal of pn:ss and although tnc popul;Hity uf
this mcthod is increasing, there is still whether or nol Ilk: consurno::r will
IIcccpt it. The induslry h.1S $t:icntific theory but the consulTler activists :lrc
concerned about the unknown ruture consequences o f the k chnolo);.v.
Another method that is not new bu has nOl been used in thc U.S. i;; stcam
steriliLation using supcrhealcd steam. This is a consumcr friendly method
that may be the solution tlmt the world is ror. The nlCth(l\l is
efti!ctive and viewed liS being safc. A comparison of the three Ill<:toods will
be presenll""d and the adv.lrltages and disndvuntages of cach method delllikd.
Spices are grown all the world, in many cases it is in areas wh..'re the cle:'lnlincss is
not closely controlled. There is 110 dian made to limit contamination of the spices wither
during growing or harvesting. While there are some specifications and regulations
concerning imponed spiccs, it is oncn necessary to process the spices to reduce the
bacteriallond prior to use in food products. The oot:h:rialload varies with thc type or
spice and total plate counts in excess of one mill ion colony unit s per 11:1\e
been reponed in some spkt:s (I). Diflerenl types of pcpp..:r are known to ()fien h:'lVC very
high bacterial counts.
The original need sterilized spices cnmc from a demand by the U.S. armed r(\Tecs.
I'rior to World War 11 , they were seeking roods that coukllx: ItI,: ld for long pcri<.. 'lls (l rtill"ll:
without spoiling. One source orlh.: bacterial contamination that rcduced the ... or
some products was the spice or seasoning being used. By reducing the \J f
in the spice, products could be made "ith significant I)' longer slk:lf-l i\"cs. Till: nll . ."thud !;lr
doing this rJCedcJ to dcstroy the bacteria without changing the Havor or color "I" th, spicc.
Sterili7.ation tcduliqucs \\ierc developed and were so successrul ror prndu<."l s I;,r th.:
I1rmcd rorces, that it soon became a gcnemlly accepted tedl!lilJllc Ii,r m' ,\ ' I'in' 'lIkl
c:r t9!17 ChCI1lK.":li S ... ;kly
St'U .. , VOlt Cl IEMtS'fllY A:"I II ,\ NTIOXIIJANT ..
"la". All 'I'to.:e" ,m.1 .,",a:mings tk, lint ro:ttuin: h:to.: lel'i;11 reJudMlI1. hUI 111I.r.: is a
l i,1.. I:,el"f wilh all ,'Ilhe '1 he "w,d in;1I111t:1llllrcr In wcil:!h lito.: ri"l..s " f
il1;1llri;.ls 111:,1 p"lentially high ha( lerial loao wil h any .. tin r-.;r(CI'1Mm
tl1l' I,' 111;1)" he ,dth th,: pr'lI.:cs.sin!!- nfthe spi..:e.
Ill<' treal n .. "t nf spices hi r"'llll(e Ihe h."lI:kri;d k'ad is "nl'lI IdeneJ h' ....
\l': ll li/"tMIII. In !<Ome c .. II ... spke nk!) ;lCt'l:llIy be IMwl'{.""er. H m,'re a<.:l"1.rat ..
h:I(I .. ,-i:11 ,-.. duetMUI. TIll: lirst mctlK.d 111;11 W;lS dc,ck'po."d 1<1 retlu .. e the
<,n "I'icl's ill\'lIl\'cd lhe u <;c o f cthykne oxide nne (omllany th;lt is using a
,,I'-.KIIl ,,(this to,by is (iriflith MicruSeicnee. The trealment (If a lillll.'lI,
I",," tnnpernt,u'e (yel<.: "r ....1eulll11 gas.sing with Cl hylcn( "xide 0' propyknc oxide. The
;:"IK'ral pr"l,ess now is below.
t . t'rodu..:t p!;J.;eu in a scak:d chaml'l(."r about the width or" standurd l l.S. palkt
(.j() x 42 in). The chamber may ...ary in k,ngth and height, hut it is critical thnt it
1\ot be too wide.
2. A \,<I": UUI11 is pulled within the chanlbcr.
3. "l'he..:h.1111hcr is heated to bring the temperature up to 110120 I:.
4. I tumidit)' is introduced. ([t sllll uld he IlOteJ thatlhe .. ombinatioll of lle,lt and
humidi ty are essential to activnte the within the sllice.)
5. I:.thylcllc oxide gas is int roduced into the ch.'lmber.
6. Th..: spice is held in Ihe dl'l mber under Ihesc conditions fu r a period
uf time which is ,Iire .. tly relah ..d to the bulk density of the spice lind the level uf
reduction thai is
1. The gas is c)(pclk.xI rrom Ih,: chambcr and it is Ilushed scveral limes with air
prior to rcl uming it to pressure.
S. The product is then ren ...J\ed to a '-Iuarantine area until it is pro ...en th.1t IMl
gas can be detected.
TIl;S pro"cn method fo r spices .. particularly wit h wlMl\e sp;"; .. and sectls. I ln"" .. \".. r.
the usc of gas for grourMl spices is eontro'-crsial today. method is e't p:lhlc of
r..:dllcing Ihc bacterial load by 31 least I x I 0'. nit' proecss may be repeated if the initial
docs IMlt llleCI customer requirement s. Using ethylenc oxide gas rC\jui .. a
luner time period than irradialion (If steal11 .. wilh a lYl'ical process requiring anyhc re
Ir01;1 12 to 18 hours. l'roduclivity is il,nucnced by the Si7,C and amount uf eham[:.,:rs
:l\"3ibblc. nle cost pcr pound of spice trellted is based upon volun .. trcated, bul l.. densit ),
ami iength ofli1ne Te(luirl'd to obtain the desired b.1cterial reduction.
Irwdiation h;l s become a popular method o f rcUueing b:leteria in spkes anJ
since it s introduction in the early 1980's, lIcc,lrding to the personnel at
SkriGeJ1ics ill Schaumberg, 1L. the cont ro lled release of Cob:llt 60, P:1SS
through a hetun or gamma energy that dcstroys b.1cteria. The amount o f energy is clil led
the radiatioll absorbed dose !IIKI is nleasureo in fads ( I rad = 100 ergs/g). Most .. es arc
tre .. ted at a Icvcl of 1.0 1.5 ntds which is less than 50% of the rate allowed by Ihe Food
and Drug Administ rat iun (FDA). TIM; gamma process requires (onlf()1 of onl y o11C
1).'lf'U1lCter and that is time of exposure. SleriGeni(s literature lK)tes thnl their process is
klh,11lH mlol! pathogenic microorganisms in spices. The following Sleps describe
the SteriGcnies method fo r usc o f irradiation to reduce bacterin in spkes.
2. U : t STKIT"l. 11 Bael cri(l/ Helillel iOlI i n Spitt{ 9
I. I'roduct is loaJnl 011tO a ( arrier and travels alo ng a .. onveyor through a series
or doors and locks int" the cdl ll r .... containing eoixlh 60.
2. The product pa.<;scs IImWJJ and then tl .. t .....o shelvcs nfisotopes where the
greatest pcrcent .. ge reduction occurs. While in the cell. the product
i: hombarded with energy lit lower as ....'CII.
3. The product rctUfl\.<; through Ihe of doors ,mo locks o n a tin){'(.( schedule
and to its origin. The entire takes between 5 and IS hours,
depending upun the initial bacterial klad and the reduction that is required.
l11is is a cost clfoxtivc method IIk!tllffers 11 number of1:ll:ocfits to the consumcr. The main
I>.:ndit is its overall effecti"encss. There is no add .. d to the spice and thcre are
no ad ...erse cfkets on the wlatilc oi ls o r rMlI1volatiJe components so the fla vor pmfile
!lOt negatively impat:ted. The irrnJilil ion wi ll completely penctrate packaging materials
so that they can be processed withoulthe need o r opening a package.
The major concern with irradiatilln cent crs upon the fear ofeOllsumer a(ti ... isl
groups, bot h domest ic and international. that nOI enough dnla hns been gathered to
dl .. terrnine the long term effects o f eoting rood product s which have been irradiated or
which contain ingr .. xlicnts tl101t have For this reason. the usc o f irradiation
or spices and has been Nmn .. d in scveral fo reign countries. Some U.S. I{)od
manufact urers ha ...e also made it a policy not to usc any ingredients which have been
irradiated. SteriGenics, the leading company offering irradiated spices is currently
these cono;crns bot h to go\,emmental ageneics and to consumer groups.
The usc or steam to reduce the b.1eterial count in spices has been gaining
momentum due to the C"oneerns with the ofbol h gas and irradiat ion. Sleam is
considered III bc a safe form orblleterial reduction throughout the workl . There arc I WO
melhods o f steam bacteria reduction being olli:red in the U.S. today. One o f tllcse is a
wet method and the Oilier is a dry pwc .. ss.
'I'hcrc are scvcrnJ mClhods o f .....et Sleam systems for bacterial reduction available
in the U.S. These include lhe MicroMaster ufoCd by McCormick. Fuchs Micro
c<) ntfoJ and the Kikkom. 1n continuous steam process. These systems use eit ner closed
chambers or continuous fced systemS. The key to all orthe:'\C syst ems is the ability to
penetrate the whole spices or Sl.'Cds with wet steam at II temperature o f 100 C or great .. r
for a period o r time. The bulk density is another critical facto r v.hk h hel ps
determine till: flow rate. moisture piek.up and e ...entually the cost per pound.
A l'(lnsistent cOllecn! with tho:: wet steam process is the anilit y to properly dry the
during with OIM; or treatmcnt after It may bei;ome ncces.o;.ary to
re.dry lhe product at a II>\\ cr temperalure 10 achieve the desired moi sture le,els. TIle
overaU .. " sls for using wet steam 11\.11ch those nfusing gas.
The usc or dry steam (superheated) is the lot esl stl'<lml echnology being used for
lind seasoni ngs. The cquipmentlo produce spices by this method was recently
installed by SpiceTec Lt d. in their plant in Carol Stream. II.. [t is the first usc of this
me thod in thc United Slales. The process tlow for system is described below.
I. Continuous precle:med product is drupped through <In airlock into a closed
chamber that contains a shaking bed.
10 SI' I( "K'i: . "IAVtllt ClI EMISTRY ANI) ANTl()XIIlANT
2. The product is sterilized (Iuiddy "t a Icmpcrulurc between 108 and 125 C.
DUring this process, the stearn is supcr+hcall:d and turns into saturated steam or
vapor. A continuous airflow system bcltins the cooling Of111<: spi<.:cs immediately.
3. TIle product drops through im airlock inlo a St:cond closed chamber where the
cooling prol:CSS is continued with ambient air withom Jhe dcvc10prncnl of
condensation. Dust is fLhered out during the entire process.
proc<''SS, which is viewed as being safe also offers a number ofbcncfits. COntinuous
flow of the produd yields exccl!cnl productivi ty. AU different fo rms of spkcs can oc
treated including whole and ground spices, herbs and ground seasoning blends. 'I'll<:
efficiency of the process oilers the potential for cost savings when compaflxl to both the
gas and wet steam process.
All t hree mcthods of o r bacterial reduction offer bot h benefits and
disadvantages. Irradiation i;; the most effective means of bacterial reduction, however,
there are COtlSlUllCr oonccms about the safety of tile product. Ethylene oxide is a proven
method that has been in usc fo r years but has recently come under the scrutiny of FDA
and the Enviroruncntal Protect io n Agency (EPA). The weI steam process is not quite as
e ffective as irradiation and the COSI is slightly higher, howevcr, it is regarded as a SlIle
method. The dry steam method is highly effective and the newest method available. It
more expensive than irradiation but k ss expensive a process than both cthylene oxide
and weI steam. No one nlCthod is perfect. The selection of thc method to be used is up
10 the individually processor and must meet the needs of the customer.
I. Reineccius, G., Ed., Source Book of Ftav(Jrs. Chapman and Hall , New York, 1994, p.
244.
FLAVOR CHEMl STRY
Chapler 3
The Principal Flavor Components
of Fenugreek (Trigonelia Joenum-graecum L.)
Imre Blank, J innming Lin, Su! phanie De\'aud, J<"unleaux,
and uwrent B. fay
Nestec Limited, Research Center, Vers_chcz_les_Ulanc,
P.O. Box 44, CJI - IOOO t.au sa nne 26, Swil7.erla nd
3- l-I ydroxy-4,5-dimethyl-2(5N)-furanone (sotolone) was established as
the character impact flavor compound of fenugreek on the basis of gas
chromatography-olfactomelry. Solo[one was found to occur
predominantly in the (5.\) enantiomeric form (95%) and to have a
olJCroo value of -19.7%0 About 2-25 ppm sotoJone were determined
in fenugreek of different origins using the isotope dilution assllY
technique. Sotolone was generated in model systems by themlally
induced oxidative deaminalion of 4-hydroxy-L-isoleucine (lUI.) using
different carbonyl compounds. Up to 24 11101% yields were obtained by
boi ling HI L and melhylg[yoxal as react ive -dicarbonyl at pH 5 for 10.
h. Strecker degradali on of HI L was found to be a competi tive react ion
resulting in the formal ion of 3-hydroxy-2-melhy[bulana1. The
of HJL, 3.arnino-4,S-dimelhyl-3.4-dihydro-2(5fl)-furanone, was found
10 be a better precursor of sotoione. It generated about 36 1001%
SOlolone in the presence of methy[glyoxal
f enugreek is the dried seed of 7i-igoneflajoelllllrl-graeclim L. (Fabaceae). The plant
is an annual herb widely cultivated in Medilerranean count ries and Asia (I) The pods
contai n about 10-20 yellowish seeds with an appetil':ing pleasing aroma. Toasted and
ground fenugreek seed is an essenti al ingredient of curry powders and is often mixed
with breadstum. It is used as a seasoning in food products such as pickles, chutneys,
vanilla extracts, artificial maple syrup, Hnd olhers.
Several volatile constituents have been reported in fenugreek (1), mainly
terpenes and fatty acids. However, 110 systematic wOlk has yet been published on
compounds that contribute to the characteristic aroma of fenugreek, 3.Hydroxy-4,S-
di mcthyl-2(5H)-furanone (sololooe) was suggested as an important volatile
constituent offenugreek due to ils seasoni ng-l ike flavor note (3, -I) .
I{) t997 American Chemical SocH:ty
J. IlI.ANK IT AL. Principal Havor of Ft nllgTUIc IJ
SOlo[one is a powerful flavor compound found in several food! and spi,,;es (.')
It contributes to the burnt/sweet note oCaged sake (6), cane sugar (7), and conce (S).
10 Ihe spicy/curry note of lovage (9) and condiments (J(J); M well as tn the
nutty/sweet flavor ofbolryt ized wines (/I) and fl or-sherry wines (/1). The lIavoring
potenlial of soto[one is due 10 its low threshold values, that of 0.02 ngll air (8).
0.3 water (detection/nasal, 10), and 0.036 water (deteclionJl'etronasal. f Jl.
The structural similarity beTween soto[one and 4.hydroxy-L. isolcllcine (111 1.).
Ihe mosl abundant free amino aeid in fenugreek seeds, was pointed oul (/4. fJ). I I
was postulated thai Ihis unusual amino acid could be the pre<:ursor of sotolune in
fenugreek (/5) (Figure I). This hypothesis has rec;ent[y been supponed by the f.1C1
that only the (5S) enantiomer of soto[one was found in fenugreek (16). This is in gond
agreement with the stereochemistry of HIL isolated from fenugreek, i.e. (2S, 31l. '1.\)
(/7). However, these authors failed to discuss possible formation pathways.
OH '1Hl

, 1 2 !!.
o
(lS,3RJ.'l)-HJL (5S) Sotolone
Figure I . STereochemistry of 4-hydroxy-L-i soleucinc (HIL) and sotolone found in the
seeds of fenugreek (Trigo/ref/a jrx!IIum-gracclIlII L).
The purpose of the present sludy was to identify those volatile compounds
which significantly contribute to the seasoning.like note of fenugreek using the
approach of sensory directed chemical analysis. Gas chromatogmphy in combinat ion
with olfactometry and mass speclrometry have been used as key sleps of this
approach (/8, /9). The formation of flavor impact compound(s) was studied in model
systems using the quantification technique Isotope Dilution Amy (10, 2/). The
mechanist ic study was based on a hypothetical pathway proposed for The formation of
sotolone via thermally induced oxidative deamination of HIL (10).
E:lperiOl cntal
Materials. Commercially available fenugreek seeds of different geographical origins
and fenugreek oleoresin were used. Sotolone was from Aldrich and diethy[ 2-methy[-
3-oxobutancdioate, L- isoleueine, Olethylglyo){al (40% in water), phenylglyoxal, 2,3-
butanedione. and 2,3-pentanedione from Fluka, Solvents and other chemicals were of
anal)1ical gradl'l'from Merck.
Synl hesis. 3 -Amino-4, 5-dimethyl-3, 4-dihydro-2(5H}-furanone hydrochloride (ADF)
and 4-hydroxy-L-isoleucine (HIL) were obtained as diaslereomeric mixtures by
photochemical chlorination ofL.-isoleucine (Figure 211), i,e. (3S,4R,5R13S,4R,5S) and
(2S,3R,4R12S,3R,4S), respectively (17, 12, 13). (S,6_
13
C]-3 -Hydroxy-4,S-dimethyl
-- ,
14 SPICFS: HAVOM CIIEMISTHY A."il) ANT.oxmANT
2(5H)-furanone (fUC1]-sotolone) was prepared by condensat ion of diethyl 2-methyJ-
3-oxobutanedioate and (I,2.
IJ
Q-acetaldchyde followed by lactonization and
subsequent decarboxylation under strongly acidic condit ions (Figure 2b) (1U) , The
structures of the synthesized compounds were verified by elemental analysis. mass
spectromtlry (MS), and nuclear magnetic resonance spectroscopy measurements CH-
NMR, IlC_NMR) (14).
Gas Chromfilogra phyOlfaclometry. GC-O was performed on a Carlo Erba (Mega
2) equipped Wilh a cold on-column inject or, FID and II sniffing-pon. Fused sil ica
capillary columns of medium (DB-OV 1701) and high polarity (DB-tFAP) were used
as previously described (9), both 30 m x 0.32 mm with a film thickness of 0.25 I-Im.
The temperature program was: 50"e (2 min), 6C/min 10 180"C, IO"C/min to 240"C
(10 min). Linear retention indices (RJ) were calculated according 10 van den 0001 and
Kratz (25). The sensory significance of each odorant was evaluated and expressed as
the flavour dilution (FO) factor (/8).
M:l$s Spectromr try
QUll litative Analysis_ Elet.:tron impact (EI) and positive chemical ionisation
(PCI, ammonia) mass spectra were obtai ned on a Finnigan MAT 8430 mass
spectrometer MS-EI were gener-.ded at 70 eV and MS-CI at 150 eV with ammonia
as the reagent gas Non-volatile samples were directly introduced into the ion source
held at 200 ec. Volatil e component s were introduced via a Hewl ett-I'ackard HP-S890
gas chromatograph (GC-MS) usi ng a cold on-column injection. DB-FFAP fused silica
capillary columns were used (30 m x 0.25 mm, film thickness 0 25 The carrier
gas was helium (90 kPa). The temperature program was: SOec (2 min), 4C/min to
180"C, lOoCJmin to 240"C (10 min).
Quantitative Aml lysis. Sotolone was quant ified by isotope dilution assay
using [
13
C:zl-sotolone as internal standard (10, 2.f). Quantification experiment s were
performed wi th a HP-597 I GCMS using the following conditions: DB-Wax capi llary
column (30 m x 0.25 111m, film thickness 0.25 1-1111) ; carrier gas: helium (100 kPa);
spli tl ess injection (250"C); temperature program: 20
e
C (0.5 min), 300C/min 10 lOO"C,
4"Cllllin to 145"C, 70C/min to 220"C (10 min); El ionisation at 70 tV; selected ion
monitoring (SIM) of sotolone (mh 128) and the labeled internal standard (mil 130).
The concemration of sotolone was calculated frolll Ihe peak areas using II calibration
factor of 1. 1 (Figure 3). A good linearity was found in the cOllcentration range 3- 150
)1g1ml. All samples were injected twice.
Fast At om Bombul'dlll f nt (FA8-l\1S). This was applied to study the
lactonization of IIi L 10 ADF. FAI3-MS was performed on a Finnigan MAT 8430
double focusing mass spectrometer. FAB ionisation was carried out with a saddle-
field atom gun (Ion Tech, Teddingt on, UK) which was operated at 2 rnA and 7-8 kV
with xenon. Glycerol was used as matrix. The positive ions at mlz 130 (protonated
molecular ion of AOF) and 148 (protonated molecular ion of HIL) were recorded.
lsotOll e Rati o Mass Spectrometry (GC-IIU\'I S), This was performed with a
Finnigan MAT delta S isolOpe MS coupled on-line with a Varian 3400 GC via a
combustion interface. Isotope ratios were expressed as Ii- values [%oj versus the 1'01'
3. Il IANK 1\1_
Prilll:ipal COl1lptmtn'.f v/ Ftllll/:reek
15
A
Lel,hv
--.
2. Hydrolysis

"
B
n 0

OE<
'CH;CHO

EtOH
AOF
HIL
J=(
o
F!gure 2. Synthesis of 4-hydroxy. L-isoh:ucine (}-IIL) and 3-amino-4,5-dimct hyl _3 4-
hydrochloride (AoF) (A), and [S, 6-13Cj-3-hydro .... y-4:5_
dlmethyl-2(5H)-furanonc ([ nC
2
1-soIOJone, used as internal standard) (D)
ISOOOO
I,.
"
Ion mh 128
0
H3C OH

sototone
100000
):0(
0
H3C o ..


SOOOO
' .00 W.OO II (1)
11.00 lJ.OO 14. 00
0 80000
" 0
Ion mh 130
ii
H]C OH
GOOOO
labeled sotolone


0

'0000

20000
9.00 1000 11.00 12.00

JJ.OO 14.00
Retention time (min.)
:igure 3. Quantificati on of sotoJone by isotope di lution IIssay using lJCl-sotololle as
mternal standftrd. The ions mlz 128 Mnd mlz ))0 wer, ',"o,d"d by GC '10 .
...... -IV opcralmg
III the selected Ion monitoring mode. (Adapted from ref. 24.)
SI' [CES: nA\'OK ANIl ANTIOXIIl" ." T
siandani having II ["CJ/I"C) mlio uf 0.011237 fur CO! yielded by
combusliun of C;lCO) (reedee Belemnite) The GC ellui llred with n 013-
FFAI' fused silica capillary column (30 m x 0 25 mrn, film thickness 0.25 J.m) using
hdium as carrier gas (100 kPa) and Ihe split injection mode (220C). The tempcralUre
pmgrnm was the same liS mentioned above.
S" lI lplc l'I"fl' 3raii oli
Qualit:.l \i lr for GC-Olraflomrlry. The gr(lund fenugreek seeds
{(OO 1;) wcre extracted wilh dicthyl elher (EtlO. 200 1111) by stirring the for
h The S(livent was separated and the extraction W;lS c(lntinlled wit h
,;"Ivent (200 1111) overnight. The extracts were combined. filtered and cuncentrat ed
(SO ml) llil Yigrellx column. Non-volatile by-products were separated by vacuum
(YS). i.e. \Tapping volatiles under high vacuum conditions mbaT)
illto 1raps cuoled with liquid N] (2/). The sample was introduced drop-by-drop into
the vaCU1I1Il to increase the yields. Then. EIlO (30 ml) was added to the
residlle and lite procedure was repeated. The condensates of the traps were
cOlllbilled and concentrated to I IllI on a Yigreux column.
Qualit ative Aualysis for GC- IRMS. Sotolone W;lS isolated from fenugreek
oit:oresin, An aqueous solution of the oleoresin (35.5 gllOO ml) was extracted wilh
containing 5 % etlmMI (5 x 100 011) . The organic phase was concentrated to
100 mland extracted with NalCO
I
(0.5 mol/L. 3 x 100 ml). After ;lcidification of the
aqueous phase to pH 2 (HC1. S moUL). sotolone was re-extracted with EtlO (4 x 100
ml). The sample WitS dried O\'er anhydrous Na!SO. and concentrated to 20 ml
Sotololle was separated from nonvolatile by-products by sublimation in vacuum (10.
1
mbar) (1/). The condensates of the traps were collected and concentrated 10 2 mi .
Qualil:Hh'e Analysis ror FAR-MS. HIL was dissolved in phosphate buITers
(0. 1 moUL) with different pH values (pH 3.0, 4.0. 5_0, 6.0, and 7.0). The soilit ions
were boiled for I h in scaled glass tubes. The samples were rapidly cooled down and
directly analysed by FAD-MS.
QUlI. llt ila l ive Analysis in Ftllugreek Serds (10). The material (5-10 g) was
homogenized in wateT"ethanol (50 ml , 95:5). [
IJ
C
2
]-Sotolone (10-20 was added
as internal standard and the suspension was stirred for 30 min. After centrifugation
(30 mi n. 10000 rpm). the supernatant was extracted with EtlO. The acidi c
components were isolated with NalCOJ (0.5 moVL). The aqueous solution was
acidified to pH J (5 mollL HCI) and re-extracted with EtlO Finall}'. the organic layer
was washed with saturated NaCI solut ion. dried over Na2S04 and concentratcd 10
ilbOiIl 0.2 ml using a Yigreux column and micro-distillation.
Quanlilati vt AII l_l ysis in Model (14). I-IIL (2-10 IIIg) and
ADFHCI (2-10 mg) were each dissolved in a phosphate buffer (0. 1 molll.... pH 5.0).
ACter adding the carbonyl reactant. the solution was boiled for I h in a sealed glass
lube. The molar ratio of precursor to carbonyl was 1:10. Water and the internal
standard ([ 13C
l
]-sotolone) were addcd to the cooled reaction mixlure. The sample
was saturated with NaCl and the pH adjusted to 4 (Hel, I moVL). Sotolone was
c."macted from the rcaclion mixture with Et20 for 8 n. The extract was dried
(Na2S04) and concent rated to I 011. All experiments were performed in duplicate.
3. RlANK ET Principal Hal'flr Cmnpnntnl.\ of Fenugreek 17
Resull s 311 d Diu uuion
Aroma Compusi li on of Fr- nugruk. Liquid.liquid ex'raetion using diethyl ether
resulted in an aroma exlract which represented the characteristic note of the origmal
product. Ihal i.e. is seasoning-like, spicy, herbaceous, and fenugreek-like. The
representativeness of the SIImples before and after purification was checked by
sensory evaluation.
GC-Olf;lctometry (GC-O) was used 10 detect Ihe odor-aclive
present in the aroma extract orrenugreek. It is II simple but effective mel hod to select
those volatiles which contribute to Ihe overall flavor (18,19). On the basis of Ge-o.
seventeen odorants were detected in the original aroma extract (Figure 4). An an.ma
extract dilution analysis (AEDA) was applied to eiassify the aroma into
three groups having different sensory relevance. "high"' (no. 17), "medium" (nos. ,t, 6,
16). and "background" (nos. 1-3, S, 7-15). Hence. identification experiments were
focused on the odorants belonging to the firs' two groups. Compound no. 17 was of
particular interest because of its high FDfactor lind seasoning-like nole.
n
FD-factor (2 )
..
- - - - .. - -- - - - --- - -- - - - - - .. --- - -- - - -- - -- - - - -. - - - - - - - - - - -- 17 -
14
12
high sellst)ly relcI'flnce
10 - - -- -- - ----- --- --- .. - , . - . -- .--- - - .- ... ------- --- . .. - -

6
",eJi,il1, S"'"S0ry rde.unce
- --- -- _. - ---- .. _- --_ ... -

4 I.
4 - backgrOlmd nm",
--
"Ie TlT
,
2

-,
BOO 1000 1200 1400 1600 1BOO 2000 2200
(Retention index, FFAP)
Figure 4. of an aroma extract obtained Crom fenugreek seeds.
The chemical (Figure 5) of the odorants were mainly elucidated by
GC-MS (Table I). Compound no. 17 was identifi ed as sotolone (Figure 6A). Sotolone
is likely the character impact compound of fenugreek as indicated by the high FD-
factor of 2". correspondingly low sensory threshold and characteristic aroma note.
,.
18 SPICES: ruvolt ANI) ANTIOXIDANT
Sowlone was detected by GC-O even after more than IOOOOfold dilution of the
original aroma c){lraCI. It s FD-faclor was than those of acet ic acid
(FD'" 2
1
), (Z)-I ,5-ocladiene-3-one (FD- 2), and 3-ami no-4,5-dimethyl-3, 4-dihydro_
2(5H)-furanonc (FO" 21) which belong to Ihe group wilh medium sensory relevance
(Figure 4).
The FO-faclOrs of the remaining compounds were lower. They mosl likely
cont ribut e to the background of Ihe fenugreek [Iavor. These odorants art short chai n
falty aci ds (nos. 10, 11 , 14), lipid degradati on product s (nos. 3, 12), and al kylaled
mel hoxypyrazines (nos. 5, 7). All odorants listed in Table I, except nos. 15 and 17,
were identified for the first time as constituents of fenugreek aroma.
Table I. Odor-a ctive Compounds Detected in a n Aroma Extract or Fenugreek
Seedli 011 the Basis of GC-O (II is the IHllllber of dilution steps)
No Compound Relention index Arol1la qlwliry FD-filClor
FFAP OV-170f (Ge-O) r1")
I Diacety[' 990
'"
BUllery
2 Unknown 1020 , d. Fruity. metallic 3
3 1296 106> Mushroom-like

4 (l)- I,S-Octadicne-3-oroc 1369 1090 Metallic, tteraniu1\l .li kc
,
,
3-lsopropyl-2 Inethoxypyr:l.7.,ue
b
14211 1145 Roasl)', earthy 3
6 aCid' 1445
'"
Acidic, pUlIgent 7
7 3 lsobuty-2- 1518 1235 Roasty. plprikalike 3
8 Linalool ' 1543 11 93 Flowery 4
9
U""""".
1554 n.d. Sulfury, roasty 3
to Butanoic acid' 1624 968 S .....eaty, ranci d 3
II Isovaleric acid' 1663 102S S .....eaty, rancid 4
12
U""""".
1760 n.d. Fatty 3
13 UnknoWll 1823 n.d. Flowery, citrus-like 3
14 Caproic acid' 1845 1! 63 Musty 3
"
Eugenol' 2163 I SIlO Spiq 4
16 3-Amil\O-'l,5-dimcthyl- 21 90 , d. ScaSOfling.like
,
3, 4.dihydro-2(5J f)-furanone'
17 SOIolone' 2210 1350 Seasoning. like 14
Idcll1ifical ion by comparison wilh the refcrCllCe compound on ,I>< basis of retention
indices, aron18 quality and GC-MS.
Identification by comparison with the reference compound OIl
""
basis of rdcntion
indices and aroma quality. The amounts were too small for verification by GCMS.
n.d. detennincd
Terpenes and terpenoid compounds do not playa maj or rolc. Only linal ool
was detected. by GCO (no. 8). Most of the Icrpenes ident ified by GCMS wefe
odorless at the concentration prescnt in the arOma extract, i.e. 0.- and p-pi nene,
sabinene, 3-carene, menthol, p-Ierpineol, cineol, anethol, p-terpinyl acetate, 11'
3. BlANK I:,'T AI_
Principal (if flIIlgruk
19
acetat e, carvone, and several sesqui terpenes. Further volatiles of low or
no sensory relevance were I-pentano!, l . hexanoJ, 2-rnethyl -2.butcnc-l.ol, 2-methyl.
2-butenal, 2-pentylfuran, formic acid, propanoic acid, and fu rther longer chuin f;nty
acids, -y-butyrolactone and several 5-alkylated -y- Iactones, )-amino-4,5-dilllethyl_
2(3}f)-fllnmone, and others.
0
)lOti
6
o


8
4:
17
0

,
DOH
II
"
.hIll
o 0
16
0

3
o

"
,
7
Fi gure 5. Sensory relevant compounds idemified in the cxt .... ct uf Ji.m,!-',cc:'
The numbers correspond to those in Table I.
20 SI'ICES: t"IAVOR Ult:MISTJt Y ANI) ANlIOXl llANT
100")
A
'"
:. 80 ":
.
" c 50
1
ID
c 29
ID
40 -

ro
20 Lj
0;
0:
72

113

'-..
43
55
83
128
20 40 60 BO 100 120 140
mass I charge
100
70
"
B
0
80
b"Jl' .
"
o 0
c 60
..
C
ID
"
85

56
-" 20 43
ID
,.1.1.
129
0:

20 40 60 BO 100 120 140
mass I charge
Figure 6. Mass spectra of3-hydroxy-4,5-dimelhyl-2(5H)-fur3none (sololone) (A) and
of :; -ami no-4, 5-d ime! hyl- 3, 4-dihydro-2(51f)-furanone (8).
Sterenisn meric Ch:ml. cterisat ion of 50101001'. In fenugreek seeds, sotolonc occurs
in the (50) enantiorneric form (95%). This is in good agreement with
the data reported by Sauvaire et aL (16)
The OI 'C values of natural and 5ynthesi7.ed sotolone were determined by
isotope ratio mass spectrometry (IRMS) using the GC combustion technique (26), As
shown in II, natural sololone was characterised by a OllCroa value of -19.7%0.
On the contrary, a racemic mi:dure of synthesized 50tolone showed a significantly
higher value (-23.3%0).
Sotolone isolated from a commercial product revealed a ,sl lCPOII value of
-22.3%0 indicating a mixture of natural and synthetic compounds. This was confirmed
by chirospeciflc GC analysis resulting in a ratio of 65:35 for (5S):(SR). Hence, both
3. ntANK ET AI . Principal Comp(mellis of Fenugreek
21
GCIRMS and chirospeciflc GC suggest that about 213 of the sotolone found in rhe
conunercially available liquid seasoning was contaminated with synthetic sOlolonc
Table II. slle Values (%e PDP) of Solo lone of Synthdic and N:lIural Origin
SQ/o/Olle
Synthetic
Natural (fenugreek oleoresin)
Natural (fenugreek seeds)
Liquid seasoning (commercial)
23.30 0. 20
-19.69O.20
-19.75 0.20
-22.28 0.20
Quantitalion or SotolOllf in Fenugreek. The concent ralion of sololone
determined by isotope dilution assay (10, 24) using labelled SOlolone as internal
standard. As shown in Table !II, the typical concenlralion range of sotolone was
about 3-12 mglkg fenugreek seeds. However, the amouots depend on the
geographical origio. Fenugreek seeds from Egypt smelled more intensely and, in
agreement with that, more sotolon was found in these samples. Some Trigoncl/Q
species do not contain any 501010ne as reported by Sauvaire et a!. (11i) and,
consequcl1tly, they lack the characteristic seasoning-like note.
The sensory relevance of 50tolone is due to ils low threshold value of 0.3
water (10). In fenugreek seeds, the concentration of sotolone is usually at least
300{) times higher than its threshold, thus indicating the sensory impact of sololooe to
the overall flavor of fenugreek and products fenugrtck . As shown io Table
III, hi gh amounts of 501010ne were found in curry powder and some commercial
liquid seasonings
Table HI. Co ncentration of 50tolone in Fenugreek and Products Cont a ining
' <enugreek
Sample
Fenugreek, seed (Egypt, 1985)
Fenugreek, seed (Egypt, 1995)
Fenugreek, seed (Australia, 1991)
Fenugreek, seed (lOrance, 1996)
Fenugreek, seed (India, 1996)
Fenugreek, seed (Turkey, 1995)
Curry powder (containing fenugreek)
Liquid seasoning A
Liquid seasoning B
S%lone /mglkgj
25.1
12.2
4.2
3.3
5.1
3.4
39.7
1.4
88.9
I
!
I
22 SPI Cf.:S: nAVOk CIIi':MI!o.'TJ.l.Y ANI) ANTlOXIUMH l 'IWf'Elnn:s
Formati on of So to lone from !' ,'tcu rsol'S
Pret ursorJ Present in Ihe Allueous EXlracls of Fenugreek Seeds. The
presence of g!ycosidically bound sUlolone was examined by treat ing an aqueous
extract of fenugreek wi th ex- and p-glucosidases. As shown in Table IV, this
enzymatic treatment did 1101 si!,lni li cantly enhance the amount s ' of 5010100e compared
10 the reference sample. On the other hand, boiling of the exlrHcl under acidic
conditions (pH 2.4) for I h led to a more than 10 fold increase.
These trials indicate the presence of precursors in fenugreek that can be
Iransfomled to sotolone using specific condi tions, particularly an acidic medium and
heat treatment. 4.1!ydroxy-L.isoleucine, known to be a characteristic amino acid of
fenugreek (14, J 5), may be Ofle of these precursors.
Table IV. Formation of So to lone from Pre( ursors Present in the Aqueous
[:ltra cts of Fenugreek Seell s.
Addition (0 the Reaclioll cOIlfiiliollS Solo/Qlle
aqueous extrocl pH Trel I {mi'" {mglkg]
None 6.8 25 120 18.9
a-Glucosidase (200 U) 6.8 25 120 19.2
(200 U) 6. ' 25 120 20.2
None 2.4 100 60 252.3
SOURCE: Adapted from ref 10.
Formation of Solololl t frolll 4-lI ydro:ly-L-isoleudne (II IL). The formation
of sotolone was studied in phosphatebuftered model systems (pH 5.0) by react ing
HIL with different mono- and a-dicarbonyl compounds at 100"C for I h. Rot h 2,3
but anedione and 2,3.pentanediont rormed only low amounts of sotolont (Table V)
Higher yields were achi eved with the u. \.:eloaldehydes methylglyoxal (7.4 mol%) and
phenylglyoxal (2.5 mol"!. ), producing about 70-200 times more sotolone than Ihe
corresponding reaction with the u-dikefones. Monocarbonyl compounds, such as
propionaldehyde and phenylacetaldehyde, generated less than 0 I molo/. 50tololle (U ,
27).
Formation or SOlolone f .. olll 3-Amino-4,S-dimcthy!-l,4. dihydro-2(S/J)-
furanone (ADF). The emciency of ADF, the lactone of HIL. to generate sotolone
was tested in the same modd system as described above. As shown in Table V,
signifi cantly higher amount s of sotolone were generated from ADF as compared to
!ilL Using methyl glyoxal, the yields were increased from 64 Ilg (7.4 mol%) to 274
(35.9 mol%), thus indicating ADF to be a more efficient precursor than the free
anlino acid (HIL).
3. HLANK t.T At-
PrillCipai Flayor Comptmtlll.f of ,.',rwgn:t!k
Table V. Fornllllion of SOlolonc from the ,'recurJo rs 4. lI ydroxy. I,,..jsolruci ne
(H and 3_Amino_4,S_dimt lhyl_3,4_d ihydro- 2(SII)_fur:lII one (A I)F)b
in Fenugreek Seeds.
(l.DiC(lf"bollyf
",ccllrxo,c StJIO/OIle<i 1"1.:'"
JIlL AJ)F {jlg/mg filL] {mo/%j
2,J Outanedione
,
0.34 0.03 <0 I
2,3.Pentanedione
,
0.30 0.03 <0. 1
Methylglyoxal
e +
64.2 0.3 74
Phenylglyoxal
+
22.2 0.3 2S
2, 3Pentanedione
+ 5.4 0.3 0.7
Methylglyoxal
+ 274.4 :t 3.4 35.9
'Control experiment (without a.carbonyl) yielded less than 0.01 molo/. sotolonc.
b Cont rol experiment (without a-carbonyl) yielded 0.03 11101% sotolone.
c The molar ratio of precursor to a.dicarbonyl was I : 10.
d Data are means of at least twO experiments, each injected twict .
e Control experiment (without HIL) yielded 0.07 sotolont.
Mechan ism of the FOl"Ill;lt ion of Sotol one (J<'igure 7, 1) ;lt hwny A). The
data reported above confinn the hypot hesis of the formation by thermally
induced oxidative deami nation of Hl L (10) Acid camlyzed cyciLzalion of HIL leads to
the corresponding lactone (ADF) which rtacts with an a-dicarbonyl (e.g.
methylglyoxal) to form a SchilT base (pathway A) and subsequent.
hydrolysis gives rise to sotolone. The data show that a-<hcarbonyls are cap.ablt ot
generating sotolone from both HIL and the lactone ADF. the
emcient in producing sotolone. Furthermore, the relatively low Yields achieved With
HlL indicate an alternative degrRdation pathway.
SO'ecker () egndation of n i L as ;) Comll eti tive (Figure ?,
Ilathway B). The lower yitlds obtained wi lh Hi t might be a pamal
Strecker degradation of the amino acid I ilL in the of an Q.dl carbonyl,
e,g. Illt:thylglyoxal. As shown in Fi gure 7, the reacti on of 1IlL
methylglyoxal results in a SchilT base which may tither cychze or decomp:Jse
decarboxylation. 11le Strecker aldehyde of IIIL .. IS
released by hydrolysis and this compound was tentatIVely Ident ,hed by GCMS ( . ./) as
a mixture of diastereomers.
In the sample based on tilL and methylglyoxal, Ihe ratio of St recker aldehyde
to sotolont was about 1:2 at pH 5 (Table Vl). Consequently, the Slr(!(:ker
degradation of liIL is a competit ive reaction to Ihe format ion of 1 II
I I of
Strecker aldehyde were detecled in Ihe sample conl ;\IIuns Ihe la<:I" "I'
on y races .' ... . . I .
ADF
bout 50 times less than in the reactIon wllh HI!. I he lurlll;ltllUl "t ",In " lit
, I.e. a
from ADF is the favoured reaction. most likely due to the hluded )" ''''')'
HAVIlK .... wn{Y ANI> ANTlnXJllANT
Tahl!' VI. Jt!lfio or Solololl e 10 3- lI ydrO:fY
M
2-1II!'lhylbu l:lllal l'Or lll ed in tll otlrl
Based 011 Mr_tbylglyoul and Ih e Prrcursors 4- lI ydro:fy- l...- isoleli dne
(III L) !llld 3
w
Amino-4,S-dimclhyl-3,4-dihydro-2(511)-rurall olle (A 1>,.' )
I,ff IIlL AJ)F
J I : 17 , 50
5 1.2 ' 50
b , 2 : ]0
7 I I.S : 40
____ '\---(NII,
'\ ,)...j,
H2
0
0 0
IlI L
o
/y"
11
2
0 ..
MG 0

.-.. ) N"0
II,
o 0 .. ./
11,-., N=CI/ - {; -CII
,
o u
,"
t-. CO, 1l
N-CH=C -(;11
'r---I' I
/-. ..... A, mJ
o 0
It
2
0

o 0
figure 7. Formati on of 5Ololone (palhway A) and 3.hydroxy-2.methylbutanal
(pathway 8 ) from 4.hydroxy.L-isoleucine (HIL) and 3-amino-4,S-dimclhyl'],4.
di hydro2(5/1).furanone (AOF) using mcthylg1yoxal (MG) as carbonyl reac:tant .
3. tHANK Kf A I M PrinciJl(JI HMor Compollents of f enugredc 25
Innuenee or Ihe pH. II is well known that Doth lactooiz.ation and the format ion of Ihe
Schiff base strongly depend on Ihe pH of the reaction medium. As shown in Figure 8,
the lactonlzation step (I-liL ADF) was favored under acidic: oondi tions and the
yields were about 50 .;. at pH ], but only 10 % at pH 5. On the other hand, Schiff
bases are readily formed under neutral and sl ightly basic condi ti ons.
The reactivity of the carbonyl c:ompound is another crucial parameter for the
formation of the Sc:hiff base. o.- Ketoaldehydes are much more effic:ient than a-
diketones (Table IV) and a-keto acids, whic:h generate only low amounts of sotolone
from HIT- (27).
7 6 5

,
}-igure 8_ Lac:loni7,..i1ti on of 4.hydroxy- Lisoleudne (!-lIT-) as a function of pH.
To fllld the optimum pH, methylglyoxal was reacted with HlL and the lactone AD!',
respectively. The reac:lion of methylglyo)(at and HIL was favoured at pH 5 whic:h
apparenlly is the best c:ompromise belween the lactonizalion slCP and the reac:livily of
the amino group to form the Schiff base (Figure 7). Once Ihe lactone (ADF)
formed the amino-c:arbonyl reac:lion was favoured al pH 56. In general, Ihe yiel ds
obtained with the lactone were significantly higher compared to the amino acid,
partic:ularly al pH 6 (40_2 molo/. ).
The reaction of HJL and methylglyoxal at pH 5 performed in waler yielded 2.8
mol% sotolone c:ompared to 7.4 mol% when using the phosphate buffered system.
This suggests a catalytic: effeC:1 of phosphate on the formation of sotolone from HIL
2. SI'I C.:S, FlAVOR CII t:MISTRY AND AN .... O)(IUMIIT l'IW"t: RTIES
!o:(ftcl of Reaclion Temperalufe Time. The sotolone yiel ds from tilL and
mcthylglyoxal were slrongly dependent on both reaction lime and temperature.
Significant amounts were generated above 70C Crable VJJ ). At II constant
temperature of 100C, the yield of sotorone continuously increased over a period of
to h, wilh no significant increase thereafter (Table VUI). About 23 to 210
sotolone were gcneraled from 30 min to ]0 h which 2.7 mol% and
23.8 rnol%, respectively.
Methylglyoxal reacted wilh 1-fTL al SOC for 48 h resulted in 3.8 mol%
5010100c, thus indicating that II long reaction lime and mild reaction conditions arc
also suitable for generating significant amounts of sotolone. Therefore, hot climatic
conditions might favour the fomlal ion of so to lone from HIL
Table vn. for mati on of Solo lone frolll n i L as a FUllcl ion of lhe React ion
TemllcrMur e Using !\Iet hylglyoxrli as Carbonyl Reacta nt
Temperature Sol()lolI'.b Yield
,
b

b
fOCI HILI {mol%{
50 0.31 0.01 0.03
60 1.150.03 0. 13
70 4.00 0 05 0.46
80 12.5 0.5 1.44
90 27.9 1. 8 3.2
100 64.2 OA 7A
Reaction condi tions: phosphate buffer (0.1 mol/I., pH 5.D), 501 OOC, \ h.
Data are means of at least two experiments, each injected twice.
Trible VIII. Fon n:Hion of Sotolone frolll H[L as Affected by the Reaction Ti me
Using Melhylglyoxal lU C;lrbunyl lleaclant :t
Time SOloloneb Yield
{II] Ipglmg HlL] {mul%]
0.5 23.4 0.8 2.7
1 64.2 0. 4 7A
2 102./ 4,5 11 .7
5 170,25,1 19.5
10 206.7 10 23.8
15 208.3 6.4 24.0
24 229.9 5.) 26.4
Reaction conditions: phosphate buffer (0. I moVL, pH 5.0), lOO"C, 0.524 h.
Data are means of at least two experiment s, each injected twice.
3. Bl ANK t>:1 AI . Principtli "lavQr Cvmponents lif f'tnu/;rtek
Concl usion
The role of as a character impact compound of fenugreek was
corroborated and its formation from 4.hydroxyL-isoleueinc (IIIL) via thermally
induced oxidative dcamination Wil S The lactone of !-IlL, 3-atnino4.5-
dimcthyl. 3,4. dihydro.2(5fl).furanone (ADF), was found 10 be lI. better precursor 11ulI\
the acid. a-Ketoaldehydes wcre more effective in generat ing sotalone frum
both IIIL and ADF than adiketoncs. The reactivity of the dicarbonyl and the
lactonization step are important parameters, particularly for the formation of the
Schill' base. The transformation yields from IUL into sototone greatly depend on the
reaction conditions., such as temperature, time, pH and amount of dicarbonyl Iligh
amounts ofsot olone were obtained by boiling methytgtyoxal with HJL for 10 h:lt pI-!
5 (24 mulOV.). Even better results (40 mol%) were achieved using ADF as precursor (I
h, pH 6), most likely due to inhibition of the Strecker degradation by the blocked
carboxyl group,
Acknuwlcdgmeul s
We are gratt-ful to V Krebs for expert techni cal assistance, Dr. H. Schierbeek for
performing the GC-IRlvIS and Dr. E. I' rior for cri tically reading the
manuscript . We also thank Prof A. Mosandl, Universi ty of Frankfurt, Germany, for
the chi rospecifle GC analysis.
Li terat ure Cited
I.
2
3.
4.
5.
6.
7.

9.
10.
Lew;s, V,S, (Ed.), tlml Herbs/ur Ihl! Pucxl 'l/dusfry; Food Trade Press.
Orpington, England, 1984; pp. 141 - \42.
Girardon, 1' .; Ilessiere, 1.M., Baccou, J C.; Sauvaire, Y. PIal/faMed., 51.
533-534
Rijkens, F.; Doelens, H. In Proc. 1111. S)lmp. Aruma Maarse, r I. ,
Groenen, 1) .1 ., Eds.; Pudoc: Wageningen, The Net herlauds, 1915; pp. 203220
Girardon, P.; Sauvaire, V.; Baccou, J -C,; Bessiere, J -M. I.eben.",". /V,.\ .....
Tedmol. 1986, 19, 44-46.
Kobayashi, A. In FI<lI"Or 01,,",i.II1)'. Trelllil- and IJew:lopmellf.\;. Teranishl .. R ,
BUllery, R.G., Shahidi, E, Eds.; ACS Symp. SeT. Amencan Ch,..nllclil
Society: Washington, DC, 19119; pp 49-59.
Takahashi, K,; Tadenuma, M. ; SalO, S, Agr. /Jiol. ('lwlIl. 1'176, ./0, 325330
TokitunlO, Y.: Kobayashi, A. , T,: MurHki. S. /'/"11('. Japall A('(/d
1980, 56/J, 451-462.
Blank, I.; Sen, A.; Grosch W. Z. Leb"mm. U'Jlers. 1992, 195,23')-245
Blank, I ; Schiebtrle, P. Plav. hagr. J. 1993, 8, 191195.
Blank, I., Schieberle, P.: Grosch W. III I'rogrtss ill Plltllf)Jlr a",1 Prenll'sur
Sflldie.f; Schreier, P.; Winterhalter, P., Eds.; Allur I'ubl.: WhealOn, USA. 19'.13.
pp. 103.109.
"
SPICt:S, ... .AVOR I.: IIEMISTIl\' ANI) ANTl OXILlANT l'IW,' ,;KTlE.-;
II Masuda, M , Qkawa, E., Nishimura, K.; Yunomc, H A.I!r. /J/(}f. Chem. 19114,
48,2707-271 0.
12. Martin, B.; Eti evarll, P.x. ; Le Quere, J.L., Sehlieh, P. J. A;:ric. Food ('//(.",.
1992,40,475-478,
lJ. Wild , H. Cirelli. Ind., 1988, 580-SS6.
14. I'owden, L. ; Pratt, H.M.; Smi th, A PhYfochemistry 1973,12, 170117 11.
Sauvaire, Y.; Girardon, P.; BaceQu, l e .; Risteruoci, A.M. I'IIYffldll.'misllY
1984, }1, 479-486.
If> Sauvaire, Y.; Srenae, P.; Guichard, E.; Fournier, N. In FOllr/h Imemllfifmal
Works/lOll all Seeds, VO/lllne 1; Come. D.; Corbineau, F., Eds., ASFIS. Paris,
France, 1993; PI'. 201-206.
17. Alcock, N.W., Crout, 0 H.; Gregorio. M.V.M., Lee, E.; Pike, G , Samuel, C J.
Phyfochemistry 1989, 28, 1835 184 J.
ll! , Grosch, W. Trends F()()(/ Sci. Teehl/ol. 1993, 4, 68-73,
19, Acree, T.E. III F/OI'(Ir Science. alld Acree,
T E.: R. , Eds : Ameri can Chemi cal Society. DC, 1993;
pp. 1-18.
20. Schieberle, P. : Grosch, W. J. Agrie. Food Chem. 1987,35,252-257.
21. Sen, A., Laskawy, G ; Schieberle, P.: Grosch, W. J. Agrie. Food Chem. 199 1,
39, 757-759.
22. Faulstich, H., [}QlIi ng, J ; Michl, K.: Wieland T. Liebigl AIIII. Chem. 1973,560-
'6'
23. Hasan, M. In Nelli Trellds ill Natllral Products Chemi.flry. Rahman, A.: Le
QUesnc, P.W., F..ds.; Studies in Organic Chemistry, Vol. 16, El sevier Sci r ubl.
The 1986; pp. 123-14 1.
2... Blank, I. ; Lin, J., Furncaux, R, Welti, D,H., Fay, LB. J. Agrle. Frw Chem.
1996,44, (i n press).
25. vll n def 0001, H.: Kratz, r . J . Chmll/alogr. 1963, I J, 463-471.
26. Bruche, G., Dietrich, A.; Mosandl, A. Z. Lebcmm. Ulller.I. ,"'or.feh. 1995,201,
249-252.
27 Blank, I. ; Lin, J. ; fay, L.n .; FUllleaux, R. In 9.5. Analysis -
Precllrsor Siudies fjiolec/molagy Etievant, X.; Schreier, r ., Eds.; Lcs
CoHoques no. 75, lNRA, Paris, 1995; pp. 385-J88.
Chapter 4
Vanilla
Da phna lIavkin-Frc nke l a Dd Ruth Dorn
David Michael & Company, Inc., l0801 Decatur Road,
Philadelphi a, PA 19154
Vanilla is the most widely used flavor in the food and confectionery induf>1ries.
Worldwide annual consumption was 1900 tons in 1995, witb 1400 Ions
imported to tbe USA aloue. . . .
Vanilla flavor is extracted from the cU(ed bean of Vamlla p/allijo/m, a
member of the orchid family originated in Mexico. A small.mount of beans is
produced from yet another species - Yani/Ja lahiteflsis. Beans (pod-like fruit)
arc produced after 4-5 years of cultivation. Fruit maturation occurs after 10
months. The characteristic flavor and aroma de-'elnps in the fruit after a
process called lasting an addit ional 3-6 mODth.s. We
tissue culture fo r vanilla, identified major components w the blOS}lIthetlC
pathway ofvanjUin and found tbat tissue culture extract contains major
components of vanilla fla vor present in the bean. Hence, plant t issue culture
may be an altemative method for vanilla fl avor production, for f>1 udying the
vanillin biosyntbetic pathway and for micropropagal ion.
With a bett er unden"tanding and ability 10 manipulate the biosYl1theti c
pat hway and to regulate cultivar quality and flowering, we lUay control
production and consistent quality of the vanilla beans.
I. Agronomic Prod uction or Va nill a
1J1troduct ion, Vanilla is tbe world's IllOSl popular flavor. It s fruit y, flora l fragrance
combined with a deep, aronatic body makes it unique and universally favored.
Vanilla is an epipbytic orchid native oftbe tropical region of Mexico. TIle
flavoring material is obtained from dry cured pod-like frui ts commercially called
"beans". The generic name Vanilla is derived from the Spanish " vauillia", a
diminutive of\.o;1Io, a pod. Its species name, plani/o/io, refers to the broad, fl at leaf
of the plant.
VaniUa became kllOW'tl in Europe following Cortexs conquest of the Aztec
kingdom in 15 19. Many centuries earlier, vanilla was. source of{l .voring and used
(D 1997 American Chcmical Society
30 SI'ICES: HAVON; l: II EMISTKY ANI} ANTIUXLIlANT PROI'EIlTI I':';
by the Aztec emperors who used vanilla \0 flavor a cocoa drink, which in the present
day is hot chocolate. The Indians called vanilla or black flower (6). The
Spanish took vanilla back to their homeland and, shortly after, began manufacturing
chocolate with vanilla flavoring. In England, Hugh Morgan recOllllllended that vanilla
flavoring should be used with chocolate to serve Queen Elizabeth I, [",Uowing the
example oftbe Aztecs.
For more than three centuries after Henlan Cortes lived, Mc.'dco was Ihe only
vanilla producing country in the world. Many attempts were made to grow the plant
in Oilier tTopical count ries but these efforts failed since the plant or vine grew and
flowered, but DO fruit s were produced. It was not UJllil 1836 thaI Charles Morren, a
Belgian bOlanist, discovered why vanilla was not able 10 produce fruit oul of Mexico.
TIle anatomy of the flower is such that self-pollination is impossible. Morren
discovered that pollination is c3rried out by a tiny bee of the Mellipone family which
lived in tbe vanilla growing region of Mexico. It is difficult for other insects to
reillace the tiny bees (3).
To achieve pollination one needs to remove the rostellum, a flower structure
that is a modification oflbe stigma lying between the male and female organs and
prevent access oflhe pollen 10 the stigma. Pollinalion is done by removing tbe
rostellum wilh a sharp objecl, so Ihat the pollen from the anther can be in contact mth
the st igma. Because Ihe blossom lasts for a very short time (less than a day),
pollination must take place as soon as tbe flower opens (6). Charles Morren was the
first to propose hand-pollination and he was the first to produce vanilla heans outside
of MeKico. This discovery laid the fOUJldalion for a new vanilla indU.1ry and broke
Mexico's monopoly.
In 1841 Edmond Albium, a slave in the Frencb-owued island of ReUllion,
discovered a rapid pollination method. With the use of a poimed tip of a small
bamboo stick, he picked up the adhesive pollen and prying up the flap-like rostellum
inside the flower, he pressed the male pollen mass onto the sticky female stigma. This
method ofpollinatioD is still used commercially today. Using the technique described
above, the French staRed vanilla cultivation on many ofthe islands in the Indian
Ocean. VaniUa plantings were established in Reunion, Mauritius, Madagascar,
Comoros, Thafila, Bra:.cil, Jamaica and other islands in the West Indies. Vanilla was
first introduced to the island of Java in 1819 by a Dutch scientist, using vines from
Madagascar. Conunercial productioll in central Java started in the early 20th century
and is known as "Java (3).
Today. the primary growing regions in the Indian Ocean are
Comoros and ReUllion, Beans from this area are called Bourbon and represent half of
the world:s vanilla production. Indonesia produces the other half of the world's
supply with an increase in production and quality. Tahitian vanilla beans, which
represent another variety, are grown mainly on the island of French Polynesia.
Vanilla is also beginning to be grown on Tonga, Although vanilla started in Mexico,
al present Mexico produces a small percent of the world's consumption (Fi g. I).
Many other countries with suitable climates grow vanilla, bnt at present remain minor
production regions (Fig.2).
4. 11,\ & DOltN
Indonesia
127,llon5
Madaga5c;or
650.3 tons
Vanilla
Costa Rica
4.6 tons
31
'-
Figure I. U.S General Imports of Vanilla Beans (U.S. Department or
Commcrce, 1995).
Tahiti
Madagascar
Fi gure 2. World Vanilla Production.
r--25"
25"
f
r
SI'[CES: FlAVOR tl.NI> tl.NTIOXII)tl.NT I'ROI'I';ltTt S
Cullh';lljon
Van.iIJa is a tropical climbing orchid. I' rom rnorc than 110 species
that have been described, on.ly 3 are of commercial value. TIl ey arc Vawfla
plallijolia Salisb Ames (VRnilla fragrans Andrews), Vanilla lahiltlleSiS, and Vanilla
pompon(l. Qnly V. plal/ijolia and Vanilla fahifellesis are pennitl ed to be used in
rood. V. pl(llli/alia. the most import ant COllllllcrcially, is cullivated in all lhc vani ll a
gTowing areas, except ror Tahiti and Hawaii . The heans arc 10 to 25 cm long and I
to 1. 5 elll wide.
V. fall/lellesis is indigenous to Tahiti and is different from V. pfallijofia by
slender siems and narrower leaves. Th e pods are al so sllOrter thall V.
pkmijolia (6). Thc V. fahitellesis beans arc perfume}" and contain anisyl alcohol,
anisyl aldehydc, anisic acid, and heliou opine which Ire absent in V. pfallijolia
The bcans of V. lahitellesis colllmand higher market price than V. p'cllli/oIw

Economically, thc imponant is Vanilla pompall(l, also known as
van ilion. It produces an inferior quality bean and in the past was IIsed for pcrfume
and tobacco fla vor. 1be plant produces less beans than V. plal/ijoli(l and has larger
and I,ider leaves. However, V. pompaflfl has important trait s, including growth under
more adverse conditions, and resist3tlce to root rot disease that may be used ill cross
Im.:cding.
Climat e. Vanilla needs II warn} and tropical climate wi th hilt
1I0t excessive rain. Under excessive rainfall, vaoilla may be aUl cked by mi ldew and
rool Under drought eonditioos, Ihe vl nilla plant can suffcr considerable
damagc, wh.ich will result in a smallllumber oflloweTS and low yicld. In gcneral ,
sloping land that bas soil with high organic 1IIltter, hi gh holding capacity and
access to irrigation in dry years will overcome thc problems caused by wcather (12).
Vanilla grows best ill 40-50% orllomlal sunlight intensity. In excessive
sunlight , the apical buds teud to lose moisture lind the gro\'\-1h of the vines is stunted.
Heavy shade \vill result in vines with thin stelllS, small beans, and reduced number of
flowcrs and fruils. A slladed nea, rainfall be1ween 1000-3000 mm and an average
temperature of 23 to 29C year round are the preferred conditions.
Soil :Jnd Nutrient!. Vanilla is a shanow rooting plantlnd grows well iu well-
drained humus-rich soil. A top layer of mulch helllS keep Ihe moimi re in and the
roots to spread. Soil plrameters, such as lexture and pH, appear 10 be more
imponantthan uutrients. In. hcavy rainfalluea, good drainage is essellliat The best
soil appears 10 be limestone with a pH of 6.0 to 7.0 and a deep layer of mulch, which
provides the nutrients, over Rnd around Ihc rools. Although the level ofcalciullI,
nitrogen, potassium, phosphorous and micronutrients in vanilla are essential for thc
and production of beaus, much IOOre allention has heen given 10 the type of
TlIulch, because mulch aJlpears to be more imponant than nutrienls (II).
4. (. IJOR.N Va"illa
.IJ
Support. '111e vllnilla pl3tlt is a climbing orcllid 3tld needs
support. Support is l lso needed to provide convenient ICcess for pollwauon
harvesting. Fast growing trees with smllil branches are preferred and also provide
shade. Howevcr, it is irnportantthat the support ing trce will not compete for water
and nutrients. Trees native to the Iropics are commonly used, such as avocado,
coffee, annatto, orange, etc. In some cases, il is possible to trees wbich.
themselves, can produce clsh crops. In MldagasclT. caSUlfUll pme trees Ife:lso
widely used. In Dlost pianlations, the support trees Ire planted before the vlllUll a
piant s ( 12). In greenhouses, vanilla can be grown faUing dowli or hanging.
P,-opagalion. Vanilla is commonly propagated from usually with 8-
12 nodes. Small cuttings will produce vines, bUI tbc larger thc cuttmg, the faster the
plant will flower. Thc situplest way to plant is lay thc 011 the and
cover it \vilb thick laycrs of mulch. However, tillS lmuts .tbe
amount of pia lit materiallhat ru.ay be used. Also, the many orlgms of the cutllugs
represcnt genetic vlriabili ty. .
A vine becomes productive after 3-4 years, depending upon the weather and
soil condit ions, and the healtb of the vine. 111e vine [Jowers ooce a year 5 or
successive years. After 8 10 10 years, productivity goes down. Drought ,
mul ch, overexposure to sun, and over- pollinatioo can eootribute to redUClion 10 the
productive period of the vine (4). WIlen planning a new plantation, one has to
account for the life expectancy of the vines. An hectare of land cln hold about 4000
productive vines. The avcuge yield is 1.5 to 2 kg green beans per vine. A nJ.;lture
good quality bean weighs betwcen 15 to 30 grams ( 12).
"loweriog, "'ruit Set, Growtb li nd Matur ation. II is well known tbat
vanilla vines need to reach a cert ain maturity before starting to flov.'tr. Appareot ly
Ibis is to accumulale growth factors or nutrients needed for blossoming. The fa ct ors
affecting the timi ng and abundance of flowers are nOI fully .. However,
according to Childers et . al. (4) drought , prumng and .the SI1.e of the .
original cutting may I n influence thc flowel1ng. Removal of 4 to 6 Inches of the aplcRI
bud 6 to 8 weeks bcfore blossoming prOUlOles flowering.
Vines usually blossom for I to 2 months. The flowers open early in Ih.e
morning and last for one day. PoUinat ion must be done Ihe same day. Also. ill a
raceme (the flowcr c1ustcr) only one f10wcr opeos per day and about I to 20 flowers
open in daily succession, although nOI aU of them \vill be fe rt ile. It is desirllble 10
pollioate 10 obtain between 5-10 beans because over-pollinlt ion wi ll result in small
beaos and a short production life.
Because self_pollination is impossible to achieve (l'i g.3), hand-pollinalioll must
be used to ohtain a comolercial crop. Hand-pollinat ion is donc as originally deS(;ribed
by Edmond Albium in 1841 in Reunion, whicb consists or. Ihe rosteliuDl
bl ck so that the pollen hearing anlher can be pressed agamst the Slig.ma. Once
pollination takes place, Ille development of the pod ensuing with the enlargemelll of
the OVlry is completed in about I and a half to 2 months (3).
SI'ICK"" nAVOR ,\ NI) ANTIOXIIiANT 1'lI.tH'ERTIES
Afler the fruit hIS amined its full size, it takes an additional 5 to 6 months to
mature and ripen. l>tlling the maturation period, the dry weight and the protein
content illcreases. Manyother physiological changes occur, U yet orunknowD
nature, Ihal resull in the productioll orflavor precursors,
.
Curing. The green, mature van ill a bean has no flavor or vauiJia aroma. If
the bean stays on the vine, the typical vanilla flavor will develop after a prolonged
ripening period. 110wever, in commercial practice, the characterist ic fla vor and aroma
of vanilla beans is due to chlnges lak.ing place during the curing process (2).
The curing process consist s offour steps: I. Killing 2. Sv.-eating 3.
Drying, aud 4. Conditioning.
Killing designates a process aimed at abolishing ti ssue aud cellular activity but
retaining heal . Tolerant enzymatic activity is imponaru for the curing I,roeess. The
ruost common processes are hot water killing, sun killing, and oven killing. Beans are
placcd in a wire basket and dipped ill 65C waler for a few minutes. Killing by heat
can be done also in an oven or by exposure to the stlU for a llrolonged time. 1bere is
a patent ror using free7Jng as IUOIher killing method ( [). TIle killing
apparently entails the breaking or cell membranes and mixing ofprl!Viously
compartmentalized subst rates aud enzymes.
Killing is rollowed by which all ows excess moisture to escape.
This reduces bacterial and rungal spoilagc, and yct Icaves enough moisture for
enzymat ic activity. Tbis is the most impon ant st ep. Glucosides, like giucovaniliin,
are broken doml to vanillin. 'Ibis is the step by which the t)Tl ical characteristics or
the vanilla bean flavor, aroma and color are devcloped. The duralion orthe sweating
process is between 7 [ 0 days.
Following the sweat ing step is "drying", which reduces the moisture content
rrom 60 70% to 20..30%. 11le rcduced moisture content is t)"llical ror a good quality
cured bean and helps to eliminate any microbial activity and to secure a long shelrtife.
' I1Le allloIDLt or moisture also affects appearauce, flexibility, and quality of the beans.
In this step it is necessary IIOt to obtailltoo hot a temperature, because some of the
flavor and aroma components can evaporate or break dov.ll . Tbe drying lakes 710
days in the sun, 2 3 hours a day. Oven drying can bc employed at 4550C.
Overdryiug of the bean will result in In inferior quality bean.
I"he list stage is "conditionillg", or aging. 'Ibis elll last from a month to
several months, Duriog this period, the vanilla bean obllins the
Chemical reaclions, such as ox.idalion and hydrolysis are the main evcnts (12).
11. Biotechnology Producti on
Recent development s in ]IISnll issue culture and the need to gai n control 011 hean
quality and yield turns our attention to this method as Rilot her approach to solve SOUle
oflhe problems associated with {he agronomic producl ion of vanilla (5). Plant
tissue culture calL be used ror three different objectives: 1. micropropagat ion
2. production of secondary products and 3. study ofbiosyllIhetic pathways.
-
4. t LWKIN t RNKEI. & UOkN Vallilla
35
Micropropagati on. In IUOst vanilla tbat is cultivated, conmlCrcial propagali on is
done by cUll iug. Several reporc s have shOWlI the Ilossibility of using tissue cuhure
tecbniques to obtain clollal propagation and disease-rreo; vanilla plants (8,9). We
used a different approacll using seeds as starting material. We were able 10
vanilla seeds ULLd ei' controlled stelile conditions in ft petri dish. 1bc seed l)rerCrs to
gemlinatc on the beau flesh (Fig. 4). YOIUl g grecn I to 2 months old afi.cr
being pollinated, lVere surface stcrili7.ed with I S% bleach. The beans wcre LntO
segment s aud placed on Gamborg's, ns basalllledium (7) rree ofhornl?lIe the
addition or cefolaxin and vancomycin It of 100 Illg per hter. Ibe
bean sections were subcultured to I clean plate every two weeks. Gennination of
seeds embedded in the bean sections staned after 36 mouths. The emerging embryos
lVere tl1Lnsferred to rifts (mesh size 25 microns, Sigllla Chenucai COUipany, St. Louis.,
MO), using the same m.:diuru without agar (Fig. 5), Each embryo was CUI illt o a rew
small pieces mt eacb subculturing. After a few mouths, plsllt lets \vere ronned and
moved to the greenhouse for hardening, using orchid mix. In an establislu:d lab, the
number ofplantlcts that call be Ilroduced is unlimiled. Seeds rrom young beans
germinat.: fhtcr than those from IIUlture beans probably because tILey have not
become dOflll\lut.
The t echniqUt: just described gives us the lool s to produce a vast nUilLber IIr
disease.free plallliets that lIUly be used to select for dcsi rable {raits, such as high yiL;1d.
early flowcring aud superior navor aud arOOla orbeans.
Secondary )'roduct. In the last rew years, there has been an altempt by I:scat!t.:LLct.il:s
Corporat iollto conune rcia!i7.c vani lla flavor by using plant cell culture (8). I\ \."..:onlmg
10 the published data, they oblained a yield of nearly 100 mg vani llin per li tel .
culture, or roughly 14 mg vanillin per one gram ofdl)' wcight ofeell ..
that a several rol d increase in vanillin yield would be necessary for pmd uctlOll to bo.:
ecoDomical. Another but alttmative biotechnological approach for the prUllucti l)LL "I"
vanilla is being investigated by Westcott et 81 (1 3). They IISC vanilla. root s to
rerolic acid to vanillin. 11Le limiting step here is where one can obla Ui lel'ull..:
acid fur the biocollversion. We used embryo cult\Jre, reasoning that dillercnti:lt..:d
tissue win be \I1Ore conducive than cell suspension to produce secondary metabolit es.
We were able to grow tbe embryo in up to a 10 L biorcaCior with a m.arine impelh: r at
100 rpLIl, because the erubiYos are sensit ive to high sllear. The culture prouIKcd a
larger number of the fl avor COIllJJonenls thl t Ire in vlnilla bean (Fig. 6 (a) and (b.
The profile of the flavor components in cuhured ti ssues is different t..han in the bcan.
It is nOI expected that undifferentiated cultured tissues will produce the same
and similar proportions ofnavor components that are produced in the bean ",lllch I.S
differentiated aud a specialized organ. We, thererore, concentrated on the prodUC\lOll
of vanillin, which has the highest ecouomie value. We found Ihat embryo ti ssue
produced up to 0, 16% vanillin and that 8 vanilla beau produced betwe<::ll 1.3%. .
vanillin ou a dry weigllt basis. At present, it appears that ti ssue culture IlroductlOIi IS
lower and may not be economical.
,
,
.j
J6
SI'ICES: n .... VOR CII EMISTR ANI) ANTIOXID .... NT PROl'ERTlES
Figure 3. Cluster of Vllnilla Fl owers.
Figure 4. Germinllting Vanilla Seeds on the Bean Flesh.
4. I IAVKI N.nu-;r.IKt: L <'II DORN Vanilla
The germinating embryos were held on noal ing fafts to maximize access to air
and nutrient s. By the end of 60 days, the resulting plantlets were transferred \0
a pot with orchid mix in the greenhouse for hardening.
Figure 5. Growth of Gcrminaling Embryos.
J7
Vanillin Uiosynt hcCic: r ath wlIY. Using the approach described previously, we grew
the embryo culture 10 study the pathway of vanillin to flnd out the
limiting step(s) in the production or the compound. After growing the cuhurc and
analyzing most of the exlracllble material, we propose that vanillin biosynthesis Slarls
from chain shortening of p.hydroxycoumaric acid, resulting in the product ion of
p. hydroxybcllzoic Acid, p-hydroxybenzyl aldchyde and p-hydroxybenzyl alcohol . We
observed an accumulation of p-hydroxybenzyl alcohol up to 10% of tbe dry weight,
suggesting thai the enzyme Ihal catal)'7,.es the hydro."<ylllion of p-bydroxybenzyl
alcohol 10 3, 4.dihydro."<ybenzyl alcohol is not active. In homogenized tissue, IDOst of
Ihe Il-hydroxybenzyl alcohol is convened 10 3, 4-dibydroxybenzyl alcohol aner a fcw
hours., suggesting that in intact tissues the enzyrne and the substrate are
companmentalized. Feeding experimeuls to the embryos with J , 4-dihydroxybenzyl
aldehyde, a by-product ofJ,4.dihydroxybenzyl alcohol, led to the accumulation of
vanillin, indicating thaI met hylat ion is not a limiting step.
GTven this ill fomlltion, understanding the limiting step call lay the foundation
for future work. fly isolating and purifying the em:.yme Ihat bydroxylates
p-bydroxybenzyillcohol and cloning the gene, we lilly obtaill a culture with I high
yield Ofvluillin that call be economically produced.
Fut ure Ollportuniti e! For Hi otcchnolgy_ We bave shown Ibat microprop'gation of
vIIl i11a is Ilossible with the use of tissue culture. This technology lUay be developed
further to select for cultivars with sullcrior traits. Illlllortantiy, it may be used to
propagate transgenic tissues tunsfected, fo r example, "ith genes to increase the yield
of vanillin, induce early nowering or disease resistance.
SI' ICES: HAVON: CII B II Sl'N:Y ANI) ANT!OXIIlANT
HPLC chromatogram of embryo culture extract.
A
1 . 00
o.so

i
,


;

;






0.16t or 0. ".
;
;
0 so
"


..
,
,
0 . 40
, 00
< .
)( H)l Minutes
Figure 6A & 613. HPLC Chromatogram. Isocratic separ-.ltion was used with the
mobile phase, 85% acidified water and 15% addilk-d melhanol (1 .25% acetic acid),
conlrollt:d by a Walers 600E system. A I'hcnomenox Selectosi l CI8 column (250 x
4.6 111m) wilh 5 pani cle size was used. The peaks were detcclcd [II 280 nm using
a Waters 994 photodiode array. ;
4. IL\VKIN-FRENU:L & l)ORN Vanilla .19
ltPLC chromatogram of Bourbon vanilla beans extract (I fold).
i

"
,


l

,
00
,


1.)l 1>.".
. .
,
,
.. 4 . 00

0
,
g


,
::I 0 0
)< 1 0
1
M'"'' I . .,. .
Figure 6. COlllimlf!1i
'"
SI'ICK.<ii, FI..AVOH ANII A ....TIOXIllANT I'ROl'EIHI K<;
/{rfcrcnCC$
I.
Ansaldi , Gwcl1sdc; Gil, Gerard ; Le PClit, Jean, u.s. PalmI 11 4,956, 192,
Sell 11 . 1990.
2.
3
4.
<,.
7.
,
9.
10.
"
12.
D.
Arana, F.E. Circular No. 15, I!Imhillglon, D. C , USDA 1945
N. r:.; H. R. Cirell/ar No. 28, Washing/OIl , D. C, USDA 1948
Childers, N. " .; Clbes, II. R. ; E. In rhe orchids a
Scil'I,lfifi
c
slin-ey; Withoer, C.L. Ed.; ROllald PrcS:S: New York, 1959
Colhn, I-I. A. ; Watt s, M. In Halldbook O/pklllf celf cullllre techlliques for
I'mfJ(lgatlOlI alld breeding Vol I Evano D A . Shan, \" " . A . P
_ . . . .., . " ,1, . " . ..... , . nmurato, .
V.: Yamada, Y. Eds; Macmillan PUblishing Co. : New York, 1983
COiTcl l, D. S. J &0,.. Bol. 1953, ', 29 1
Gamborg, O. R. A.; Ojima , K. Exp. Cell Re:;. 1968, 50,148- 151
Knuth, Mark E.; Sahal, Om P., U.S. P(I(ellf #5,057,424, OC(. 15, 199J and
U.S. Pa/t'II/ 1I5. 068,J84, NOII.l6, 1991.
KOll 01l0\IIiCl., 1-1.; Janick. I. lIor/sciellce. 1984 19 5g
Philip, V. J. ; Naillar, S.A.Z. J. 1'1011/ Physiol. 1986: 122,2 11
PUlesegJove, J.W.; BrowI!, E.G.; Green, C.L.; Robbins, S.R.J Spices Vol 2
1981
In Spices, herh$(l1/d edible jilllg; ; CharalamOOus, G. Ed. ;
ElscVler SCIence n. V.:AmstcrdalH. 1994
Westcott , R.I.; Cheet ham, P.S.J.; Barraclough A.I. Phy /ochem. 1994 35(1)
135138 " ,
Chapter 5
Onion Flavor Chemistry and Factors
Influencing Flavor Intensity
William M. Randle
Vcparlment of liorticulture, 111 1 Plant Science Building,
Unhersity of Georgia, Athcns, GA 30602-7273
Although onions are an importam vegelable and halle nUlriliollal IIalue
ill diets around the world, Ihey are pri marily consumed for Ihcir
distinctille flavor or their ability to enhance navors in ot her foods .
Onion flavor is dominated by organosul fur compounds arising from the
enzymatic decomposition of Salk(en)yl-L-cysteine S-oxidc navor
precursors following tissue disruption. Compounds ari sing frolll
prccursor decomposition, such as the lachrymatory factor and various
thiosulfin.ates gille onions their characteri stic navors. Sul fate is
absorbed by lhe planl and incorporaled Ihrough cysteine 10 glutathione.
From glutalhione, sulfur can proceed Ihrough several peptide pathways
and terminate in Ihe synthesis of one of three flavor pre<:ursors. Plallor
intensity is governed by genetic factors within the onion and
crwironmenlal conditions under which the onions grow. Onion cul tivars
differ in the ability to absorb sulfate and in the efficiency of
synthesi7j ng fla vor precursors. Increased sul fate fertilit y, higher
growing temperatures and dry growing conditions all contribute to
increased navor intensity in onion.
Onions (AWl/lIZ CI!(IO t .) hallc world-wide imparlance, ranking second among all
IIegetahles in economic imporlanee with an estimated value of $6 billion (I) .
Although onions cont ribute significantly to the human diet and howe therapeutic
properties. they arc primarily consumed for their unique nallor or for their ahilit y
10 enhance the of other foods. Onions arc an ancient IIcgelahle and can be
traced back through archeological records and early wri tings. OniOllS were used as
food, as medicines. in mummification. in art . and as spiritual objects (2) .
Interestingl y. hallc been domesticated for so long that they no longer have the
1997 American ChcmlCilI Socicly
42
SI'ICK"' : flAVOR CII EMISTRY AND ANTIOXIO,\NT PROPERTlFS
ability to exist in [he wild and require human intervent ion for survival In
contempor.try society, onions weave (heir way through OUf diet and are consumed
daily by most people. United States Pl;: f capita consumption of oniom in 1995 was
approximately 18 pounds compared to 9 pounds in 1975 (per. com., National Onion
Association).
Onion Flavor Chemistry
While compounds such as the watersoluble carbohydrat.:s (sugars) and organic acids
can contribUle [Q the sensory experience when consumi ng onions, onion flavor is
dominated by a special class of biologically active orgaTlOsulfur compounds (3-4),
Intact dry-bul b onions have lillie onion flavor or aroma, Flavor and aroma develops
only when the onion is damaged or cut and flavor precursor compounds undergo
enzymatic decomposition to form a variety of volatile sulfur compounds which give
onions thei r taste and aroma, The first sulfur compound associated
with onion flavor was identified in the 1890's (5). Pioneering in this
century among scientists such; as Chester Cavall ito, Authur Stoll, Ewald Sl:ebeck,
nolx:l laureate Artturi Vinanen, Sigmund Schwimmer, Mendel Mazeli s, George
Freeman, Jane Lancaster , and Eric Block, has many of the sulfur
compounds contribut ing to the biosynthetic pathway for flavor precursor
development , and the process by which flavor develops once the Onion is cui or
cooked,
Onion flavor precursor formatiOn begins with the uptake of sulfate (SO/) by
the onion, its reduction to sulfide, and suhscquent assimil ation to cysteine by light-
dependent reactions in the leaves of the plant (6), From cysteine, the sulfur can be
further metabolized to foml other sulfurcontaini ng compounds, Sulfur'S
proposed entry into the !lavor pathway is via glutathione, Early studies using
radioactive isotopes suggested that sulfur passed through glutathione and was
incorporated into S-2carboxy propyl cysteine or S-2-carboxy propyl glutathione,
eventually terminati ng into Spropenyl cysteine sulfoxide (7), Using radioactively
laheled sulfate in pul se chase experiments, Lancaster and Shaw (8) demonstrated that
sulfur was first incorporated into ), -glutamyl peptides as biosymhetic intemlediales
prior to terminating in the S-alk(en)yl cysteine sul foxide precursors,
There are 3 S-al k(en)yl cysteine sul foxide !lavor precursors in oniom: S-(E)-
I propenyl cysteine sulfoxide is usually found in hi ghest concentration and is
responsible for tearing and pungency assoc iated with onions; S methyl cysteine
sulfoxide which nomlaJly occurs in lesser concentrations; and S propyl cysteine
sulfoxide which is gener,dly found is the lowest concentration (9), A founh
precursor, S-2-propenyl cysteine sulfoxide, is the predominate precursor of garlic,
and found in other Alfiurns, but it is not detectable in onions, The pathways leading
to the synthesis of each flavor precursor are not fully understood, nor do we
understand the regulation of sul fur through these various pathways. Sulfur is
thought to be transformed through several peptide intemlediate pathways, unique to
each of flavor precursors COmpounds (f0; Fig. I),
The precursors are synt hesized and stored in the cytoplasm of the plants's
S. KANI)U; OniIJTI Chemistry
cells (l }) . Al1 iinase is compartlllentalized in Ihe ct:JJ's vacuole. When lhe
membrane of the vacuole is broken, allii nasc is released and the precursors ,IrC
broken dowll producing a chain of events. Primary products include shon-Jived
imemlcdiatc compounds, such as the I-propenyl cys!cim: sulfoxide dcrivcd
lachrymatory faelOr (LF, propanethial Soxide), and other sulfeni!; acids from the
different precursor specics. Other primary products arc compounds pyruvate and
ammonia which are marc stable. The LF, common to only Ipropenyl cysteine
sulfoxide a(;(;umulating AlliulIIs, produces the tearing, mouth burn, and pungency
scnsat ions (12). A series of thiosulfinmes arc then formed which chMaclerize the
unique flavors and aromas of onion. Early reports of di- and polysulfidcs and
thiosulfonates were shown to Ix: "artif,lcts" of hot injection port and gas
chromatographic column l13). 111C different flavor precursors give risc to diffen:nt
thiosulfinates which impart dist inet flavors to the sensory experience (14). ror
example, the propenyl/propyl thiosulfinates have green raw fresh onion !lav{)r, and
the methyl/methyl thiosulfinates have a cabbagc note (Table I).
The thiosulfinates then serve as the progenitor species of vinually all Iht:
snlfur compounds formed from the cut plillll. These compounds are amI
undcrgo disassociation and rearrangcment to form primary headspace (initial
products formed from cut onions at room tcmperature such as the thiosulfinmes),
S(."Condary volatiles (secondary products produced from the thi osulflnatcs ,11 rUOIll
lCmperature), and secondary solution components (products formed when
thiosu lfinates stand in so\tllion at room temperatuare) (15), Cut onions sitting on a
kitchen counter or on a salad bar, therefore, change flavor over time. For a
thorough discussion of cut onion compounds. see Block, 1992,
Innucncing FJaYOr Intensity
Onion tlavor hltensity is governed by plam genetics and the environmental condilions
in which the onion grows, In the marketplace, oni ous vary widely in sulfur-based
tlavor intensity,
Culti var Factors and Flavor Int ensit y, Some onioll'l arc very pungent 3nu
aromatic while others can be eaten raw hy the average person with relativi! case.
Onion color has very little 10 do with how pungent it might be , Red and yellow
ski nned Onions can range from being mild \0 very pungent, White onions, howel'cr,
are usually only pungent to very pungent. Some onions, such as the Grano and
Granex cult ivars can be grown mild with lillie flavor. Other unions, ;lS the
Danvers and Southport cuhivars, are very pungent while some, such as the Sweet
Spanish eultivars, falllx:tween mild and pungent. The gt: netic potential of a Cl!itiv:n
to ahsorb sulfur and synthesize the flavor precursors great ly determines how
flavorful an onion will be,
While the her itability of !lavor prceursor accumulation inl)nions has not been
determined, it is most likely a quantitatively inheri ted trait. As descrihed e:lI'licr,
the biosymhetic pathway is compit:x with many peptide intenlleliiates, [n ;ldditioll,
11 proposed enzymes regulate compound synthesis in the llavur pathw:, y (/7;.
Further empirical evidencc to support of the l1avo)"
SI'ICt:S: F1AVOIt CHEMISTRY ANIl ANTI OXIIlANT I'IWI'ERTIES
g Glutamylcysleine > Glutat hione
GIYCC /
S.2.carboxy
propylglulalhione
g.g I myl . S pro pyl

g.glll lamyl-S-2-c arboxy
S mel hylg!u1 .. thion.
propyl cysteine
1
1
cystel no ...:;: -
g -glutnmyl S ' .
propcnylcys teinc
g. glutamyl S-mllt hytcysleine
1
g.
cyst ei ne S.oK1<I",
g. glutamylS" propcnyl
cystcine S oKide
01 -01 rnyt . S m ethyl
cysllline S oxl dc
S. pr opylcys' e ine
SoKide
S (E) - ' .Propeny
cysteine Soxide
Smethylcysleinl! S,oKide
Fi gure I . The proposed biosynlhctic pathway for S-al k(en)yl cysteine sul folt ides in

Tallie I . Sensory experience of different onion t hiosulfinat es and lachrymat ory
factor
ComI22!!Jl!1
Methyl methane thiosulfinate
Methyl propyl thiosulfinate
Propyl methyl thiosulfinale
Dipropyl thiosuHinate
Propenyl methylthiosulfinate
Mctbyl propenyl thiosulfinate
Propcnyl propylthi05ulfinatc
Thi opropanal S-oxide
Flavor NQte
cabbage
cabbage/onion
cabbage/ollion
green onion, chive
cabbage, onion, metallic
cabbage
raw frcsh onion
pungent, heat, mouth bum
S. RANIlU: Ollion CllcmLI'f ry
45
precursors is the continuOtis variat ion found for the different navor precuf .....JrS
withi n onion and the faCl that the growing environment greatly influences precursur
accumulation (18). Qualitatively inherited trai ts, which are governed by one hI
several genes, usuall y fonn di screte classes and arc unaffected by the environment.
Cuhivar differences can begin 10 be explained in several ways. Some
eulti vars accumulate 2 to 3 times the amou nt of sul fu r as others, and sulfur
accumulation can be a key element in detenni ning onion fl avor intensilY (/9). If we
were to assume that cuhivars assimi late the avai lable sulfur inl o the fl avor precursor""
with the same efficicncy, di fferences among cultivars CQuid be explaincd simpl y on
how much sulfate enters the plant. However, culti vars differed in their efficicncy
at moving sulfur through the fl avor pathway and in synthesizing the
precursors (20). The correlati on, however, between total bulb sulfur accumulati un
and bulb pungency was poor among onions of broad genetic background, ranging
between r = -0. 05 and 0.36 (2/ -22). Some mild cult ivars accumul ate large amount s
of sulfate, but do not accumulate correspom)ing concentrations of flavor
while pungent cuhivars will assimilate more of Ihe available sulfur into fla vor
precursors (23). As an exampl e. Sweet . a mild eultivar, had OAO% total
bulb sulfur (dry weight basis) and 4, 05 mg/g fresh wei ght of total flavor
On the other hand. Rio Grande, it pungent cult ivar. had similar total bulb sulfur
accumulat ion (0.43%), but accumul ated 5.73 Illg/g fresh weight total flavor
precursors (a 42% inerease in total precursor accumulation).
The rat io of the various fl avor precursors also differs among cul tivars.
Flavor precursor composit ion and concentration can significantly influence fl avor and
navor intensity. Some cultivars synthcsi1.c and accumulated mostly I-propenyl
cysteine sulfoxide, the lachrymator and pungency producing precursor, but little
methyl or propyl cysteine sul foxide (24). Other cult ivars accumulate significant
amount of and propyl- cysteine sul foxide relative to the I -propenYl precursor
(25). The different precursor ratios give rise to different tastes and aroma . For
example, if [otal precursor concent rations were equal, a culti va r with a rat io of
12:4:2 I-propenyl :methyl:propyl cysteine sulfoxide would be highl y pungent and [ear
producing, while another eulti\'ar with a 6:9:3 rat io would be less pungent , and have
signi fi cant fresh onion flavor. Why eul tivars differ in partit ioni ng sulfur into [he
various pathways leading to the 3 cysteine suI fox ides or how environmental signals
regulate synthesis among the pathways has yet to be dctennined.
Onion cultivars differ in alliinasc concentrat ion which may affect fla vor
devel opment and intensity. Pukekohe Long Keeper. a pungent cuhivar, had a 4 fold
increase in bulb a11iinasc content compared to the mild Granex 33 cullivar (per.
com . J.E. Lancaster, Crop and Food, NZ). In additi on, onion al1iinase decomposes
the three cysteine su lfoxide flavor precursors al different rates (26). The kineti!': S of
I-Propenyl cysteine sulfoxide decomposition is quickest, followed by the methyl and
propyl cysteij!e /II vivo. I -propenyl cysteine sulfoxide decomposilion was
almost and complete, whereas methyl cysteine sulfoxide propyl
cysteine sulfoxide were decomposed to a much lesser extent (27). Betwcen 60 and
80% of the methyl cysteine sulfoxi de, and approximately 50% of the propyl cysteine
sul foxide was left intact in onion macerate , even after 2 hours foll owi ng macerati on
46 SPICES, FlAVOR C1IEI'oI ISTRY AN[) ANT/OXIIlANT l'IHH'nfI'lES
3
-
-
"'
'"
2.5
'"
"'

2
'"

'"
-
'"
1.5
.s


1 0


"
U
'"
0.5
CL
0
0.10 0.48 0 .85 1. 60 3.10
Sulfur Fertitity (meq/titer)
Figure 2. S-alk(en)yl cysteine sulfoxide flavor precursor accumulation from mature
onion bulbs in response to increasing SU](Uf fe rtility from appl ied nutrient solutions.
Solid bars are I -propenyl cystei ne sulfoxide. Lined bars are met hyl cysteine
sulfoxi de. Dotted bars are propyl cysteine sulfoxide.
S. Onitm 1'lal'or Chemi.l'lry 47
of bulb tissue. Culti vars ais() diffe red in the degree of decomposition of each
precursor (28). -nlcrefore, onion cuhivars differing in fl avor pretursor composi/ion
will vary in flavor intensity do 10 their precursor conccll1ralion and degrt:c to whidl
lhe precursors afC decomposed.
Flavor Distribution. Flavor is unequally dimibU!cd within the onion planl :md is
dependent on onion ontogeny _ Wit hin the bulb, Ihere is a flavor gradient. The
highest concentrat ion of precursors occurs in the inlerior hase of the bulb, while the
lowest concemrat ion of precurSors aTC in Ihe top outside scales (29) . As the pl ant
and bulb grows, flavor first increases in intensity, but then decreases as the bul b
swell s to maximum size and mat ures (30). Onion flavor and intensity also change
during Storage. Some cuUivars increase in flavor during storage while
others decrease (31). The ratio of lhe diffcrem precursors can also change duri ng
storage . In one onion cuhivar, the l -propenylthiosulfinates significant ly increased
in concentrati on during six months of storage while the met hyl methane lhi osul
completely disappeared (per com. Norman Schmidt, Dep\. of Chem., Georgia
SOuthern University).
Environmental Factors and Flavor (nl ensity. The environment in which an onion
grows and develops will greatly infl uence how mild or pungenl an individual cul livar
will be . Yearly and locational flavor intensity differences among onion cultiv(ifs
have been known and reported (32). When isolated individually, environmental
factors such as sulfate availability, temperature, and water supply afree! flavor
intensit y.
Sulfate Fertility. Sulfate availahility greatly influences onion flavor inrcnsi ty
(33) . When onions were grown with sulfur fenility levels rangi ng from deficient to
luxuriant, the concentration and ratio of the three flavor precursors changed (34: fi g
2). At sulfur fert il ity levels which sulfur defici ency symptoms in onion plant s,
methyl cysteine su lfoxide was the dominant precursor while I -propenyl cy"wine
sulfoxide was a minor precursor. As sulfate feniJity increased to ndequat e Jt: vcl s,
I-propenyl cysteine sul foxide increased in concentration and methyl
sulfoxide dt.-creased in concc: ntralion. At luxuriant sulfate fcrlility , i -propt: nyl
cysteine sul foxide was the dominant precursor while methyl cysteine sul l"o xidc
became a mi nor precursor. Interestingly, total precursor concelliration was similar
from those onions grown al defi cient sulfate fenility and those grown at luxuriant
levels. I! appeared that when onions were stressed for sulfate, the methyl cyst d nc
sulfoxide biosynthetic pathway became a st rong sink for the available sul f"J[ and
large amounts of methyl cysteine sulfoxide aeculllulmed, even at the cxpcnse of plant
growth and bulb developmem. The prupanethial S-oxide and !hiosulfina te
concentrations followed the same pallem as the cystei ne sul foxide precursor
concentrations when responding to increasing sul fate feni Jity (35). Scnsory
eval uati on from professional tastt: panels confi rmed the chemical analysi s. As
available sulfate levels increased in the growing envirOlUnent, se nsory notes ;IS
SI'ICES; HAVOI{ ANI) hNTIOXI ,)A"'T I'ROI'EI{Tr ES
Tallie II. Aroma and navor notes from 'Granex 33' onions grown at increasing
.\III"ale fertili ty 11' 1'1'15. Hi gher numbers represent more intense nOll'S.
Auribute
low S mod S high S exeess S
Green aroma
6A 6.0
5.' 6.6
Pungency aroma
7A 8.0 7.2 8.4
Total
8.3
8.' 7.7 9.5
l1iuer
3.3 5.0 4. 1 6.7
GrCCIi Flavor
6.8 6.' 6.2 7.3
Ileat
4.9 7.2 6.6 8.1
Pungent navor
5.7 8.1 7. 2 9A
Sweet
7.6 6.0 6.2 4.8
Total Fl avor
8.2 9.5 9.2 ll.l
Total Sulfur Flavor
5.2 7.8 7.0 8.2
.' .
Uoilcd Onion Flavor
4A 3.1 3.7 3.
Cabbage Flavor
3A 3.1 3.1 3
Fn:sh Sulfur Hwor
7.0 7.5 6,7 7.5
Fruit Sulfur Flavor
4.7 4.6 4.6 4.7
Hydrogen Sulfide
1.4 1.8 1.5 22
Rubbel)' Sulfur
2.3 4.0 3.1 4.4
S. IthNI>I,E 011;011 n avor Chemistry
a5tota l aroma. biuerness, heat, pungency and navor also increased (Table
11).
Grou-i ug TemlK!rat ure. Tempentture is important for onion growt h ill lll
development. Minimum temperalUres for bulbing around IO"C and reach a
maximum around 35"C. In an early slUdy . although increasing tempcratun:
increased volatile sulfur compounds in onions. the author was unable 10 specify if
navor differences were due to temperalUre related plant growth (i.e., the faster tILe
plant grew, the more pungent the plant became) or to the direct effect of temperature
on nilvor development (36). Recently, a study was completed where plalUs
to 4 different growing tcmperatures I) grown for a specific length of time and
harvested when the plants were of different developmental age. or. 2) grown 10
maturilY which took increasing lengthli of lillie as the tempcracure dttrea.'ied. In
both cases, increasing the growing lemperamre from IU'C to 31"C increaSCtl the
pungency ( measured by enzymatically fonn pyruvate) of the onions. and the
increase was lincar in response to increasi ng temperature (Figure 3). The hotter
the growi ng conditions. the more pungent the onions will be.
Water Supply . Growing oni ons under dry conditions will also increase bulb
pungency to onions grown under well irri gated conditions . When onions
werc grown under natural rainfall or supplemented with irrigation WOlle r, the non-
irrigated onions produced a higher volatile sulfur content compare<! to irrigated
onions (37). or produced increased flavor strength as measured hy volatile headspace
analysis. enzymatically developed pyruvate, and .'\Cnsol)' evaluation (38). II.s bulb
size was small er in the non-irrigated plots. it was thought that increased flavor
strength was due to a concentration of the flavor prc<: ursor compounds in smaller
cell s. The exact for flavor increases in water-stressed plants. however,
is yet to be detennined.
Water usage and sulfate uptake by onions was al so poorly correlated
(r = O.09; Randle, unpublished data). Whcn plants were grown hydroponically to
detemline sulfate uptake requiremcnts o\er lhe course of the growing season. water
usage was greatly affected by dail y differences in rolar radiation while sulfate uptake
was unaffected. TIle greater the solar radiation. the more water was transpired
through thc leavcs.
Summary
The chemistI)' of onion navor from sulfur compounds is quite complex. The
hiosynthetic pathway leading to the three of S-alk(ell)yl -L-cystcinc S-o.lide
navor pre<:ursor synthesis is eomplicatcd arKI still being developed and proven.
Onion cultivars differ in thei r ability to synt hesize the navor precursors and differ
in the ratio of,precursors synt hesized Each precursor gives ri se to sulfeni e acids
and thiosulfinates which define different fla vor experiences and navor intensity . The
environment in which the onions grows is al so important in detennining navor
int ensity and composition. Increasing sulfate fertility. increasing the growing
temperature . and/or decreasing the water supply will increase onion flavor sirength.
-OJ
'-'
:J
u
0(3
5 -
ro 3 --
o!d
>
:J

>-
"-
SI'IC":'" CHEMISTRY ANI) ANTIOXIIJANT l' IIOI'i<IO"I ES
35 days of bulbing
To maturity

10 17 24
31
Growing Temperature (Celcius)
.3. pungency (umo! pyruvale per Illl of oni on juice) in response II)
mCICaSltli; growlIlg temperalures. Solid line reprcsenls planls grown allhe diffe rcnt
Icmrer3111res for 35 days in a bulbing photoperiod. Dashed line represenlS planl s
grown 10 IllHlurilY at Ihe different lemperatures.
5. MAN!)I.!': 011;1)11 FIMor Chemi!;try 51
Li terat ure Cit ed
(I) FAD, Prodllc/ion Yellfbook; Food and Agricuhure Organizalion, Rome . ... 7,
1994.
(2) 1landl , P. Ill: Olliolls and Alliell Crups; J. L.; Rabinowitch, II .D .. Ed.
CRC Press, Inc. l10ca Ralon, FL, 1990, Vall ; .
(3) Block, E. Angew. Chem. 1m. Ed. 1!)92, 31, 1135 1178.
(4) Darbyshire , n.; Steer , U: r. In: Oniolls and Allied Crops; Brewsle r. J.L.;
Rabioowi tch, 11 . 0 ., Ed. CRC Inc . Boca Raton, FL, 1990, Vol 3; 1-16.
(5) Semmler , F.W. Arcll. PJwrm. 1892, 230,43 ... 443.
(6) Laocaster, J. E.; Boland , MJ . In: Onions allli Allied Crops; Brewster. 1. L. :
Rabinowilch, I-I. D., &I . CRC PreSS, Inc. Boca R:lIon, FL, 1990, Vol 3; 3372.
(7) Granroth, B. AntI. Acad. Sci. Fel/n. Ser . 1970, A2 154, 1-71
(8) LancaSler, J.E.; Shaw, M. L. Phytochemi stry. 1989,28,455-460.
(9) Randle, W.M. ; L:mcaslcr, J.E. ; Shaw, M.L.; SUllon, K.H. , Hay, R. L., Buss:ud;
M.L. 1. Amer. Soc. lIort. Sci. 1995, 120, 1075 1081.
(10) Bl ock, E. Alisew. ('hem. /111 . Ed. 1992, 3{, 1135 11 78.
(1/) Lancasler, J. E.; Coll in, 1-I .A. Plant Sci . l..ell. 1981, 22, 169 176.
(/2) Bl ock, E.; Naganathan, S.; P1.Jl man, D.; Zhao, S. 1. Agrc. Food Chell!. 1992.
40, 2418-2430.
(13) mock, E. Angt' w. Cllt'm. 1111 . Ed. 1992, 31, 1135 1178.
(14) Dlock, E.; NaganallulII, S.; D.; Zhao. S. 1. Agre. Food ('hem. 1992.
40,2418-2430.
(15) Block, E. AlIgf'll-'. CJI t' m. 1m. Ed. 1992, 31, 1135 1178.
(16) Bl ock, E. A"Kl'w. Cllt' lII . 1m. I:.i l . 1992, JJ, 1135 1178.
(/7) Whilaker , 1. R. Adl-'. F/JOlI Rn. 1976, 22, 73- 133.
(/8) Randle , W.M.; LancaSler, J. E.; Shaw, M. L. ; Sullon, K.H., Hay. R.L;
Bussard; M. L. J . Amer. Soc. Nort. Sci. 1995 , 120, 1075- 1081.
(l!1) Randle, W.M. Euphyrim. 1992,59, 15 1- 156.
(21) Randle, W.M. Ellphylim. 1992,59, 15 1 156.
(22) Randle, W.M.; Bussard , M. L. 1. Amer. Soc. lion . Sci. 1993, 118, 766770.
(21) Randle, W.M.; Laneasler, J. E.; Shaw, M. L.; Sulton, K. H.; Hay. R.l.. ;
Bussard; M. L. 1. Allier. Sot'. Hon. Sci. 1995, no, 1075-1081
(24) Thomas, O.J.; Parkin, K. L. 1. Asr. Food Chell! . 1994, 42. 1632- 1638.
(25) Randle, W. M.; uneasie r, J. E.; Shaw, M. L ; SUlion, K. II., 11 3Y. H. L. :
Bussard; M. L. 1. Amer. !We. lion . Sci. 1995,120, 1075- 1081.
(26) Schwimmer, S. ; Mazcl is , M. Ar('/I. lJi(x:hl'm. and lJivphy . 1963. tOO, 6373.
(27) Lancaster, J.E.; Shaw, M. L. ; Randle, W. M. PrOf' . Natl. Oniun Res. Conj.
1995, Madison, WI , 53 58.
(28) Lancaster, J.E. ; Shaw, M. L ; Randle, W.M. /'roc. NlIIl. Olliull COli/.
1995, Madi son, WI, 53 58.
(29) Freeman, G.G, J. Sci. FOUlI AKrir. 1975,26,471-481.
(30) unCasta, J.E.; McCallion, B. Il . ; Sh:IW, M. L {)hysioll'lalll. 1986. 6r,. :! lJ3
297.
(3{) Kopscll, D. E. ; aandle. W.M. l/ortSril'llCl! . 19')5 , 31.766.
-,
,.
SI' I t:K>.;, F1AVOIt CHEMI STRY ,\NIJ ANl"JOXII)ANT F,S
J. P. .: Reay. 1' . 1' . ; J. D. : Bcnncn, W. O.: Scc.lcoic, J. R. New
Z"o/(Illd J. '.1'. ARrie. 1988, 16, 279-285 .
cJ3} I:rccman, G,G,: Mossadeghi, N. J . Sci. Food Agrk. 1970,11, 610-61 5.
(N) 1{'II1Ulc . W. M.; mock. E.: UUlcjohn. M.; !'ulman. D.; Bussard. M. L. 1.
,lgri<.', Fond (.1rt!lII. 1994, 42. 2085-2088.
US) Randle, W.M.: i...1rlCaSlcr. J. E. : Shaw, M. L. ; Suno n, K. II. ; Hay, R. L. ;
M. L. J. Amer. Soc. /lor/ . Sci. 1995, 120, 10751081.
U(j) 11. 1. Agrie. Nu. 1944.62,371 -379.
(J7} I'lalenius, II . J. Agric. Rr:s. 19.t4. 62. 37 1-379.

(3S) Freeman, G.G.; Mossadcghi, N. 1. 11Qrt. Sci. 1973,48,365 378.
Chapler 6
Contribution of Nonvolutile SulfurContaining
Flavor Precursors of the Genus Allium
to the Flavor of Thermally Processed
Allium Vegetables
Tung- ll s i YII
Department !If .' ood Engi neering, Du- Yeh Ins titute of Technology 11 2,
Shan-jea u Road, Da-Ts uen, Chan g- UwlI , Tniwun, Republic of Chiml
This article di scusses the contributions of nonvolat ile sulfur-containing
flavor precursors of the genus Allium to the flavor of thermally processed
Allium vegetables through two approaches. In the first approach we
have analyzed the volatile compounds generated from thermally
processing blanched Allium \egctalJles. In the second approach we have
analyzed the volatile compounds produced during thermal degradaTion o r
thermal interaction solut ions of S-alk(en)ylcysleine sulfo)(ides, the major
nonvolatil e sulfur-containing flavor precursors of Allium vegetables. A
la rge number of volatile compounds can be generated from thermally
processed blanched Allium vcgetables and from the thermal degradation
or thermal interaction solutions o f these sulfoxides, and many of these
volatile compounds can be fO\J nd in thermally processed Allium
vegetables. We have demonstrated that the nonvolatile sulfur-co n(aining
!Javor precursors of Allium vegctables can not only generate volat ile
compounds through thermal degradation or thermal interactions but also
t hrough the thermal degradati on or interaction product s of these
sulfoxides 10 make important contribut ions to the flavor of thermall y
processed Allium vegetables.
Leaves or bulbs of Tile genus Allium, as represented by garlic (AI/illm .5Olil1ll11 L.), onion
(Allilllil C"lXl L_), shallot (Al/iulII {l.sallulliclillI auct ), lind ....'Cish onion (A llillm ji.WII/OJIIIII
1.. ), have been widely used aT home and in the food industry as vegetabl es or flavoring
materi als because of their strong flavor properties. The characteristic aromas of the
Allium species are mainly allribUied to the sulfilr.containing volatile compounds. Unlike
the preformed volatile compounds, such as ester s and terpene compounds in fruits and
some spices which are biosynthesized as plants develop, thc volatile components of the
genus Allium are released from their nonvolatile sulfur-cont aining flavor precursors,
especially S.a!k(cn)ylcystcine sulfoxides, by lin enzymatic-mediated degradation that
takes place when the plant s are disrupted. Depending on the species. in the alk(en)yl
54 SI'I O:S: FlAVOR l:I IEM[STHY ANI ) ANTII) XLIIANT I' ROI' t:RTl ES
groups the sulfoxides are mainly a combination of allyl. methyl. propyl, and I propenyl
group (1-3).
Most of the research on the flavor of the Alliulll vegetables before 1991 focused
mainly on the analysis of the volatile compounds of these vegttables, the formation of
the volati le compou nds in the genus Al lium through the enzymic de,gradati on and tht
transformation of the nonvolatile sulfur-containing flavor precursors, such as the S-
alk(en)ylcysteine sulfoxides, and the stability of the thiosulfinates which are the enzymic
reaction products of the S-al k(en)ylcysteine sulfoxides. There has been very li tt le
research on how amino acid characteristics affect flavor precursors. Since S-
alk(en)ylcysteine sulfoxides are amino acids, they should undergo thermal degradalion
or interactions with other food components., especiaUy reducing sugars. The focus of this
research has been on lhe contributions of nonvolatite fl avor precursors of Ihe genus
Allium to the characterist ic flavor processed Allium vegetables. especially
gartic. shallot, and welsh onion We approach this research in two ways. First, we used
blanched Allium vegetables to see if significant amou nts of volatile compounds could be
generated from the nonvolatile sul fur-contai ning flavor precursors which wcre retained
in Ihe tissues of All ium vegelables ancr blanchi ug these vegetables during Ihennal
treatment, especially baking and frying. In our second approach. we synthesized four
important S-alk(en)y]cysteine sulfoxides: S.allyl-, melhyl-, propyl-, and 1-
propenylcysteine sulfoxides. These sulfollides were Ihen subjecled to thermal
degradation or interaclions wilh other food components in the Allium vegetables to
determine what their conlributions were to the characteristic fla\'or of themlally
processed Al lium vegetables. 111e following reviews some of our research.
Volatil e Compounds Generaled frolll Thermally Tru ted Blanched Allium
Vegel:lblu
To deter mine the potenti al contribution of the nonvolatile sulfur-containing fl avor
precursors of the genus Allium to the navor of thermal ly processed Allium veget ables,
we IIsed blanched Allium vegetables to which we applied Ihermal treatment. BlanChing
Ireallllent can deactivate the Ilavor enzymes and retain most or Ihe flavor precursors in
Ihe tissues of Allium vegetables. Since no signiftcanl amount of volatile compounds
existed in the intact tissues of Allium vegetables, volati le compounds detected in
thermally treated Allium vegetables could have been generated by thermally degradi ng
Ihe nonvolali le precursors in the tissues or Ihrough the Ihermat inleractions of these
precursors and olher components. especial ly sugars, in the tissues.
We were able to identi ry some impon ant volatile compounds in blanched and
Ihermally trealed blanched garlic sli ces, shallot slices and Welch onions as found in
Tables I, 11 and III, respectively, As you can see in these three tables, no significant
amou nts of volat ile compounds were idenlified from blanched garlic (BG), bl anched
shallot (BS), blanched green leaf of welsh onion (BGL), and blanched white sheath of
welsh onion (BWS), These results prove thai blanching treatment of Allium vegel ables
can deactivate the Ilavor elllymes in these vegetables efficiently and inhibit the enzymic
format ion of volatile compounds from the flavor precursors in Allium vegetables.
Hllwever, you will note that in Tables I, 11, and III , baking or frying treatment of lhe
6. YU Jo1a'l'or of TI,ermally All ium
Table I , Some imponant volatile compounds identified in garlic samples
Yield, ppm
Compound BBG F13G
Compoullt!s Probably Generated from TIJerill al D" grlld:llion of
Nonvolat ile FlJn'or Precunmrs
I-propene
acetaldehyt!e
melhyl allyl sulfide
ally sulfide
methyl allyl disulfide
I ,2 -dithiacyclopenl-3-ene
allyl disulfi de
methyl al1yl trisullide
3vi nyl-4 H- l ,2-dithiin
2-vinyl-41'1-1.3-di thiin
allyllrisulfide
712
1.85
0.06
1.16
0,80
0.77
9. 14
153
OA9
0.31
0.52
10.55
25.24
1.31
8.99
1.26
2. I S
50.74
4.34
1.44
0.72
0.66
liG'
2.20
0.70
nd'
0.G9
0.04
006
2,40
0.17
0. 14
ed
006
1,2,3,4-tetrathiepane 221 S.98 nd
dimethyltrithiepanes 2.42 7.63 003
Compou nds Probably Generat ed from Thenll allnltrll Cl ioll $ of Sugars
and Nonvolatile "' avor Precursors
2,5 -dimet hylpyrazine
0.51 0.69
elhylmethylpyrazines 0.35 0.72
3,S-diethyl.2-methylpyrazines 0,47 OA5
Comll oulltl s I'robably Generat ed from Thermallnteractiolls of Upitl s
and Nonvolatil e f.'[:Jvor Precursors
4-heptenal
2-ethylpyridine
2-pentylfuran
methylethylpyridine
phenylacetaldehyde
buytlbenzene
pentylbenzene
hexylbenzene
bem:othiophene
0, 11
0. 12
ed
0.06
0. 10
0.06
0.25
043
0,70
0.68
0.20
0.22
0.19
OA4
0.19
0.34
0.62
OAO
li RG: Baked Bl anched Garlic; FBG: Fried Blanched Garl ic;
BG: Blanched Garlic. Data regenerated from ref. (-I ).
nd: not delected
nd
ed
ed
ed
,d
,d
,d
ed
, d
ed
"
ed
51> F/.II\' OK ANIl ANT1(IXIIlANT
II . Some volatile idenlified in shallOI sameles
Yield, 2Em
COll12ound BBS' FilS' .S'
Compounds P.-obably from Thermall>egradati OIl of
Nonvolatile lilavor l'rf<' lI r5ors
I-propene 0.43 0.86 nd "
melhanethiol 0.58 196 M
dimethyl disulfide 0.46 1.91
"' dipropyl di sulfide 1.66 2.00
"' l_propanethiol 0.29 0.30 0.03
3-mclhyhhiopheltC
"'
0.78
"' methyl propyl disulfide 2.49 8.50 0.01
dimethylt hi ophenes 4.83 15.44
"d
I-propeny] methyl di sul fides 0.75 5.46 008
dimet hyl trisulfide 3. 52 7.28
"' I- propenyl propyl disulfide
"'
1. 42
"' methyl propyl trisolfidc 19. 12 1644
"' dipropyl trisulfide 7.29 3.94
"' Compounds Probably Gentra lt d from Thf rmallnt f rllcli ons o(Sugan
a nd Nonvolatil e Flavor prtcunon
pyridine
"'
0.41
2-pentylfuran
"'
5.80
methylpyr31.ine 1.14 5.48
dilliethylpyrazines Jl64 59.24
2,3 -di met hylpyridi ne 1.23
"' ethyl melhylpyrazines 6.11 15.38
Irimethylpyrazine 4.53 10.77
S-ethyl-2-methylpyridine
"'
0.82
2, 6-diet hylpyrazine
"'
338
et hyl dimethylpyrarines 33.85 29.38
methyl propyl pyrazi ncs 443 331
dimethyl propylpyrazincs 2.70 12.83
2,3 -dimethyl-5-[ I-methylpropyl]pyrazine 1.51
"d
j -eth>:l-2. 6-dimeth:r: IE>:ridi ne 0.94
"d
BBS: Baked Blanched Shallol: Fl1S: Fri ed Blanched Shall ot;
ns: Blanched Shallot . Data regenerated from ref. (S) .
nd: not detected
"'
"'
"'
"d
"d
"'
"'
"'
"'
"'
"'
"'
"'
"'
,.
YO
Flal'or "f TJrermall)" l 'mcn,\td Al lium Vegt'labll''\"
57
Table III . Some volatile idenlified in welsh onion
Yield, ppm
Compound
BDGL' FflGL'
DGL BBWS' FBWS' BWS'
Co mpound Probllbly Gener :Hf ti from Therma l Drgradati on of
t'IIonvolllli le Flavor Pfe<:urson
methanethiol
002 0.04 trace" 0.02 0.07 trace
I_propancthiol
0.08 nd ' 0.02 0.08 "d
0.02
dimethyl disulfide
0. 10 0.58 trace 0.05 0.67
,,'
2_methylthiophene
0.02 001
"'
0.01 0.01
"'
methyl propyl disulfide
0.06 0.04
"'
0.05 0.03 rnI
dimethylthiophenes
2.02 0.72 0. 16 U8 078 "d
I-propenyl methyl disulfides
0.63 2.64
"'
054 2.25 "d
dipropyl di sulfide
0.02 0.03
"'
0.02 0.05 "d
dimethyllrisulfide
0.63 0.96
"'
0.72 1.44
"d
I_propenyl propyl disulfides
0. 14 0.15 "d
0.11 1.01 "d
methyl propyl trisulfide
0.23 0.39 "d
0.33 0.69 "d
I-propenyl methyltriSlllflde
0.08 0.2 1
"'
0.08 0.20 "d
dipropyl lrisulfide
0. 18
"' "'
0.25 "d
M
I-propenyl propyl trisulfide
0.08 0.21 trace 0.07 0.20 0.01
Conlpounds Probably Gt ll trll lM f.-om Thtrmallnl eracli ons ofSugfln
lind Nonvolatil e Flavor Precunon
2-penty! lUran
"'
0.13 "d
"'
0.09
"
dihydro-2-met hyl-3 _[2I-i J_furanon
"'
"d
0.28
"'
"d 0.24
dimet hylpyrujnes
0.04
"' "'
0.05
"'
"d
3-(met hylthi o) propanal
"'
0.08 "d
"'
0. 11
"'
2-ethyl-3 ,S_dimethylpyrazine
0.03
"'
"
0.03
,,'
"d
dimelhylpyridine
0.04 "d
"'
0.01
"'
M
2-acetylfuran
0.26 0.40
"
0.34 0.49
"'
2-acelylpyrrole
0.08
"'
0.04 0.20
"'
0.0 1

0.27 0.Q7 004 0.03 0.02 0.05
,
llllGL: Baked Bl anched Grcen Leaf; FllGL: Fried Blanched Green Leaf;
,.
...
DGL: Blanched Grcen Leaf; BBWS: Baked Blanched White Sheath;
FBWS: Fried l11anehed Whit e Sheath; BWS Blanched White Sheath.
Data regenerated From ref. (5 ).
Trace: < O.OOS ppm
nd: not detected
58 SI'ICES; t'i.AVOl{ CII EMLSTR't' ANO ANTIOX([lANT !'IUll'mIT!I':"
blanched AlliUm vegetable slices generated significant amounts of volatile compounds.
These volatile compounds arc proposed to be mainly generated from the thermal
degradation of nonvolatile flavor precursors ill the Allium vegctabks and the interactions
of these precursors and other components, especially sugars and lipids, in thcse
vegetables.
The major volatile compounds listed in Tables I, II and III separated into two
groups; Ihose that were probably generated from thermal degradation of nonvolalile
flavol' precursors and those generated from thermal interactions of nonvolatile flavor
precursors of garlic and sugars. The major volatile compounds generated from thermally
degraded nonvolatile flavor precursors of garlic listed in Table I were allyl disulfide, 1-
propene, acetaldehyde, allyl sulfide, methyl allyl trisulfide, and other polysulfur-
containing cyetic compounds. Thc major volat ile compounds generated from thermal
interactions of nonvolatile flavor precursors of garlic and sugars were pyra7.ines,
Formation ofpyra1.ines in these thermally treated blanched garlic sliccs are proposed
through the Maillard type reactions of the amino-containing precursors and the reducing
group-containing sugars existing in the garlic tissues. The major volatile compounds Ihat
were identified in baked blanched and fried blanched garlic slices which were probably
generated from thermal interactions of nonvolatile flavor precursors of garlic and lipids
were pyridines and alkyl benzenes.
Table II lists the major volatile compounds from baked blanched and fried
blanched shallot slices generated from thermal degradation of nonvolatile flavor
precursors of shallot were methyl propyl trisulfide, dimethylthiophenes, mcthyl propyl
disullide, and dipropyl trisulfide, The major volatile compounds that were probably
generated from thernlal intemctions of nonvolatile flavor precursors of shallot and sugars
were pyrazines, especially ethyl dimethyl pyral.ines, dimethyl-pyrazines, ethyl methyl
pyrazines, and trimethylpyrazine.
The major volatile compounds listed in Table llf are from baked blanched and
fried blanched welsh onion slices. The first group, generated from the thermal
degradation of nonvolatile flavor precursors, were dimethylthiophenes, and alk(en)yl
disullides and trisulfides. The a1k(en)yl group mentioned above could be methyl, propyl,
and I-propenyl. Those major volatile compounds generated from thermal interactions
of nonvolatile flavor precursors of welsh onion and sugars were pyrazincs, especially
dimethylpyraz.ines, 2-methoxy-4-vinylphcnol, 2-acetylfuran, and 2-acetylpyrrole
Volatile Compounds Generaled from Thenn:11 Degradati on or Thermal
Int enlc(ions of of Alk(en)yl Cystei ne Sulfol: ides
To delennine the potential contributions of the nonvolatile flavor precursors of the genus
Allium to the flavor of thermally processed Allium vegetables, in our second approach
we analyzed the volati le compounds generated from the thermal degradation of the S-
alk(cnhoJcysteine sulfoxides and fi 'om the interactions of these sulfoxides with other food
compllnelUs existing in Allium vegetables. Four S-alk(en)ylcysteine sulfoxidcs were
by us and were subjected to thermal degradation or thermal interactions; the
VOI:tlilc compounds generated WCfe analyzed to determine what contribution these
$ull,'xities made tll the navor of (hermally processed Allium vegetables. The su!foxides
6.
'"
Hal'or oj ProCfHeu Allium Vegfllab/es
Table IV. Some important compounds identified in the [henna! degradation
or them!al intcmction solutions of S-all\'lcyslcinc sulfoxide (alliin)
Yield, rnllLrnolt: of ;'Illi;n
Compound A' A'I G*
I -propene 39.90 "do,
acetaldehyde 1199.60 16.50
allyl alcohol 93 ,)0 340.70
ethyl acetate 23230
0'
acetic acid 72.70 4.30
thiawlc 9.90 13.30
acetal 26.40
0'
dimethyl disulfide Il.ID nd
2-methyllhiazole 7.50 0,70
1,2 dithiacyclopc:ntanc 1.70
0'
methyl elhyl disulfide 61.QO
,,'
methyl allyl disulfide 1.70
,'"
dimcthyltrisulfidc 11.50
,,'
formylthiophcnc5 3.50 29.90
2-ace\yllhia:role 359.50 140.30
methyl-l ,1 ,3-triuliacyclopentane 188,00
0'
dimethyl tctrasulfidc 530
"d
ethyl
0'
I 10
methyl- I ,2,3,4 -tctrathianc 57.50
,,'
1 ,2,3, 4-tdralhiepane 9.70 10.60
dimcthyltetrathianes 20.90
,,'
pyra1.ine "d
4. 10
methrlpyrazine
,,'
43.80
dimclhylpyrazines
,,'
10.00
cthylpyrazine ",I 15.10
ethyl mctllylpyrazincs od 9.40
2-acetylthiophcnt
""
13.40
ethyl dimcthylpyrazincs ,d 2.20
mctllyl propyhhiazolcs
,,'
30.60
bcll1.othiophcne
,,'
20.80
phcnylacetaldchydc
,,'
1.90
acetyl methylthiophcnes
0'
6.00
2-blllylthiophene
,,' 0'
2-pcntyhhiophene
0'
,d
2-pentylpyridine
0'
"d
2-hexylthiophcnc
0' ,,'
2-hcxanoyllhiophcne "d
ud
2 -pentylbcnzaldchyde ,d
,,'
2
,d od
A: Alliin was themlally degmdcd in a "H 5 aqueous sotut ion.
Data legener31ed from ref (6).
A+G: Alliin was thermall)' inlera(;ted willl glucose in 3 pH 7.5 aqucous sotUlion,
Data regenerated from ref. (7),
A+D'
4.50
0'
2193.60
od
"d
42, 50
nd
,,'
1.30
15.90
od
,,'
0'
6.00
16.20
4, 00
,,'
24.90
"d
16.50
"d
,,'
"d
,d
,,'
0'
od
od
"d
od
,,'
,,'
3 1.50
6. 10
18.40
til [0
31 .50

110.20
A+D: Alliin was interac1ed wil1l2,4-i"1ecadien31 in a pH 6 aqucous solution.
Data regeneraled fro'" ref. (8)
nd: nOl dctocted
:;9
SI'I CI':S, FlAVOR CHU,II STRY ANIl ANTIOXIOA:'I.T I'ROl'ERTtK"
Table V. Some imponant volatile compounds identified in the thennal
degradation or thermal interaction solutions of S-methy1cystcine
sulfoxide
mg{mole of
Comeound
methanethiol 8.20 0.90
dimethyl disulfide 3320.80 2598,90
dimethyl trisulfide 21.60 59.80
lhiaz:ole 18.20 2,90
thioeyanie acid, methyl estcr 65.30 1.30
Illcthylthiol furan 157.50 86,60
pyrazine 69.80 114,60
methylpyrazine 8 90 264.00
ethylpyrazine \.70 99.90
di lllct hylpyrazincs 4. 60 336.90
trimcthylpyrazi nc n.d.
u
40.5Q,
ethyl methylpyrazine 23.80 115.20
2,6.diethylpyrazine 19.30 6.30
ethenylpyrazinc 5.40 53.20
ethyl dimethylpyrazines 6.50 74.90
ethenyl methylpyrazines 330 176.90
2-mcthylpyridine 4.30 3.10
2acelylpyridinc n,d. 8,30
J -( mcthylt hio )pyridine 17.40 5.70
t H Elrrole 11.00 40.20
McCySO . S-methyl-L--cysteine sulfoxide
..
MeCySO+G S-methyl-L-cysteine sulfoxide + glucose
nd : not dctectcd
6_ YU Flavor of Tilf!rll1nify All ium Vegclnhll'.l'
Table VI Some important volatile compounds identified in the thermal
degradation or thenna! interaction solutions of S-propy1cysteine
sulfoxide
mglmole of PrCySQ
ComEound
I-propanethiol 88.80
1,1 'thiobis propane 7.20
methyl propyl disulfide 8.60
dipropyl disulfide 6246.40
dipropyl trisulfide 7088.80
ethanethioic acid, S-propyl ester 44.00
pyrazine 450
methylpyrazine nd
dimet hyl pyrazines 3.60
ethylpyra1;inc ed
2-ethyl-3-methylpyrazine
e'
2, 6-diet hylpyrazine od
ethenylpyrazine 4.60
ethyl dimethylpyrazines 5.50
methyl propyJpyrazines 12.60
isopropenylpyrazine
0'
dimethyl propylpyrazinc
0'
mcthyl -{ l -propenyl)pyra7.inc 7.50
2-ethenyl-6-methy!pyrazine ed
3_(propylthio )pyridinc 10.30
methyl (propylthio)pyridinc 5.40
od
PrCySO . S-propyl-L-cysteine sulfoxide

PrCySO+G : S-propl-L-cystcine sulfoxide + glucose
nd: nbt detected

5.60
33.40
38.70
8367.80
3853,00
2.70
60.90
126,80
11 9.40
70.90
1330
9.20
21 10
65.30
86.30
42.40
9,30
65,80
87.50
11.90
4.80
18.40
', I
62 CII EMISTln" ANIl Ai'I'TIOXJI)lI N'r 1'IHlI'EIHJES
Table VII . Some important volatile compounds identified in the thermal
degradation or themla! interactions of S-(+)-cis- ! -propenyJcysteine
sulfoxide
mglmoie of PrenCySO
Compound PrenCySO* PrenCySO+ G'
(I-propenylthio) acetaldehyde 7.20
2-methyl-I,3-dithiane 15.80
methylthiophcncs 33.60
2,4-dimethylthiophene 1.40
tetrahydrothiophene-3 -one nd *
thiophenecarboxaldehydes 2.40
S-methyl -2( SH J-thi ophene
0'
formyl methylthiophenes 16,20
2-acet yl-S-methylt hiophene 10.40
thi azole 3.20
4-methyl isothiazole 5.20
2,4-di mehtylthiazole 2.80
2-acetylthiazole 499,80
pyrazine
0'
melhylpyrazine ' .00
2,6-dimethylpyrazine 4.60
ethyl methylpyrazines od
tri methylpyrazine od
3 -ethyl-2,S-dimethylpyrazine od
3-methylpyridine 10.60
3,S-diethylpyridine 7.6<)
3 butylpyridine od
PrenCySO. (+)-S-cis-l-propenyl-L-cysteine sulfoxide
PrenCySO+G (+)-S-cis-l-propenyl-L-cysteine sulfoxide + glucose
*. nd: not detected
' .6()
3.00
253.00
16.00
4.40
58,00
9.80
84,00
g.20
29.40
5.80
2.00
158.40
52.60
86.40
51.40
10.00
5.20
10.20
32.60
15.60
16.20
6. YU Flaw}r of Thermally I'rlJCl.'s,mi All ium Vegettlb/",\'
were S-;llIyl. , S-mclhyl-, S-propyl -, and (+)-S-d.\'-l -propcnylcysleinc slllfoxidcs.
Some imponanl volatile compounds identified by thermal degradation or lhr nnni
interaction solutions ofS-allylcystcinc sulfoxide (alliin) afC shown in Table IV. Thermni
dcgradalion of alliin mainly generated acclaldehyde, 2-acctyhhiawle, ethyl aCC1 <l le,
mClhyl . l ,2,)-lrilhiacyc!opcntane, and acetic acid. Thermal interaction of alli in :lIld
glucose mainly generated pyrazincs, especially melhylpyrazinc, dimcthylpyrazines,
cthylpyrazinc, ethyl methylpyrazines, methylpropylthiazolc:s, 2-acctylthiophene, and
benzothiophene. l1lemlal interactions of alliin and 2,4-decadienal represented the
products of oil containing linolenic acid, and were mainly
alkyhhiophenes, especially 2-formyl-S-pentyllhiophene, 2-hcxyhhiophenc, 2-
bUlyllhiophene, and 2-hexanoylthiophene, 2-pentylpyridine, and 2-pcntyl-bcnzaldchydc,
In Table V we havc identified some imponant volatile compounds from the
thermal degradation or thcrmal intcraction solutions of S-mcthylcystcinc sul fo>: idc.
Thellllal degmdation of S-methylcysteine sulfoxide mainly generated dimethyl di sulfi de
and methylthio furan. Thermal interactions of S-methyl-cysteine sulfoxide and glucose
mainly generated pyrazines, especially dimethylpyrazines, methylpyrazinc,
mc!hyJpyrazine, cthenyl methyJ-pyra1.ines, pyrazinc, and ethylpyrazine.
Table VI lists some important volatile compounds idcntilied from the thermal
degra(btion or thermal interaction solutions of .'I'-propylcysteine sulfoxide. Thermal
degradation of S-propylcysteine sulfoxide mainly generated dipropyl Irisulli dc and
dipropyl disullide. Thermal interactions ofS-propylcysteine sulfoxide and glucose m:\i ' lly
generated pyral.ines, especially methylpyral.ine. dimethylpyra7.ines, met hyl
propylpyrazines, ethyl dimethyl-pyrazines, and 2-ethenyl-6-mcthylpyrazine.
Some important volatile compounds identiflcd in the thermal degradati on or
thermal interaction solutions of S-L'is-l-propenylcysteine sulfoxide arc shown in
VII. As shown in Table VII, thermal degradation of sull'oxide
mainly generated 2-acetylthiazole and methyl-thiophenes. The addition of glucose to the
aquoous solution of S-ci!i-I -propenylcysteine 'sulfoxide increased the yield of thiophenes
but decreased the yield of 2-acctylthiozole, Thermal interactions of S-cis-l-
propenylcystcine sulfoxide and glucose mainly generated pyrazines, especiil lly
melhylp)'razine. pyrazine, 2,6-dimethyl pyrazine. elhyl mcthylpyrazine, and
dimethyl-pyrazine, 3,S.diethylpyridine, and 3-bUlylpyridine.
We have demonstrated that a large number of volatile compounds genemtcd
frolll the thermal degradation of sulfoxides or from
intemctions of these sulfoxides with food componems can also be found in the thenn:\lly
processed Allium vegetables (5,\1). proof that the nonvolatile sulfur-containing !lavor
precursors of Allium vegetables can generate volatile compounds through thcrmal
degradation or thermal interactions but abo that the thermal degradation or thennal
interaction products of these sulfoxides make an importam contribution to the fbvor of
thenllally processed Allium vegetable.
Literatllre Cited
Block, E.; Naganathan, S.; Putman, D, Ziao, S.H. J. Agric. Food 011:111. 1992, 40.
24 IS.
s l'le)';''': FIAVOW: ,\NIl ,\NTIOXtllANT I'tIOJ'EIITt K"
2. Bl ock, E.; I'ulman, D.; Zhao, S.-H. J_ Ab,,.ic, FQotlChem, 1992,40,2431 .
3. IJIock, E.; Naganathan S ; Putman D.; Zhao S.-H. l'u/'t & Appl. Clwm. 1993,65,
625,
4 YI!, T.H.; Lin, L.Y.; 1-10, C.-T.J. Ag/'ic. Prxxl Chem. 1994,42,1342.
5. ellen, Y.N. Sludy 011 Ihe CO/llrihllliOIl 0/ jlm'OI' PI'<'cli/'sor.t I(J Ihe jlm'Or /orlllolioll
a/ l IreI'll/oily IN'OC/!!i$l!d jj,nllol alld ",eI,." onioll, 1996, M.S. dissertation of DI!. -Yeh
Institute of Technology, Taiwan. ROC.
6. Yu, T,B ; Wu, C.M.: Rosen, R.T ; Ual1man, T G. ; Bo, c. -T. J.
Ag,.ic. FoocJChem. 1994,42, 146.
7 Y\!, T.B.; Wu, C.M.; 110, c.-T. .J. Agrk. FoodClU'III. ' 994, 42,1005.
::; VII. T.H.; Lee, M.H., Wu, C.M.; Ho, C.-T. In Upid!; ill Fo<xl Ho.c.r.,
Han man, T_G_ Ed.; ACS Symp_ Sec 558; American Chemical Society:
Washington, D_C , 1994,61-76_
9. Yu, r.I!. ; Wu, C,M, ; 110, C.-T. 1 Agric. FoodChcm, 1993,800.
A NALYfICAL T ECHNIQUES
Chapter 7
Characterization of Saffron
by Aroma Extract Dilution
Flavor
Analysis
Keith R. Cndwall oder, Hyung lIee Raek, ond Min Cai
Depurtment or .'nod Scie nce and Techn ology, Mi ssin ippi Agric ult urlll
lind Experimen t Slation, r-,li ssinippi State Un ive r sity,
lIer.r:er BUlldmg, Stone Buuleva rd, Mi ssissippi Sl ll te, MS 39762- 9805
Volatile compounds were isolated from Spanish "Mancha Superior"
by simult aneous steam distillntion-sulvcnt (SDE) nnJ
dlrect solvent extraction ( DE). Extmcts wcre analyzed by gas
chromatography (GC)-mass spectrometry, GC-ol fnctometry (OC-O), and
aroma extract dilution analysis (AEDA). Total ion chro!JJ:llograms of
SDE.and DE extr::lcts were different with respect to levels andtypo:s of
\'ohlllies prescnt; however, both e;(trncts had distinct saITro!l-like
quniities. A tOlal of 25 uroma-active components were consistently
detected by OC-O und AF.DA, with 18 common to both SDE lind DE
One compound telllatively identi lied as 2-hydroxy-4,4,6-
trlmethyl-2,5.cyelohcxadien-l.one (saffron, dried hay-like) had the
hi.gllest 10g,(Oavor dilution fac tor) in hath extracts, followed by 2,6,6-
,3-cyclohexadiene-l-carboxaldehyde (safranal) (saffron, tea-
like) alld an unknown compound having a saffron, drit.:d haylike aroma.
Saffron, the dried dilrk-red stigmas of CmCIll" salil'w L. flowers, is used hI impart
both col.or flavor to foods. Considered to be onc of the most expcnsive spices,
saffron IS pTllnaril y produced in Spain U.l). but is grown commercially in :;(\'eral
countries. Production, chemistry, adulteration, und quality assurance aspeets of
saffron ha\'e been reviewed (1. J). The imense yellow color of samon primarily is dUe!
to .the presence crocin, a of water solublc carotenoid glycosidcs (./-8). The
taste-actIVe component of saffron is picrocrocin, a bitter glycoside of 2,6,6-
tflmethyl.4-hydroxy-l-formyl-l -cyc!ohexene (./-8). Jlydrolysis of picrocrocin leads
to thc formation of the lllnjor component of the essent ial oil of samon (9).
Safralllil. having :I typical saffron-like aroma, Ims been generall y consi dered 10 t>e the
character-i mpact component of saffron (10). Rlklel and Pclrlika (II), using gas
ehronmtography.olfuctometf)' (GC-O), confirmed the importance of safranaito saffron
:lroma: !lOwc,er, they detected several unidemified aroma-aeti\'<' components.
@ 1997 Chemical SIJdet)"
"
GC.() is used for the of llromu-m;livc cOinponcnls in a volatik \"\;r:11:1
(12). The aroma intensity of each (;[\Il be determined hy C1Wrll:l
extract dilution armlysis (AEDAJ. which involves OC-O cV:lluntion of [\ serial d:itl \i"n
series of a voblilc (!tImet (I J). rrolll these results. <I Ilavvr dilution (FO) Ih<'
highest dilution :II .....I\i.;h [\ spetific aroma compound was last UClcch:d b)' GC-O,
obtained for each aroma-acth'e COlllponent of Ihe (xlraCI. ro factors arc then u'C"\!
10 arrange the aroma-aclive compounds according to their significance in \11\: ("lr.l("1.
prudding a bc!ncr understanding of the rule each compound pl ays in the on'ra!l
!Javor of the food. A critical c"uJuiltion 01' ,\ ED.I\ can be found ( J.:. J 5) .
AEDA hilS been cmployed for of irnportml\ nroma in
wid..: varkty uf foods, e.g. lobst("r (16), crab (17), beef (18), and brc:1d cno<! (/9).
Thc prcsenl study dt"aJs wi lh th..: identification of predominant arorn:I-:ICli\"e
compounds ill saffron by AEDA. In ordcr to "ssur..: that [he result s of AED,\ \wrc
representat ive of s,'1ffron aruma, ,olatile components w..:re isol:.ih:d hy IWO cxtr:u:lion
tcchniques.
Mat eri:l ls & MethOds
i\.1aterials. Dried Sllffrull, type "Mancha Superior", W:lS purchnseti from twO
spice suppliers. A total of thf\.'C I 01.. samples were ohtain..:d for this study. Orilnal
packing date of Ihis material was 1995. Samples were stored in the dark .:11 room
temperature until analysis.
St:lndard fl avor compounds were purchas.."'<i from Aldrith Chemical Co.
(Milwaukee, WI).
Simull ancous Sttalll Distillation-SohNII E:).:tradiun (Sin:,. \Vh,)le saffron stigmlls
(U.5 g), 50 IIlL of deodorized-distilled ....'tIter. and I 0 of internal standard SI)hltion
(3.07 mwmL 3- 11cptanol in met hanol) were continuously exlrael.;d \\ith \0 l1IL of
r.:distilled dichloromet h,lI1c in a micro SCille SOE apflaratus (c:l talog 1\0 . 16415 J.
C'hrompack, Raritan, NJ) for 3 h under atmosph..:ric condi ti ons. snE colJ tinger
temperature was mainl3incd at _4C. Elich extract was dried over. 2 g.
sodiulll sulfate and COnCCnlmted to 0.5 III L under a gentle stream tol punll..:J nltrugen.
Duplicate exlrm:tions prd"urmed fllr each sample.
Birc: ct Extrat:lion ( BE). Who!.: salTron (2.) g) were ground for 3()
s in a spice milt (Model K74)OC, Reg:!1 Wilfe, Inc., Kewaskum. WI) and then
Ihrough a no. 60 nylon mesh screen. Saffrnn powder (0.5 g), 10 of i11lcmnl
stnnJard !IOlutiun {as atM,,cJ, and 3 mt of redistilled t1ichl oromclharle were ....
sh:lken (or 5 min in a 16l' 125 mill serew capped testlut-.e scalt.'I.I with 11 I>TI L-li ncd
cap. The suspension allowed to stand ;n the dark for 2 h, shakt:n ugnin. <I nd
t:entril"llgctl al 3,OOOxg ror 30 min. Alkr recovering the supernatant, the ,'us
extractetltwu more times with 3 IllL of dichlo[mneth:me ahove I
was dried over 2 g of anhydrous sodi um sulfate .. nd concentratcd to U.s Ill !. unt !.: r i\
gentle stream of purili ..."li nitrogen. Each saffron sample II'US extracted in lillpli..;atc'.
Gas Chromat ography-Mass Spect rometry (GC-MS). <.;C-fl.\S sy,,!em " f
an HI' 51190 Series II GCI Il I' 5972 mass sdccti\"c ,!ettxlor (fI.'SI), Itcwlcll I';,,k;ud.
I
"
SI'le..:", YiAVOR CIIEMI STRY ANI} ANTI UXlDANT j'KOI'ERTl ES
Co. , Palo Alto. CA.) One of each extract was injected (spJill ess mode; 30 s valve
delay; 200"C injcctor temperature) into a capillary column (DB-wax or DO-5ms, 60
m Icngl h x 0.25 mm i.d. x 0.25 ).lm film thitkness (d
r
); J & W Scientific, Folson,
CAl Helium was used as carrier gas at a constant fl ow rate of 0.96 mUmin. Oven
IcmpcralUre was programmed from 400C 10 lowe 81 a rate of 3Clmin with initial
and finn! hold times of 5 and 60 min, respcrlively. MSn condi lions were as foll ows:
capillary direct interface temperature, 280"C; ionization energy, 70 cV; mass range.
33350 a.m. II.; EM voltage, 1956 (Atune + 200V); scan rat e, 2.2 scansls. Each SDE
or DE eXlraCI was anal yzed in duplicate.
Aroma EJl: trAct I>il ulion AnAlysi.! (AEllA). GC-olfactometry (OC-O) system
consistro of a Varian 3300 GC (Varian Instrument Group, Walnut Creek, CAl
cquippe<l with a flame ioni7,.8tion detcctor (FlO) and sniffing port. Column crouent
was split 1: 1 between FlO and sniffing port by using deactivated capillary columns
(I m length x 0.25 mm Ld.). Serial di lutions (I :3) were prepared from SDE and DE
extracts using di chloromethane as diluent. Each dilution (1 ].IL) was injected into an
FSOT column (DU.wax or DIl-Sms, 30 m length x 0.32 mm i.d. x 0.25 ].1m dr. J &
\V Scientifi c). GC conditions were same as GC-MS except that oven temperature was
programmed at a rate of 6C/min and with initial and final hold times of Sand 30
min, respectively. FID and sniffing port ....'Cre held at 250C.
GClO wa.s performed by two trained panelists familiar wilh saffron aroma. Each
panelist evaluated a dilution series for only one SOE or DE el(lraet for each saffron
sample. Because of the si mi larity between GClO results for SDE elltraels from all
three saffron samples, and likewi se DE elliracts, final AEDA result s were summarized
by calculating average 10glFD factors) (n '" 6) for each type of extract.
Compound hl entifi ca lion . nd Qua ntil ali on. Compound identifications wcre based
on compari son of GC retention indices (RI )(2O), mass spectra, and aroma properties
of unknowns with those of authentic standard compounds analyzed under identical
conditions. Tentative ident ifi cations were based on matching mass
spectra of unknolVns lVi th those in the Wiley 138K mass speclral database (John Wilcy
and Sons, Inc. , 1990) and literature or on matching Rl values and aroma properties
of unkllOwos with those of authentic standard compounds.
Relati ve concentrations of positivcly identifi ed compounds ....'Cfe dct cnnined using
their MS rcsponse factors compared with the internal standard. Response factors wcre
determined by analyzing standard compounds at three level s under identical GClMS
conditions. ll1e relative concentrati on of each tentatively identified compound was
estimated from it s peak area relative to the internal standard.
Results & IJi scussion
Vola ti le co mponenU or saITron. Typical total ion chromatograms of SOE and DE
volatile cxtracts of saffron arc shown in Figures I and 2. respectively. Quantitat ive
result s arc prescnted in Tablc I. A total of 46 core volatile componcnts wcre
identificd. or 30 compounds were found common to both SDE and DE
extracts. Sarranal (2,6,6trimethyl-1 ,3-cyclohexadicnc- l .carboxaldchyde, no. 31) was
the most abundant volatilc component , followed by isophorone (3,5,5-t rimethyl-2-
cyc!ohcxen- l.one, no. 26), and 2,6,6-trimethyl-2-cYclohclIenc-1 ,4-dione (no. 36). Thc
7. CAJ)WAUAOEK lIT Analysis of Saffron f7a.,()r
.9



,

.-=
.c'
U
-:
..
r
, .. ,
I'
, , .
r
a ,, 36
.. U

. i
..
1.5.
1
"

,
II
I
!..f

j
I .. 1 "
,
'"
,
1 0 . 00 20.00 3 0 . 00 4 0 . 00 50. 00 60. 00 7 0 . 00 80. 00 90.00
Time (min)
figure: 2. Total ion chromatogram of saffron volatiles isolated by dircct solvent extraction. Peal.: !lumbers correspond
\0 those in Tables I and II and Figure I.
Table l. Volatile Co mpounds in Saffron
No. ' Compound JtJ& Cone. (u glL:)'
DB-wax DB-Sms SDE
d
DE<
966
'"
6.OJ.9 2.0O.2
975 <700 9.31.7
,,'
2 5-Tcrt-but yl-I.3-cyclopt:ntadi ene
r
3 2,3-Butanedione
J
'
5 3-Hydroxy-2-bUlanoncC 1285 711
nd; 7}30
6 j. Heptanol (I.S.') 1295 899
10 3,5,5-Trimclhyl-3-cyclohcxen-l-one
f
14 12 104' 61=10
,d
11 Acclic acid" 1446 <700 ,d 754108
13 Megasligma-7,9. 13-tricnel'" 1452 1262 112
""
14 2.Furancarboxaldehydcl" 1463
'"
,- 23:t9
""
15 2 -Methy1enc-6.6-dimclhyl-3 -cyclohcxenc- l-carboxaldehyde' 149K 1112 12022 34:::4
18 2,6.6-Tri mclhyl-3-oxo-3-cyclohexcne-l-carboxaJdchyde' 1541 \ I!I 26J
19 LinalooJ;' 1546 J 104 265 2.4:tO.3
20 2-McthyJpropanoic acid" 1569 , d 9.41.6
21 5-Memyl-2-furancarboxaldehyde" 157; 7.32.8
"" 22 1577
'"
,d 21:t2
23 Sulfi nylbis methane" 1582
'44
,d lS2
('muimlc(lon IK'XI/)(lgI'.


,
,
c
"
;(.
.
Ii

g

:;,
<


:t
g

"
"
"

. ..::

tl

c
,


>

,
"
>



.,.

!;;"
...
S:!>


'"

;
,

' LI Dle I ('un ! j ll il et l
No.'
24
26
27
28
2.
)]
32
33
J5
36
37
J8
J9
41
44
45
46
47
48
Compound
Mcgasti gma-4,6,8-trienc isomcrf'
3,5,5-T rimcthyl_ 2 -cyclohexen-l-one (isophorone)'
Megastigma-4,6,8-trienc isomer!'
Mcgastigma-4,6,8-tricne isomer"
Oihydro-::!(J /1)- fur:monel"
2,6,6-Trimcthyl-I,J-cydohcxadiene-I-carboxaldehydc (safr:maJ)1
2 -Hydrox),-3,5.5-trimethYI_ 2-c)"d ohexcn_l _onef
3Mcthylbuulnoic acid"
4-Hydroxy. 3,5 ,5-trimeth}' 1-2 -cyc!ohc)(en-]-{)nei
2, 6,6-T rimelhyl-2-cydohcxene_l. <I-di oncl
2-H ydroxy-4,4,6-trimcthyl -2.5-cycJohexadien_ l_one'
2,6,6 Trimethyl-2, 4-cYdoheptadien_ l_
onc
f'
2(SfI)-Furanom."
2.2,6-T rimethyl.j ,4-cyclohexilnedionef)
l-PhenethvlaceullL" -
DihYdro-n-iononef"
Hexanoic :lcid
i
'
(E)-Gcraniol"
6.10-Dlmelhyl-S,9-uooo:cadicn_2 -ono: isomer"
Tab le I Continued
No.' Compound
49 Benzcnemcthanol
l
'
51 Sulfonylbis methanc'"
52 2,4,6-Trimcthylbenzaldehyde
f
>
53 2-Phenylethanol
'
54 2,4,6,6-T etramclhyl- ! -cyelohexenc-I-earboxaldchyde
i
55 /l-Jononel '
56 2,6,6-T rimethyl-3 ,4-eyclohexadicne-I-earboxaldchyde
i
58 2- Hydroxy-3,5,5-trimethyl-2-cyclohexcn-1 A-dione
r
;
60 2,6,6-T rimethyl-4-oxo-2 -eye lohexen-I-earboxaldehyde'"
61 2,6.6-Trimethyl-5-oxo-1 ,3-gelohexadiene-I-earboxaldchyde
i
62 4- Hydroxy-2,6,6-l rimethyl-2-eyclohexen-I -onef'
63 4- H ydroxy- 2.6,6-1rimethyl 3-oxo- l-cyelohexcne- I-carboxaldchyde
i
64 5-( BUla-1 ,3-dicnyl)-4, 4,6t rimethyl- 1 ,S-cyelohcxadien-l-ol'
B.! '

1586 1323
1602 1129
1620 1363
1628 1365
1633
1656 1210
1668 1152
1672 869
1683
1698 115 1
1734 1163
1740 1229
1761 912
1785 1179
1820 1258
18J9
1849
1849
1857
Bt
DB-wax DB-5ms
1880 1043
1903
1904 1321
1917 1121
191 7 1313
1947 1484
1970 1312
2023 1240
2083
2127 1369
2141 1258
2152 1346
2179 1501
l:Qnc.
SDE" DE'
DS oct
1270130 95-k::130
9. 7 1.9
"d
80:!:18 7A11
7.S:rJ A
525<>::390 3-180=300
20739 9712
od 90I J
5521 2J6
1280:::170 921::1 20
258211 848
21:::! 18
12815
od 400S5
-07135 554:t 70
4.50.7 4.4'OA
7.k!: 1.0 3.610.6
od 337
114nd
7.ILO 3.90.8
-
Continued on n(XI
I);;Qnc.
SOE' DE'
4.81.5 6.9L2
od 7.S10. 9
58iO u
23163 264129
N/A' NIA
8.61.1 4.71.1
7012 427
281130 37S53
3724
" 1 SiS 2810
15S 93137
I 2975 7418
401J 3910
' Peak numbers I;orrespond to those in Fi gures I and 2 and Table II . retention index. <Avenge relative concentration =: standard
deviation (n=12). simultaneous sleam distillation-soh'ent extral;tion. ' OE, di rect solvcnt e)(lraction. 'Compound tcntativcly
identified by comparing its mass spt{:\rum 10 Wiley 138K mass spectral database. ' Compound nOI previously identified in saffron.
' Compound positively identified as des<:ribed in materials and mdhods. "tr. trace. 'Compound lent:lti vely identified by I;omparing its
mass spectrum with published literature ( II) . indo not dctt:cted. .. internal stMd:ud. 'N/A, not available, peak I;ould not be resolved.




"
;{.
"
>
<


0
:"'l
<


<
>

>
,
g
"
;;
>,
"
i
0


;;



S


0
"
"
>
r

c-

'"
'"

'"


,

Sl' t O:<o;: n AVON ANI) ANTIOXIIIANT t' tHlI'liKTJES
cOllccmmtions of safnmal in Tabl.: I were similar tu pub1i!;lwJ values (6).
Likcw;s<:, the rci:uive (lnundllnees (wi th resl)<:ct to s3franal) for prominent volatiles
(e.g., nos. 15,26.31,36.38,41,53. and 511) were in reasonnble agrt"Cm..:nt with
previous studies (9,11).
Dcspite agrccll1ent with published literature, volmile profiles of SOE and DE
t:xtracts dillcred frOIll eilch other with respect 10 levels of major volatiles liS well as
in I) I'<s of compounds identified. DE allowed for the isolation of acidic COmponents
(nus. 11, 20,34, 3nd 46), whil.: no aciJs were delttted in SDE extracts. On tht: ot her
hand, SDE extracts cont:aincd higher levels of major saffron volat;l.:s (e.g .. nos. 26,
31 , (lnd 36). Iligher le\'els of thes<: components may hilv,;: been a resilit of thcrnmUy
induced hydrolysis of thei r glucoside pr\.'(:ursors during SDE (9). FUrlhermlJfI: , il also
is lil. el)' that other tht:rmal1y generated compounds were form..:d during SUE. Fllr
example, the sugilr hreakdown products, 2-furancarhoxyaldehyd<.: (110. 14) and 5-
llletllyl-2-fllnlncarboxaldchyde lno. 21), wcre oilly detected in SDE SOt::
also contained four megastigmatriene isorrn.:rs (nos. 13, 24, 27, ami 28),
which have been previously reported 115 \'olatile const ituents of starfruit (]f).
In (lJJition tn the above mentioned eOlllJXlUnds, IllUIlY (lthers wert' idenli tied for the
firsttilllt: as saffron components (see compounds indicated by an asteri sk ill Table I).
Of [lo1rticular interest and importance were the following compounds detected hy both
GC-MS and r.C0: 2,3-hutanedinne (no. 3), acetic 3cid (no. II ), linalool (no. 19), and
3-mcthylbutunoic acid (no. 33). Le\'i.'ls of these compound were cOInJXlmt ivd) low
with respect \0 previously ideillili cd enillponelllS.
comiloncnts of saffron_ The characteristic aroma of s.1Jfron has been
previously defi ned in the lit erature as "sweet, spicy, l10ml odor wi th a fany,
herhac..:ous ulldertone (I)". Despi le quantitative and qUlllitutivc differences
SDE and DE ext racts, both had di stinct saffron-li ke aromas, although their :lromas
were not identic.11. DE extr:lCts had sweet, spicy, and l10ral nol1.:5 ami were
representative of dried saffron; whcreas, SDE extracts hac! nUlly, cooked ri<:c- and
hay-like notes but st ill maintained a d..:finite saffron-like quality.
A \otal of 25 aromaactivc compounds were consistently detected by GC-O and
AEOA in SDE and DE extrl1ets (Table II). More compounds (23) were detected in
SDE extracts lhan DE eXTracts (22), wit h 18 common to both extf:l<:IS. Aromas
deleclec! during GC-O could be grouped into Ihe following general catcgnries:
Silfrron-likc (nos. 10, 30,3 1,34,37, nnd 57), falty, stale, biller (nos. 16,17,42,43.
and 50), sweet, floral (nos. 19, 25, 53, and 59), sour (nos. I. II, and 33), nutt y,
cooked (nos, R, 9, and 12), eilrthy, plilstie (nos. -1 and 7), buttery (no. J) and
onion (no. 40). In general, compounds that were not COmmon 10 buth extracts had
low 10gj(FO-factors) I). For compounds detected in both ext racts, most had higher
loglFD-raclor.;) in SDE extracts. Thi s was csp<.ocially true for nUlly, cooked aromas
whieh may have been thermally generated during SDE exlraction.
Compounds having high 10gj(FD-factors) (>2) in oot h extmets were described as
S;lffron, drit:d hay (unknown, nos. 34 ilnd 37), saffron, tea (s.1franal, no. 31) and
cooked rice, baked bread (no. 9). COtllPOU!U! no. 37 had the highest log)(FD-factor)
in both t:xtraets. Interestingly, this compound and no. 3.1 were detected at only high
dil utions (> I :9), i.e., these compounds were not detected Juring GC/O of concentr:lled
extwcts. This phenomenon was observed previously (1-1) ami may hal'c h..:en due to
1, C \l)WAU.AllEH ET At. AI/alysis uf Saffrun nlWflr


<

.2
]
Q
"

"'

,
e
<

o
I
l

'"
"

__ "t:l_-';:_
V V "'....: v ... i vv ", v -;:. vv
VV
0-'" -.:> r 0' ",
- ,., ....... Q() Qo - - - _. , ,
1'0 JO ;}!oI!lt '(60 :lU!UO!l{l:JW WOlJ 3P(4:JP1"e J3)j:l;}JlS Il 'f I ' OU pm: '(gO
l:mp01d UO!PC:lJ PJemeW Il '8 ou se 11:>115 spunodluo:J pJl"eJJU;}ji AIII:WJ:J41 ' AIJCHW!S
'AI:J,,!p:Jlls;)J '{al -11811 LO'O pUll '(9Z) 1'0 '(fll 1'0 JO 5:Jnlll" P\04
s
:J
J
lll ;),\C4 (I'll
f1' rue S"Z ' L 'SOli sl'.)n\)O.Id UO!lllppW P!d!1 341 ';)ldw"eX;} JO:;l SPlutIS.)J1{1 MOl ;}AC\{ 01
IL\\OU:>\ ;}lC puc spooJ Aumu 01 SIlWOlC "l!ISUJ1JilJIl4J lJedw! spunodwo"l ;loSJ41 '{O'oo
'\lIlUJ!PC"l;}P- \,Z-{3':\)!leJ P:JpJ 'AlleJ puc 'lS"z 'OU '\euJ!pcuou-9'Z-(Z"::J>! JJqIUll"lnJ
'\JJ, ....5 ' \(1 ou '\llulldllJd(o!41IA1]Pwl-(\ o)ll\od P"l\il<{ ,(S ou ';)Il!1
oJJ
'(d-l-1AI;}JC
-Z) UJo:;xlod 'AIlIlU '(L ou 'JUO- (-lIJI:"lO-I) .(II\JC:J 'mooJ4SIlW sc P-X"!!l:"lS;J II JJ;) ..... SIlIUQJC
J!:l4.l '(spunodwoJ pJcpue)s ISU!1l8e p:;u1l<lWO) "J'!) S3!lJ.xIOJd eWOJIl pull s:Jnlc"
HI J!JI!I uo p:!Seq pJ!j!IUJP! JJ;}.\\ spunodwoJ lCJ;J.\:!S s Icu'iI!S SI"l" .\\1) Jna '(l-U)
SllIllld JO S\U;)U)!lSUOJ punoq .(lICJ!P!sOJ.( \8 U,)IUIUO:"l :JJI! IOm14\JIAU:JI\d-c: PUl1j
OO
IliUn
'(H 'ou 'IOtIC'l;;JIAU;JI\d-Z) ")SOl 'IUJOU puc ([[ ' OU 'IIPC '.)!OUC\IU.IlA'lP1U-() I!IU)
P3!JP 'lUllS 'UJIlOJ '(6 \ ou 'IOOIIlU!O J]lI)1lsii.JuOtI 'ICJOU SI! p-X"!!JJ')Jp JJJ.II S!4
1
U!'\)!,II spunodwo:"l PJL)l\UJPl ,{pA!)!SO,1 ! t\IOlj 10 ]U JO ]OS U'lI!J U! I < (SJU)JI!)
O:I/llOII s:J!l!su:J\U! 1l1U0Jil ;))BJJpOlU 'l11.\1 spUnodlUOJ j O l-X"!wnu C ;)JJ, .... JJJIIl
'11 pm: I SJ\'-lll.l. puc Z 1)IlC ! SJJnS!,1 U! Lt ' 0\1 lInd JO w!u)XIds SSC)'-i T ;u n':i!.'I
7./W
1\ -rT'rr-rrlf[- -

-
..
...
'(II-af) UOJ.LII1S )O )u:)uoduru'.) 1'.)1I<I1U! -J:)PCJlllP sc
r:JPJc1bJ u:JJq Suol Sl1lj ICUI1JJl.'S ;}:lIl!S l'Ju!s!JdJns IOU jiU!PUU S!LJ.l ' "eWOlll UOJ.LI
1lS
01 J:lU1!llodw! S\l iJU!lllJ!PU! 'L( ou JOJ :;)SOIll Sil 4iJ!4 SC AtJCJU ;}Jl.II (If 'ou) ICIlCJjllS
J ()J "4.1. 'SJn\l1t1 HI l'Jtqpuods:';IJJOJ J!;}tll 111 S\IlU:J!S SVIl
0\ ;"I!l\l Pl!j!IUJP! ;,q )OtI PIIlOJ \7 [ pue 6 'sou spunodulO) '({ {) Cl!ZJIJJ PIIC
) 0 Il ICP ICJ1:).)l.\s SS:IlW .\'1 pJIJoddlls S! JJIlI:MlJ IS J4.l ;;>uo-I-UJ!P
I1XJ
4
U
ll,('.)
jU JJI"lpnJ1S pue to:l l1;) JO IllnUlJoJ JClrI:lJl
tu
:;Illl " I:;I,'!)Il)l!J) SII.'<\ LE 'ou plll10dwu:l '( "J.InS!J) Ul1lJIJ:xis SSI!IU sl! uo p:lSCU
' SU
UIlI
10"l SCA'-Ua UI> ICUIIJjllS JJUC p;))np puc suwn\oJJ StuS"-{](I uO ( I [ 'ou) ICUCJjBS
P;) lrll;) punodwoJ J'l) .lJU!S P;}!J!)UJP! :II.{ PItH):) L( ou punodwoJ 01
IUllJ p;)ds s<;CUI puc lIc.xI ;}11.l 'IlWOJ"C 8U!lnp lJ!lJIlJ lIC .{'1 RU!'ISIlIU JO
(tl JU Ii ' \JU punoduJOl il u!IIlIJ JJ!IJIl;l UI1 0\ SIS!I:J\lcd JII\ )0 uO!lc)dcpl! S100JJ
"
,lU,OUl:l uwffu!,' fu - IV.I.:i 'L
O'f;: i
;!"''''m co
a;::-::5- 1(3
(.>'<:0"'-;:::0-
1'> 10 !!.
,.;;-,;'< n < ...
(") ::r _. - 2.. n
o '" "'- '" 0
3;;;Jg-;:;
'8 =>-<"'"
co fr3g 5''8
"' .. ... _oo=>
... o Co.
-o:o":!()-
... _ ,." 0
:O='S""O-
=!g,. g-
-. I: ... 0 ... );:
a.. E. g ? :: -.
'< ._. ... '" '"
_. () => n <.l ..,.,
Co. 0 "" 0 ijQ _ .

:::"''8 0 '80" <;
- I: '"
Co.5Q.::> --. -
(:1..:1 Co.'Tl",

os-2..;:o,,"''>
3 :::". a "<' 0
;i '< !!...:;!.. Q.
:1'-';;0'<0,""",
:::> - . () _ . ......

;::;" ;:l 9;:l .., "
",:::". - = . .g:-
3
F:1 " 0 "a n' ;z:;
<>,.,

"1? n i:: Q. =.
,.; 0 ;:l " 0
-3"" 0:>
2' "0 - tTl-'
3 ;,; g ::! . - 5.
:;:: 5';;:; a::'! n'

::r(i;!!.:> ;::;-
"O:::>-...:Q.o.>-
i:: - _ '" :::> ...,
0'''''',: .... 2''

l'T"'-:o"'''.

<"> " :::l . <> :;>

;;I:a ..
";3'<",,,,,-
-0 <> :> ;;;. 0
co O(:l..-()

,..... -''''' <>
S. g.

S y;



'" (:I.. co :;'
""",;::>",..0

o :::"...
;;32g:o:
- '" - .. <
....' v, ..... v. .jO. "" .f'o
'-0-> ...'0 .... ''-'0
...., ..., ......
-> "" ....'
w"'
o
t's ';-'E,@E!"";-'E't'g!"'
" iT; ' X'q ' s: g g: 9
Q.o<">oYoo-"':(:I..o!!;><' 0

'" :::> - ;:s "" 0 '" 0 '" ;::I _ ;:;;: _ . ::>
..., !l' ::r,< a .. 3
, ::r 0 .. , CO::r n
!'" f,> ><f- ;!-...:s-
..... 2 f,> Ii"" 0"," ,<
, (:I.. _, . 0 n -
9-, ..;;;. g 'f' ;:; . ...... .!..
3 0 '5, <> i:: to,
!!. '3
::r - 0 ... ""0'<
K. <>!2
, ""'< co
7'''ir
N "
9 IS.
2 n'
" ,
'!'

.0

\0000---.12 ..... ","N,,,, ..... 00

e,
v' N
" ::'>:::o:::><;>!!t!P:
2
o:-c-:;<-':::>

I i; /"i t: r. - .
6.._ . ;:s-Q...., 0
-'
g'
'"
" "
, ,
if


6.. ...... I\I\ -=>-
:'---0(:1.."00
w
1\1\ _0_ 1\ _
.... Q. 0 - ".
e'
e,
v '



o
t <3

o ,
, , .
:.. ::.. :s
- " 0
ri'_"
.
2:
2.
_ 2
. -.
'< -
..... , .........
v, :.- <=>

..... .f'o_
v. v. 0.

,0
o
e,
" .

P ?
;:j
.. 2:
-
0

'JO , .....
t"o
"'
'- 1 o-

AH.L"II 11:IID HO,\VH 'S::l:llolS
z
0
n

1l
0
,

Z
0

o
'i'


'"
yo
3

>
"
3
0
0
n
"
.g.
o
' .
0[0' m.

,

"
;;:
"
,
=
a.
'I!
78 SI 'I CK'>, FL\Vt)1{ <.:JJEMlSTltY ANlJ ,\ NTIOXlIMNT J'I{OI' E,; UT1K"
(27) and 0.2 Jig/L (30), r.:spt:\:lively. A compound willi a saffron, !Ioral, hay- like
aroma \ViiS tenHllivdy identi fied by its mass J,5,S-trimcthyJ.).cyclohcxcn-
I-one (no. 10). Unidcntilied compounds within this group described as stale,
bill ...r (nos. 16 and 17), green union (no. 40), lmu flowl, rose, saffron (nil _ 57).
Compounds to have only minor roles in s.a llron aroma 1I0)(FO
factors) < II werc ,]cscribcd as sour, dark chocolme (unknown, no. 1), bultery, crcam
(2,3bulanedionc, no. 3), plastic water boule (unknO\\11, no. 4), vineg;!f, acidic
(acetic acid, no. II), SInk, soapy (unknown, no. 42), fruity, stale (urWlown, no. 50)
lllld COllon candy, strawberries [4.hydmxy2,5-dimet hyl-3{2!-l) furrulOne, Ihl . 59[. With
its low threshold compound no. 59 is <HI important component of scveral foods
(18.3/) .
Conclusion
Through usc of AEDA and two extraction tcchniqucs it was to indicmc
important arorna-:lctive components in saffron. Aromas contributed by 2- hydroxy-
4,4,6.t rimelhyl -2,5-cyc1ohexadicn- l -one (telllatively identilied), safranal, and an
unidC'ntificd compound (saffron, dried hay Moma) were predominant in Mffron.
I [o\\'cver, it also was app:lrent that other componenlS contributc 10 saffron aroma,
sumc of which are thermally gcncmted during cooking of s..1.ffron. Many compounds
were identified for Ihe firstlimc as constituents of saffron. Results uf this sludy may
Ix: useful in devclopment of analytical strategies [or monitori ng s..1.ffron Oav(>[ quality.
Acknowledgemcnts
Approved for publication by the Mississippi Agricult ural and Forestry [xjJ\:riment
Station as manuscript No. PS&919. Support for this study was provided by the
Agricuhl!r:ll and Forestry EXjJ\:ri m('!lt Station.
Literature Cited
I. Ft'II(1l"o/i's llondhook of MaYor Ingrcdil:nl.l". Burdock, G.A. Ed., CRC Press . Bt)Ca
Raton, FL, 1995; Vol I, 3rd cd., pp 247-24K.
2. Sllmpathu, S.R., Shivashankar. S.; Lewis, Y.S. Cdt. Rev. Food Sci. NlIlr 19!i4,
20. 123157.
3. Oberdk-ck, R. D/wh. Lebensm.-Rulldsch. 1991, 87. 246-252.
4. CastcllM, M.R., Montijano, H. ; Manjon, A; lborra, J.L. J. Chruma/ogr .. A 199J,
648. 187190.
5. lborra, 1. L; Caslcllar, M.R., Canovas, M., Manjon, A. J. Food St.i. 1992, 57. 714
731.
6. Sujata, V.; Ravishankar, G.A.; Vc[[kataram;}n, LV. J. Chroma/ogr. 1992, 624,
497502.
7. Tarantilis, P.A.; Polissiou, M.; MiUlfait, M. J. Chrolllli/ogr .. if 1994, 664,55-61.
8. Tarantilis, P.A.; Tsoupras, G., l'olissi ou, M. J Chromu/ogr .. A 1995, 669, 107-
11K.
9. Zarghami, N.s.; lIeim:, D.E. Phylochem. 1971,10, 2755-2761.
10. Nanlsimhllm, S.; Chand, N.; Rajalakshmi , D. J Food QUill. 1992, 15. 303-3 14.
..
II Riidel, W.; 1'1'1. J fligh Res. Chmnw/ogr. 1991 , 1-1, 771-771.
12. Auee, T. In Hm'or M"aSllrelllt'lJI; 1-1 0, ('.-T., [I.-lanley, (;.1-1 . Ells., Dekker: :\"C\\
York. 1993; Vol. I, Chapter 4.
13. Grosch, W. Trends ill Foml Sci rrrllllol. 1993, 4.61::-73.
14. Abbott , N.; Etievant , 1'. , lss.:lIlchou, S; Langlois, D . .I. Agric. Fo"d ('hem 1')')3 .
4/. 1698-1703.
15. Guichard, 11. ; Guichard, E.; Langlois, D., bsandlOu, S.; Abho!!, N. 7..
/'cb..".ml . UIIICIT /<,,'ol"sch 1995, 201. 344-350.
16. Cadwallader. K.R.; T(lIl , Q.; Chen, F., 1\kycrs. S.I' J Agric F"od Ch,.lII. 1')95,
43. 24322437.
17. Chung, H.Y. ; Cadwallad.::r, K.R. J. Agrh-. Food e"1'III. 1994, 42. 28672869.
18. Guth, II. ; Grosch, W 1I.r"elll/o!. 1993,26. 171-177.
19. Schieberie, P., Grosch, W Z I.ebell.l"m. Unlen. Fondl. 1987,185. 11111 3.
20. van d.:: n 0001, II. ; I'. D . .I. Chroll/olOgr. 1963, II, 463-471.
2 1. MacLcod, G.; 1.M. J'hywch"m. 1990, 29, 165-172.
22. VasscrOt, Y; Arnaud, A.; Galzy, 1' . AclO 8io/edlllol. 1995,15.77-95.
23. Shu, C.-K.; LawrcncO', I1.M. J. A,;ric. Food e "e",. 199-1, 42. 1732- 1733.
24. Lipid Oxida/ioll in Food, 51. Angdo, t\.J." ACS Symposium Series No. 500 ..
Amcrkan Chemical Society: Washington D.C., 1992.
25. Whillicld, F.13.; Freeman, D.J.; HI. AII.llr. J. Chelll. 191::2, J5, J73-.18:; .
26. Forss, D.t\., Dunston.::, E.t\.; [hullshaw, E. II.; Slark, W. J. Food Sci. 1962, 27.
90-93.
27. Bullery, RG.; Seifer, R.M., Ling, I.. C . .I. Agric. Food Chem. 1988, )6. 100G
1009.
28. Schicberle. P. Z. l.ebl:lISIII. Unrcn. F()1"Sch 1990, IYi. 206-209.
29. Fmss, D.A. J. /)airy 1979, 46.691-706.
30. Glladagni, D., l3ut1cry, RG.; Tllrnbaugh, J.G. J. Sci. fi,(,,1 Agric. 1972, 23. 1 .. :;)-
1444.
31. Meyerl, F, \{ cgula, N.; Thomas, A. "111'/0<"111'111. 198
1
), 28. 631-633
"
'i
I
.,
i
I
I
Chapter 8
The Charactel"i7 ..at ion of Volatilc and
SCl11 ivnlatile Components in Powdered
Tunncrk by Direct Thcrmal Extracti on
Gas Chromatography- Mass Spcct romctry
Ric hard II. lIi scrudt
l

l
, Chi-Tan g li n', and T.
' I)CI':trtnlen t of Scie nce and ICent er ti lr ,\dv:l nced F fllMI
Tt ... hnulu!U', Couk Culkge, Ru tger s, The State Uni" er si ty or New J e r sey
New NJ 0890J - 02:l 1 '
Fivc commercial powdered turmeric samples werl' unalYled to identify
vulallk and semi-volat ile Components. analysis did not include
lIoll-volatile eurcuminoids. Structural infomlution was obtainoo by
direct tl.lennal extracti on Kas Chromatography-mass spectrometry (GC.
MS) uSing the electron ioniz.l1ion (EI) mode. Semi-qulIIttitativc values
lire also rcportcd.
Turmeric tJClongs to the famil y Zillgiheruceae along with the ot her noteworthy
memhers ginger, cardamom, and galallgal root. It belongs to the genus CIJrn/;',u
of of species of plant s th"t grow from undergf()ulld
roolhkc stems. EconomIcally, the most importallt species is l/I)/l/1'.wicll. Tunneric is
grown in rainy regions of the world such !IS China. Indonesia. India, JHllI"ie:l.
antll'eru (/).
Tunn<:rie has been used since carly times 10 cure cverything frOIll leprnsy to
c<.>mmon cold but is probably best known for its properti es as a carminat ive. It is
used to provide aroma and taste to foods and as a dye for f.lbric. lIS as a dye
arc poor because tunneric is not light stable. Turmeric is also used in some cultures as
a to lightell the skin.
. The prim:,ry II ses ofturmcrie today arc liS a component in curT}'. us a colnring
111 J \'ilrtety o f drted and frozen foods and as a po\\'dercd spice. Tht."" powdered spire is
prcp;I: t.""d by harvcsli ng tbe fresh rhil..omes and boil ing them to gelatini7.e the stareh
,IUd the color. 1he rhil..ome$ arc dri ed in the sun for 10 to 15 days and then
grollnd IIItO a powder (I). The work prescnted in this report is on the anal ysis of
volD.tilc and semi-voll1tile components in powdered turmeric. Thi s work does not
include the non-volatile curcuminoids
. . The primary focus oftunneri c research today is based on its properties as an
antluxl.dant (1.3) and as an anti earei nogen (4,5.6). The anti oxidant of
tunnerte <Ire based on the abi lity of eurcumin to form complexes with metal s and to
fOrln.3 resonance stabi li zed free radical. CUTCumin, along with Ihe other eurcumilloids,
@ 1!197 Amelican Chemical Society
8. 1I1SEIWllT in I'tJlI'df'rf'l/ Turmeric
delllethQxycureumin ami bi.nlemclhoxycureu111in, arc nOllvolalile eompOllcnls of
turmeric that impart its orange-yellow color. Curcumi n l,olllp1cxes" ilh
metals, such as copper and nickel. thilt init iale or catalyze free mdie:ll ox idation. Thi s
property enables cureumin to act synergistically with other antioxidantS in inhibiting
free radical oxidation (7).
The bond dissociatiOll energy for the phenolic hydroxyl groups in eurcumin is
low and a phenoxide free radical is easily formed (Figure J). This free radical is
resonance stahili1..ed. The odd electron can be delocali"led through the entire molecule
hecausc this molccule is a completely cOlljugated system. Figure 2 shows curcumin
acting as a free radi cal SCa\'enger, reacting wilh a pero.'(y free radical , fonning a
hydropcroxide and the eurcurnin free radical. The other property ofeurcllrnin thai
m:lkes it a good antioxidant is that once the free radic31 is fonncd it will not go on 10
react with unsaturated falt y acids and initiate or propagate oxidation reactions.
Tunneric has been shown to inhibit the initiation and progression of cancer.
Azuine and Flhide (4) demonstrated Ihal turmeric inhibited forestomach tumors
induced by benzo[ajpyrene (01') and skin tUlll(lrS induced by 7.12- dimethyl-
benz[ajanthraeene (DMBA) in mice. These studies were conducted al the 2% and 5"10
levels. BenzolaJpyrene. which is a component in eigarellc smoke and barbecue foods,
is a procareinogen that is o.'(idized during I'hase I oxidation in the liver to an ultimate
carcinogen (Fi gure 3). The ultimale carcinogen forrned is 11 good aikylating agent and
can react with DNA to fonn a mutat ion. Tunneri e h:ls been shown to inhibillllmors
formed by this mechanism by inhibiting Ihe P450 monooxygenase system. Tunnerie
also increases levels of glutathione and glutllthi one S-transferase :lelivit y. GIUlilihioile
is an emlogenous nuc1eophillie chemical that can react with ultimate carcil1\""Igens
formed during l'hllSC I oxi dation without forming tumors. Turmeric has als(I been
shown to inhi bi t the progression of cancer by inhibiting ornithine decarbo xylase. Thi s
<: nzyme dccarboxylates ornithine, an endogellous :Imino acid, fomling a polyamine.
This polyamine sti mulates cell and consequently tumor gro\\1h.
In another study. Mukundan, e/ (If (5) showed that as little as 0. 1 % IUrmnie in
the diet signifi cantly removed UP-DNA or inhibited the binding of DNA with
01' formed in rat liver. They also observed 0.03% eureumin to he more effective in
inhibiting OP-DNA adduCIS. They did not propose:l mechanism for thi s obse .....-ation.
n ac kgrnund
Preliminary work by these authors involved the analysis of components in lht.""
methanol extracl of powdered turmeric using thennospray and particle beam li quid
chromatography-mass spectrometry (LC-MS). Thermospray is a soft ioni1..ation
tcchniquc producing protonaled and dcprolonated molecular ions. providing mokcular
weight infonnation about analytes. The therrnospray LC-MS chromatogram for the
methanol extract of powdered turmeric is shown in Figure 4. The eureuminoids were
identified based on their molecular weight. Also detected were three maj or late eluting
components as wel1 as numerous minor components. The molecular weights ror these
components were obtained from thc thermospray Illass spect ra. Particle beam LC-MS
was attempted 10 obtain structural infonnation from EJ-mass spectra, for the
C1Ui.MISTJ.tY AND ANTII'XI1)ANT "!lOI' ERTIES
01-1
H
o ' 0


.0 4- I "" Oil
H
o '0
01-1
Figure I. Resonance SI2biliutioD or tbe Curcumin Pre''' Rltl iCliI
l - OO + HO - C
P, ro,v
Fr" R' lIlul
Cu,curnln
l -OOH +<O - C
C .. ,cumln
f, u Radle.1
l-H + O -C---7(-
Un" !",,,t d C';,cumln
f . tty Acid Fret Radlc.1
"' igure 2. Reaction Mechanism for Curcumin ACling as" !'ree
ScavenJ:er
8. UISEIHlllT \,:'1'
bentO lil! pyrenc
procarcinogen
ONANt-I:.!


""
ullimatc carCInogen
Fil,;ure 3. Phase J O).itialiOI1 IIf Bcn'l.oluhlrrcnc
I
,
5


..

"

c
"
.,
,
"
15
limf\min. )
I
"

c
c
"
,
,
;:

2S

c
-t
,



,
,
I
311 35
r igu r ... -I. Thl' rmosll rll), LC- jI,'l S Chrum'tlngra m of a l\IeCh:UlOl E\tnld "f
Powtl en:u Turmeric
"
j,
" ::
i
!
SI'ICt::S, HAVOM CII EMISTRY AND ANTIOXIIJANT I'ROPEKTI ES
components detected by thennospray. The problem encountered with particle beam
LCMS was that none of the latc eluting peaks detected by thcrmospray were
observed. This phenomenon was attributed 10 the volatility of the components that
were lost during removal of the mobile phase: in the particle beam interface. Structural
infom131ion for the volatile components in the methanol extract oflUnneric was then
obtained by dinxt thennal extraction GCMS.
Materi a ls
Samples of powdered turmeric, consisting of 5 different eommerciallabcls, wcre
purchased locally. Chromosorb WHI' (8011 00 m) was purchased from Supelco, Inc .,
Bellefonte, PA, USA. Chromosorb WHP was used to dilute the powderec tunnerie
and was conditioned at 300 GC for 2 hrs. prior to use. d,-Naphthalene
(98+ atom %D) was purchased from Aldri ch Chemical Company, St. Louis, MO,
USA for use as an internal standard. The internal standard solution WIIS prepared by
diluting 10.0 mg ofdlnaphthalene with 10.0 mL methanol. A 60 m x 0.32 mm (i .d.)
dj = 0.25 ].1m, 00-1 capillary GC column was purchased from J&W Scienti fic,
Folsom, CA, USA.
Experi ment a l
There are many IIltematives to direct thelTnai extraction methodology. For instance
compounds could be isolated using the Likens Nikerson system, which provides for
steam distillation of the aqueous sample wi th ether extraction on a continuous basis in
onc apparatus. The disadvantages of this system are that the steam distillation utilizes
high temperatures ( IOOGC) which may dccompose labile compounds or cause other
thermal induced reaetiOI15 to occur in the product being llI'Ialyzcd thereby complicating
Ihe data interpretation. Additiona1Jy, the ether extract from the Likens Nikerson
system m:eds to be concentrated prior to GCMS analysis IlI'Id there is always a risk of
losing highly volatile components from the sample during the concentration step
which is usually performed in a spinning band still or Kudema Danish concentrator.
Another alternative is to usc vacuum steam distillation combined with
cryotrappi ng. In thi s technique the sample is steam distilled under reduced pressure so
that lower temperature isolation is permissible. The steam distillate is then coooenst.'d
in a series of four or more eryotraps cooled with liquid nit rogen and or dry ice acetone
slurries. 11le maj or disadvantage of thi s tcchnique is that the first several cryotraps
rapidly become filled wilh icc from the large volume of distillate. After thawing the
traps, the water and traps must be extracted with organic solvent such as ether or
mcthylene chloride to isolate the volatile organics. There are additional problems with
cxtraction efficiency for polar organics that are not easil y extracted from the water.
This results in large volumes o r organic solvent that must be concentrated, once again
risking thc loss or highl y volatile compounds in the process. The extracting solvents
used must be or ultra high purity. Frequentl y the extracting solvents must be distilled
before use because concentrating large volumes of solvent will concenlrate solvent
impurit ies as well. Dhmk rons of solvent, with no sample, are often needed to identify
8. IIISEROOT ET
Volotifl'.$ in Powderrd Trmneric
85
artifacts and impurities arising from the solvent concentration step. Extraction
experiments arc lime consuming and costly since elaborate glassware setups, vacuum
systems. and concentration equipment are employed. . .
Di rect lhemlal extraction involves the extraction of volati le and scnlH'olall lc
components from solid samples using heat (8). Since volatile compo.ncnts are ..
desorbed directly onto the head of a GC capillary column at subamblent conditions,
samples shoul d contain less than 510% water to prevent icc formation at the of
the column. Direct thermal extract ion is a single stcp process that invol vcs mlmmal or
no sample preparat ion making it advantageous over other extraction techni ques.
Additional benefits are no solvent disposallll'ld no anifact peaks in the sllI'Ilple
chromatogram from the concentrati on of large volumes of extracting solvent.
The instrument used for this experiment was the Shon Path Thermal
Desorption (SPTD) system ID]. This is a commercial system developed join!!)' by
the Center for Advanced Food Technology (CAFT) at Rutgers University and
Scientific Instrument Services, lnc_ (SIS), Ringoes, NJ, USA. The TD-3 was
connected to a Varian ]400 GC (Sugarland, TX, USA).
Powdered turmeric samples were prepared for quantitati ve analysis by
thoroughly mixing 20.0 mg with 180.0 mg of Chromo sorb (80/100 Fifteen
(15.0) mg of this mixture was weighed into a 10.2 em x 4 mm (I.d) glass
stainless steel thermal desorption tube (Stientific Instrument Services, Inc., Ringoes,
NJ, USA). Si lani7.ed glass wool \\-"as added to both ends to contain the sample. Both
the glass wool and thermal desorption tube were conditioned at .
300 "c for 2 hrs. prior to usc. Five (5.0) ofdl-naphthaiene solullon
(1.0 mg! mL CH,OJI) was spiked into each sample. This was equivalent to 3,333 ppm
based on the weight of turmeric in the turmeric-Chromosorb mixture. A of
nit rogen (80 mUmin.) at room temperature was used to purge the desorptIon tube (30
min.) of methanol from the internal standard solution.
The thermal desorption tube has a needle at one end and is threaded at the
other for connection to the SPTD system. When an injection is made, the thennal
desorption tube ucedle pierces the GC injection port septum. Heating blocks close
around the thcmlal desorption tube aud provide rapid heating. Helium passes through
the desorption tube at I mUmin. and the volatile components are desorbed onto a 013
I capillary column held at20 c with dry ice (see Figure S).
In thi s analysis the volati le components were trapped onto the head of the
capillary column at suOOmbient (-20 "S) and .separated by temperature 0
programming the col umn from -20 to 150 Cat 20 CJ mm. ISO to 280 C
at 5 CJ min. Quantitative results were obtained using a flame 100l7.ation
(FlO). Structural information was obtained using a Finnigan MAT 8230 magnellc
sector mass spect rometer, operating in the EImode, and interfaced to a SS300 data
system (San lose. CA, USA). TIle details of the instrument condit ions lire listed in
Table I.
86 FI ....\ \'OI! Cll E,\ JI S'I"ln' '\/'1 11 ANTI CIXI IL\ NT ]' ll( ) I'FnTIFs
_ _ - C:",-i" Gil) flow
Desorption Tube
Heale r Blocks
Needle
1 _ __ GC Injection Port
ToMS
Figurc 5, Schematic ofthc Shorf Parh Thermal Desorption System
II. III "; UIII I)"[" K f AI.. VII/lItHIC,I' ill l'llwll end 'J''' fllll'fir
'1' iI bll' I. I liS! r UIIlCll1 Cond jl
InslrunlclIl: Shurt P"th Thcrmlll Dn orpt ion Sysh-m TI}-.\
Carri.: r Ok)
Hc;lling bhwk
Desorptio n lime
1 mUmi!,-
220"C
5 min.
Varian 3-l0U GC
Column
Temp. prog.
luj. h: l1lp.
I'll) temp.
DU-I 60 In x 0.32 mm (i.1I.) capillary colullln, 0. 2:' ,!ill
2U to I SO "c I@ 20 "Clmin., 15010 21S1J "C@ 5 "CI mi n.
220 "C
J2j (Ie
Instrument: "' inni gan MAT 11230 r. b ss Sp{'ctnJll1{'h' r
Sl-an!led
1.:1 -111odc
GC MS inh:rfacc lin.:
MS inkll.;mp.
10 11 source IClllP
35350 amu
7IJeV @ II1lA
280 "c
240 "C
2!S0 "C
Figur;; (, shows ;l GC chromatOgrllll powd"r oblaincd by dir..,..:1 lhcrm::1
extraction. The major laic eluti ng pt:aks in Ihis I:hrornalognlill were corr..:I"h;u lI'i\h lhe
]all' eluting peaks JClcd<.'d by lhnnwspray LC-MS based on lll<>lccular wei ght ,llId
n.:lalivc intensity.
The data oblained by (j( : a flame ionizalion Jctcctur were II) ,1bl ;lil l
for the \ulali k c,>mpvnellb in fivc hmlleric
C;lkul"tious wcrc- on Ih.; respunse "f eaeh the
response uf
ds-n"phthalen.: uscd as an inkrnal standard. The fullowing cqumion usnl tor this
Results af.; i ll T"bll' II .
where
ppm '" 1(ISTDWT I' g x AREAs1'1) I (AREA
s l
" x SPLWT i ll x
lSTJ)WT is Ihe weighl oCthe illlcrn,,1 standard I" c;H:h 1)1
)II:
AREA
sPI
. is Ihe arCH lor rUllIl'O)lent "I" illtn ..\t
ARE-Ann is till: pt:ak ar<.'a li'T thc ,1""d;"d
SI'L WT Ih..: w.:ight orlh..: !mHII"Storh )\1ixl"''' ,,,'!'k :
inlo thc luh.; in g!'allls


"

e-
"
a"



3
..
'"

,
,
"
"

d
3

"


3






-!

"
3

,.

j '
" ,



;;


"


,
"
L_.,..
..--

E

,
c
"
;;:;;.:;;;;;.;:: n .p'\", ...

111!,::"a a:.

.:
=::: 0. ..
. "
i
,

.,
i
1
,
.. "' "
, ;; c

i :
I. :0 ;
4
I
i
- ...
,. r '!; ..
ii '" i-
n .
'1 ' !
,
'. ;:;
'i
\! f
, .
i
!
."

,


"
,
n


"
o


" <

o
>

o



o
"
,


"
-
90

o

"
~ ~
B
~ 0<
"" ::;:
,
- __ -=i
r

1


" -.
- -CO.!


,,--=/


=

"
.,
,;-
v >
~ ~
."
...; ' .....
~ c
-

J
1-
-,
,
"



'>.
/
- ~
.,
:i : ~ :
~
.
'-
;: "'
E
/.
.
:... -:

1
I
'"
.,
1

'"
,",II .. ,,"
"
"IW .
"
105
!II
,,' \ I
,,' i';1. , I" , [, .... ,J,," ' ....,...,....,-, " Ir
"
'"
.. ..
"" '"
''"
1611
""
l llU HO
amu
Figure 9. 70 eV [ 1Mau Spt'l;"lrunl orCurlone from ref. \ 8. Copyri ght 1996
with kind p.:rmis."ion from Scien.::e NL. S:lm Burg<!"rhartst rlal 25. 1055 KV Amsterdam. The

g
;;
o

'-" =--"0

>= 0-
a g
(> _ . -
fr f. ;:;
::: g :
....: ::;>?"

ri :;: D
z
=.
;; .., CI2.
:::.. n

_ 0
!" ;;. c..

z
,-

" ,

=:
'\1

a

"
e'
< -
" i" .\ .J"
.: i

,

,
.'
,
t
>

,
2

,
"
?: ..

;: j



-
;
,
,


;;
"
,
,
"
,
,
t
,
,

2
-.
;! l
H
i "
,
\

j
"1
,
}

,
l
,

"
,

,
,

' 0
H
o !
, ...
.

,


,



"

...
,
..
"
,
1
"

.; L
i .
,


"1
,
,
i
"
,
"


,
,
,
,
i
,

, -
,j
l
-
,

,
;::


,
t
.,
,

,

,

2

,
\
,
?
,
1
>

,

,
2
j
i
C o
"
H

.,

,
,

,
i
,
,
z
,
"1

.


,
,
,
C

- i

"
,
1
"
,



,
,

,
,

,

-.
;: ..
"j
E j
0.
,
l
,
,

,
1
,

,
"
,
...
"
,
!
i
!
,
I
,
,


:
'f
,
,
r
,.
,

,

,
1
, -I-
", ,.
Hi
;: = !
,

,
;:
,
...
"
I'
, ,
!
\
,
"
...
,
,
..
"

"
"


r;
...
o


:;


f
::: ...
... "
::: ;.
... .
. ;;
, .
... ...
=:
... -.
".

...; '
...
-. ;; ...
... -
:;; . :::

...
i-
;;
...
"

,";

"

...
>
...

2
;;:
...
"
...
>
>
c
>
z
...
x

I.'
...
,



z

c
...
::.;
,
..
t

;:.
?

a

.
...
,
S I'I I ' I':": 1'1 A H)1l CrrEMI SI'Il\' ,\ NII ,\ N"I'10XII\,\1\ r 1'111 11' 1'.11"1
" .. , ... 'h)).2-,:, ..... ,."""". )"""'"
IN)
... ;11,"
IMW- IS1)


""nl"'''"''

" .. m .... '
(MW- 11U)
'"""""'"

...

a,
(MW_11&1
"'_ 'TO
U
l'h)J,u,) 1 :..",1 .. , ......
1\1 \\' 0) '01
.... ,.j 1 ()"O .......

do:h)'.ll ...... """" ... ,_d)
1.\l W-JOIJ)


;
"",,...
),"\\ 1 ...."
f'igllrc 10_ KU(llI'u :mtll'roposed S'rurluns for Vnl.lli le ami SCllIi' \'o!alilc
Componcn ts in Tur meric
8. 1lI:o; I': IHlU' f ET ,\1..
........ m""""
(MW-!16)

t)ll

(M\\'u21u)
..

[ Jnid..:milit"d
(' ump.'JlCIll
, ...."1 ,""19

1'"IMilc.I ill Tllrlllrrif:
''''''''''''''''
-1 1)
Ullidcnti!i..:d
Cornpon<! lll
''''"I.",nJ S
1.\IW 216)
.... JI

'''' '' I''",''J IU
-11 11
!-"igllr'C 111. (',m/ill/wel
, .. ,!<>n<-
(.I'II _lUI
"""'I" ..."d 6
(\IW- 2IS1
.......... .....,


i
!
I
!
I
I
,
,
,
,
I
,
,
,
!
I
i
,

,
!
,
!
1
2
\
,

.,
,
,
,

i
"

"
"
i
,
,
,
,
96 SJ' ICES: HAVOK CHEMISTRY AND ANTlOXIDA..Yl" I'ROr ERTl F..s
PERCENT is the percent, expressed as a decimal , oftunneri c in the
tumlcric-Chromosorb mixture.
EI-mass spccu<t were obtained for each componenl. figures 7, 8, & 9 show the
El-mass spectra for the three major components detected by di rect thermal extraction
GC-MS. Table 1II contains a [i sl of the eight most abundant ions for the minor
conlponcnts in luoncric powder and their relative intensities. Figure 10 lists the
structures for the components detected. Some ofthe!;e identificat ions are based on
correlation with mass spectra in the NISTIEPAlNI H Mass Spei: tral Library, However,
the mass spet:t ra for most of the components could not be found in the library and
many of ,the structures in Fi gure 10 are proposed structures. For proposed struct ures,
the Identlfications nre based on similarities with components for which there was a
good librolry match and by evaluati on of the fragmentation patterns.
Figure 7 shows the EJ-mass spectra for ar-tunnerone, the character impact
compound for turmeric. The identifi cation of this component was based on
of the spectrum and correlation with previousl y reported data
(9, /0). FIgure 8 shows the mass spectrum ofturmerone. Mass spectral data for thi s
was also reported by Su, el 0/ (9). They reported a base peak at mil. '" 121
WI th a probe temp equal to 70C. ll1ey also reported the appearance of an ion at ml7. -
119 when the probe temperature was increased which they attributed to the
aromalization of the eyclohexadienyl moiety at the hi gh temperatures of the GC MS
interface line. Figure 9 shows the mass spectrum of curl one. Thi s agrees with the data
reported by Kisco, et al (/2).
Addit ional research on the analysis of volatile components in Curcuma species
can be found in the ]itcrdture (10, /), 14,15).
, ..
Conclus ion
Extraction of natural products yields complex mixtures of volatile, semi-volatile, and
nonvolati le components. No one technique for obtaining mass spectral data for these
components is generally applicable. Direct thennal extTllcli on GCMS proved to be a
valuable tcchni que for the ident ification of volati le and semi-volatile componcnts in
powdered turmeric.
Acknowledgemen ts
We acknowledge the Center for Advanced Food Technology (CAFT) mass
spectrollletry facility for providing instrumentation support. CAFT is an initiative of
the New Jersey Commission on Science and Technology. This is New Jersey
Agrieullural Experiment Stalion (NJAES) publication #D10570 196.
.-
8. JIISEROOT ET ,\ L VtJiatiles in Turmeric
L.ilenture Cit ed
I) Govindarajan, V. S. Crit. Rev. in FoodSci. andNldr. 1980, 12, pp. 199]01
2) Masuda, T.; lsobe, J.; Jitoc, A.; and Nakatani , N. Phytochem. 1992,] 1,
pp. ]645]647.
97
3) Toda, S.; Miyasc, T.; Arichi , II.; Tanizawa, H.; and Takino. Y. Chem. Pharm.
Bull. 1985, 33, pp. 1725-1728.
4) Azuine, M. A. and Bhide, S. V. Nutr. Cancer 17, pp. 7783.
5) Mukurnian, M. A.; Chacko, M. C.; Annapuma, V. V. ; and Kirshaswamy, K.
Carcinogenesis 1993. 14, pp. 493-496
6) Azuine, M. A.; Kayal, 1. J.; and Bhide, S. V. J. Cancer Res. Clin. Oneal. 1992,
118, pp. 447 452
7) Tonnesen, H. H. In Phenolic Compounds in Food and Their Effects Oil flea/lh
, . Analysis, Oceurance. ondChemiSIrY Editor, Ho, C.T.; Lee, C. Y. ; and
Huang, M-T. ACS Symposium Series 507, American Chemical Society:
Washington, DC, 1992; pp. 144- 153.
8) Hartman, T. G. ; Overton, S.; Manum, l ; Baker, C. W. ; and Manos, J. N. food
Teehnol. 1991, 45, pp. 1041 05.
9) Su, H. c. F.; Horvat, R.; and Jil ani , G. J Agrie. Food Chem. 1982, ]0,
pp. 290292.
10) Rao, A. S.; Rajanikanth, B.; and Seshadri, R. J. Agrie. Food Chern. 1989,37,
pp.74074].
II ) Kingston. D. G. I. ; Bursey, J. T.; and Bursey, M. M. Chemical Reviews 1914,
74, pp. 2 15 .
12) Ki so. Y.; Suzuki, Y.; Oshima, Y.; and Hikino, 1-1. Phytochem. 1983,22,
pp.596597.
13) Gtwlap, A. S. and Bandyopadhyay, C. J. Agric. FoodChem. 1984,32,
pp.5759.
14) Dung, X. N.; Tuyd, N. T. B.; and Leclercq, P. A. J. $senJ. Oil Res. 1995,7,
pp.261.264.
15) Ky, I>. T.; van de Ven, L. J. M.; Leclcrcq, P. A.; and Dung, N. X. J. Esse"t. Oil
Res. 1994,6, pp. 213214.
16) Majiat , P. ; Erdos, Z.; and Takacs, 1. J. Chromatogr. 1974, 91, pp. 89103.
17) Khurana, A., and Ho, CT J. UquidChroma/ogr. 1988, II , pp. 22952304.
18) Hiscrodt, R.; Hartman, T.G.; Ho, C.T.; and Rosen, R.T. J. Chroma/ogr. 1996,
740, pp. 5164
Chapter 9
Pungent Flnvor Profiles and Components
of Spices by Chromatography and
Chemi lumincscent Nitrogen Detection
E. 1\1. Fujinari
An lek Instruments, Inc., 300 Hammel Weslficld I{uad,
Huuston, 'rX 77090- 3508
The pungent characteri st ic of hot is ort en dUl: TO the
presence of a class of nitrogen (;onlainillg compounds s u(; h as
capsa ki noids. These compo llent s ca n he an'l lyzcd by
c hromatography wit h c he milumincsccnt nilrogen detect ion
(CI. ND). Thi s nilrogc rH;peci fie deled or l: an
s impl ify complex analyses by eliminating non-nilrogenous
compone nt s in the snrnplc. Thi s allolVs the chromlilographer
to easil y focus on the scp'Hali oH of the nitrogen ('umaining
compo nenTS responsibl e for Ihe "hotness" of spices.
Many kinds of nitroge n cont aining compounds res pons ihl e for
pungent or "hot" navors in s pices hn ve heen reported . Struct ures of
hoi component s (I.y) in horseradis h oil :Ire s hown in Figure I. The
red hot c hi l i pe ppers ('onl:l ill nil rogcnou.c; compounds known as
which nrc quite si mi lar in struc tllre (Figure 2)_ Hyhrid
peppe rs possess diffe re nt degrees of hotness, j:l lapcno> Indian
birdseyc's> Mexi can habanero vnricti es. Piperine ( I X) is the hot
component in hlack pe ppe r. Si nce Ihese nnal ytes contai n nitrogen,
l'l lro mnt ographi c det ection usi ng the c hemi lumincscent
de tect or is inhe re ntl y s ui l:lble. Simpli fied l: hrolll:ltogra llls arc
o btai ned s ince no n-nitrogenous in t he sample s nrc
tTa nspnre nt to the detector.
Hi s toril:ll lly, capsaic inoids in foOl.!!; have bee n anal yzed hy
organ ol epti c e valuation CD. colo rimelry (). nnd llY
spectrophOloll1c tri c me thods Q, i). Chrom;ltogmphi (' methods have
also been used, incl uding thin layer c hrummogmphy ('1' 1,( ') (,5,. and
gas c hrumntography (Ge) (79). Typically, GC metlu>\ ls require n
"
III
IV
V
II"", It
"

/ II II
BUI . ) Cllv"ihilc
II"", II
C= C--C- S- C= N
If /I II

AII)I
elhyl irorhiocpnate
Figure I. Stnlclme of hot compoll nds itt horSeradi sh oil.
'19
I
100 51' ICES: ANI) ANTIOXIDANT 1' ltol'ER1U:S
eH,o):)N
VIII I H
#
I

IX = CH- CH =
Pipctille
rigurc 2. Structure of c<lpsaicinoids and pi perine.
, .
9. H1JINAIU Pungent Profiles & ComfXJnelfiS of Spicn
101
derivatization step for these compounds prior to analysis in order to
make them more volat ile. However. capsaicins have been analyzed
without derivatizati on by high performance liquid chromatography
( HPLC) with UV detection using norma l- and reversed-phase
techniques (10- 17). Better resolution of capsaicins (11) has been
reported using reversed-phase (RP) rat her than normal-phase (NP)
chromatography. Reversed-phase HPLC separati on of piperine
followed by UV detection has been report ed earl ier (H. lID Usi ng
mass spectrometry (MS) as a means for GC detection provides a
powerful tool for st ructural characteri zation of Oavors. e.g. ginger oil
(W. However, quantitati on of ana[ytes by MS detection in
chromatography may be diffi cult because of the response variation of
the detector due to sample andlor solvent induced matrix; effects. On
the ot her hand, the CLND response is stable and not affected by
complex sample matrices. Quantit ati on of ca psaicin and
dihydrocapl>aicin in red pepper by HPLC*CLND was previously
report ed aID. The linear response of the CLND is shown in Figure 3.
Corantnt lun
capsaicin blg)
r-0.99952; m- O.00189; b--O.00202
Figure 3. HPLC*CLND calibration curve of capsaicin.
Reprinted with permission from E. M. Fujinari, in "Spices,
Herbs and EUible Fungi". G. Charalrunbous (Ed.), 1994 ,
pp 36n79.
with kind permission from Elsevier Science * NL, Sara
Burgerhartstraat 25, 1055 KV Amsterdam, The Netherlands.
102 s ri CES: )'IAVOR CII EMISTRY ANI} ANTIOXlnANT PIWI' EKTIES
Benn et. al. (ill demonstrated the use of GC-CLND for the
detection of nitrogen containing compounds in navors and esscnl ial
oils. The advanlage of this techni que is illust rated in Figures 4a-b
using a CP-Si l 5 CB (Chrompack, 25m x O.53mm 10, 1.0 film
thi ckness) column. Fi gure 43 is Ihe GC- FID profile of a green pepper
flavor containing nitrogenous components from the horseradish oil
and (he three pyrazine compounds. As Figure 4b shows, using GC-
CLND. the Ili. ragen-cont aining compounds are eas il y deJected
wilhout interference from the sampl e matrix. Peak identificati on was
achi eved using the horsemdish oil and the pyrazi nc siandards.
SUI>ercrilical flui ds (s uch as supercritical CO
2
) possess simi lar
viscosi ti es to those of gases, yet their diffusivities are much greater
than liquids. These phys ical prope rti es together provide hi gher
separation effi ciencies for SFC. with sharper peak. s than for HPLC.
Since hi gh molecul ar weight and thermall y labi le compounds can be
anal yzed by this technique, SFC also provides an added advant age
over Gc. Taylor et. al. (22) reported a feas ibi lity study for
supercritical Ouid chromatography - chemil uminescent nitrogen
detect ion (SFC-CLNO) with open tubular columns. SFC-CLND of
hot mustard ext ract is shown in Figure 5. Hot components were
ident ified as all yl isothi ocyanatc (peak. A) and butyl isothi ocyanate
(peak B) usi ng corresponding analyt ical standards.
This papcr will focus on nitrogen-speci fi c detection for liquid
chromatography incl uding HPLC-CLNO profi les of chili powder.
paprika oleores in. black pepper, and capsai ci ns i n oni on and garli c
navors.
Experimental
Appa r-dtus. Hi gh performance liquid chromatographi c separations
were achieved on a binary gradi ent mi crobore HPLC system: primary
pump (A) Model 305, secondary pump (B) Model 306, manometric
module Model 805. and a dynamic mixer Modcl 81 IC from Gi lson
Inc. (Middl eton, WI). Sample injections were achie ved with a 20!-!L
loop on a Model EQ-36 inj ection valve from Valco Instruments Co.
Inc. (Houston, TX). A stainl ess steel V-splitter also from Valco was
used in order to achi eve a post-column split of Ihe mobile phase now
10 Ihe CLNO. A Supelcosil LC- 18S anal yti cal HPLC column was
purchased from SUPELCO Inc. ( Bellefonte, PA). The V-s plitter was
all ached to the analytical col umn by a SLIPFREE connector,
ava il able from Keystone Scientific Inc. (Bell efonte, PA). Anal yses of
nor-dihydrocapsai ci n, capsaicin and dihydrocapsai ci n in spi ces as well

9. Pungent flavor Profiles &: of Spices
.)
'b)
,

'" '" " 1ln .. In,lnl
"

1
,I
,
, , ,

" "
" Tlmt (mini
f igure 4. GC profile of green pepper navor with horseradish
oil and pyrazine mixt ures. .
a) FIJ): pCllks E :;:: and
G = 2 methoxy-3-clhylpyrazme.
b) CLND: peaks A :;:: B =:' all yl thi ocyan<ltc,
C:;:: allyl isothiocyanate, D :;:: 2: butyllsotllLoc
y
anate.
E:;:: 2_mclhyl _3_ lllcthoxypyrazlllc, F :;:: 2-lllet hyl.-5-
melhox ypyrazi ne, G :;:: 2_lll cl hoxy_3_ethylpyrazlIlc;
<lnd H :;:: phenyl et hyl isothiocyanatc. Rcpn nted .
permission frolll S. M. Bcnn, K. Myung, E . M. h lJlnarl.
in "Food Flavors, Ingredients and Composition, G.
Charalambous (Ed.). 1993. pp65.-73., .
with ki nd permission frolll ElseVier SCience - NL, Sara
Burgerhallstraat 25.1055 KV Amsterdam, The Net herlands.
103
r
,
Ii:
,
j
\1
I:
"
,
j,
"
I
"
".
!,1
". ,
,
,
\,
, .
,
,
104 SI' ICES: FlAVOR Cl l EMISrRY AND ANTIOXlnAl'iT
,
o
A

00
c
Time (min)
20
, .
Figure 5. Capi llary SFC-CLND profile of hoi mustard extract
LO. I gram of mustard powder extracted in I mL water (30%)
and solution I. Peaks A = allyl isothi ocyanatc.
B = 2.-b.utyl lsotluocyanate. and C = unknown nitrogen
cOl1laltlll1g compound. ChrOlnatographi c conditions:
pressure program from 80 aIm (hold 5 min). ra mp 10 150 atm
at 10 aIm/mi n, then to 200 atm al 15 atm/min; Cyano (20 m x
10, 0.25 mm film thi ckness) column; time split
IIlJCCtl OI1 0.2 sec. Reprinted wi th pcrmission from H. Shi ,
J. T. B. Strode III. E. M. Fujinari and L T. Tayl or,
J Chrornatogr. A, 734 ( 1996) 303.
with kind permiss ion from Elsevier Science - NL Sara
Burgcrhartstraat 25, 1055 KV Amsterdam, The Netherl ands.
9. I'UJINAKl PUflgelll Projifts & ComfKmenu of Spices
lOS
as piperine ill black pepper were accomplis hed wit h the nitrogen
speci fi c detector. model 7000 HPLC CLND, from Anlek Instruments
Inc. TX) and Della chromatography software from Digital
Solutions (Margate, -Australia) run on an IBM 486 compatible
computer.
Reagents and Standards. The nalUral capsaicin standard mixture
was composed of 65% capsaicin and 35% dihydroca psaicin. Piperine
(97%) and citric acid (99+%) were also purchased from Aldrich
Chemical Co. (Mil waukee, WI) . HPLC gmdc methanol (99+%) was
obtained from Fisher Scientific Co. (Fair Lawn, NJ). Sodium free
distilled wate r was obtained from Q7.arka Drinking Water Co.
(Houston, TX). All standards and reagents were used without further
purification. The HPLC mobile phase was filtered through a
Millipore Corp. (Bedford, MA) !-IV filter with a 0.45 Ilm pore size.
Paprika oleoresin, oni on, garli c. black pepper. and chili powders wcre
obtained from commercial sources.
Analytical method. The capsaici n/dihydroca psaicin and piperine
refere nce standards were prepared as 13. 19 mg/ mL and 4. 16 mglrnL
solutions in methanol, respectively. Sampl es were prepared by
separately weighing the following and bringing the volume to 25 mL
wit h methanol: red chil i powder (2.5 136 g), black pepper powder
(2.5154 g), onion powder ( 1.5 126 g) . and garli c powder (1.5297 g).
Each sample (5 mL) was concentrated to I mL final vol ume with a
gentle st ream of helium. Samples for the hot garli c and oni on flavor
profiles we re prepared usi ng a ( I + I v/v) mixture of the capsaici n
reference standard and 100 IlL of the concent rated fl avOrs. HPLC-
CLND anal yses were accomplished using a 15 ilL partial fill ed
injection to a 20 J.lL sample loop into a Supel cosil LC- ISS anal ytical
col umn: 250mm x 4.6mm 10, 5 1lm particle size, 100 A pore size. An
isocratic mobile phase, methanol/water wilh 0. 1 % citric acid at pH 3.0
(65:35 v/v). was utili zed with a fl ow rate of 0.650 mUmin. A Y-
splitter was configured post-column and used to deliver a fl ow of 100
to the CLND. The CLND condi ti ons werc: 1100
0
C pyrolysis
temperature, PMT voltage 780. range x 10. and I volt detector output.
Results and Discussion
Thi s chemilumincsccnt nitrogen detector fo r HPLC was first
described in (W. The detecti on mec hanis m for nitrogen
determination is shown below:
I
106 SPICES; Fl.4.VOR CHEMISTRY AND ANTIOXIDANT PROPERTIES
- - - - - - - ~ . . . 'NO + other oxides
Ni lrogen cont aini ng analytes (R-N) are oxidized in a furnace al high
temperatures 10 nilric oxide (NO). Chemiluminescence as shown in
the second equati on is detected by a photomultiplier lube (PMT). The
photons detected are proport ional to the amount of nitrogen in the
analyte(s).
The HPLC-CLND profile of red chi li pepper is shown in Figure
6. The capsaicinoids lIordihydrocapsaicin CAl, capsaicin (8), and
di hydrocapsaicin (C) are easily observed without interference from the
red chi li powder matrix in this chromatogram. This separati on was
achieved via isocratic reversed- phase chromatography using a mobile
phase consisting of MeOH/water with 0.1 % citri c acid at pH 3.0
(65:35 v/v). A Supelcosil LC- ISS (250mm x 4.6mm ID) column with
a mobile phase now rate of 0.65 mUmin was used. A post-column
split was used to direct 100 l.d .Jmin to the CLND. The three
capsaicinoid compounds (A, B. and C) are also very easily detected in
onion (Figure 7) and garlic (Figure 8). Several pola r nitrogen
contai ning component'; eluted at the solvent front for the onion, garlic
and red chili powder samples .. Figure 9 shows the HPLC-CLND
profile of black pepper. Paprika oleoresin was ex.t racted by the
method reported by Coopcr (m and analyzed by HPLC-CLNO
(Figu re 10) showi ng the presence of capsaicin (A) and
dihydrocapsaici n (8).
Conclusion
Pungent (hot) components can be separated using gas chromato-
graphy, s upercriticat nui d chromatography, and high performance
liquid chromatography and detected wilh the chemiluminescent
nitrogen detectors (CLND). These nitrogen-specific detectors provide
a means of analyzing nit rogen-containing compounds rree of matrix
interfcrencc.
9. FUJINARI Put/gent Flavor Profile,f & Components of Spice.1i
, .. .. -"
.. ... "' ..
B
c
A
-
-
Time (min)
Figure 6. HPLC-CLND profile of red (;hili powder.
Peaks: A = nordihydrocapsaici n, B = capsaicin, and
C= dihydrocapsaicin.
107
-

108 SI' ICES: FlAVOR C1H;MISTRY ANIl ANTiOXIDANT
" . .... '" f-
" ........ f-
., ..
- -
-

B
A
-
"
Time (min)
c
Figure 7. I-I PLC-CLND profile of capsHicins with onion navor.
Peaks: A = nordihydrocapsaicin, B =: capsaicin, and
C= dihydrocapsaicin.
9. FUJINAIU Pungent Profiles & CrJmponenls of Spices
109
.... ... .,." -
B
........ " -
c
... "' ...... f--
- -

,;
Time (min)
Figure 8. HPLC-CLND profile of capsnicins wilh garli c navor.
Peaks: A = nordihydrocapsaicin, B = capsaici n, and
C= dihydrocapsaicin.
I
110 SI'I CK'l: HAVON. CHEMI STRY ANn ANTIOXII>ANT I'KOI'ERTI\';s
A
-."--
.. .......
I
_. '------.J
.
- -
- -
-

,
Time (min)
Figure 9. HPLC-CLND profile of piperine in black pepper.
Peak: A = pi peri ne.

9. niJINARI ProjiJu & Components of Spices
III

. ,
... ..... ..
A B
. ,
.. .......
A
-... ..... ..
I 0 0 , I
..... - .......... - ..
";..;!!
Time (ruin)
Figure 10. HPLC-CLND profi le of paprika oleores in.
Peaks: A = capsaicin and B = di hydrocapsaici n .
Literature Cited
I. Govindarajan. V.S.: Narasimhan, S.; Dhanara, SJ. J. Food Sci.
Techno!' 1977 .M. 28-34.
2. Van Fedor, K.; Unters, Z. Lebcnsm 193 1, hl. 94- 100.
3. DiCecco, J.J . J. Assoc. Qff. Anal. Chern. 1979,@..998- 1000.
4. Trejo. G.; Wila, A. J. Food Sci. 1973.lH. 342-344.
5. Spanjar. P.; Bla'l.ovich. M. Analyst 1%9,21. 1084-1089.
6. Andre, L.; Mi le, L. Acla Alimen. 1975,1. 113- 121 .
7. HarIman, K.T. 1. food Sci. 1970. Ji. 543-547.
8. DiCecco. J. J. Assoc, Off. Anal. Chern. 1976.22. 1-3.
9. Todd, P.H.; Bensinger, M.G.; Biflu. T. J. Food Sci. 1977.11.
660-668.
10. Woodbury. J .E. J. Assoc. Ofr. Anal. Chern .. 1980. Ql. 556-558 .
11. Saria, A; Lembeck, F.; Skofitsch, G. J. Chromatogr. 1981.
208,41-46.
12. Hoffman, P.O.: Lego, M.e.; Oalello. W.O. J. Agric. Food
Chcm .. 1983.21 1326-1330.
13. L'lIv, M.W. J. Assoc. Ofr. Anal. Chern. 1983, M.. 1304- 1306.
14. Weaver, K.M.; Luker, R.O; Nealc, M. E. J. Chrornatogr. 1984.
301. 288-29L
15. Kawada , T.; Wfl tanabc, T.; KUl surn, K.; Takami, H.; Iwai, K 1
Chromatogr. 1985, 329, 99- 1 05 .
.,
112
16.
17.
18.
19.
20.
21.
22.
23.
SI' ICF$: FlAVOR ANn ANTlOXlDANT I'ROPERTIK.,
Weaver, K.M.; Awdc, 0 .13 . J. Chromalogr. 1986. 367. 438-
442.
Cooper. T.H.; Guzinski. 1.A.; Fisher. C. 1. Agrie. Food Chern.
1991. 39 2253-2256.
Wood, A.B.: Barrow, M.L.; James. DJ. Flavour Fragr. J. 1988,
J.55-64.
Yemin. G.; Parakanyi, C. In Spices, lind Edihle Fllngi;
Charalamhous, G" Ed. ; Devel opments in Food Science Series
34; Elsevier Science Publishers, Amsterdam The Netherlands
1994. 34579-594. "
E.M. In Herh .... (llfd Edihle Fllllgi;
Charail.llnbous, G., Ed.; Developments in Food Science Series
34; Elsevier Science Publ ishers, Amsterdam, The Netherlands,
1994, 34 367-379.
Berlll, S. M.; Myung, K.; Ftrji nari , E. M. In Food Fll1vor,f,
(llId COII/pMifioll; Charalambous. G., Ed.;
Developments in Food Sciellce Series 32; Elsevier Science
Pu?li shcrs, Amsterdam, The Netherlands, 1993, lZ 65-73.
Stu. H.; Slro(k, J.T.B. III ; Fujinan , E.M.: Taylor, L. T. J .
Chromatogr.A, 734 ([996) 303-3 10. -
Fuj inari E, M.: Courthaudon L. O. J. Chromatoer. 1992,592
209-214. - -'
Chapter 10
Supercritical Fluid Extraction
of Allium Species
E1iUlbeth M. Calveyl and Eri c Blockl
' Center for Jo' nod Safety Bnd Applied Nutrition, U.S. Food and Drug
Admi nistration, 200 C St reet Southwest, Washington, DC 20204
l Department of Chemistry, State University of New York- Albaoy,
Albany, NY 12222
Supcrcri tical fluid teclmologies are viable alternatives for the extraction
and analysis of natural products because of the heightened awareness of
the cost and safety hazards with the use and di sposal of
conventional organic solvents. Supercritical C02 (SC-C01) is of
particular interest to the food industry. Because of its low critical
temperature, SC-C0
2
exlrllction can provide an accurate representation
of the taste, color and odor of naturally occurring materials found in
spices, fla vors and foods. The apptication of supercritical fluid
extraction technology to the analysis and/or production of spice and
flavor cltracts is discussed. The use of the technology in the extraction
of organosulfur compounds found in Allium species is emphasized.
Applications of supercriticat fluid extract ion (SFE) technologies are being investi gated
extensively by the food industry (1-9). The extensive patent activity related to food
processing applications since the early 1970s is evidence of interest in SFE. Over 50
patents have been issued or applied for in the US or elsewhere. Table I lists some of
the patents related to the fl avor and spice industry (10-23). Supereritical C02 (SC
COl) with it s low critical temperature (3 t .3 0c) is of part icular interest to the food
industry because many of the components in food matrices react or degrade at elevated
temperatures. The advantages of SFE for the food industry include potentiall y higher
yields, better quality products and the use of a nonflammable, nontoxic solvent.
Current Food and Drug Administration regulations [21 Code of Federal Regulat jons
184.l24O(c)] list COl as "generally recognized as safe"' when used as a di rect human
food ingredient. For this reason, CO
2
can be used in food with virtually no limitations
other than current good manufacturing rractiee. In contrast, the use of traditional
organic sol vents requires the removal 0 residual solvent to permitted levels. This
removal usually requires some distillation, which can cause off-flavors due to
decomposition of components at the elevated temperatures used. The chance of off-
navors resulting from residual solvents is eliminated when extracting wi th SC-C0
2
because of the ease of removing C02 from the food matrix. Other motives for the
investigation of SFs include the potential for new product development and compliance
with more stringent pollution control regulations that increase the cost of waste disposal
of traditi onal solvents (24).
Although SC-COl can solubilize non-polar to moderately polar compoundS, the
use of entrainers or modifiers (e.g . ethanol) enhances the solvating properties of the
1997 AmeriCiln ChemiCil I Sociely
I
"'
SI'ICES, t' IAVOR CII EMI!>-rR't' AN)) ANTIOXlIlANT " RO!' t; RTIES
TABLE I. Represc:ntalivc Palen!s Related to SFE of Aavol'S and Spices
Patent ##
us 3 477 856
us 4123559
DE3133032
us 4 400 398
us 4 470 927
us 4 474 994
US 4 490398
JP 84 232 064
EP AppJ. 206 738/
US Appl. 746 607
EP Appl . 154258
l P 63 87 977
JP Oll17761
JP02 135069
JP 02 235 997
us 5 120558
1111 If69
IOl3Ins
3/3/83
8123/83
9111184
100/8'
12125/84
12126/84
1113006
6119/85
12111/85
4119188
5110189
512"""
9118/90
6/9/92
Title Reference
Process for the Extr.lCt ion of Fl avors 10
Process for the Producti on of Spice 10, II
Extracts
Apparatus and Methods for Extr.lCtion 12
Method for Obtaining Aromat ics and 10
Dyestuffs from Bell Peppers
Method of Extracting the Flavoring 10,13
Substances from the Vanilla Capsule
Purification of Vanillin 10, 14
Process for the Preparat ion of Spice 10, 15
Ex",,"
Manufacture of Flavoring Substances 16
Process for the Production of Citrus 17
Flavor and Aroma Compositions
Flavoring Extracts 18
[solation of Spices, Peroxides and 19
Other Useful Components from
Cruciferous Plants
Extraction of Flavor Components of 20
Alcoholic Beverages
ManufactufC of Spice Extracts 21
Extraction of Flavoring from 22
Seaweeds
Process for t he Supercr iti cal 23
Extract ion and Fractionation of
Spices
10. CALVEY &. BLOCK
SCFE of Allium Species
115
fluid. Whereas a neat compound may be soluble in 5C-C02. it may not be extractable
from its matrix without the addition of an entrainer. This phenomenon is demonSll1Itcd
in the dc:caffeination of coffee: ( 10); neal caffeine is soluble in dry SC-C02. bUI moist
5C-C02 or moist coffee is necessary for the e.uraclion of caffeine from coffee beans.
This same phenomcnon occurs with decaffeination by traditional organic solvents.
Investigators have hypOlhcsiud thai water frc:es the: "chemically bound" caffei ne in the
coffee matrix.
An SFE system contains five basic components: pump(s), extraction vessel.
temperature controls. pressure controls and separator{s), For processing, a variety of
recovery strategies is viable: a) change of temperature; b) change of pressure; andlor c)
use of a suitable adsorbent material. The complexi ty of the proceSSing SFE system
depends on the desired application and the mode of product recovery. Rizvi et al. (8)
divided food-related SFE applications into three basic categories: a) total extraction, b)
deodorization, and c) fT".lctionalion. Total extraction is the removal of a component or
group of related compounds from an insoluble matrix. This type of process is
exemplified by the extraction of vegetable oi ls with SC-C02. Deodori1.ation relates to
operating the extract ion system at less than the maximum density of the solvent. The
extraction condi tions are usually held constant while the components of interest,
gencrally the more soluble ones, are preferentially removed from the matrix. Thi s type
of application is appropriate for the removal of objectionable aromatics or the extraction
of desirable odor components such as from spices. Whereas SFI! in gencral
fractionates a matrix, Rizvi et al. (8) use the term fractionat ion to describe the
of coextracted components from each other. They also usc the term to
describe concentration of components either as the ext ractant or in the residual material.
In sample preparat ion, the SFE systcm can be direct ly attached to a chromatographic
system and used as an on-l ine inject ion technique. If SFE is performed off-line, the
resulting ext ractant is int roduced into a chromatographic system via conventional
injection techniques. Mul tiple examples of on-line and off-li ne analytical scale SFE
related to food products can be found in the literature (9.25-44).
SFE of SpictS and Flavors
Although many compounds have been extracted by using SC-CO:!, the majority of the
work rdated to food productS can be divided into three broad categories: flavors/spices,
herbicides/pesticides and lipids. Table lI lists the extract ion conditions which use SC-
CO2 for a variety of spice and flavor applicat ions, The loss of the volatile components
desired in spice extracts is reduced because C02 is a gas at room temperature and the
higher temperatureS required in di stillation are not needed. Hubert and Vitzthum (47)
indicated that the best spice extracts have all the organoleptic factors of the whole spice
even after dilution of the extract. They investigated black pepper, nutmeg and chilies
and indicated that the products obtained from the SC-C02 extracti ons were
organolept ically simitar to the commercially obtained extr acts. In 1981 Caragay (48)
revi ewed the extraction condi tions for cloves, cinnamon and vanilla pods. Stahl and
Gerard (45) studied the sol ubility and fractionation of essential oils in SC-C02. They
were able to obtai n quantitative recovery of these volatile oil s free of undesirable
substances without fractionation of the essenti al oils themselves. Once the essential oil
components wefC extracted, further fracti onation into ccrtain substance groups (i.e"
monoterpene hydrocarbons, hydrocarbons, oxygen cont aining
monoterpenes and oxygen contai ning scsquiter penes) was possible. Several
laboratori es employcd off-li ne extraction techniques to fractionate the navor
components of ginger, pimcnto berries, apple essence (29) and lemon peel (30). By
using gas chromatography (GC) as the means of analysis. these laboratories were :Iblc
to show some fractionation of the fl:lvor components as a function of thc extmclil'll
densi ty. Other laboratories have investigated the use of SFE in flavor analysis hy din.'..:!
coupling of the extractor to a chromatographic technique such as GC (25,26,3J,.IX)
and SFC (27.34).
..

'"
SI'I CFS: n AVOR Cm ;"m'TKY AI''W ANTI OX1OANT I'KQI'IO.RTIES
TABLE II. Representative Applications of SFE Related to Flavors and Spices
C01l1l11odit:t Conditions
Reference
Rosemary
45 C; 300 atm; 10 min; cryolrnpping 25
-6S or 10 "C
Eucalyptus leaves. li me peel,
45 C; JOOalm; 10 min; cryogenic
26
Icmon peel,
trapping -SO to 30 "c
grapefruit oil
70 "C; 0.1767-0.8579 glmL; 12 mi n; 27
cryotrapping -65 or _10 C
Garlic and onion
35 "C: 240 aim; 30 min; cryOirapping 28
().] "C
Ginger, pimemo berries
SO GC; 1500-5000 psi
2'
Lemon peel oil
30-58 "C; 90-250 kglcml 30
Mexican spices (OrigcU!um
and Pimpinella ulli.sum)
55 "C; 167 bars; separation vessel:
55 "C; 26.7 bar,
31
Chamomile essential oi l
40 "C: 90 bar: separation vessels in
series: 0 GC, 90 bar; -5 "c, 30 har.
32
Clovcs
40-140 GC; 82-400 atm; 4.5 h
33
Tunncric
60 "C; 250-280 bar; 15-20% McOH 34
, .
Shiitake mushrooms
40 "C; 3000 psi; fractionation
through $C\'eral
35
Hops essential oil, biller principles
50 C; 0.05-0. 1 glmL; 15-30 min
36
Lavender essential oil, waxes
48 GC; 90 bar; separation vessels in
37
series: - 10 "C, 80 bar; 0 GC, 25 bar.
Caraway fruits
50 "C; 9.7 MPa
38
Ginger
40 "C; 0.85 glmL; 3 min static, 30
min dynamic; I mUmin
39
oi l components
(limonene, carvone, anethole
40-120 "C; 20- I 20 bar
45
cugenol , caryophyUene,
valcranone)
Orange juice {treatment used to
35-6O "C;7-34MPa; IS- 180min 4.
deactivale
10. &; BLOCK SCFE 0/ Allium 117
SFE of Allium Specle5
The natural flavors from garlic (Allium sot;vum), onion (A. ct!pa), and ot her Allium
species, like those from many other common vegetables and fruits, are noc present as
such in the intact planls but arc fonned by cntymatic processes when the plants arc
chewed or cut (49). Addi lionai flavors, also considered natural, are formed during
cooking as a result of the thennal breakdown of the initial enzymatically produced
navoraots in either an aqueous on nonaqueous (e.g. , cooking oi l) medium. If the
breakdown products are unstable, other compounds can be formed, which can
contribute to the aroma and taste of the food.
Supercritical fluid extraction of Allium with C02 provides an effective
and environmentally friendly alternative to tradItional organic extractions. Miles and
Quimby (SO) extracted garl ic products by using SC-C02 under mi ld conditions, They
analyzed the extracts by GC with atomic emission detection. Sinha (SI) extracled
onions with SC-C02 and analyzed the extracts by GC-mass spectrometry (MS). The
use of traditional GC methodologics in the analyses of these SF extracts by the above
research groups precluded the identification of those compounds primarily responsible
for the charactcristic fla vorll of freshly cut members of the genus Allium (S2). We have
previously shown that the chromatographic profiles of extracts of Alii14m species (garlic
and onion) obtained with SC-C02 at thc low temperature of 3S
G
C were simi lar to the
chromatographic profiles of corresponding organic solvent extracts (28), The flavor
qualities of the SF extraclS were j udged 10 be comparable to Ihose of fresh garlic and
onion. Our experience with SFE of garlic contrasts with that of Wagner and Breu (S3).
They reported the complete dccomposition of garlic compounds, such as allicin (S-2-
propenyl 2-propenethiosuUinate, AlIS(O}SAII ), after garlic juice was stored at room
temperature for 3 h and then extracted wilh CO
2
, It is di fficult 10 evaluate their work
because they did not report Ihe SFE condit ions employed. We observe very little
change in the liquid chromatographic profile of SF extracts of aqueous homogenates of
garlic that were extracted at 35C following storage for 10 min or 2 h at room
temperature (27.5C) (Figure I). Our observations agree with those of Lawson (54),
who has shown thai the hal fli fe of allicin at room temperatun: is 4 days in water.
Supercritical fluid extraction of the major garlic flavorant, allicin, is 25% more
efficient than the best procedure which uses organic solvents if the SFE is done at or
below 35C. Liquid chromatography (LC)-MS under thermospray conditions
confirmed the identity of allicin from the garlic (28). Analyses of SF extracts of
commercial garlic products which have been reconstituted by addition of water showed
LC profiles similar to those seen with fresh garlic. Al lici n and related Ihiosulfinales
(RS(O)SR') have been identified in SF eXlracts of garlic (Figure 2A), freeze-dried
ramp (A. tricoccum) and frozen ramp (Figure 28), The ratio of the thiosulfinat es
foond are signifi cantl y dirferent in the garlic and ramp. The major constituent in the
garlic extract is allicin, peak (pk) 4. The major const ituents in the frozen ramp e;1.t ract
are allyl methyl dtiosulfinates (AIIS(O)SMe, pks 213) with a signi fi cant contribution of
the dimethylthiosulfinate (McS(O)SMe, pI:. I). The I-propenyl isomers of allicin are
also present in both the garlic and ramp extracts. The chromatographic profi le of the
freeze-dried ramp extract more closely resembles lhe garlic extract than the extract from
the frozen whole ramp. Ajoene (AllS(O)ClhCH=CHSSAlI), a major compone nt
found in oil-macerated garHc products, has been found in small quantit ies in SF
extracts of garlic and (tentatively) in those from ramp. The identity of these
compounds has been verified by LC-MS employing atmospheric pressure chemical
ionitation (unpubl ished data).
A simi lar study employing onion j uice showed SFE to be ca. 69% as efficient
as conventional organic solvents for extracting the major organosul fur compounds
(28). The diminished efficiency of SFE for onion compared with gart ic may be due to
1I8 SI'ICKS: .. LAVOR AND ANTIOXIDANT I'}I:C)PERTI ES
3.0
A

2.0
0



-<
1.0
3.0
B
t! 2.0
J
1.0
0.0
5 10 15 20 25 30
Time (mi n)
Figure I. LC chromatograms (UV, 254 nm) representing: (A) SF extract of a garlic
homogenate stored at room temperature for 10 min and (8) SF exnact of a garlic
homogenate stored al room temperature for 2 h.
10. CALVt,Y & 8LOCK sen: of Allium SpKieJ.'
119
0.8
4
A
0.6

of 0.4
0
"
-<
0.2
,
2
,
I
3
,[
0.0
,
:1
e
,
,
I
0.8
2
B
0.6

0
<

4
.. 3
<:
0.2
0.0
,
2 4
6 8 10 12 14
,
Time (min)
Fi gure 2. LC Chromal Oarams (UV. 254 om) of SF extracts: (A) garlic and (D)
frozen fresh ramps. Pc identification: ( I) MeS(O)SMe; (2/3) AIIS(O)SMc; and
(4) allicin.
':'
,
:
1-
,
,

,
,:!
!\
,
120
SI'ICl';S: FlAVOR CIf EMI.!o'l' RY AND ANTIOXIDANT "IWPEKTIt.:S
2.5
A
2.0

u
:;iu

"
1.0
0.5
0.0
2.5
II
2.0

g I.S


0
llLO
<
0.5
6
1
00
5 10 15 20 2S 3D
Timr (min)
3: LC .chromatograms (UV, 210 nm) of garlic SF extracts: (A) 35 cC, 3
renod 240 atm. (B) 50 "c, 28 min slatic period at 240 atm. Peak
Idl;Ull fi eat lon: (5) aJoene(s); (6) I ,3-vinyldithii n; and (7) diallyl trisulfide.
10. CAI.V,,-Y &: nWCK SCFE of Allium Species 121
the IOO-fold decrease in levels of flavorants in the former andlor our failure [0 detect
non-volatile flavoranls present in the SF extracts of onions. The profile of the
navorants in SF C1{tracts of onion, as detcnnincd using GC-MS, was very similar to
that previously reponed for ether extracts of onion (55). Anal ysis of SF extracts of
onion by LC showed many more componenlS than seen by GC- MS but peak overlap
makes component identification difficult In brief. analysis of an SF extract rapidly
prepared from 10 IS g of fresh yellow onions and inuncdiately analyzed showed the
presence of var ious thi osulfinales. propilnelhiai S-oxide (LF), bi ssulfine(s),
zwiebelanes, signifieant quantities of a series of eepaenes (RS(O)CHEtSSR') and a
few related compounds (56,57). Because many of these compounds were non-
volati le, GC-MS may have significantl y underestimated the efficiency of the SFE
procedure. Moreover many of these I;ompounds had weak UV absorption so that
monitoring LC analyses by UV detection at 254 nm was problematic. Instead, the
ident ificati on of compounds was achieved by LClMS techniques that resolved
overlapping conlpounds of different masses (58). This analysis employed synthetie
standards that had been authenticated by OIlier spectroscopic: methods.
The Erred of SFE Conditions on the Di stributi on of Organos ulfu r
Compounds Fouud in Garlic
Yields of biologically active rearrangement products (ajoenes, dithiins and alkyl
sulfides) derived from all ici n vary with the type of solvent used. lberl el a!. (59)
reponed on the transformation products from crushed garlic that had been stored in
ethanol for 9 days (216 h). Approximately 10% of the ini ti al allicin was converted to
the ajoenes and 8% of the allicin remained: the major product fonned from the initial
thiosul finates prescnt was the diallyl trisulfide (>70%). This pattern contrasted wit h
lherl et aI's results for crushed garlic stored in oil. In this case, no allicin remained
after 24 h, and the ajoenes, dithiins and diallyl disulfide const ituted ca. 25, 60 and 10%
of the transformation products, respectively. At 48 and 216 h slight changes in these
ratios of the transfonnation products were found in the crushed garlic in oil.
We have continued our SFE st udies of Allium species by investigating the
effects of di fferent SFE conditions (i .e., temperature, stat ic extraction time and
modifiers) on the distribution of organosulfur compounds in the result ing ext racts .
These extracts were analyzed by reversed-phase LC. Initially, we investigated the
effect of temperature IlJId the duration of statIc extral;tion ti me on the distribution of the
organosulfur compounds in garlic SF ext racts by comparing chromatographic profiles.
An SF extract obtai ned at 35C with a 5 min static extraction time was used as the
reference. Our J)Climi nary results showed that an SF extract obtained at 5O"C with a
30 min static period produced the most dramatic changes in the chromatographic
prori les. These c hanges were observed after IS min (Fi gure 3). By comparing LC
retention times and on-line UV spectra with synthesized standards (Figure 4), two of
the peaks in this region can be assigned as ajoenc(s) (pk 5) and diallyl trisulfide (pk 7).
The most intense peak in this region can be tentatively identified as I,J-vinyldithiin (pk
6) by comparing Its UV spectra with that found in the literature (59). The other peaks
in this region have UV spectra characteristic or al.kylsulfides.

Extr action. Allium extracts were obtained by using a PrepMaster and Accutrap
(SuJ)Cx, Inc .. Pittsburgh, PAl. Whole cloves of garlic (2-4 g) or- ramp bulbs (2-5 g)
were homogenized in 10 mUg of water at ambient temperature with a Ti ssumizer
(Tekmar, Cincinnat i, OH). Freeze dried ramp (0.5 g) was homogenized with 20 mL of
water. The solutions were allowed to stand at room temperature for 5-10 min to ensure
complete enzymatic conversion to the thiosulfinates and related compounds.
Hydro matrix (Varian, Palo Alto, CAl was mixed wit h 15 mL of homogenate and
122 SI' ICES : nAVOR CII MISTRY ANI) ANTIOXIDANT
"



"
Sol:

N
:j
.
"
U .
N o :.=.g

0"'..:

<> gS
8


II.. "C .-

.. K.r::'
;:::<l ........
E

"
'"
WI.!:: a

.: ::I.s N
'f,
c
119::s
u :J E:.s

.25>-
:5
0
8..,5

c ,
N
g ... <"'I.

e:i:o
"
Ol-
8
"
..
.... 'E'-'
N
-8


"

.... N
..... N

. :.... . 3
,,'

uo
8
.
N
U. ;;; . _
"
N
"
6 6
"
6
"
:):)Ul!qlOSqV

10. CALVEY & Il LOCK SCF of Allium Sp;t.f 12.1
placed in a 50 mL extraction vesseL Unless olherwise indicated, .he homogenates
were extracted at 35C and 240 aIm at a flow rate of 2 mUmin unt il 60 g of C02
(SFE/SFC grade. Air Products, Allentown, PAl were consumed. A static eXlratliOn
time of 5 m.in was used to ensure [hennal equilibrium. The effluent was trapped on
glass beads at Joe and dcsorbcd with I mL of methanol, mcthanoUwater (50150). or
ethanol. The ramp exlfaClS were analyzed directly, whereas the garlic extracts were
dil uled to 5 mL with water.
LC Ana lyses. LC analyses were perfonned by usi ng a Waters 600 pump equipped
with a Rheodyne Modd 7725i manual injector (50 ilL loop) and a Waters Model 991
photodiode array detection system. Chromatographic separation was achicved on a
YMC-Pac!r:: PolymcrCI 8 (YMC Inc, Wi lmington, NC) column wit h a OPTI-SOL V
mi ni fillcr (Opti mize Inc., Portland, OR). Thc linear grati ient was 60%
H
2
0 (sol ... ent A)J40% acetomtrile (solvent 8) which was held for 10 min, then
increased by 8% acetonitri le/mi n to 20% H20/80% acetonitrlle, and held for 15 min .
The fl ow rate was 0.8 mUmi n,
Acknowled gme nt s
We gratefully acknowledge support from the National Science Foundation (E.B.), the
NRI Competi ti ...e Grants ProgramlU.S. Department of Agricult ure (Award No. 92-
37500.8060), and McCormick & Company, Inc. (E.B.). We thank R. Epply for
walk ing the banks of the Potomac in search of wi ld ramp and P. Whangcr for
pro ... iding the freeze dried ramp.
Literature Cited
l. Grimmell , C. Cliemistry and Industry, 1981, 359.
2. King, I. W.; l ohnson, I .H.; Friedrich, J.P. 1. Agric. Food Chern. 1989,37, 95l.
3. Hoyer, G.c. ChemTech 1985,440.
4. Banlc, K.D.; Cl ifford, A.A. Spec. Publ . . R. Soc. Chern. 1994, 160, I.
5. King, J.W., INFORM 1993,4, 1089.
6. Eckert, C.A.; Van Aislen, J.G.; Stoicos, T. En ...iroll . Sci. Technol. 1986,20,319.
7. Dzie:zak, J. D. Food Teelinology 1986, June, 66.
8. Ri:z...i, 5.S.H.; Daniel, l A.; Benado, A.L.; Zollweg, I.A. Food Techllology 1986,
57 .
9. Hawthorne, S. 8 . AI/al. Cllern. 1990,61, 633A.
10. McHugh, M.A.; Krukonis, V.J ., Eds. Supercritical Fluid Extroction Principles (IIu/
Practice, Butterworths, Doston, 1986.
11. Vitzthum, 0.; Huben, P. US Patent 4 123 559, 1978: Chern. Ahstr .. 1973, 79,
306780. .
12. Strauss, K. Ger. Offen. DE Patent 3 133032, 1983: Cllern. Abstr .. 1983, 99,
7546k .
13. Schil tz, E. ; Vollbrecht, Il. -R. ; Sandner. K.; Sand, T.; Muhlnickel , P. US Patent 4
470 927, 1984; Chern. Abslr. 1983,99, 4363n.
14. Makin, E.C. US Patent 4 474 994, 1984; ClielTl . Abslr. 1984, 101, 210742y.
15. Behr, N.; Van der Mei, H.; Si rtl , W.; Schnege1berger, H.; Von Ettinghausen, O.
US Patent 4 490 398, 1984; Chell! . AbSlr. 1985, 101, 111886n.
16. Japan Oxygen Co., Ltd, Jpn. Kokai Tokkyo Koho JP Patent 59 232 064, 1984;
Chem. Abslr. 1985, 101, 202893d.
17. Japikse, C. H.; Van Brockin, L. P.; Hembree, lA.; Kitts, R.R. ; Meece, D.R. Eur.
Pat. App!. EP 206 738, 1986; Chern, Abslr. 1987, 107, 6048w.
18. Calame, J.P.; Steiner, R. Eur. Pat. App!. EP 154258,1985; Chern. Abstr. 1986 ,
104, 33288u.
19. Kobayashi, T.; Taniguchi, M. lpn. Kohi Tokkyo Koho JP Pa.ten! 63 87 977,
1988; Chern. Abstr. 1988,109,2075275 .


,
",
.,
I
.1
124 SI' ICt:. ...: FlAVOll ANI) ANTlOXI1>ANT I' ROI't; RTIES
20. Sugyama, K.; TOlllohiro, Y. Jpn. Kohi Tokkyo Koho JP Patenl 01 J 17 761.
1989; Cluml. Abslr. 1989, 1/1, 76540g.
2 1. KaZll Yuk..i , Y.; Kobayashi, M. l pn. Kokai Tokkyo Koho JP Patent 02 135069,
1990; Abstr. 1990, 1/3, 96385e.
22. Kobayasbi. M.; Shiraishi. S.; Matsukura, K.; Yamashi ta, K. Jpn. Kokai Tokkyo
Koho JP Palent 02 235 997, 1990; Abstr. 1991 ,1/4, 60706s.
23. Nguyen, U.; Evans, O.A.; Berger, 0.1.; Calderon, J.A. US patentS 120558,
1992.
24. Coenen, II .; Kriegel. E. Ger. Chem. lIg. 1984,7, 335.
25. I-ia .....lhorne, S.B.; Krieger, M.S.; Miller, 0 .1. Allaf. Chern. 1988, 60, 472.
26. li awthorne, S.B.; Mil ler, 0.1.; Krieger, M.S. 1. Chroma/ogr. Sci. 1989, 27,
347.
27. Anderson. M.R.; Swanson, J.T. ; Porter, N.L. ; Richter, B.E. 1. ChromalOgr. Sci.
1989, 27,371.
28. Cal vey, E. M.; Matusik, J.E.; White, K.D.; Betl, J.M.; Bl ock, E.; Li nlej ohn,
M.H.; Naganathan, S. ; Putman, D.J . 1. Agric. Food Chem. 1994, 42, 1335.
29. Krukonis, V.J. ACS Symp. Su. 1985.289, 154.
30. SUf;iyama, K.; Saito, M. J. C/lromalOgr. 1988,442, 121.
31. Ondarla. M.; Sanchez, A. CflrOll1alogruphia 1990, 30, 16.
32. Reverchon, E.; Senatore, F. 1. Agric. Food Chem. 1994,42. 154.
33. Huston, C. K.; Ji, H. J. Agric. Food Chem. 1991 ,39, 1229.
34. Sanagi, M.M.; Ahmad, U. K.; Smith. R.M. J. Chromatogr. Sci. 1993, 3f, 20.
35. Charpenti er, B.A. ; Sevenants, M.R.; Sanders, R.A. in The Shelf of Foods
(!l I d Charalambous, G. (Ed.), Elsevier. Amsterdam, 1986, 413.
36. Verschuere, M. ; Sandra, P.; David, F. J. ChrOllla/ogr. Sci. 1992,30,388.
37. Reverchon. E.; Della Porta, G.; Senatore, F. J. Agric. FoO(/ 1995,43,
1654.
38. Kal lio, H.; Kerrola, K.; Alhonmaki, P. J. Agric. Food 1994,42,2478.
39. Bart ley, J.P. J. Sci. Food Agric. )995,68, 21 5.
40. Taylor, S.L.; King, 1.W.; Snyder, 1.M. J. Microcofumn Sep. 1994, 6, 467 473.
41. Blanch, G.P.; Ibanez, E.; Herraiz, M.; Reglero, G. Anal. 1994,
892.
42. Nam. K.S.: King. J.W. HRC 1994, f 7, 577582.
43. Lembke, P.: Bornet, J.; Englehardt, H. 1. Agric. Food Cll elll. 1995,43, 38 55.
44. Blanch, G.P.; Reglero, G.; Herraiz, M. J. Agric. Fomi 1995,43, 1251
1258.
45. Siahl, E.; Gerard, D. &: FfOl"oris/. 1985, /0, 29.
46. Arreola, A.G.; Balaban, M.O.; Marshall , M.R. ; Peplow, A. 1.; Wei, C.I.; Cornell ,
J. A. l . Food Sci. 1991 , 56, 1030.
47. Hubert, P.: Vitzthum, O.G. Cllen!. Inl. Ed. Engl. 1978,17,710.
48. Caf3gay, A.B. &: FfavoriSI. 1981,6, 43.
49. Vi rtanen. A.I. PIrYlochem. 1965,4, 207.
50. Miles, W.S.; Quimby, B.D. Am. Lab. 1990, Jll fy. 28F.
51. Sinha, N.K.; Guyer, D.E.; Gage, O.A.; Lira, C.T. J. Agric. Food Chem. 199 2 ,
40, 842.
52. Block, E. J. Agric. Food Chem. 1993, 4/,692.
53. Wagner, H.; Breu. W. DISch. Apolh. ZIg. 1989, 129, 21.
54. Lawson, L.D. in Human Medicillal Aserlls f rom PfOIllS, Kinghorn , A.D.;
Ba13ndrin. M.F. (Eds. ) ACS Symposi um Series No. 534, American Chemical Society,
Washington, D.C. 1993, 306 330.
55. Bl ock, E.M.; Putman, D.; Zhao, S. H. J. Agric. Food Chern. 1992,40,243 1.
56. Bl ock, E. Angell'. Cllem. fil l. Ed. Engl. 1992.31. 1135 1178.
57. Bl ock, E. ; Calvey, E.M. in Sulfur Compounds in Foods, Mussi nan, C.1. ; Keelan,
M.E. Eds. ACS Symposium Series 564, ACS, Washington, D.C. 1994,6379.
58. Matusik, J.E.; Whi te, K.D.; Bl ock, E.; Calvey, E.M. "Supercritical Fluid
Ext raction and At lllospheri c Pressure: Chemi cal lonilation MSIMS of Cepaenes. "
presented al44th ASMS Annual Meeti ng. Port land OR. May 19%.
59. lberl . B.; Wi nkl er. G.: Knobloch, K. Pfarl/a 1990, 56, 202.
Chapter 11
Determination of Glucosinolates
in Mustard by High-Performance Liquid
Chromatography- Electrospray Mass
Spectrometry
Carol L. Zrybko
'
and Robert T. Rosen!
INabi sco, Inc., 200 AVC!nue, East Hanover , NJ 07936
for Advanced Food Technology, Cook College, Rutger s,
The Slate of New New Brunswick, NJ 08903- 0231
A method was developed to detemli ne glucosinolat es in mustard
seeds by reverse phase HPLC using volatile ion pairing reagent s
followed by UV at 235 nm. Two external standards.
phenet hylglucosinolate and sinigri n used to quantify results.
The identities of individual mustard glucosinolalcs were confi rmed
by negative ion elcctrospray mass spectrometry as was LC peak
puri ty. This LCIMS method may be used to identify speci es which
are not commercially available as pure standards since mass
spectrometry can be used to check for all known glucosinolat e
anions_
Three mustard types were chosen rrom the Brussica species.
(Bross;ca hirfo), brown, and oriental (8rassica j lll/Cca). Eleven
mustard samples representing harvest areas of southwestern Canada
were anal yzed in triplicate for glucosinolale content. Percent
coefficient of variat ion l>ctween triplicate samples or the same batch
was often less than 10"10.
In mustard and other members or the species of the Cruciferea family of
vegetables, the import ant nonvolatile precursors are the hydrophil ic glucosinolates.
The structure of glucosinol ates. as seen in Figmc I , impart s strong acidic properties
to the compound duc to the sulfat e group. Other import ant structural moieties
include the glucose and cyano groups, vary due to differences in
the R group. The side group can be alkyl, branched alkyl. aromati c. or
unsaturated_ The differences in the si de chain impart different l1avors to rood, AJ I
glucosinolates are formed in pl ants from Lamino acids and sugars through a
common pat hway.
<D 1991 American Chemicat Society
126 SPI CES: HAVOR CII KPtIlSTRY ANn ANTIOXIDANT I'ROI'ERTIES
R

s-c/

Oli NOSO)'
0/-1
Il
Fi gurt I The structure of glucosinolates.
Glucosinolates breakdown dut to the action of enzymes and heat to give many
fla vor compounds. Myrosinases (thioglucoside glucohydrolases) are found
compar1menlalized within the plant, and are released when lhe cell walls break
through bruising, sample grinding, Of during normal growt h as a result of damage
and death of cells_ Olucosinolates are easily degradt:d by the myrosinase enzyme at
both acidic and basic pH to give such products as isothiocyanates. nit riles,
thiocyanates, m;:azolidine2-thioncs. hydroxynitrilo:lS, and epithionitriles. all of which
impan flavor to mustards (1).
Kalt. kohlrabi, broccoli, cabbage, mustard, ClIuliflower, rapeseed. turnip. and
rutabaga. of the Brassi(""(1 group of cruciferous vegetables, are best known for their
pungent flavor. In ancient limes, these crops were thought to have medicinal power
over headaches. goul, diarrhea, deafness, and stomach disorders (1). Olucosinolate
breakdown products, specifically )-i ndole carbinol, )indole acetonitrile, and 3,)'-
diindolylmtthane, produced by myrosinase degradation of giucobrassicin. are
considered Phase II anticarcinogens (2). These compounds have also been provtn
to increase enzymatic activity of glutathione-S-transferase (OS H), ethoxycoul11arin
O-detlhylase (ECD), aryl hydrocarbon hydroxylase (AHH), and epoxide hydratase
(EI-/), all of which in tum increase the polarity of lipophi lic carcinogens so that Ihey
art: more readily excreted in the urine. Olher benefits from the glucosi nolates in
plal"llS include decreased growth of microorganisms. fungi, and Ihe ability
to repel mOSI insects (3).
To date, most research into gl ucosinotale idenlificalion has been dOlle by
isolating glucosinolate breakdown products and then determining the original
precursor (-I). This method of back tracking was favored because il was
inexpensive and fast. Analysis of glucosinolales has also retied heavily on Gc/MS
techniques on eit her gJucosinolate degradation products such as the isothiocyanates
or on the desutfated and trimethylsilyated (TMS) derivit ized gtucosi notates (S) .
Mosl often, in the case of MS, the isolhiocyanates and other volalile compontnts
are identifi ed rather than the g!ucosinolate itself(6}.
In this st udy, negative ion dect rospray mass spectrometry is used because il
detects molecular anions and at!ows for direct identi fi cation ofgJucosinotates, most
of which are unavailable as pure standards. We apply a novel method of reversed
phase liPLC ulilizing volatile ion-pairing reagents followed by negative ion
electfospray mass spectromelry to semi-quantify and ident ify the glueosinolates in
mustard seeds. Formic acid and triethylamine (TEA) were used to produce the ion-
pairing agent , triethylammonium formate which interacted with the glucosinotates to
increase retention and prevent their elution in the void volume.

1I . ZRYBKO" Glucosillolaln in Mustard l27
Material s & Method
Analytically pure sinigrin (98%) was purchased from Aldrich (51. Louis, MO). as
were the ion-pairing reagents, formic acid and triethylamine. OluconaslUni in
(phenethylglucosinolate) was obtained from LKT Laboratories (St Paul, Minn)
Sinapic acid (98Y.) was supplied by Aldrich (St. Louis, MO)_ HPLC-grade
methanol was purchased from Fisher Scientifi c (Pittsburgh, PA)_ HPLC-grade
watcr was generated in the laboratory through the lise ofa Milli-Q UV Plus Ultra-
pure waler system by Millipore (Milford, MAl.
The externat standard. phenethylgtucosinolate, was weighed (10.27 mg)
and dissolved in 10 mt m'LC'grade water 10 obtain a solution with a conctntrati otl
of 1.027 mgfml. Si nigrin monohydrate (10.27 mg) was also dissolved in HPLC-
grade water to obtain the same ooncentration as the phenethylglucosinolMe
Sinapine bisulfate (98%) was purchased through Spectrum Cht:lI1ical (Gardenia.
CA) and was used as an external standard to quantify sinapine found in all three
mustard types. All standards were filtered using a 0.45 micron Acrodisc purchased
from Gelman Scientitic (Philadelphia, PA). Standards were stored in
autoS.1mpler vial s under refrigeration until needed_
Drown, oriental, alld yellow mustard seeds were obtained from various suppliers.
Continental Grain (Albena, Canada), North Dakota Mustard and Spice, and
McConnick & Company (Hunt Valley, MD). Samples were prepared by grinding
-25 grams of mustard seed in a 250 1111 IKllypropylene co:lntrifuge lube, Thomas
Scientific (Swedesboro. NJ), with 100 ml methanol. A Deckman Polytron was used
for three minutes 10 insure a homogeneous grind. Tho:l sftmpl es were then
centrifuged at 2700 rpm for 5 minutes. The supernatant was poured off into a 250
mt round bottom flask and then evaporated using a Rotovap set at 55C and an rpm
of 55. Samples were evaporated to a volume of approximately 2 ml methanol. The
samples were Ihen dissolved in 100 ml HPLC-grade water. All samples were
filtered usi ng a 0.45 micron nylon Acrodisc as used during 51andacd prcparalion
Analyses of mustard samples were performed by injecting a 20 ul aliquot of the
sample into a Waters (Milford, MA) high' perfonllallce liClUid chromatograph linked
to a Waters 490 ultra-violet detector set at 235 nm. A Waters 600E System
Cont roller and a 712 WISP were also used. The tillire f-IPLC system
was controlled using the Walers Millenium data acquisition software program
version 2. 1. The column used was a Phenomenex (Torrence, CA) 5 pill 00S(20)
column measuring 250 mm X 4.6mm with a 55 mm X 4.6 mm Walers C-18 guard
I;olumn attached
The two mobile phase solvents, methanol and water. were prepared with O. I 5%
tri ethylamine and 0. 18% formic acid added as ion. pairing reagents_ Both solutions
were filto:lrcd using a 0.45 11m filte r, sonicated for 30 minutes, and degassed with
helium throughout Ihe chromatographic procedure at a sparge ralt: of 50
ml/minute. The initial mobile phase was 100% HPLC-grade waler. At ten minutes.
the mobile phase was switched to a linear gradient of loew. water to 100%

128 SI'ICt;S: n .... VOR t: llliMISTRY ANIl ANTIOXIDANT !'IW,' EM.Tl K'i
methanol over 60 minutes After each run the illilia! mobile phase conditions were
set and the system allowed 10 equilibrate. The now rllle was kepI conslant at I
mUminute. The column temperature was held al 40
6
C
Confinuation of peak identity was accomplished through the use or a Fisons
(Danvers. MA) VG Platform II mass spe<:\romeler connected 10 a YG Mass Lynx
data system. The system was operated in the negative ion eledrospray mode. The
ion source temperature was sct al 150C while the cone voltage was -20V. The
HI'LC conditions were identical to those obtained off-line except that a Varian 9012
HPLC (Fernando, CA) was used al ong with a Varian 9050 UV-Vis detector.
Quantilicat ion of the nonvolatile flavor precursors, sinigrin, progoit rin, si nal bin.
was determined using external standards. A solulion containing seW.
and 50% si nigrin monohydrate was analy..:ed to determine
the response of both aromatic and alkenyl glucosinolates, respectivel y. The
ratio of response factor was 1 19, indicating that the aromatic glucosinolate was
more sensitive. The concentration of the aromatic compounds, sinalbin and
sinllpinc, were calculated using this response factor since pure were
commercially
li nd I)i scussion
Glucosi nolat e QlIlIntil1ClI tion. Ye!low, brown, and orielllal mustards were
analYl.ed for their glucasinolate content. The major compounds identified by mass
speclromelry can be seen in Table I. A chromatogram of ycllow mustard can be
seen in Figure 2, whereas Figure 3 shows a chromatogram of brown mustllrd
Many ofl he minor peaks seen on the HPLC chromatograms were nol visible by the
mass spectrometry method utihed. Thi s indicates that these compounds are
um; harged and therefore are not gillcosinolates. Result s for the of
compounds in yellow mustard (Brassica hirta) identified by mass spec. can be seen
in Table 11 .
Ta bl e I. Non\'olaciles J' ound in Yellow, Brown, a nd Orient ll ll\luSlards
RT "til: hlelltijictJtiQII
1# (obseM'CtJ)
10. 5 minutes 358 AJ lylglucosinolate (Sinigrin)
II 2 388 2- hydroxy-3-
butenylglucosinolate(Progoi trin)
21 3 42. l2-hydroxybenzylgluCQsinolate (Sinalbin)
33 4 35. 3.5-dimethoxy-4-hydroxycinnamoyl
choline (Sinapine)
35 5 586 Sinalbi n + glucose
II . ZRYBKO&ROSEN GlUC05inoltltl.f in MtJ/j'laNI

3
;;
L
,1 ,
0
-
"
1
, , , , ,
0 20 '0
60
Time (min)
Figure 2- [fPLC chromatogram of methanol extract from yellow mustard seeds
(See Table 1 for idcmilications)


1
0
l.wL
0 ,

, , , .
0 20 '0
60
Time (min)
Figure 3- HPLC chromatogram of methanol extract of brown mustard seeds
(See Table I for identifi cations)
Tabl e II. Gluc(lsinolate Contellt of Veil ow i\"lust:ml Seeds
Il) /I Htln'!!st Ar!!" Progoitr;" SimI/bin Silltlpine
Amou", (mg/g)
4-6 Calgary, Canada 0.51:,::0. 14 27.46:!: 1.20 27.68:!: 1.41
10-12 Manaloba, Canada 0.60:,:: 0.17 27.1!7.:t..!.09 32.09+ 0.69
16- 18 Alberta, Canada 0.56:,:: 0 10 24.81 1.92 30.60:'::2.21
19-21 Saskatchewan, Canada 0.45.:':: 0.04 23.00+ 2.85 28.05+ 4.1 5

12.
130 SI 'H': E."'; fL\ VOI( CIIK\II STRY ANI) ANTlOXIUANT
I' roguilrin. 2.hydroxy-J-butcnylslucosillolah.:, (figure 4) is Ihe precursur I() g,l i]rin
whie'h lms been shown 10 hypothyroidi sm in ani ln:!ls fed large of
cnlCi rerOIiS vegetables. The avenge Ieve! of progllilrin tillllld in yell ow ll11btilrd
was 0 53 mgtg! 0,06. The identification of progoiuin was (ktcrmiucd directly
lioml hc I1l.:gali,'c iOIl C"lC<:l fOSllray mass spectnml The mass spccrnnn CIUI De seen
in 5. The molrcular aniun can be seen al 11111. J:S1! The mulo..'(:ulnr weight of
I' rngoilrin is ltct ualty 427 (a<ldi . iol1 o f potassium (lIIion) The llIass spl-.:: trolHelcf
sees o nl y the lInion when in lhe negl1!ive iOIl mode,
Oil
I

" .
Oil NOSU}
'"
Oil
Figure 4. StrucTure ofprogoilrin
Si oi.lbill, (Figure 6), is found al level s in
yellow mustard The coocentf',Hion of sinalbin was 25.79 mglg 1: 2.30, which is
consiSU!n1 wilh Fenwi ck's Crilical Review ofGlucosinol:ues and thei r llre:lkdow/l
l' rulhl!,;lS (1982) and The USDA I'hytochcmiC;J1 datll available on the Int ernet
Si nalbin is n01 found ill 1I11{\Jl C >I S a !lOtassium
sail. but mlher has sillapinc as Ihe posi tive iOll The stn lcture and informati on lu r
silmpinc is Illentioned below bt."ClIusc sinapine is present liS :I peak Oil the
of brown, and OI'icnlalliluSlards.
Si nalbi n was idemified by 1ll1l SS specn omcl ry The molecular aniun is seen at
nih 424 7).
Figure 6 StntCt UfC of si nalbin anion
The cOlllpound wit h a retenlion lime of 35 minut es has been hypot hesizc!! as
sina!bin plus I:\IUC05C. The m/z of 586 shows 424 mass unit s from lim! 162
from gilleose have rCllOrled Ihe presence of glucosi nn!ates wit h II
second sugar in Illustard lind OI lIer vegetables liom the Hm ....,inl spccies (I)
FUr1hcr res ...arch, illcludill!; fractionation and NMR, must lJe done on Ihis cnlllJ'lOuml
in order 10 determi ne Ihe exact stl1lct ure
..
I I. ZIH' UIW &.

<'l 0
'" 0;
e _ ... ''l.

8

. --_._._------
"
"
o

e

o
-
N
o
N
N
o


HI

-"
o
/.
1.32
SPI CE.Ii: HAYOR Cm : M/STltV ANI) ANTlOXIJ)ANT
i
I
I
j
11. ZRvnKO & ROSl<;N Glucosillll/ale.f in Muslard 133
Brown and oriental mustard are from the same species, Brassica jlfllcca. and
therefore. had very similar gluoosinolate content. The individual results for each
mustard sample can be seen in Table III.
Tahl e UI. Glucosinolat e Cont ent of Brown Mustard Seeds
1/) /1 1)"pc Hlln'Q/ Arca Sinigrin Sinnpinc
Amount (!lIgfJIJ.
1--3 Orienlal Calgary. Canada 9.59 0.94 22.05 J..'j0
79 Oriental Saskatawan. Canada 9. 10+0.19 24.50 1.09
22--24 Oriental Saskatchewan. Canada 9.3 1 0.52 25.00 1. 97
25--27 Orienlal Alberta, Canada 8. 16 0.36 20.24 1.77
13-15 Brown Saskatchewan, Canada 8.29 0.26 24.20;t 0.92
28--30 Brown Alberta, Canada 8.S0 1.97 14.36 2.70
31--33 Brown Saskatchewan, Canada 8.20! 0.95 12.98;+;; 1.78
The major compound in brown and oriental mustards is sinigrin. commonly known
as allylglucosinolate. (Figure 8) with a retention time of approximately 10.5
minutes. Sinigrin degrades to allylisothiocyanate (AITC) by the myrosinase enzyme
in the presence of water to give mustard its pungency. The concentration of sinigrin
was found to be 9.04 mg/g 0.62 in oriental mustard and slightly lower than that
for brown. The CV of the sinigrin results for all twenty one samples in the 8rassicn
jill/ceo group was 1069' 10. The identity of sinigrin was proven by negative ion
c1ect rospray mass spectrometry when the MS showed a peak wit h a mass to charge
ratio of)S8 as seen in Figure 9_ The negative ion is seen by the mass spcctrometry
method used. whereas the positive ion. potassium. is not observed. The actual
molecular weight of sinigrin is 397.
Figure 8- Structure of sinigrin anion_
Sinapine. a choline ester ofsinapic acid (Figure 10) was identilied in yellow, brown.
and oriental mustard. This compound, 3,5-dimethoxy-4. hydroxy cinnamoyl chorine.
is common in the 8rassica group. Researcher.; have shown that although sinapine is
the positive ion to sillalbin, the glucosinolate and phenolic choline ester contents are
not correlat ed wit h one anot her (7). For this reason sinapine is seen in both brown
134
SI'ICt;S: FlAVOR CHEMISTRY AND ANTIOXIIlANT .'ROJ'EN-TIES
o
"
. .,

."
'0
E
2
I

11. ZltYDKO & ROSt:N Gluco:tinolaU5 ;11 Mustard 135
lind oriental mustards although sinalbin is not present. The neglllive ion
eleclTospray mass spectrum of sinapine using a low cone voltage of 20 V, showed
an ion al mJz 354, The peaks seen above the mlz 354 peak in Figure 11 al'e 46 mllSS
units higher, which is the compound fonning adducis wilh formic acid, the mobile
phase modifier. Since very liule fragmentation occurs at such II low cone voltngc.
8 second mass spectrum was run at II higher cone vol tage, 120 volt s, to give
fragmentation. This data and II NMR of the fractionated peak showed Iha\ the
compound was an adduct of sinapine and formate, 3 10 mass units coming froln the
sinapine calion and 4S amll frOIll the anion. The [M-Hr of this comple ...
occurs at nlfz 354. Interestingly. becKuse of the ion-pairing, the sinapine did not
give a spectrum in the positive ion mode.
o
II
CH - cu -CO-CI 12CJ I1 N(CI'!lh
HOT

Figure 10- Structure of sinapine.
Conclusion
To date, the major nonvolatiles in mustard seeds have been isolated. identifio:'d.
and quantifi ed by reversed phase HPLC using ion-pairing reagents and negative iOI1
etcctrospray mass spectrometry. Accuracy and I)recision of the method was
determined by bolh spiking studies and replicate analysis of individual smnples.
respectively. The accuracy was detennined by spiking both brown and yellow
mustard samples with sinigrin and calculating the percent recovery. Recovery was
olien greater than 100-/ . . Precision was detennined by running all samples in
tripl icate and calculating the standard and coefficient of variation. CVof
most SIImples was less than 10"10. This data indicates that this method works well
for identifYing intact glucosinolates and therefore eliminates the need for desulfating
and derivitil.ing glucosinolatcs as seell in previous (5), Use of mass
spCl,;trometry allows identification withoul glllcosinolate startdards. most of which
caullot be purchased.
Ackll owlrdglll tlits
The author gratefully acknowledges the work of Dr Elaine M. FukudH, CAFT,
Rutgers Universi ty and Dr. Janet M. Deihl , Nabisco, Inc. ntis is New Jersey
Agricultural Experiment Station Publi cation /I DI0570-96-2 .

136 SI'I CK'), FlAVOR CH):;MISTRY AND ANTIOXIDANT PROI'I' )l.Tl ES
; -




M M H::
________________________________________ ___
"
o
o
"

"

lL ZRYRKO & ROSEN
Glm;os;lIoiate!l" in Mustnrd
Literature Cittd
Fenwick, G. R.; Heaney, R. K. . Mullin, W. J" eRe CriT. Rev. Food Nul., 1982,
18, 12320 1.
2. Bradfield, C. A , Bjcldanes, L. F Vtl. Chern. Toxic. 1984,22,977-982.
3 Delaquis, PJ; Ma7..za G. Food Tech, 1995,11,73-84.
4. Oleson, 0 ., Sorensen, H . 1. Allier. Oil Chem. SOC;. 1981 , 857-865.
5. Helboe. l'. Olsen, 0 ., Sorensen, H., J. Chrolll_. 1980, 197, 199-205.
131
6. Kjaer. A., Ohashi, M.: Wilson, J. M., Djerassi, C. Acta Chemica Scal1dillavica,
1963,17,2143-2 154
1.Bouchereau. A , Hamelin, 1; Lamour, T., Renard, M; LaTher, F; Phyux:hemis/I)'.
1991, 30 (6),1973- 188 1
Reasons for
of Some
Chapter 12
the Variation in Composition
Commercial Essential Oils
Chi Kuen Shu and Brian M. Lawrence
R. J. Reynolds Tobacco Company, 950 Reynolds Boulevard,
Winston.Salem, NC 27102
The reasons which cause the compositional changes for some
commercial essential oils are due to (1) the effect of exlrinsic
(2) the effect of interspecific and infraspecific
dtffercl:ces, (3) the effeci of ontogeny, (4) the efft(;l of
processmg parameters, and (5) the effect of adulteration
Using examples, each effect is discussed_ .
Over the years it has long been known that commercially available essential oi ls
vary from to and from geographic source to geographic source_
However,. thIS questt?n of variation has not been carefully examined across a
range of oils to detemune why such variations take place. Usually it is explained
that one se.ason to another result in compositional changes;
however, thiS IS a little too Simple to explain some of the composi tional changes
that are encountered.
As a result, we would to discuss some basic reasons why compositional
are encountered usmg examples of the magnitude of such changes The
effects that will be discussed are: .
I . The effect of (climate/geographical origi n).
2. and mfraspecific differences (between and
wlthm species).
3. The effect of ontogeJ.ly (growth stage of plant material harvested).
4. T.he. of processing parameters (dry 'Is. weI plant material on
dlSl1llal1on changes)
S. The effect of adul teration (the addition of synthetic or other natural
components to "stretch" an oi l.
t997 American Chemical

12. SHU & lAWRENCE
Variations ill Compoljtion oj Some Essential Oils
The EfTect ofClimlite and Geographic Origin on Oil Composition.
In 1989 Lawrencc et aI . (I) showed that peppermint oil, which is produced from
clonally reproduced plams that are grown in different regions of the United
States, could be readily differentiated based on Iheir area of cultivation and oil
production. The areas of commercial production oflhese oils are the Midwest ,
Idaho, Yakima-Kennewick Valley in Washington, and the Madras and
WiHamette valleys of Oregon. It was proved that geographical location and
microenvironment were influential faclors which affect, albeit slightly, the
composition of the oils produced in each region. Examination of the quantitative
data of the most constituents revealed that it was extremely dimcult to
differentiate the oils because of the similarity in the data. As a result, selected
component ralios were used to prove differences between the oils. To simplify
the component ratio data obtained from ca. SO samples of oil obtained from each
growing region, Ihey were presented as polygonal representations. Such a
presentation of data can be thought of as being pictorial pallern recognition of
the oils produced in each region.
139
The components ratios selected to differentiate between the oils can be
seen in Table I. Many of these ratios have multipliers that were used to ensure
1hat the polygons formed would lit the scale of the circular graph. What this
means is that only oils produced in a specific region will have their component
ratio data fit within the polygon. Oils whose component ratio dala does nOI fit
within the polygon cannot be a pure oil produced in that specific region. The
pallern recognition of the five different types of peppermint oil can be seen in
Figures 1-5.
To further usc this concept to differentiate between other oils produced
in the United States from different regions we attempted to differentiate between
Native spearmint oils (ex. Mentba spicata 1..) grown in the Midwest and in the
Farwest. Similarly, we also attempted to differentiate between Scotch spearmint
oils (ex. Menlha iracilis sole), which are also produced in the Midwest and the
rll.lWcst We could not use lighter regional differences between the oils because
the oils are not sold according to the state or valley in which they are produced,
only by the broad regions of Midwest and Farwest.
For both speannint oils the component ratios used can be seen in Table
n. Again, multipliers were used to ensure that the component ratio data for bOlh
spearmint oils would fit on the same circular graph. The results of the
component ratio data that were obtained for 20 samples of each oil type plus 10
samples of a second culling oil obtained from Farwest Native spearmint can be
seen in Figures 6-10. The compounds chosen for component ratios were those
whose raw analytical data varied so that a differentiation between Scotch and
Native spearmint oils and 2"" cutting Native spearmint oils could be made. For
example, Scotch speannint oil is slightly richer in limonene, 3-octanol, menthone
and carvone as compared to Native spearmint oil, which in tum, is slightly richer
in mycrene, 1-8-cineole, 3-octyl acetate. trans-sabinene hydrate, P-bourboncne
and trans-carveol than Scolch spearmint oil. However, because it is difficult 10
look at raw analytical data and readily see slight differences, the use of
.i
,
'.
','
I
'I

I
,"
'J
,
-
140 SI' ICt;S, H A VOR CIIEMISTRY ANU ANTIOXIOANT l'IW Pt;RTl ES
Table I. CO"' IIOnl'; lI l Ral lo) Selected for Peppermin t Oi l OifTt rt nl ili l ion
According 10 Origin
I. 1,8CineoleILi monene
2. 1,8CincoldM:enthofu ran x 112
J. 1,8Cineole/Menlhyl AcelBle
4. 1,8 CineoleIMenlhol x 50
5. 1,8 CineoleIMenthone x 25
6. MenlhoruranIMenthOllC x 100/6
7. Ment hofuran/Mentho] x SO
8. Meru hofuranlMcnt hyl Acet ate x 5
9. McnthofuranILi monene
10. MenthoneiMcnlhyl Acetate x 112
II . Menthol/Menthane
! 2. Mellthol/Ment hyl Acetate x 112
,
12
Midwest
Fi gUi' CS 1-5 I'atter n Recognitioll s of tht Five Types of Peppermint Oil

12. SII U & LAWRF",,"CE
Vtl riations in Composition of Some OiLf 141
7
4
12
2
Wilamette
Figure 2.
7
,
Madras
fo' jgure 3 .
.
142 SJ'JCf;S: HAVOR CII EMISTRY AND ANTIOXIDANT !'U.O\'ERTIES
7
12
9
Kennewick
Figure 4.
4
12
Idaho
Figure S.
12. SIIU & Lo\WRENCE Varia/iolls in Composition of Some w 'rn/iot Oils
Table n. Componenl R:ui os Ustd to Differentiale between !\'Iidweslll nd
Farwesl Native Spearmint Oils !Iud Scotch Spellrln int Oils

I . LimoncneIMyrcentx In
2. Limoneneli ,8Cincoie x 112
3. i.irnonenel) Qc\aool x 115
Limonenelp.Bourbonene x lIS
MyrceneJj3.Bourbonene x 2
I ,8-Cineolc/Myrcene x 5
CaIVone/Myrcc!!c x 1110
CarvonelJ3-Bourbonene x 1/ 10
3.OctanollJ-0cIyl acetate x 112
4.
5.
6.
7
8
9.
10.
II
12
I ,8-Cineo]eltrans.Sabi nene hydrate x 112
cisCarveoVlrans-Carveol x 1/ 10
Carvone/trans-Carveol x 11150
6
"
"
2 "
3
..
9
' 12
Midwest
Native
Speannint
Figum 6-JO " I Uern Recognitions oflhe Five Types ofSpcllrmill1 Oil
143
144
srICES: HAVOM CIIEMJSTRY AND Al'/TIOXII)ANT I'WQI'!:: H.T IF,S
Figure 7.
3
Figure 8.
Far West
Native
Spearmint
Midwest
Scotch
Spearmint
12. SII U & LA .... 'RENCE Jlariatiotu in Composition of Smne Esselltial Oils
Figure: 9.
3
I-igure 10.
Far West
Scotch
Spearmint
2nd Cutting
Far West
Native
Spearmint
i
145
-
146 5I'ICES: FlAVOR CII EMIS1'RV AND ANTIOXIDANT PROPERTU:S
component rat ios magnifies these differences so that they can both be readily
discernible and can be easily displayed pictorially.Examination of the
limoneneimyrcene ratios for each different speanninl oil , we have 1.81-3_39
(Midwest Native), 0.88-2.68 (Farwest Native), 1.60-2.02 (2"" cutting FarweSI
Native), and 6.99-9,27 (Midwest Scotch), 8. 12-1 1.48 (Farwest Scotch). From
these data it can be seen that it is easy to different iate Native spcannint oil from
Scotch speannint oil. Using the eleven other component ratios merely makes
Ihis differentiat ion task easier. When Ihis data is presented as a polygonal
representation. this differentiation is readily evident from these pattern
recognition figures (Figure 6-10) that:
(a) Native and Scotch spearmint oils can be readily differentiated.
(b) Farwest and Midwest Scotch and Native spearmi nt oil can be
readily differentiated.
(c) A second cutting of oil ofFarwest Native spearmint can be
readily differentiated from all orthe other spearmint oils.
The Effect oflnlerspecifi c and Infraspecific Differences on Oil
Composition.
In the previous above example, it was shown that two spearmint oils, which
were produced from don.aUy reproduced distinctly different species of bkmhA.
produced oils whose compositions were similar but could be different iated;
however, the detailed oil composit ional data were not presented. To address
this point, the analysis of oils that are known commercially as cedarwood will be
the next subject of discussion.
Around the worl d a number of oils are produced that are described
as being cedarwood oil. Each of these oils is produced from a different species
as can be seen in Table m. The oils produced from Atlas cedar. t-limalayan
cedar, Incense cedar, Lebanon cedar, East African cedar, Oregon cedar, Mulanj e
cedar, Alaskan cedar, Eastern white cedar and Western red cedar are known to
possess different chemical compositions. The maj or components of these oils
can be seen in Table IV. 10 contrast, the Texas, Virginia, Chinese, and Atlantic
cedarwood oils, although produced from different species, all possess
composi tions thaI are qui le similar (Table V) Using the hypothesis that it is
easier to look II component rat ios rather than raw data, the components that
show most variation between the four cedarwood oils were a-cedrene,
thujopsene, j}- himachalene and cedro!. As a result, the component ratios were
examined rather than the raw data (Table VI), A comparison between these data
reveals that Virginia cedarwood oil can be clearly differentiated by the
thujopseneia-cedrene and j}-himachaleneicuparene ratios. It is more difficul t to
differentiate Texas li nd Chinese cedarwood; however, from the
thujopsene/cedrol ratios it is possible. Finally Atlantic cedarwood oil possesses
a unique j}-himachalene/cuparene ratio, thereby different iating it from the other
oi ls.
To !Unher discuss this qUes1ion of int erspecific differences in
composition. lemongrass oil wi ll be used as II second example. There are two
forms of lemongrass oil produced commercially, one produced in Asia from
Cymbopo8QO 0txU0SU$ (Nees ex Steud.) Wats. and the other produced in the
1.2. SHU & IAWRENCJo:
Vllriat;ons in Composition of Some Oiu
147
Tabl e In. Various "C,dan'" that have heeD 1I .'ed 19 produce ceda rwood oi l
I Virginia Cedar Juniperus yirginiaoll L.
2. Texas or Mexican Cedar Juniperus ashej Buchholz
3. Western Red Cedar Thuja pljClIllI Donn ex. D. Don
4. Northern or Eastern White Cedar Wa occidentalis L,
5, Kenyan or East African Cedar Hochst ex. Endl
6. Chinese Cedar Chamaecyparjs funebris (End!,) Franco
7.
8.
9.
10.
II
12.
IJ
" 15
Alaskan Cedar
Oregon Cedar
Mulanje Cedar
Atlas Cedar
Lebanon Cedar
Himalayan Cedar
Atlantic Cedar
Incense Cedar
(syn. CUPfeSSUS funebris End!.)
Chamaecyparis nootkatensi s (D. Don)
Spach
Chamaecvparis lawsoniana (Andr
Murray)
Widdringtonia liupressoi des ( L ) End!.
Cedrus a!lal11ica G. Manett
Cedrus libani A. Ri ch.
Cedrus deodara (Roxb.) Loud.
CbAtllilecypads tbYQj des (L) B,S.P.
l .jbQccdrus decurreos Torrey
Table IV.
!\bjor Coml,Qnents in Vari ous CcdltI"WQod Oil s
3.
4.
5.
7.
8.
9.
met hyl thujale 10. a-himachalene
T-muurolol j}_hi machalene
tropolooes l" -hi mllchalene
occidentatol
occidol
cedro!
l"-cadinene
I ( IO)-cadinen-413-ol
o-cadi nene
a-pinene
borneol
thuj opsene
cedrol
B-himachalene
II .
12.
14.
a-torosol

(E)_allant on_6.0!
(Z}-a-atlantone
(E)-o.-atlanlone
p. tnethoxythymol
carvacrol
tropol ones
I
!
I
..
r
.'
'i
14. SI'ICES: HAVOI{ CII EMI!t-rRY ANt) ANTIOXIDANT I'ROI'I-; RTmS
T" blt'. V. J\lflin Comll oncnlS of Four Crd!!nvood Oil s
C(l nJpolJ ud V T C A
-cedrene 38_0 226 33.8 20.5
j3-cedrene 9.2 5.5
' . 1 5.0
thujopseoe 23.4
".
360 31.7
cx.himachalenc 0.' 07
j)-chamigrene 14 L5
(3-sclincne
0.'
p. himachalenc 2. 1 II 0.7 0.2
a.-charnigrcne L5 1.6
ar-Curcumene trace 0.3
cuparene
0.'
I.,
22 J5
cedrol 12_3 122 85 lOA
widdrol
I.'
1.1 I I 14
a-bi sabolol Irace trace Irace 0.)
Ccdanvood Oil s: V . Virginia
C - Chinese
T - Texas A - Atlantic
Table VI. Component Rat ios Used to Differentiale Between Th ujoluene-rich
Ced:uv.ood Oi ls
Compound Ratios V T C A
tllujopseneJa-cedrene 2 - 3 - 9
4 _ 5
6.5 - 7.5
thujopsene/cedrol 1.7-21 3.6 - 4.0 4.0 - 4.4 2.9 - 3. 1
13- 2.2 - 2,1; Q - Q 7 Q 2: - Q 4 001 - Q, I
Cedarwood Oi ls: V Virginia C Chinese
,.
T - Texas A " Atlantic
nhle VII. Chemical Composition of t;, Indian Lcmongran Oi l
Comtlound ludiau I Indi HlI II I1hytanese
Il.-thujene trace 2.8
Il.-pi nenl;: 0.4 5.4
camphene 0.5 1.0 10. 1
myrcenc Irace 10.3
limonene 2.7 3.8 ' .3
4-nonanone 0.5 0.3 trace
6-mel hyl- 5 -hepten-2 -one 2.6 2.5 32
citronellal O.
0.' 08
linalool 14 I I 2.1
Il-caryophyllene 13
0' 0.4
neral 32.6 34. 1 18.5
geranial 40.8 44.3 '1.2.7
geran)'1 acetate
3.' 2.2 3.4
cit ronellol 0.5 0.6 0.6
geraniol 4.2 )8
22
12. SHU &: LA''VIU:N<':E VariatioftJ in Compo.vitioR of ScIM Esuntial Oils
New World from CyroboPOiQO cjlratus (DC) Stapf Authentic oi ls obtained
from each species were subjected 10 analysis and by GC/MS. A summary o r lhe
differences betwun their compositions can be seen in Table VII and Vii t
Although the differences between these oils appear to be major be<:ause of the
high limonene content of the West Indian oils, il is easier and more prudent to
use component ratios to differentiate between the oil as shown in Table IX. It
is believed thai examination of the hydrocarbons and oxygenated compounds
separately is II good way to examine Ihe authenticity of an oil.
The existence of morphologically identical plants which possess oils
that have differing chemical compositions is not a new concept . In fact, the
existence of infraspecific differences is widespread in the Labiatae and
Compositae families, but it is IIOt limited to them. A few years ago Lawrence (2)
showed that Qcimum basjljcum L. can contain oils thai possess a variety of
compositions. It was also found that oils contained constituents that were
biosynthesized either via the shikimic acid pathway, or the mevalonic acid
pathway, or both (Figure I I ). A summary of the data obtained during the
analysis of more than 200 separate Q basilicym plants revealed that they
possessed oils that, IIOt only contained components from single pathways or dual
pathways, but also wi thin these groupings a wide quantitative variation of
constituents was also observed. A summary of these dala can be seen in Tables
X and Xl .
The Effect of Growth Stage of the lJar"l'uted Plant on the Composition of
Its Oil.
Although the effect of growth stage of the plant on oil composition should be
well known, oils produced commercially are not always produced year after year
from plants that were harvested at their same growth stage. Two examples will
be used to illustrate this point.
First, oils produced from Coriaodrum satjyym L differ chemically quite
drastically (3) as can be seen from the dala presented in Table XII. The oils that
are known as cilantro oil are generally those produced from plants at stages I
and 2; however, the oi l yield is higher the later the plant is harvested.
Nevertheless, using the ratio between decanal & (E)-2-decenat and linalool and
(E)-dodecenal , a reproducible cilantro oil can be obtained (Table Xl1I) .
A second example of a change in chemical composition as the plant
matures can be found with the oil ofTagetc:s mjnula L. (4). E:tamination of the
main acyclic hydrocarbon and the acyclic ketones (Table XIV and Figure 12) will
allow the user to select an oil that is most liked. From these data, component
ratios can be used as a quality check for oil reproducibility. The reason why oils
can fall outside the desired component ratio can be understood based on the data
presented in Table XlV.
T he Erred of Processing Parameters on Oi l Compositi on
In 1985, Schmaus and Kubeczka (5) examined the influence of oil isolation
condit ions on the composition oflinalyl acetate-rich oils. They found that the
increased water content in plants distilled when fresh, resulted in a corresponding
II
14'
ISO SI'I CES: n AVOR CIIEI\USTRY AND ANl'I QXIIlANT PROPERTIES
Table Y in.
COlUlJ OU lid
Chemic!! 1 Composi li on ofW. lndi an umongrll$$ Oil
Gualtmllin I Guatemala n
limonene
4-nonene
6-methyl-5-heplen-2-one
citronella!
linaloo1
/3-caryophyllene
!leral
geranial
gerany! acetate
citronellol
geraniol
12.3 15.9
1.8 2.4
3.0 44
0.' 0 .
1.8 2. 1
2.2 1.3
30.5 28.6
37.3 37.1
L1 1.1
04 0.2
46 26
Table IX. Componerll Ratios Used to Differentiate between E. Indian lind
W. Indi a n J.emongrass Oil
CllmpOnt ntlRfi tio E. Indi lln
methyl heptenonel4-nonanone 5.0 - 9.5
geraniollgeranyl acetate 1.2 - 1.8
citronel1allcit ronel1ol \ . 1 - 1.7
neralllimonene 8 5 - 12 5
W, India n
I 0 - 2.0
2.2 - 3.0
2.0 - 3.2
! 5 - 3.0
Table X. Major Components of Various OC;lIIum bm"ilicum Selections
Rios!nthetk Pl!thwfI:\:s)
Compound I(a)
Hbl 2{a) 2(b) 3(/1) 4(a)
1.8-cineole 3.2 4 . 1.4 2.7 2.4 3.4
(Z)-p-ocimene 0. 1 0.3 0.2 0.1 tracc Irace
(E)-p-ocimene 0 . 5.4 1. 0 2.2 0.4 Irace
camphor 0.2 1. 8 OJ 4.3 0.' 1.8
linalool 0.5 4.8 83. 51.6 0 . . .2
methyl chavico! 85.3 49.0 0.4 0.7 1.5 5.0
geraniol Irace trace 0.3 22.8 Irace
,""
methyl eugenol I.. \2.2 trace trace 67.7 0)
(E)-methyl cinnamate Irace trace trace trace 03 640
I[l!et Q,Z Q.l trace
ZO
Legend:Chemolypes of OcimulII basilicllnI wi th a Single Bi osynlhetic Pathway:
Type! (.) met hyl chavicol-rich
(b) methyl chavicol>lI1cthyl eugenol
Type 2_ (.) linalool-rich
(b) tinalool>geraniol
Type 3_ (.) methyl eugenol-rich
Type 4. (.) methyl cinnamate-rich
Type 5. (.) eugenol-rich (not found)
12. SHU & lAWRENCE Variatioll\' in Compm'ition of Some El'senlial Oi/.f 151
Tll ble XI. Major Components of Vari ous Odmum bllSifi(.uIII Sd cctions
(Dual Biosyn lhelic PalhwllYs)
Compound He) IIC)' zee)
1,8-cineole 5.7 3.8 4.8
(z)-I}-ocimene 0. 1
(E)-I)-octmene 2.0
camphor 1.5
linalool 22,9
lCc)'
I..
0 .
1.0
0.2
51.0
2(dl
3.8
'4
2.5
4.'
30.2
methyl chavicol 52.1 21 0 2.1
geraniol trace 0.2 0.4
methyl eugenol 0. 1 0 I 10.2
Z(e)
2.8
0.2
1.0
LS
34.2
1.2
0. 2
0. 1
Ji b)
7.7
0. 1
3. 1
0.8
21.2
0.7
Irace
38.4
(E)-methyl cinnamaletrace trace trace 27.8 trace
eugenol trace lrace 37 139 148 42 5 I
Legend: Chemotypes or OcinllIm basilicum with a Dual Biosynthetic Pat hway:
Type I. (c) methyl chavicol>li nalool
(c)'melhyl chavicol>linalool>methyl cinnamate
Type 2. (c) linal ool>mcthyl chavicol
(c) 'Iinalool >methyl chavicol>eugenoJ
(d) linaloot>eugenoJ>methy\ eugenol
.... (e) linalool >methyl cinnamate
Type 3. (b) mcthyl eugenol>linaloo\
Ta ble XII. Conlpa r ative chern ital co mposi ti on of Cor;fIIltlru m snt;I'uIII 1_ at
Vari ous Sugts 2fMaturit!
Stages of Plant Maturity2
!::ol11 lH!!!nd I
2 3 4 5 6
Octaoal 1.20 1.20 0.85 0.66 0.44 0.35
Nonanal 0.5\ 020 0. 11 0.05 0.05 0.08
Oecanal 30.0 18.09 11.91 6.30 6.24 1.61
Camphor 0.08 Irace 0_52 1.26 2,18 2.44
(E)-2-0ecenat 20.6 46.5 46.5 40.6 30.2 3.9
Oodeeana!
3.)0 ,.7 0.26 064 0.52 0.41
(E)-2-Undeeenal 2.56 2 1.7
13'
Tridecanal 3.07 J.87 2.02 0.92 1.08 0.46
(E)-2-Dodecanal 7.63 8. 14 5.95 4.59 4,78 2.19
Telradecanal 0.68 0.30 0, 12 0. 15 O. I! 0. 15
(E)-2.Tridecena! 0.42 0.21 0. 14 0.09 009 O. i3
(E)-2-Tet radecenal 4.45 2.57 1.73 1.53
15'
1. 73
Linalool 0.34 4.27 17.47 30.05 40.88 60.37
Geraniol 0. 12 0. 1 I 0.35 0.7 1 0.93 1.42
! 11 Q Q 16
0,69 Qli:2 Q 26
l Stages of maturity: 1 - noral initiation; 2 - nearly full nowering; 3 .. full fl owering,
primary umbel young green frui t; 4 '"' past full fl owering 50010 nower, SOOIo fruit ;
5 - full green fruit; 6" brown fruit on lower umbels, green fruit on upper umbels:
-- - nOI detected
I
I.
.!
"
1
.,
,


. /
[52 SI' ICES, H.AVOf{ ANU ANTIOXII)ANT j'ROl'ERTIES
Table XUI. Component Rat io! of Ci llUll ro Oil Consti luenlS as II I1I C11suremt ll 1
of oil gualily
Componen t Rll iio Siages of Matu rit y
,
3

5

dccanaV(E)-2-dccanal 1.5 OA 0.3 0.2 0.2 OA
lill1\loo!/(El2-dodecenal O,Q Q,52 294 6.55 8,55 24,J
Table XI V. Chll. nge in per(cnl nge composil ion of T(lgclc$ mi nI/til oi l dur ing
m, S;xs:I, of pl nnt
Development Dihyd ro- (Z)- (E)- (Z)- (E)-
t:li:;etone Tnl:,etone TIIKetenoue
vcgetativc 16.9 46.4 22.4 3.2 L5 2.0
buds visiblc 11.8 51 .3 18.5 14 44 3.0
buds opening 21.9 36.0 24. 1 LJ 6.9 L5
!lowers opening 33.3 30.0 17.5 20
, .
1. 9
full flower 35.3 30.5 16.9 22 ' .9
" immature fruit 41.3 24.1
' .2
03 18.2 2.0
fruil 2 148 04 20,4
J2
Table XV. Compllrllt ive ChemiclI l COlli position of French a nd Russian
Clllry Sage Qib
Compound Frt nch Oi ls Russian Oil . ,
A {dried} n {fresh}
a -pinene 0.20-0.21 0.08 0.02
p-pinene 0.34 - 0.J8 0.98 0.22
limoncne + 1,8.cineole 0. 19-0.20 0.29 0.41
(Z)-I}-oeimcne 0. 16 0.40 0.09
(E)-I}..()Ci mene 0. 2 1-0. 23 0. 72 0.06
linalool 8888.98 28.48 14.43
a-terpineol 0.20 - 0.21 2.64 2. 11
11<':1"01 0.050.06 0.56 0.42
geraniol 0. [ I 1.54 0.98
linalyl acetate 72.41-14. 18 50.86 63. 18
lI<.:ryl acetate 0.13-0. 17 0.75 1.03
1;;eranyl acetate 0.30 - 0.40 1.42 2.01
(3-caryophyllcne 1.85 1.89 1.65 202
Cl"1ll3Crene D 3.29 - 4.09 3.54 0.36
caryophyllene oxide 0.35040 0.16 0. 16
121-1.27 0.91 1.812
12. SHU &
Varia/imll' in Compru'itiorl of Some Esnmlial Oils
increase in distillation time and a decrease in pH causing partial hydrolysis of
linaly! aCdate followed by partial acid catalyzed degradation orlinalool result ing
in an increase of (E)- and (Z).p-otimene, limoncne, terpinolene, a-terpineol,
geraniol. ncryl acetate, and geranyl aeclale. In 1985, Dore and Jaubert (6)
presented an example ofthesc effects as it applies to clary sage oil (Table
Similar results were presented by lawrence (7) (Table XVI) where the U.S. 011
was produced from fresh plam material, while the other two oils were produced
from dried material. The effects described by Schmaus and Kubeczka (5) are
very evident in these results. A processing parameter that needs to be mentioned
is storage. If an oil is stored under poor conditions such as in a halffilled drum
in which some dissolved water has deposited causing some minute rusting in the
drum, oxidation of the oil can and does occur. 1\ self-explanatory example of the
oxidation of coriander oil can be seen in Table XVII.
The EfTect of Ad ult era ti on on the Composition or a n Oil
Two examples of oil adulteration will be discussed namely, lemongrass oil and
peppermint oil. Examination of the data presented in Table XVIll shows the
analysis of three commercial samples oflcmongrass oil. Sample Nos. 2 and 3 are
grossly adulterated because they were found to contain ca. 20% of either
diacetone alcohol or triacctin as an adulterant . In contrast, Sample No. I has to
be examined from the standpoint of component rati os. A comparison between
the component ratios ofE. Indian, W. Jndian lemongrass oi l and Sample No. I
(Table XIX) reveals thai the oil is not a troe example of either lemongrass oil;
therefore, it must be adulterated.
Over the past few years the production of Indian peppennint oil has
increased. both in the Punjab area as well as in the Dareilly region of Uttar
Pradesh. As this oil is substantially lower in price than North American
peppermint oil, an economic temptation exists 10 mix Indian peppermint oil with
one oftne U.S. peppermint oils and sell it as U.S. oil. I\s a result, it was decided
to determine if this practice al the 10'10 adulteration level could be readi ly
detected.
[53
Initially, 18 oils from both regions were analyzed by Ge/MS 10 ensure
that there were TIQ unusual constituents Ihat would interfere with further studies.
The component ralios shown in Tabl e I were calculated, and these were plotted
on a circular graph to show the typical pattern of an Indian peppermint oil
(Figure 13). From the individual Indian peppermint oil samples, a composite
was made, and it was added ala \0% inclusion rale into a composite oil of the
Midwest, Idaho, Kennewick, Madras, and Willamette peppermint oils that were
analyzed earlier. These adulterated oils were then analyzed by GC and the
important peak identities were confirmed by GClMS. The component ratios of
each adulterated oil were calculated and .",ere plotted against the original pattern
recognition polygonal representations of the oils from each region oil (Figures
14-18).
I\s can be seen, the data plot for the U.S. oi l containing 10% of the
Indian composite oil fell outside of lhe polygons, thereby confinning that each
oil was adulterated.
I
IS4
SI'ICES: H AVOR ClIEI'tIl:nRY AND ANTIOXI OANT
SHIKIMIC ACID
PATHWAY
VIA
PHENYLALANINE
MEVALONIC ACIO
PATHWAY
VIA
GERANYL PYROPHOSPHATE
_ CHAVICOL 1
l:
I ----UNALOOL
METHYL I
I 7 GERANIOL
EUGENOL
CINNAMATE
' igure 11
OUIII Biosyntheti c Pathway Found in Ocimum JJilsificum
(Z)- ( E)-
tll getenones or J!-oclmenones
(Z)- (E)-
tagetones or 2.3_dlhYdro_p_odmenones
dlhydrot agetone
Figure 12 Tagetes Oil Ketones
12. Sil l! & LAWRENCE Variations ;/1 Composition of Some &sential Oils IS5
Table XVI. ComQllralive Chemicll l Composi ti on of Commtrical Cl:u), Sage Oil
Compound
U,S. Oil Frc lI ( h Oil Russia n Oil
a-pinene
myrcene
limonene
(Z)Il-ocimene
(E). p-ocimene
linalool
0.17-0.22 0. 1-0.3 0.2-0.3
1.25-1.71 0.1-0.2 0.3-0.5
040-0.77 0.1-0.2 01-0.2
linalyl acetate
p-caryophyllcne
a . terpi neol
germacrene 0
nery! acetate
geranyl acetate
ocrol
0.42 - 0.70
0.43 - 1.36
20.29 - 28.6]
44.9-S34
0.86 - 1.3 1
1. 05 - ] .05
2.6] - 3.55
1.00 - 1. 67
19]-].24
0.62 - LI S
1.67-326 geraniol
caryophyllene oxide
sclacco\
0.2 1 - 0.27
017-044
"''''
0. 1 - 0.2
9.0 - 16.0
49.0 - 73.6
\.4 - 1.6
02 - 0.6
\.6-20
0.2 - 0.]
0.]-05
trace - 0 I
0. 1 - 0,]
0.3 - 0.5
0 1 - 0 2
trace - 0.2
0. 1 -0.4
10.4 - 19.]
45.]-61.8
1.1 - 1.8
1.2 - 2. 5
0.7 - 2.0
0.4 - 0.6
0.8 - 1.2
0.] - 0.5
0.6 - 1.2
0.5
0. 1 - 0.2
Tahl e xvn. The Effeci of Stnrfli!.e (Ot ici alion) on Coriander Oil
Compound
p-cymene
cis-linalool oxide (furan)
trans-linalool uxide (furan)
camphor
linalool
4-met hyl-5-hexen-4-olide
cis-linalool oxide (pyran)
trans-linalool oxide (pyran)

Good Oil Oxidized Oi l
1.7 -], 8 1.9
0.2 - 0 5
0.2-04
2. 1 - 4 4
70.3 - 77.1
14. 1
12.]
63
]8.1
3. 1
1.5
14
06
Tll ble XVIl!. ChemiclIl Comllosi ti on of ,\duhtra ted Lt mongrass Oil
Compound I 2 3
myrcene 1.8
limonene
4-nonanone
6-met hyl-4-heptenone 3,4
diacctone alcohol
triacetin
citronella! 0.6
linalool 2.9
I}-caryophyl lene 0.5
neral 36,2
geranial 45.5
geran)'l acetate 0.7
citronellol 0.3
geraniol 26
trace Irace
I.'
1.6
0.6 0.4
1 9 1. 6
21.3
25.8
0.5 0.5
J.l 1.5
1.0 0.9
26.3 24. 1
32.7 J 1.0
4. 1 3.8
0.3 0.3
6,0 56

i
j
I
>
}.
'.
,
t
i
i
"
I
t I
1
J
.,
.,
1
i
I
156 FlAVOR CJl EMISTI{Y AND ANTIOXIDANT I'IWI' ERTIF.5
Table X[X.Componen t R:lIi os Used to Differenliate Bel ween 1<:. indian
and W. Ind ian Lelllollerass Oil lllnt an Adult rated Oil (Sample No. L)
Compollent Rat io E. lndian W. lndia n Samllh: No.1
methyl heptenoncl4-nonanone
geraniol/geranyl acetate
citronellal/citronello1
n"rall1imOIlCIlC
6
5
1
2
5.0 - 9.5
1.2 - 1.8
L1 - I. 7
8.5 - 125
3
1.0 - 2.0
2.2 - 3.0
2.0 - 3.2
L5 - 3 0
9
10
12
11
India
3.7
2.0
Figure 13 Pattern Recogn iti on or all Indian l'cpperl11in t Oi l
12. SHU &
Variations in Composition of Some Es.wntial Oils
Midwest
3
Figures 14-18 Pattern \{ecognit ions or the Fi ve Adulterated Pellperminl
Oi ls

Willamette
3
Figure 15.
157
15'
SPICES, FlAVOR CHEMISTRY AND ANTI OXIDANT ,'ROI'ERTIES
Madras
3
Figure 16.
12
Kennewick
Figurt 17.

12. SHU & I.AWRENCE
Yar;aliollS in Compositiot1 of Sorm &selltial Oils
Idaho
3
<
Literature Cited
I. Lawrence, D, M,; Shu, C-K.; Hams, W. R. Per/11m. Flav. 1989, 14(6), pp.
21-30.
2. Lawrence, B. M. In Flavors alld Fmgrallcts: A World Perspective.
Lawrence. B. M.; Mookerjee, B. D.; Willis, B. 1., Eds., Elsevier SeL Pub!.,
Amsterdam, 1988, pp. 161-169.
3. Lawrence, B. M. In New CropS; Janick, 1.; Simon, J. E., Eds . 1. Wiley
& Sons, New York, 1993, pp. 620-627.
4. Lawrence, B. M Progress ill Essential Oils. 1992. / 7(J). pp. 131-133.
5. Schmaus, G.; Kubeczka, K. H. In Essential Oils and Aromatic Plams.
Baerheim Svendsen, A.; Scheffer, 1. 1. C., Eds., Maninus NijhoffJDr. W. Jank
Publishers, Dordrecht, Netherlands, 1985 pp. 127-134.
6. D O T l ~ , D.; Jauben, J-N, Par/11m. Casmef. Ar6me.r. 1985,61, pp_ 7985.
7. Lawrence, B. M. In Advo/J(:t:s ill LAbiatae Science; Harley, R. M.,
Reynolds, T., Eds.; p, Royal Botanic Gardens Kew, 1992, pp. 399-436.
15,}
Chapter 13
Component Analyses of Mixed Spices
C. K. Cheng
l
, C. C. Che n
l
, W. Y. Shu
l
, L L. Shih l,
a nd U. II . Feng
l
Lrood Industry Reseatl,:h a nd Developme nt Institute, P.O. Box 246,
Ihinchu 300, Ta iwa n, Republic of Chin a
l lnstitute of St a tisti cs, National Tsing- Hua Uni\'crsity,
lIs inchu 300, Tniwan, Rc public or China
A qualitative and quantitative technique for recogniling spices as
well as the mixi ng ratios in an unknown spice blend is established
via numerical analysis based on a database consisting of 355
difTerent spices. Formulating the spice recognition problem as a
constrained opt imitation problem is a key step. A simi larity index
measuring the degree of similarity between two spices is proposed,
which is very useful in the process of spice recognition. For
purposes of testing, an existing mixture (basi l, 47.9%; clove,
12.8%; mint, 9.4%; sage, 23.8%; wintergreen. 6. 1%) and a trial
spice blend with known composition were analYled by the .<
techni que. The results are surprisi ngly accurate.
Humans have been consuming spices and spice blends for several thousand years.
Even to this day, spices and spice blends are highly valued commodities from all
over the world (J - 2).
From the flavor point of view, the major attributes of most spices are
contribut ed by essential oils. These can be extracted by steam distillation (or
similar process) of ground With the advent of gas chromatography (GC)
and the combined technique of GCI mass spectrometry (MS), analyses and
identifications of volatile components of essential oils extracted from different
spices have been very thoroughly and well documented (3-5). However, most
of the published reports were based on the analysis of an individual spice. When
it comes to food processing, usually it is a spice blend instead of thc individual
spice which has been added to the fonnulation.
Cl t997 Amcrican Chcmia] Socicly
13. ET AI..
ComfHJmmt Analps of Mird Spices
Detection ofan individual spice in a spice blend can be achieved through
sensory analysis, instrumental analysis, or the combination of both. When
analyz.cd by a sensory method, which requires trained panelists, only quali tative
result s can be obtained (6). flor semi-quantitative analysis, there have been two
types of approaches to invest igate the content of an individual spice in a spice
blend. One is based on calculation of major components in a mixture as
compared to published results. A good example has been dcmonstrated by
Lawrence and Shu (4). The more advanced approach is based on a computer
pattern recognition method. Chien (7) has shown a very genius mathematical
method to analyze the presence of an individual essential oil in a mixture. In that
report, four out of six essential oils with contents ranging from 5% (geranium,
cedar wood, patchouli) to 15% (bergamot) were correctly identi fi ed. The other
two oils, galbanum and Dourgeon de Cassis, were not picked up by the computer
because of their low level (0.25%).
This paper presents a qualitative and quantitative analysis of an individual
spice in a spice blend. The ratios of each spice were calculated by using a
computerized numerical analysis wi th a data matrix composed orJ55 spices and
accumulated total of922 volatile compounds. The principal part oflhis method
is the solving of a constrained optimil.3tion problem.
Th w ry
161
Thc problem of rC(;ogniling the individual spice as well as the mixing ratios in
a spice blend is first fonnulated as a constrained optimization problem. We then
solve the problem by numerical methods. The procedure is dcscribed below.
Vector r cprcsenta lion or spi ces. Each spice can be characterized by its
components and composing ratios. Mathematically. each spice can be treated as
a vector X=(x/.x
1o
".,xJ. where x, represents the ratio of the I-I h compound in
that specific spice. In the present study. 355 different kinds of spices were
coll ected and 922 volatile compounds (n=922) were counted among these 355
spices. Therefore. n is equal to 922 and x/+x
2
+ .. +xm" l. The 355 spices were
represented by 355 vectors ... XJl7 Using Anise # I collected in the present
study as an example, this is shown in Table t .
In Table I, each code number represents a single volat ile compound
identified in the essential oil. e.g., code no. 236 represents estragole, code no.
162
n AVOH CIIJ.;;\USTRY AND ANTIOXIDANT " ROI't; RTIES
Table I . The volat il e compounds o r anise #1.
Code No. of Ratio el.)
compounds
"6 2.513
187 1,487
236 4.975
250 0,121
292 0.312
29J 0.221
'"
0010
454 2.382
511 0.784
5J2 0.030
540 86.683
666 0.22 1
667 o 76J
Som 100
13. CII"NG);'f COmpollt 1lt AI/(llyses of Mi:ad Spit'S 163
Equations 3-5 indicate thaI if spice blend X is compared with itself, then
JlIlier(J(X)=I , if component s in spice blend X can not be found in Yand \lice versa,
then '"dex(X.I')-o, otherwise, the index shall be ranged between I and 0 Therefore,
Jlldex(X, Y} can be used to measure the similarity between X and Y.
Simi lari ty index has ils rnalhematical meaning. Any spice vector can be
considered 15 a probability distribution on the sel {I.Z, .... ,922}. The well kllown I
squared Hellinger dist ance between two probability distributions X and Y is defined I
"
2 911
'jj E: - 50) = 2 -2 I E:5o = 2[1-lndex(X, y)], (6)
,.,
It is very clear that the larger the bldex(X, t) is, the smaller the Hellinger distance
bctween X and Y wi!! be. Therefore Iml ex(X.t) directly refl ects Their similarit y. For
example. Anise#2 (9) and Anise #4(10) have many volatile compounds in common
(Table II) . The similarity index between these two spices is 0.964 indicat ing high
degree of similarity. On the contrary. when volatile compounds of Anise #2 are
'I
!
540 represents trans-anethole. A spice blend X ofindividual spices contained in compared wi th those of coriander # I (/I) as shown in Table III , an index value of 0
our collection can be viewed as a complex combination of vectors X,.X}o .... Xm: is obtained, indicating no similarity between these two spices.
where tJ i is the mixing ratio of the ;-th spice in the spice blend _ For a given
unknown spice blend Y, identiryi ng the spices contained in Y amounts to finding a
t;orwex combination X'" fj /X,+ fj :X}+ ... + fJ J.uXJJJ as "close" to Y as possible
Methods for measuring the ctoseness between spice blends X and Y wi ll be
discussed next.
Simil arity index For any two spice blends X=(x, .xz, " '''9}}). and Y=(y,.yZ, .. . Y9U),
the similarity index. denoted '"de.x(X.Y). is defined as:
'"
llldex( X ,Y) = I .r;;.;y;,
Basically. Inde.x(X,Y) has the following propenies.
For any two spice blends X and Y.
1.0:i l mlu:(X. y) I
2. If X- Y, l llllex(X, 17- 1
3. U cOlll llo nt nls in X CAn not bf fo und in Ya nd vice Yfrs a, thf n
ltUlt! I (X, Y)=O.
(2)
(3)
(4)
(5)
Table 1[. Comparison of the volatile compositions of anise #2 and star
#4. The similarity index " 0.964
Codo: No. Rallo Code No. Ratio
Anise'2 SUir Anise .2 Star anise u
20 000000 0.00081 436 000000 0.00081
40 0.00000 0.00092 437 0.00010 0.00581
60 0.00000 0.00 102 440 000000 0.00112
IS. 0,00567 0.04563 450 0.00000 0.00051
236 0.00304 0,00346 454 0.00000 0,00112
257 000000 0.00509 467 0.00010 0.00051
292 0.00608 0.02954 480 0.00365 0.00428
29J 0.00263 0.00428 499 0.00020 0.00346
lOJ 0.00000 0.00092 51' 0.00010 0.00000
331 0.00000 0.0006 1 531 0,00000 0.01039
315 0.00000 0.00 122 512 0,00010 0.00183
175 0.00010 0.0005 1 540 0.97376 0.87044
386 0.00020 0.00387 552 0.00000 0.0014)
38'
0.00010 0.00000 666 0,00405 0.00000
425 0.00000 0.()()()41 745 0.00010 0.00000
164
SPICIiS: FlAVOR Cn EMISTRY AND ANTIOXIIJANT 1'R:Oi'ERl'll':S
Table 111 Comparison of the volat ile compositions of anise 112 and coriander
III . The simi larit:r: index - O.
Code No. Ra1io Code No. Ratio
Anise 112 Coriander H\ Anise #2 Coriander
' I
,
0.00000 0.465 19 211 0.00000 0.00505
7 0.00000 0,09284 231 000000 0.01 6 15
,
0.00000 0. 10394 236 0,00304 000000
II 0.00000 0.00303 267 000000 0.00303
IJ 0.00000 0.05853 292 0.00608 0.00000
14 0.00000 0.00706 293 0.00263 0.00000
J5 0.00000 0.0565 1 348 000000 0,00202
32 0.00000 0.00303 349 0.00000 0.00202
50 0.00000 0.04339 353 0.00000 0.00505
58 0.00000 0.00202 357 0.00000 0.00101
61 000000 0.00]0) 375 0.00010 0.00000
62 0.00000 0.00 101 386 0.00020 0.00000
64 0.00000 OJ)()202 388 0.00010 0.00000
65 0.00000 O.c)0202 390 0.00000 0.00706
"
0.00000 0.0010 1 399 0.00000 0.00 101
67 0.00000 0.0 1413 402 0.00000 0.00505
71 0.00000 0.00404 437 0.000\0 0.00000
83 0.00000 0.00706 467 0.00010 0.00000
87 0.00000 0.00605 480 0.00365 0.00000
9 1 0.00000 0,00908 499 0.00020 0.00000
"5
0.00000 0.00101
51'
0,00010 0,00000
143 0.00000 0.01312 5J2 0, 00010 0.00000
144 0.00000 0.00807 540 0.97376 0.00000
145 0.00000 0.00 101
'"
0.00405 0.00000
156 0.00567 0.00000 745 0.00010 0.00000
210 0.00000 0.04440
tJ. CHENG lIT AI. Component AnalJsts of Mind Spica
165
Numerical Analysis. Given an unknown spice blend Y, we consider the spice
recognition problem a.s that of finding the best rat io in wltich one can mi){ the 355
spices X',xl .. . ,xJ55 so as 10 achieve the maximal si milarity to the target spice
blend Y. This problem is e){actly equivalent to the following constrained opt imization
problem:
(
maximize S<P)
P
o (7)
SubjeCI lo p , + p,+ .... +Pm = 1 , , ~ ,
where .fl... ~ ( / 3 ,,/3 ]> ... , /3 JJJ) represents a set of mi ){ing rati os and the objective
function S is defined as:
(')
To determine the best ratio J!.... ' so as to achieve the ma){imal value of S(p), we
employ optimization theory and numeri cal methods (see Luenbcrger (8 a n ~
computer programs are designed accordingly. The nonzero component s of .Ii...
indicate the presence of the corresponding spices.
EXllerimenlal Seclion
Pupara tion of mixed spi ces for test ing. A testing mixture of spice blend was
prepared by mi){ing powdered spices purchased from local trading company. The
formulat ion is shown in Table IV .
Table IV. The formulation of mi){ed spices for testing.
Spices
Wt(g)
Bl ack pepper 27.00
All spice 27.02
Celery seed 13.52
Clove
1. 62
Garli c 27.01
Star anise 6 1. 03
Sum 157.20
Isolat ion of volatile component s from spi ces. One hundred grams of ground spice
blend shown in Table IV were added to a three-neck bottle containing 300ml of
disti lled water. Thi s slurry was steam distilled and extracted with dichloromethane
(Fisher Scientific) for 2 hours in an apparatus si milar to that described by Likens-
Nickerson(l2). The organic: layer was separated, dried over anhydrous sodium sulfate
(E Merck). evaporated 10 mini mal vol ume in a Vigreu){ column, and then
concentrated to about 0.5 ml under a gentle stream of nitrogen.
166 SYI CES: nAVOR ANI) ANTIOXIDANT I'ROI'F.RTIES
Gc. GC analyses were carried out on a Vari an 3400 chromatograph equipped
with a 30 m x 0.25 mm i,d. fused silica capillary column (DB-WAX, J&W
Scientifi c, USA). 1be linear flow rate of the carrier gas (H
1
) was 43 emfs. The
oven temperature was programmed from 40C to 200 C at a rate ofloC/min
with initial holding at 40C for 2 min. The injector and detector (FID)
temperatures were set aI200e and 210C, respectively.
GC/l\1S. GCIMS was carried out on a Finnigan Mal ITO. ac conditions were
the same as those mentioned aOOve. Identification of volatile compounds was
based on comparison of retention index and mass spectra with those of authentic
compounds.
Sel up of the data malril. The raw data were collected from published
Jitcrarure such as those cited by TNO-ClVO (3), ESO (5), or those listed in the
present study. Take essential oil ofbJack pepper (I J) as an example, the original
data format is shown in Table V, the volatile compounds were then arrangcd into
coded mlmher as those of Anise #1 shown in Table I. So far, a 355 (spices) by
922 (volatile compounds) data matrix has been establ ished. Of the 355 spices,
there exists repeated coll ection of data from the same spice from different
geological areas or differcnt publications. f or examples, 4 doves, 5 star anises,
9 anises, 6 corianders, 8 allspices, 2 black peppers li nd 27 basils, are included in
the present data matrix.
Results a nd Discussion
Test nm of known spice from published li terature.. The spice blends
which cont ain basil, cinnamon leaf, peppermint, sage and wintergreen as
reponed by Lawrence and Shu (4) were used to test the effectiveness of the
present tOcory. The ratios of indivi dual spice are shown in Table VI, the coded
I\llmbers of identified volatile compo1Jnds in the spice blends are shown in Table
Vll. Computer output of numerical analyses of compounds shown in Table vn
are li sted in Table VIII, the similarity index of the testing result is 0.959,

indicating a high level of confidence. In Table VI II, there arc repeated
identifications of the same type of spice, e.g. , the appearance of basi Is #3, #19,
1124,1126 and #27; cloves III and II 4; and mentha # 1 and II 12. In fact, the
repeated appearance of the same spice indicates the closeness of these spices.
It was also confirmed that the similarity index within the same spice listed in
Table VIII usually ranged from 1.0 to 0.90 (data not shown). In ordcr to
13. CIl t:NG t;'f '\1_ Component Analysts of Miud Spict.f
Compounds
%
Sabinene 22.50
alpha- Pinene 17.50
Limonene 17.00
del ta-3-Carene 13.50
beta-Caryophyllene 5.00
alpha- Humulene 1.00
Carvone 0. 10
Eugenol 0. 10
Linalool 0. 10
Myrist icin 0. 10
Nerolidol (unknown isonler) 0. 10
Pi peronal 0. 10
cis-Sabinene hydrate 0. 10
trans-Sabi nene hydrate 0.10
Safrole 0. 10
Terpinen-4-ol 0. 10
alpha-Terpineol 0. 10
TOIal 17 components 77.60
(Source: Adapted from ref . J J )
Table VI . Reported composition of test ing mixture of essential oi ls
Essential oils "I.
Exotic basi l
Cinnamon leaf
Peppermint
Sage
Wintergreen
42. 1
IS.8
15.8
21.1
' .2
(Source: Adapted from ref . .J)
167
,j


;i
,
1611 8I' 10 :S, HAV()U CHEMISTRY AND ANTIQXIUANT ,'ROI'fo; RTIt:s
Table VII . Volatile comEounds identified ITom Tabl e VI .
,
Code No. %
4J7 alpha-Pinene 1.3
182 Camphene Il
467 beta-Pinene 0.7
m Myrco.::nc 0.5
292 Limonenc \I
.0 l.lI-Cineolc 4.'
444 alpha-Thujonc , .
470 1x..1aThIUone I..
'"
Camphor 4.7
297 Mcnthonc 3.0
762 Mcnthofur.an 0.4
.-
293 Linalool 0.9
296 Menthol
4.'
23. Methyl ehavieol 34.7
177 IJornool L7
m Methyl salieylale
..,
197 Cinnamaldehyde
0. '
250 Eugenol 9A
25'
aectate OA
(Source: Adapted from ref . -I)
13. CIIENG "1' AI_ Compom",t Ana/y.Wl' 0/ Mi.ud Spica
Table VI1l . Ratios of essential oils IS analyzed by numerical analyses,
Ihe similari ty index. - Q 252
Spice II
Basil 113
Basil #19
Basil #24
Basil #27
Basil #26
Ray leafNI
Cinnamon N4
Clove #1
Clove #4
Commint #8
Mentha #1
Mentha #12
Roscmary #3
Sage #2
WinteriTeen 1# I
% Assisnmem
2.7393
2.6\87
2.8070
12.8562
18.7463
2.7733
0,4658
0.5672
10.6274
7.4766
0.5556
2.6013
2,7350
19,4858
59446
Basil #26
Clove #4
Commint #8
Sage #2
WinterSTeen #1
169
simplify tbe result, thc ralios of those spius and herbs which appeared repeatedly
were summed up and assigned to the spice with the highest ratio, as shown in the
third column of Table vm. Table IX shows the summed up ratios of spices
from Table VllI, as compared with tbose proposed by Lawrence and Shu (4).
In Table IX. clove 114 and commint 118 were not reported in tbe previous study
(.f), instead, cinnamon leaf and pcppennint. respectively, were assigned. When
the similarity index between these two seemingly differenl pairs of spices (clove
#4 vs. cinnamon leaf; commint #8 vs. peppermint) were compared, an index
value of 0.95 was observed in tbe former, indicating a high degree of closcness
(similar in composition of essent ial oil), the latter pair showed a less satisfactory
index valuc ofO.83, slillthis was a good matcb wi th acceptable confidence It
should be considered that the results in Table IX were from data in reference 4
that was used in the data matrill of the present study. The ratios of spices
analyzed by the numerical mctbod were very accurate when compared to the
original data. This is a good demonstration that tbe approach used in the present
study can be used to examine the ratios of any spice blend, as long as the
individual spice is included in tbe data matrix (355 spices).
"0 SI'ICES: nAVOR CHEMISTRY AND ANTIOXIJ)ANT i'ROI'EHTII!:S
Tahle IX. Compari son of the numerical melhod with previous publications
Spices assigned % lI(cal) Spicesb Yo
Basil #26 47.9 144 Exolicbasil 42. 1
Clove 114 c
12.7725 Cinnamon leaf
15.8
Comminl #8 d 9.3919 Peppermim 15.8
Sage #2 23.8251 Sage 21 I
Wintergreen #1 6.0962 Wint ergreen 5.2
asummed up values from Tabl e VIII .: b: data from ref 4.: c:index (Clove
#4, Cinnamon Leaf 114 )- 0.95: d:lndex (Cornmint 118 , Pepper #8 )- 0.83
"
,,-L
li\I1
fl
~ ~
, ,
0 '0 ' 0 . 0 _0
' 0
, ,
'0
_ 0
.0 .0 ' 00 " 0
..... C( .... "u,. )
Fig I Gas chromat ogram of the volatile components of the mixed spices from Table X.
13. CHENG ET AL Component Analysts of Mixed Spices 17l
Allll lysis ofte5t ing spice blend prepa rw in the labontol")'. The formulation
of a real spice blend used in the present study is shown in Table IV. Volatile
compounds from this spice blend were isolated by simuillneous steam distillati on
and solvent extraction. Identification of volat ile compounds was acromplished by
compari ng the retention index lind mass spectra with those of authentic samples
and publ ished publications (/4 - IS) . Figure I shows the gas chromatogram of
isolated volatile compounds from the testing spice blend, Percent ratios orthe5e
volatile compounds identified are shown in Table X. Compounds in Table X were
then coded by number in the same manner as those of Table VII. The orig inal
resllits of numerical analyses are shown in Table XI, similar in fomlat to those of
Table VIU. Recognition of spices in this spice blend showed a very high similarity
index (0.981), indicating high level of confI dence. Those spices thAt appeared
repeatedly with a high similarity index were assigned to the spice wi th highest ratio,
i.e. the ratios of anise and star anise were combined, as ment ioned earlier, that is
because both anise and star anise listed are similar in composition. The same holds
true for the combinations of black pepper and allspice. The only spice which is
determined by numerical analysis but not added into the testing mixture is
marjoram (ca. 0.43%).
The comparisons of ratios of spices as determined by numerical method am!
those calculated from the content of essent ial oil are shown in Table XII. Sensory
analysis of a 0.4% solution of the actual spice blend as compared wi l h that of a
predicted spice blend have shown lillie difference. The accuracy of Ihis test is
surprisingly good for star anise, black pepper and garlic. The mli os of allspice and
clove are quite good, although some deviations from the actual data exists. Celery
has been correctly identified but the ratio determined in the present test (only 0.03
%) is qui te low, probably caused by technical difficul ty of our MS database in
assigni ng the correct volatile compounds of celery. As to the appearance of
mrujoram, although the ratio assigned to this spice is smllil (0.43%), the similarity
index between mrujoram and other identified spices has not shown any high degree
of similarity. Probably this deviation is caused by the low val ue assigned to celery
(0.03% vs. \.78%).
Conclusion
The present study has shown that the ratios of individual spices in a spice hlend can
be accurately solved by a numerical met hod when compared with previous refKln5
(4. 7). Spices in a spice blend with an essential oil coment as low as 0.370/. can be "
I

,
,I
, )
, I
Ii
d
! '
i
'.
il
'I
"
I
I
i
' /
l
rj
"
,
172
SI' I CF.s: nAVOH CIIP'}'U!>'TRY ANI) ANTIOX.IDANT I' ROI'ERTI I;:S
Table X.
Identified volatile compounds from the s ~ i c e blends from
Table IV
Peak NO a
-6. Compollnds
0.29 alpha-pinene
2 0.67 beta-pinene
)
0.19 sabinene
4 1. 50 della-J-carene
5 0. 16 alpha-pheJiandrene
6 0.21 myrccne
7 0.05 alpha-terpinene
8 2.28 li monene

0.10 1,8-cincole
10 0.12 gamma-terpinene
II 0.18 para-cymene
12 0,04 altyl methyl disulfide
IJ 0.07 lerpinolene
14 0.0 1 methyl trans-I.propenyl disulfide
15 0.0 1 dimethyl trisulfide
16 0.03 penl ylbenzenc
17 002 4-isopropcnyt-l-melhylbenzcnc
18 0,02 lioalool oxide
I. 0.25 alpha-copaenc
20 0.34 linalool
21 0.02 trans-alpha famesene ,..
22 3.37 bela-caryophyllene
23 0.49 terpinen-4-o1
24 0.45 deha-cadinene
25 0,08 cSln.gole
26 0.25 alpha-terpineol
27 0.09 carvone
"
0.<15 cis-anethole
2' 0.90 diallyl tri sulfide
) 0 70, 16 trans-anethole
1I 0.06 para-cymcn-8_01
32 0.1<1 3 -methyl-I-phenyl-I -butanonc
JJ 0.39 beta-caryophyllene oxide
l4 3.02 anisaldehydc
35 2. 17 methyleugellOl
~ ~ 11.46
CII8SO
n
Q!
TOIa! JOO 00
a, Number refers 10 Fig I
13. CHENG t:r AL CQmponent Analyses of Mixed Spices 173
Table XI , Computer output of ralios of essential oils as
anal:tzed b:z: numerical anal:tses. The similarit:t index " 0.981
Sl!ice N Calculated % Assignment
Anise #1 3.92<15
Anise #2 17.8]<11
Ani se 11<1 9.5 1<19
Anise #5 7.0813
Anise 116 9.6]66
Anise 117 0.7293
Star anise 1i2 9.92<11
Star anise 1i4 19.3589 Star anise 114
Celery #5 0.0305 Celery 115
Clove 11<1 <15803 Clove #<1
Garlic iii 0.<1170 garlic #J
Marjoram #2 0.4275 marjoram #2
Black pepper
"
0.278<1
Black pepper #2 5.7291 Bl ack pepper #2
Allspice #<1 0.3164
Allspice liS 1.9528
Allslli,,1I:8 J 2518 Alislli",1I:8
Table XII. Ratios of individual spice in the spice blends after adjustment
3CCQrding to closeness of simiiarit:t inde)!.
spices WI (g)1 essential oil % oil in the Yo oi l in the
contcnt (%)b mixture (cal)C mixture (exp)d
black pepper 27.00 1.76 6.<13 6.29
allspice 27.02 2.58 9.4<1 6.31
celery 13 .52 0.97 1.78 0.03
clove 1,62 7,67 1.68 4.80
Slar anise 6 1,03 9,72 80.31 81.70
garlic 27.01 0. 10 0.37 0.<14
marjoram 0.<13
I ' Dala from Table IV.;b: Data obtained from this s\udy.;c:Nonnalized per cent
distribution; d:Ralios combi ned from Table XI.
174
SI'I CES: HAVOR AND ANTIOXIIlANT I'IWI'ERTlK'i
accuflltely ,Ietermincd without significant deviation from the actual value. It is the
utmost hope of the authors that applicati on of thi s technique can definitively
shorten time and effort used to determine the ratio of individual spi ce in a spice
blends
Acknowl e(lgrmenu
This work was supported by the Ministry of Economic Affai rs, Taipei, Taiwan,
Republi c of China. We uc also gfale!ul lo our laboratory slaffs for technical assistance.
Liarllltur( Cited
I_Farrell ." T. In SpiCt!l'.COlldiIllI!IIIS,(llld Van Nostrand Reinhold, New
York. U, gO.
2 Giese, f. Food Techllol . 1994. 48(4),88.
J . Volmill Compol/llds ill Foods H . Vis5(;her, CA., Eds., TNO-CI va,
Zeist , tile Netherlancs 1989, Vol 1-2.
4.Lawrel,ce, B., Shu, C. K. In Plavor Measurtmelll. Ho, C.T.; Manley, C.B , Eds ,
Marcel P ekker, New York, 1993, pp 267328 ..
S. ESO: 'lhe COli/plett Database of wenfial OiLf. BACIS, the Netherlands, 1995.
6.Salzer,!JJ . CRC Crit. Rev. Food Sci. Nutr;., 1977,9, 345.
7 Chien PA. A/KII. Ciltm. 1985, 57,348.
, ,
8 Luenbt!:rger,D.G. Optimizatioll by Yeetor Space Methods. John Wiley & Sons,
New y p rk. 1968.
9 Tabacctu, R., Gamero, 1 ; Buil, R. Rivista Iial., 1974, 56,683.
IO.Cu, J Q In /Olh Congress Ess. Oils, Fragr.& Flav., Washington, 0 C . 1986,
pp231 , 2 4 1
[I.Poller, TL., Fagerson, 1.5. J. Agric. Food Chem., 1990, 38, 2054.
12. S. T., Nickerson, G n ., Am. !XJ<:. Brewing ChemislS Proc. 1964, 5.
13.WrolstjJ'dt , RE., Richard, H.M.; Jennings, W.G. J. Food Sci., 1971, 36,584.
14. Jenni nls, W.; Shibamoto, T. In QII(I/itath'e AI/aly:l'is of Flavor alld Fragrallce
Vola/ii' s by Glass C(If/ilIary Gas ChromalOgrt/phy Academi c Press, New York
1980.
15. II/fit/red, Mass, HNMR, (md CI J NMR S/Hctra alld KOI'rIfS
II/dice.f. Swigar, A. <\ .; Silverstein, R. M. Eds.; Aldrich Chern. Co. Inc.,
Milwaufkee, WisconJin, 1981.
ANTIOXIDANT P ROPERTIES
"
"
,
;
)
i
..l
I
j
"
I
I
I
,
,
Chapter 14
Antioxidative Activity of Spices
and Spice Extracts
lIelle Undberg Madsen. Grete Bertelsen. and l..eif U. Skibsted
Department of Dairy a nd I<ood Science. The Royal Veterinary
and Agricultura l Univer s ity. Rnli ghednej 30,
DK- 1958 .' rederiksberg C, Denmark
The alllio,,;dative activity of spices and spi ce extracts can usually be traced
back to their content ofpheDolic compounds. Plant phenols may scavenge
free radicals involved in lipid peroxidation .s has beeD documented in
several model systems, although other mechanisms should be considered
especially in relation to the carty stages of oxidative deterioration. Phenolic
compOlwds isolated frOIll spices have been found to react with hydroxyl
radicals wiOI nearly diffusion controlled reaction rates. An assay based on
a combination of detennination ofpbenol equivalents and determination of
radical scavenging capacity by the ESR spin trapping technique confiTU.1S
the nearly diffusion controUed reaction rates and may prove useful for
exploration of new plaut materials and for adjustment of e"'Taction
procedllfes, including selection of solvent. This and otber assays based on
oxygell deplet ion measurements are recommended prior to final testing in
real foods. Co-extracted cbJorophyUs in spice extract present a problem as
photosensitizers in food cxposed to light during storage IIld use.
of spices to food is ao establ ished procedllfe in most cult llfes. The seasooing
contl'lbutes a pleasant flavour, and at the same time a number of compounds possessing
amioxidative activity are added. Scientific investigatioll of the antioxidative activity of
spi ccs has been initiated duriug the last few decades and an understanding of the
mechanism of the antioxidative activities is emerging. Early work by Chipauh and
coworkm (1-3) in the ftfties documented significant reduction in oxidation in lard and
e.dible The most pronounced cffects were found for rosemary aod sage, observa.
lions which lIave been verified io most later investigations (4-6). However, a great
variety of other spices invest igated show similar antioxidative activity and notably their
relative efficiencies are strongJy dependent on the actual food.
1997 ChcmiCilt Society
14. MADSEN ET AL ,v,tioxidati,, of Spias lind Spicl Extracts 177
OJidation and Antioxidants
During oxidation free radicals are generated cOnlinuously in the propagation phase. In
this phase a lipid radical (R') reacts very quickly with (tri plet) oxygen in a reaction
controlled by diffusion. The formed peroxyl radical ( ROO') can abstract a hydrogen
from an unsaturated lipid molecule (R' H). This reaction lead5 to generation of a new
lipid radical (R') and continued acti vity in the chai n reaction. The hydroperoxides
(ROOHl, the products fonned in the chain reaction, are without smell and taste.
Hydroperoxides are easily decomposed to yield alkoxyl radicals (RO). These radicals
can funher abstract a hydrogen from a lipid molecule in effect starting a new chain
reaction contributing further to the propagation phase (7) (Figure I). React ion products
from the degradation of the initial fonned hydroperoxides, the secondary oxidation
products, are carbonyl and carboxyl compounds which give rise to the unpleasant taste
and smell of oxidized food (8).
An efficient way to delay oxidation is scavenging by antioxidants of the free radicals
generated in the propagation phase or during the break down of the hydroperoxides, i.e.
scavenging of either the peroxyl radicals or the alkoxyl radicals. The critical level needed
of such primary antioxidants to be effective in a given product corresponds to the
concentration necessary to inhibit all chain reactions started by the ;nitiation process. M
long as tlte concentration of the antioxidants is above this critical concentration, the total
number of radicals is kept at a constant low level, a time period which is defined as the
induction period. During the induction period the antioxidant is gradually depIcted and
when the critical concentrat ion is reached, radicals will escape from reaction with the
antioxidant, and the concentration of hydroperoxides will increase_ The high level of
hydro peroxides wi ll funher increase the concentration of radi cals, and the rcmaining
antioxidant will be used up completely (9). With all the antioxidants consumed, the
oxidative processes will aceelerate, and the increase in the production of secondary
oxidation poducts will lead to a progressing deterioration of the product.
Phenolic compounds act as primary antioxidants by donating a hydrogen atom to
either the al koxyl radicals (equat ion 1) or the peroxyl radicals (equation 2) in irreversible
reactions. The reactions lead to generation of an antioxidant free radical (A) with a
lower energy compared to the energy ofRO and ROO, the reaction is enthalpy
driven.
AH + RO
J\H + ROO -
A + ROH
A- + RooH
( I)
(2)
The relatively high stability of A- reduces its ability to abstract a hydrogen from a lipid
molecule in effect breaking the chain reaction. The efficiency of phenolic compounds as
scavengers offree radicals is due to the high stabi lity of the generated pheno)(y radical,
whieh is caused by delocalisation of the unpaired electron in the aromatie ring.
However, phenols are not act ive as antioxidaots unless substitution at either the
onho or para position has increased the elect roo density at the hydroxy group and
lowered the oxygen-hydrogen bond energy, in effect increasing the reactivity towards
the lipid free radicals. Substitution in phenolic compounds at the meta pOsition has a
rather limited effect , and compounds like resorcinol with a hydroxyl group in the meta
178
Initiati on
Propagation
Termination
SI'ICES, n.AVOR ANI) ANtiOXIDANT I'ROrERTIF.s
ROOH
RH
ROO
RH

1"'----- XH
R
ROO,
-yRO'
Non radi cal products
0 ,
J-igure I. Lipid Olcidation wilh an iuiliatioD pbase, propagation pbase and termina-
tion phase.
-.
14. l\lAIlSJ.:N AL An/ioxid(Jt;vt Adi!'i" of Spias and Spice Extracts 179
position are thus poor scavengers offree radicals compared to similar compounds with
substitution al the 011110 or para position. The preSCllce of additional hydroxy groups,
either 31 tile onho or para position, further iucreases the antiolri dative aClivily of the
corupowld IS illtn -molecular hydrogen bouds stabilize Ihe phenoxyl radicals (9).
Antiolidative Compou nd, in Spices
Phenolic comlloUJl ds constitut e the IHgcSl proport ion of know nalUral antioxidant s. It
should, however, be noted that only H minor part of the compounds in spices which have
antioxidative aetivity has been isolated and identified. Nakatani and coworkers (1 0./1)
have reported the structure aud characterized the antioxidative prop,mies of cveral
phenolic diterpeno.'S isoLtted from roclllary (RosmarillllS officinalis L.). From the SlIme
plillt , eompounds like eamosic acid, eamosol (12-/4), rosnuridiphenol and rosm3Ti-
quinone ( />./6) have been identified (Figure 2) as antioxidants. Moreover a number of
the cOllvouuds fOlU1d UI rosemary have been found in the botanically closely related s.1ge
(Sa/via officil/a/;s L.) ( 12. / 3) Hnd summer savory (Soll/reja harle/ISis L.) (/7./8).
In other spices, fiavolloids aloll e or together wilh ot her phenolic compounds have
been found 10 contribute to the antioxidative activity. In oregano (Origanum vulgare L.)
various autioxidative compounds have beeD isolated, and among the act ive component s
four flavonoids were identified (/9). From thynJe (Thymus vulgaris L.), which like
oregano also is a ruelJlbcr of tbe LabialfU! family, antioxidBtive compounds have been
isolated and identified as dimers oftllyruol and f1avonoids (20). The antioxidalive aClivilY
of pepper (Piper lIigrum L.) can, at least partially, be ascribed to the presence of
glycosides of the f1avolloids kaempferol, rhamlletin and quercetin and at least fivc
different phenolic amides (21.22).
A variety ofdilli::rent volatile compounds such as the terpeuoids thymol , caT\'ocrol,
eugenol., carvone and thuj one. which are character-impact compounds for imponalll
spices (23.24), have 3ntioxidative activity hili the use of these compounds as amioxi
dants for different foods are limiled by the cbaracteristic flavour of the plnicular
compound.
Evaluation of the Uadic:l1 Activity
Screening of IIltioxidative activity in various mode! systems is imponan! prior 10 te!>ling
or application of IIltioxidanls in foods. Such model systenl$ are more rapid compared
to food storage experiments. and the model systcms might eveu be more illfonn:uive in
relation 10 Intioxidant mechanisms.
Electron spin resonance (ESR) spectroscopy is curre1Llly being introduced to
different areas of 311plicatioll ill food science. The lechniquc may provide valuable
information conceming thc elemeutary processes in lipid o)(idal ion and the nature of
reaction A number of ant ioxidant s iocluding extracts or compounds
isoLaled from spices have been investigated by differelll ESR technique. The amio)(ida
live activity of the fractiou containing tll e essential oils from oregano (OrigallulIl \"Ignre
L.), summer savory (Sawreja horUltti$ and thyme (Thymus vulgaris L.) has been
documented in system where oxidative stress and free radical reactions were induced
either by UV irradiation or the supcmxide radical. Free radicals Cill be detected directly
,
,
i
r
,
1,<
'J
.1
,
.1

.J
;.
J
,
,
j
l
.
i
,
.,
,
,
1
OH
Carnasel
OH
Epirosmanol
HO
OH

Rosmaridiphenol
9
H
, .... \
Carnosie acid
Too
HO
' OH
Isorosmanol
OH
o , ~ /
-,/
CHO
'CHO
Rosmadial
OH
,
"" '\
UH
Rosmanol
~ O H
Hoae 0
o OH
~ O
H
Rosmarinic acid
., ....
Rosmariquinone
F.gurt 2. Chemical structure for compounds isolaled from rosemary (Rosmarinus
officinalis L.).
---- ---------
~
~
R
~
-.
):
<
0

9
,

~
<
~
0
~
0

~
z
.;


0

~
,
,
..
~
;:
~
~
~
~
~
~

s'
".
~
~ .
~
,;.'
...
.g>
S'

it
.g>
~ .
r
~
-

182 SI'ICES: flAVOR CIn1l-mi'TKY AND ANTIOXIDANT PROI't:RTl ES
0.6
0.5

""
0.4



0
0. 3
0


0.2
"
c
0 . 1
0 .0

E
0 0 .-
,
c c

,
c
c 0
0 0 0
0 0 0

E
,
" "" ""
CD
> >
c
0 0

, , ,


0
0 0 0 , ,
"

c
E 0

c

E
"
.-
,
0
0
"'
U
Figure 3. The hydroxyl radical sc.vcngiug activity of nine different spices all
belonging to the family Lobimae. The activity was measured by using elect ron spiu
resonance technique, and Ihe indexes were calculated on the basis orlop height
from oata of reference (28). A low value of the index indicates an efficient
selvenger. Columns with different letters arc significantly dilTereDl at 5% level
-
14. MADSt:N F:r AI . Alltiandative ActMty of Spices tllld Spiu Extrar:l$ 183
by ESR spect roscopy, and have in the spice extract of oregano, thyme and summer
savory been ident ified as free radicals derived from thymol and carvacrol. NOlably,
thymol and carvacrol radicals were also present in untreated essential oils indi cating thc
high stability of the radicals and the potential of the compounds as scavengers of
oxygen-centred free radicals (15). Experiments using spin trapping technique are based
on compet ition between the compound expected to have antioxidative propenies and
a "spin trap" in reaction with free radicals. 1be it! vitro anlioxidant properti es of rebam;
pide, a nove! antiulcer agent, have thus been documented by the spi n trapping technique
(26). and the scavenging activity of carnosine and relat ed dipeptides have been fun her
investigated (27) In an investigation with focus on the early stages of oxidat ion in foods,
the ability of spi ce ext racts to inhibit the initiation of free radical process was evaluated
by adapting the ESR spin trapping technique (18). The cxtracts or tile spices obtained
by ethanoVwat er extraction were freeze dried and redissolved in waler. The water
soluble compounds, panly phenol s, were found to reduce the signal int ensity in the ESR
spectrum ofthe spin trap/free radical adduct when hydroxyl radical s were generated by
the Fent on reaction, indicating a high scavenging potential of t he spice extracts against
hydroxyl radicals. Ext ract s of each of nine different spices investigat ed, all belongi ng to
the Labiatae fami ly, were capable of scavenging hydroxyl radicals (Figure J ) Most
efficient was winter savory. rosemary, sage and the two oregano species. Kinetic studies
gave an estimate of the rate constant between hydroxyl radicals alld phellols in the
extract of Turlcish oregano. The value of the second order rate constant was slightly
above that for a diffusion controlled bimolecular reaction in water, and the high value
indicated that compounds other than phenols were involved in the compeTit ion with the
spin trap in the reaction for the hydroxyl radicals. Rate constant s for reactions between
compounds isolated from spices and different radicals have, however, o nly been
determined for a few substances Ilnd funher studies should be encouraged. An estimate
of the rate constant of 4.810!0 M ! S I between hydroxyl radicals Ilnd eugenol is,
however, an example (29). TIris corresponds to a veT)' fast reaction wit h a rat e constani
similar 10 the rate constant found for the reaction between hydroxyl radicals and phenols
in extracts of Turkish oregano (28). These result s together document that hydroxyl
radical s are very reactive, and model systems based on other radicals should be
developed for a differentiation between different groups of frce radical scavengers. The
second-order rate conslant detennined for lhe superoxide scavenging act ivity of eugenol
(a volat il e present in clove, cinnamon and basil (29 has been used for comparison of
phenolic antio,odants and the enzyme superoxide dismutase. Superoxide di smutase was
most effective in deactivating superoxide, rat c constant s were found for
both eugenol and guaiacol, while the acti vit y of phenol was rather poor (30) .
At later stages in the oxidation process the scavenging activity of antioxidants
towards peroll.yl radicals is of relevance. Aruoma et al. ( 31) investigated the rate of
reaction or a number of ant ioxidanu wit h the trichloromethyl peroxyl radical . Carnosic
acid had a r1lteconstant of 2.7 10
1
M' 5', about 10 times higher than the rate const Jnt
of camosoI. The react ivity of carnosic acid with the peroxyl radicals was comparable 10
propyl gallate, and it could be concluded that camosic acid is effective as a scavenger
of trich1oromethyl peroxyl radicals, although the reactivi ty of carnosic acid was not as
high as the reactivity of ascorbic acid and the wal er soluble tocopherol analogue, trolOl(
C. Determinati on of the rate constant s for the reaclion between hydroxyl radical s and
"
1
-I
.;
'84 srICI:t.'i: HAYU\{ (; II I::MISTRY AND ANTIOXIDANT "IWI'ERl"I F:S
either camosic acid or camosol again showed ute constants whicb indicated that the
reactions were essentially diffusion controlled (31).
AnlioI.idativt ICl i"it y in model systems
While dcterolination of rate constants fo r scavenging of free radicals by antioxidants is
highl y rel evant for establishment of structure/activity relationships. it is essential for
food application to constnlct betefogeooous model systems with phase transfer.
Frankel et al. (31) evaluated rosmarinic acid, caruosic acid and camosol in such model
systems by mel5Uring both primary and secondary oxidation products and showed that
the antioxidat ive activity depends strongly upon the substrate. The compounds
coluaining a carboxy group (rosmarinic acid and camosic were the IliOst effective
aotioxidaut s ill bulk COllI oil whereas the antioxidative activity of camosol WIS only
limited in this oil. However, in an olw emulsion a bigb antioxidathre activity was seen
fOJ" camosol and camosic acid while only slight autioxidative or even prooxidative
activity was found fOl" rosmarinic acid. The observation tbat tbe more polar antioxidants
aTe more active in pure lipids, and the nOli polar antioxidaniS most acthre in a polar
substrate, and for wh.i cb the tenll ')lOlar paradox" has been introduced, confimls
previous findings for tocopherol s .and ascorbic acid derivatives as antioxidants (33). In
the bulk oil the hydrOI)hi lic ant ioxidant s are oriented in the oillair interface providing
optimal protection of tbe lipids aglin!.t oxygen radicals while the hydrophobic
antioxidants dissolves in the hOlllOgeneous lipid pbase. The opposite situation with
hydrOI)hobic antioxidant conccntntion in the oil/water interface is encoulllered in the
emulsion syst;om, where the hydrophobic antioxidants, like camoso!, are most efficient.
These obsclVRtions may at least paniaUy explain tbe variation of antioxidative activity
seen for different spices in different foods (3), and the need for investigations in such
bet erogenous model SySlemS and. in the actual food products prior to prtlctical use is
obvious.
Antiol.idative activity in food
Thl: results obt ained by Frankel et a1. (32) using food models mig\Jt explain the
(IUalilalive di.fl:erence observed between antioxidative activity of summer savory and
rosemary in a meat storage eX]leriment and in a storagc experiment with dressing. In
precooked pork meat balls no sig;nificant difference between the antioxidative activity
of rosemary and SllllI.lt1er savory was seen, and both spices were able to retard the
development of oxidative offOavour, measured by determination ofthiobarbituric acid
reactive substances (Madscn et al/. TIl e Royal Veterinary and Agriclilrural Unhrersity,
Fredcriksberg, OK, unpUblished da1\a). l11csc results were, however, different from those
fouud for a dressing with a high fat content (50%). In the dressing the addition of
rosemary was found to be signifi cantly more efiective in proteclioll against oxi dation
compared to summer savory (M:adscn et al. The Royal Veterinary and Agricultural
Uni ....ersity, Frederiksberg, OK, wl:published data). Although the nature and quantity of
the compowlds prescnt ill the two spices were not determined, it may be
speculat ed that differences in lhe hydrollhilic or bydrophobic properties of the
anlioxidants lUay be imponall t. Ie is well-known, tbat rosmarinic acid is ao important
14. MADSEN t.T AL. AnJioridativt Activity 0/ Spices and Spia Extracts
185
antioxidative coqJooenl in summer savory (/8). The low efficiency of this substance in
bulk oil might provide I rationale for the reduced antioxidative activity observed for
summer savory in the dressing. Other factors such IS differences in pH in the two foods
may also oontnout e to the different activities. Frankel el 01. (32) have also st udied the
a"tioxidative activity of camosa!, clmosK; acid and rosmarinic acid in emuJsions with
diJrerent pH. At pH 4 and pH 5, where the carbo1C}'1 group is protonatcd to a significant
extent the aotioxidative activity of camosol and camosic acid were much higher than
the .ctivn
y
at pH 7 where the carboxylate form dominates. This pH dependency might
be explaioed by the cbange in polarity and oxidat ion potential or by a higher stability of
the compounds at lower pH. An increased knowledge of tbe effects of pH will be
beneficial for a more rational use of plant antioxidant s.
Further perspectivu and development5
A thrce-step prOCedllfe as the one out lined (i) determination of radical
scavcnging activity of spice extract, (ii) ill model systems, and (ill) fuJal
application in foods may prove useful for future work on natural antioxidants. Different
solvents should also be explored for extraction of spices. Different polarity of the
solvents might be used for "tuning" of extracts to differcnt products of different
hydropbobidbydrophilic balance. A higller antioxidative activity was found of a hexane
extract of rosemary and sage compared to other solvents by using successive extraction
of solvents of increasing I,olarity (34'). I-:I owever, most. investigations report a high
antioxidative activity in extracts obtained from polar solvents (/. 35,36). The use oflarge
amount s of org:anic solveuts is in general unwanted for environmeotal resonance, so
altemllive extraction methods should be considered. lne use of supercritical COl
extraction, which has already shown positive results (37), might be getting more
attention since the tecbnique is efficiellt and II.so protects the phenolic compounds
against degradation during extraction.
The interaction between different antioxidants such as tocopherols. ascorbic acid
and spice componenls and possible synergisms should also be explored. Fang & Wada
(]8,39) already invCSl:igated the synergism between a-tocopherol and extract of
rosemary in ethanol The results support tbe hypothesis that rosemary antioxidants
regmCTIte oxidized a-tocopherol. Similarly Ternes & Schwarz (4'0) found an increased
stability oftoeopherols wben rosemary extract was added to meat sampl es. To obtain
more information about the antioxidathre compounds a determination of the reduction
potential of key component s (6,4' /) would be of interest for ranking of antioxidants io
order to determiue \o\"IDch oft be spice Illtioxidants can regenerate the tocopherols. 11le
problctllS encountered with photosensiti7jng ehlorophylls co-ext raction from the spices
making the food sensitive to lig\Jt (Madsen el 01. Royal Veterinary and
Agricultural University, Frederiksberg, DK, unpublished data) also needs further
investigation.
Acknowledgments
Thi s research was part of the frame program "Natural antioxidants from plants"
sponsored by the F0TEK-program t.hrough. National hod Agency and LMC-Centre for
Advanced Food Studies.
I
186 SPICES: FLAVOR CIiEMISTItY AND ANTIOXIDANT \,ROl't:RT1ES
LiaralUre Cited
I. Chip.uh, 1,lL; Mizuno, G. lL; Hawkins., 1.M.; Lundberg, W.O. Food Res. 1952, 17,
pp, 46-55.
2. ChipIUIt , J.lL; Mizuno, G.R. ; Lundberg, W.O. Food Res. 1955,10, pp. 443-448.
3. Chip.ull, l lL; Mizuno, GR; Lundberg, W.O. Food Techno!. 1956, 10, pp. (5)
2092 11 .
4. P. litzsch, A. ; Schulze, H. ; Metzl, F. ; Baas, H. Fll!iscJnYirtsch. 1969, 49. PI'.
13491353.
5. Gcrhardt , U.; Schr6ter, A. Gordimr 1983, 9, pp. 171 176.
6. Gerhardt , U.; B6hm, T. Fleischwirtsch. 1980,60. pp. 15231526.
7. Kanner, 1.; German, J. B. ; Kinsell a, lE. CRC Crit. Rev, Food Sci. NIl/f. 1981,15,
PI' . 3 17 364.
8. Shahidi, F. III Upidoxidation illfood:. SI.Angelo, A. J. Ed. American Chemical
Society:. Washington DC, 1992, pp. 161-182.
9. "okomy, J. In AU(Qxidorofl ofullsotllrared lipids;. Chan, H.W., -5. Ed. Academi c
Prcss:.London, 19M1, pp. 14 1206.
10. Nakatani, N.; (!latani, R. Agric. Bioi. Chelll . 1981,45, pp. 238S2186.
I I. Nakatani, N.; Inatani, R. Agric. Bioi. Chem. 1984, 48, PI). 208 12085.
12. Bri eskom, C.B. ; Fuchs, A.; Bredenberg, J.B.; McChesney, J.D,; Wenken , E, 1,
Org. Chl!m. 1964,19. pp. 22932298.
13. Bri eskoru, C. H.; 06mling, H.1. Z. ubmsm. Unters. Forsch. 1970, U/. pp,
1016.
14. Wu, J.W.; Lee. M. H. ; Chang, S. J. Am. Oil Chern. Soc. 1982, 59, pp. 339-34S.
15. Houlihan, C.M.; tlo, C.-T. ; Chaug, S.S. 1. Alii. Oil Clrem. Soc. 1984, 6/. pp.
1036- 1039.
16, Houlihan, C.M.; Ho, C.T. ; Chang, S.s. J Am. Oil Clrem. Soc. 1985,62, pp.
96-98.
17. Reschke. A. Z. Ll!benslll. ,Ulllers. Forsch. 1983,176. pp. 116 119.
18. Bectelsc:n, G.; Ouigopberseu, c.; Nielsen, PH ; Madsen, H.L. ; Stadel, P. 1. Agric.
FoodChelll. 1995,43, pp. 1272127S.
19. Vekiari, SA; Oreopouiou, Y.; Tzia, C. ; Thomopouloll, C.D. J. Am. OJJ Chern.
Soc. 1993, 70. PI' . 483487.
20. N. 10 Plrenolic compounth illfood and their effl!cts 0/1 heallh 1/;. HUIlIg,
M. T., Ho, C.' T., Lee, C.Y. Ed. American Chemical Societ y:. Washlogtoo DC,
1992, pp. 72-86.
2 1. V6sgen,D. Hernnano, K. Z. Lebensm. -Ulllers. -Forseh. 1980, 170, pp. 204 207.
22. Nakatani, N.; Inalani, R. ; Ohts, H. ; Nishioka, A. EIIyiroll. Healtlr Pl!fspect. 1986,
67, pp. \ 3S- 142.
23. Far.g, lLS.; AZ.MA; Hewedi, F.M.; EIB.roty, G.S.A. JAm. 011 Clle",.
Soc. 1989,66, pp. 792-799.
24. Kramer, lLE. J. Am. Oi/Chem. Soc. 1985,62, pp. 111 113.
25. Deighton, N.; S.M. ; Deaus, S.G.; Goodman, B.A. J &i. Food Agric.
1993,63. pp. 221-225.
26, Yoshik.lwa, T.; Naito, Y.; Tanigawa, T.; Kondo, M. An"tim. FlYsch. IDrug Res.
1993, 43. pp. 363366.
14. MADSEN AlilioxidDtive Activity of Spit.'tS twd Spice Extracts 187
27. Chan, W.K.M.; Decker, E.A. ; Lee, J,I1.; Butterfield, D.A. J. Agric. Food Chf'lII.
1994, 42. pp. 1407- 1410.
28. Madseu, I-lL; Nielsen, JUt. ; Derl elsen, G.; Skibsted, L H. Food Chemistry. III
press.
29. Nagababu. E.; Lakshmaiah. N, Biocllem. Pharmacol. 1992,43, pp. 2393 2400.
)0. Tsujimolo, Y.; Hashizume, tl ; M. /111. J. Biochem. 1993, 15, pp.
491 494.
31. Arnoma, 0 .1. ; Halliwell , 8 .; Aeschblch, R. ; Loliger, J. Xe"ohiotica 1992, 22, pp.
257-268.
32. Frankel. E.N.; Huang, SAY.; Aeschbach, R. ; mor, E. J. Agric. Food Chem. 1996,
44, pp. 13 1-135.
33. EN.; Huang. S.-W.; Kawler, J. ; Gennan, lB. J. AgriC. Focx1Chem. 1994,
42, pp. \ 054 1059,
34. Svoboda, K.I'.; Deans, S.G. Flavour alld Fragrance Jor/rnal 1992, 7, pp. 8 1-87.
35. Chang, 5.5.; "Ostric-Matijuevic, B.; Hsieh, O.A.L. ; Huang, C.-L. J. Food Sci.
1977, 42, pp. \102-1106.
36. Palit zsch, A.; Schulze, N.; Lotter, G.; Steichcle, A. Fleisc/nvirlsch. 1974,54, pp.
63-68.
37. Z.; lankov, lLM.; Schwirtlich, E.; Djulinac, D.; Djordjevic, A. J. Am. Oil
Chem. Soc. 1991,68. pp. 73 1-734.
38. Wada, 5.; Fang, X. JOllrl/ol of Food ProcesSing alld Preservatioll 1992, / 6. pp.
263-274.
39. Fang, X.; Wada, S. Food Research /lllerl/o/iOlraI 1993, 16, pp. 405-411 .
40. Ternes., W.; Schwlrz, K. Z. Lebellsm. _Utllers. -Forsch. 1995,201, PI' . 548 550.
41. Palic, A.; KrizanC(;, D.; Oikanovio- Luean. Z. Fleischwirtsch. 1993, 73. pp.
670-672.

"
"
I:
"
.,
,
,
"
,
,
"
'.
, '
I
Chapter 15
Antioxidative Effect and Kinetics Study
of Capsanthin on the Chlorophyll-Sensitized
Photooxidation of Soybean Oil and Selected
Flavor Compounds
Chung-Wen Chen, Tung Ching Lee, and Chi-Tang Ho
Depar tment of Food Science, Cook College, Rutgers, The State
University of New Jersey, New Brunswick, NJ 08903-023 1
The antioxidat ive effect of capsanthin and lutein on the chlorophyll-
sensitized photooxidation of 2-ethylfuran, 2,4,5-trimethyloxazole, and 2,5-
dimethyl-4-hydroxy-3(2H)-furanone (DMHF) was studied using a
spectrophotometric method. Both capsanthin and lutein exhibited the
antiphotooxidative effect on these flavor compounds. The
antiphotooxidative activity of capsanthin was higher than that of lutein. The
antioxidalive effect of capSlnthin, p-carotene, and lutein on the
photooxidation of soybean oil containing chlorophyll was studied by the
Rancimat method, in which the induction time and the anti-photooxidation
index (API) were measured. The induction times of soybean oil which
contained the carotenoids were longer than that oflhe control sample which
contained no carotenoid. As the number of conjugated double bonds
increased, the API of soybean oil increased. The capsanthin, which contains
II conjugated double boods, a conjugated keto group and a cyc10pentane
ring, had higher API than Il-carotene, which contains II conjugated double
bonds but neither a conjugated keto group nor a cydopentane ring. The
steady-state kinetics study of capsanthin on the chlorophyll-sensitized
photooxidation of soybean oil was conducted by oxygen depletion method.
The result showed that capsanthin quenched singlet oxygen only and its
quenching rate constant was 5.746 x 10
9
M-
I
s -I in methylene chloride,
Some food constituents are sensitive to photooxidation, Oil and fats are the most
sensitive food constituents for photooxidation becausc thcy cont ain a large number
of double bonds in their structures and probably due to the greater solubili ty of
molecular oxygen in the lipid phase as compared to the aqueous phase (1,1) . Flavor
compounds especially five membered heterocyclic flavor compounds such as
alkylfuran, alkylpyrrole and alkyloxazolc compounds easily undergo pholooxidation
reaction (3).
Cl I997 American Chemical Sociely
I S. CIIEN f::T AI... Effoct o/OIp.fanthin 011 Plwtooxidotion 0/ Soybean Oil 189
The whole photosensit ized oxidation reaction can be described as follows (4):
A sensitizer, SlIch as chlorophyll , can absorb light energy and becomes an
excited singlet sensitizer ('SenO), which is then rapidly converted to the excited
triplet sensitizer eSen
O
) by an intersystem crossing (JSC) mechanism, a process of
electron rearrangement. The )Sen
o
transfers its energy to triplet oxygen ('0
1
) to
generate the singlet oxygen (10
1
) by a triplet-triplet annihilation reaction, a process
which occurs via a collision of two excited )0
1
Singlet oxygen is a very reactive form
of oxygen and easily attacks the substrates (A) which have double bonds to generate
oxidized products (A01),
The process of photosensitio>:ed oxidation can be minimized by quenchers,
which can quench singlet oxygen andlor excited sensitizers. The antiphotooxidative
effect of nickel chelates on the singlet oxygen oxidation of soybean oil have been
reported by Lee and Min (.5), Ascorbic acid and tocopherols are not only free radical
scavengers but also good singlet oxygen quenchers (6.7). len-Butyl hydroquinone
(fDHQ) as weI! as ot her phenolic compounds can also acl as singlet oxygen
quenchers (8-/2). Rosemary oleoresin and rosemariquinone, a rosemary ext ract, can
inhibit light-induced oxidation (13./4) . Among these quenchers, carotenoids
appeared to be the most plausible candidates for quenching the singlet oxygen. The
quenching mechanisms and kinetics of several carotenoids. including f3-apo-8'-
carotenal, Il-carotene, canthaxanthin, lutein, zeaxanthin, lycopene, isozeaxanthin, and
astaxanthin have been well studied (/5-17) .
Capsanthin (Figure I) is the most abundant carotenoid in the paprika spice.
The concentration of capsanthin is about 1590 mg per kg of dry mailer and 41 -
54% of total carotenoids in paprika ( /8,/9). The antiphotooxidative effect and
kintics study of capsanthin on soybean oil has not been well studied. In addition, the
study of the antiphotooxidative effect of carotenoids mostly focused on lipid or fatty
acids. The antioxidative elTect of carotenoids on the photooxidation of flavor
compounds has not been well studied, either.
This paper reports the antioxidative effect of capsanthin on the chlorophyll-
photosensit ized oxidation of soybean oil and selected flavor compounds including 2-
ethylfuran, 2,4, 5-trimethyloxazole, and 2,5-dimethyl-4-hydroxy-3(2H)-furanone
(DMHF), which are unstable when exposed to light in the presence of sensitizer
(3,10). The quenching mechanism and kintics of capsanthin on the photosensitized
oxidation of soybean' oil are also reported. Il-carotene and lutein were used as
cont rols for comparing the antiphotooxidat ive activity with capsanthin.
EXPERIMENTAL
Mat erials. Soybean oil was purchased from a local supermarket and purified in our
laboratory according to the method of Lee and Min (/5). Chlorophyll was extracted
and purified from spinach according to the method of Ornata and Murata (1/). 11-
carotene was purchased from Sigma Chemical Co. (St. Louis, MO). Capsanthin,
190 St>I CES; CII F ..MISTRY AND ANTIOX!I)ANT "KOt'ERTl E.."
Compound Structure
Capsanthln
BCarotene
Lutein
u
on
0 "
No. of conjugated
double bond
11
11
10
I. Structures of cap sa nIh in, J3-carotene, and lut ein.
coolant ol.ltl d +--
condenser
oxygen Inld ..
coolant lold
reaction n ask
light 5Oun-e
water light
Figure 2. Apparatus for the photosensitized mtidation of flavor compounds.
I S. CII IN AL Effect of Capsollthitl 011 P//IJlO(Jxidation oj Soybean Oil 191
which was extracted from paprika spice, and lutein were obtained from KALSEC
Co. (Kalamazoo, Ml). 2-Ethylfuran, 2, 4,S-tri methyioxazoie, and DMHF werc
purchased from Aldrich Chemical Co_ (Milwaukee. WI), Ab$Olute alcohol was
obtained from Pharmco Products Co. (BrOOKfield, CT).
Evaluat ion of the IIn li oxid.uj"t effect or carol enoids on the chlorophyll-
sensi liud phOlOOl idllion or 2-ethylrunn, 2.4,5-lrimel hylollzol e, lind DMHF
by 5pectrophotometric method. Mixtures of I x 10
03
M 2-ethylfuran, 2,4,5
lrimethyloxazoie, or DMHF in]OO rol absolute alcohol containing 10 ppm
chlorophyll and either 1.0 x or 2.5 x M carotenoids were prepared in 250
ml of flask. The antiphotooxidative effect of carotenoids including capsanthin and
lutein on these flavor compounds was carried out under light exposure at 20"C whi le
bubbling with oxygen It 260 ml/min according to the method of Chen lind Ho (3) as
shown in Fi gure 2. The light intensity at the sample posit ion was 67,000 lux. The
effect of carotenoids was detennined using I Hitachi U-31l0 spectrophotometer
(Danbury, Cf) in which the percentage of subst rate remaining was calculated by
measuri ng the decrease of optimum wavelength absorbance of test compounds. The
optimum wavelength absorbance of 2-et hylfuran, 2,4,S-trimethyloxa.tole, and DMHF
was 215, 225, and 29Onm, respect ively.
Evaluati on of th e antioxidative effect of carot enoids on the chlorOI)h), lI -
sensit ized photooJidati on of purilied soybean oil by the Rancimat meth od.
Samples of 100 or 200 ppm carotenoids in 2.5 g purified soybean oi l containing 200
ppm chlorophyll were prepared in 2.0 cm i.d. x 15cm glass open cylinders. The
cylinders were then placed in the light box, which was modified from the device as
reported by Lee and Min (15). The light intensi ty at the sample level was 10,000 lux.
The temperature of the light storage box was kepI al 25C. After 4 hrs storage in the
light box, the cyli nders were placed in a Model 679 Rancimat instrument (Metrohm,
Switzerland). The oxidative stabili ty of this soybean oil was measured by determining
the induct ion time at ]()()OC. The ai r supply was maintained at 20 IIhr.
Quenching mecbanisol and kinetics study of capsanthin by olygen depl etion
met hod. The quenching mechanism and kinetics of capsanthin were studied using
the steady Slate kinelic method of Foote and Denny (22) and Lee and Min (16).
Samples of 0.03, 0.06, 0.10 or 0. 16 M purified soybean oil in methylene chloride
containing 4 ppm chlorophyll and 0, 0.25 x 10-$,0.5 x 10-$,0.75 x lO-s or I x
M capsanthin were prepared. Fifteen mt of prepared sample was transferred inlo
30m! serum bottles and sealed air-tight with rubber septa. Duplicate samples were
placed in the light box for 2hrs at 25C and the percentage of heads pace oxygen was
analyzed using a Hewlett-Packard 5890 II gas chromatograph (Ge) equi pped with a
thennal conductivity detector (TCD) and a Hewlett -Packard )396A int egrator. The
GC column was a two-layer concenlric column (Alltech CTRI, Avondale, PAl; the
outer column was 6 ft x 1/4 in. o.d. packed with an activated molecular sieve; the
inner column was 6 ft x JIB in. o.d. packed with a porous polymer mixture. The fl ow
"
,

,
"
"
"
,
i
f
1
I
I
,
!
,
I
I
[
,
I
,
j
'I
i
,
l,j
I
;,

,
..
,
"
" i
192 SPIC.;S: HAVOR Cm :i\W,TRY ANI) A.NTIOXIDANT I'ROPER1H:S
rate of helium carrier gas was at 20 ml/min. The temperature of thc inj ection pon,
oven and detector wcre 50, 35 and 100C, respectively.
RESULTS AND DICUSSION
Antipholooxidative effect of ca rol enoids on 2-elhylfuran, 2,4,5-
trimethylolnole. and DMRF. The effect of 1.0 Ie iO-
s
or 2.5 x lO-
s
M capsanthin
or lutein on the photooxidative stabili ty of2-ethylfuran, 2,4,S-trimethyloxuole, and
DMHF in absolute alcohol containing chlorophyll was conducted under light
exposure while bubbling with oxygen. The percentage of substrate remaining was
determined using spectrophot ometry and was used as an index of ant iphotooxidative
activity of carotenoids. The gre<lter the percentage of substrate remaining, Ihe higher
the anlipootooxidative activity exhibited.
The ant ipholooxidative effect of capsanthin and lutein on the 2-ethyJfuran,
2,4,5-trimethyloxazole, and DMHF is shown in Figure 3. Both capsant hin and lutein
showed the antipholooxidative effect on these flavor compounds. The
antiphotooxldative activity of carotcnoids increased when the concentration
increased from 1.0 x 10-
5
to 2.5 x lO-
s
M . Comparison of the same concentration of
capsanthin and lutien, the anti photooxidative activity of capsanthin was higher than
lhal oflUlein.
AnliphotoOlidlitive elTed of carotenoids on the induction time ofsoybean oil a5
measured by the Rancimat method. The effect of 100 ppm or 200 ppm
capsanthin, p-carotene, or lutein on the pholooxidative stability of soybean oil
containi ng 200 ppm chlorophyll was measured by the Rancimat method after 4 hours
light cxposure at 25C. Preliminary studies showed that Ihe inducti on time of
purified soybean oil containing 100 Of 200 ppm carotenoids wi thout chlorophyll was
slightly lowef than Ihat of the blank sample which contained no carotenoids (data nol
shown). Therefore, there is no anti-autoxldative effect of carotenoids on the soybean
oil. Table I shows that as the concentration or the carotenoids increased from 100 to
200 ppm, the induction time as well as the anti-photooxidation index (API)
increased. The induct ion lime of soybean oi l containi ng the carotenoi ds was longer
than that of the control sample which containcd no carotenoid; however, it was still
shoner than that of the blank sample which contained no carotenoid and no
chlorophyll .
Table II shows that as the number of conjugated double bond increased from
10 to II, the API of soybean oil increased at the concentrations of 100 and 200 ppm
carotenoids. It has been reponed that as the number of conjugated double bonds of
carotenoids increascd, the peroxide values of chlorophyll-sensit ized photooxidation
of soybean oil decreased significantly (16.1 1). Therefore, the si nglet oxygen
quenching abi lity of carotenoids was dependent on the number of conjugated double
bonds of the carolenoids as reported by Lee and Min (16), Jung and Min (17). and
Hirayama et a!. (23).
Table II also shows that the soybean oi l containing capsanthin had a higher
API than the oil containi ng j3-carotene ahhough both of the carotcnoids have the
L5. cm:N ET At... EJ/tcl o{ on Pho(ooxiriation o{ Soj'briJn Oil
"3

II' 100
no chlorophyll and


carotenoid
:
80
c

e
'"
capunthln ( 2.S J: 10-
5
M)


lutein (2.5 x 10.
5
M)

40

capsanthin (1.0/10.
5
M)


20 lutein (1.0 J: 10' M)
.Q
chlorophyll only ,
'"
0
0
,
3 4 5
Time (min)
II'
100
no chlorophyll and

carotenoid

.5
:

80
capsanthln (2.5 J: 10.
5
M)
e lutein (2oS J: 10.
5
M)


.0
capsanthin (1.0 I( 10.
5
M)


lutein (1.0" 10J 1\1)


40
(B)
.lI ehlorophyll only
,
'"
20
0
,
3 4 5
Time (min)
II'
'00

no chlorophyll
=
and carotenoid
c
80
.;
e
"
capsanthin (2. S x 10.
5
M)



.. lulein (2.S I( 10.
5
M) 5

capsanthln ( 1.0 J: 10' 1\1)


"
(C) lutein (I.0 I( 10.
5
1\1)
.Q
, chlorophyll only
'"
0
0
"
12'
""
Time (min)
Figure 3. Antipootooxldative effect of capsanthin and IUlein on (A) 2-
et hylfuran, (8) 2,4,S-trimethyloxazole, and (C) 2,S-dimethyl. S- hydroxy-3(2H}
furanone (OMHF) in absolute alcohol conlaining \0 ppm chlorophyll at 20 C.
I
19'
S I ' I C ~ ; S : n AVOR CHEMISTRY AND ANTIOXIDANT " ROI't; RTl ES
Ta ble I. Effect of carotenoids on the induction time detected by the Rancimat
method of soybean oil containing 200 ppm chlorophyll during light
exposure for 4 hours.
Carotenoid
(ppm)
Control)
Capsanlhin (100)
Capsanthin (200)
Blank)
Controf
p.Carotene (100)
pCarotene (200)
Blank
1
Cont rol
l
I.utein (100)
Lutein (200)
Blank}
Induct ion lime OT)I
(hIs)
7.96
8.94
9.64
11.20
7.96
8.90
9.60
11.20
7.96
8.74
9.39
11 .20
Ant;-phOlooxidation index
(API)I
0.303
0.518
0.292
0.505
0.241
0.440
1. API - (carotenoid IT control IT) I (blank IT - control IT)
Each value is the means of duplicates
2. soybean oil + 200ppm chlorophyll
3. soybean oil only
Ta ble U. Effect orthe number of conjugated double bond or carotenoids
on the anti -photooxidation index (API) of soybean oil during light
exposure.
Carotenoid
(100 ppm)
Lutein
pCarotene
CapsBnthin
(200 ppm)
Lutein
!3Carotene
Capsant hin
Number of conjugated
double boud
\0
I I
II
\0
II
II
API
0.241
0.292
0.303
0.440
0.505
O.S Ig
15. ell EN ET At. Effect o/Cnpllonlhin on I'Jlf)tooxidation of Soybean Oil
195
same number of conjugated double bonds. It has been reported that the singlet
oxygen quenching ability of carotenoids was determined not only by the number of
conjugated double bonds.. but by the functional groups of the carotenoid (23). As
reported by Uirayama el II (13), conjugated keto groups and the presence of a
cydopenlane ring in the carOienoids stimulate quenching efficiency. Comparison of
the struclUre of capsanlhin wilh that of Il-carotene reveals that there is one
conjugated keto group and one cydopentane ring in the SlNClure of capsanlhin but
!}-carotene comains neilher of them (Figure I). Therefore, the anliphotoo)(;dat ive
act ivity of c.apsanthin is higher than that of p-carotene, as shown in Table II This
result confirms Hirayama's prediction.
The Rancimat method has been used to measure the antioxidant activity of
synthetic and natural antioxidanu (U.}6) and has correlated well with oil stability
measured by the Active Oxygen Method (17) and peroxide value measurement (28).
Our study showed that using the Rancimat method to study the antiphotooxi dative
effect of carotenoids on the soybean oil was in agreement with the results using the
headspace oxygen depletion method (16) and the peroxide value method (16,1 7) .
Quenching mecha nism a nd kinet ics of caps!lnt hin. The following steadystate
kineti c equation would be established if carotenoids reduced the chlorophyll
photosensitil:ed singlet oxygen oxidation by singlet oxygen quenching (4,22).
{.d[02Vdt}t - {d[A0211dt}\
- K'\ (1 + (kq[Q] + kox-Q[Q) + kd)l1:r{All
A02 : oxidized soybean oil
K . rale of sing let oxygen formation
kq : reaction rate constant or physical singlet oxygen quenching by capsanthi n
Q . capsanthin
kox-Q : reaction rat e constant or chemical si nglet oxygen quenching by carotenoid
kd : decaying rate constant or singlet oxygen
kr : reaction rate constant or soybean oil with singlet oxygen
A . soybean oil.
The intercepts and slopes of the plot s of (-d[02Vdt)-\ vs [AJ-l at various
concentrations of quencher (Q) are K'\ and K-\ {(kd + kq[QJ + kOX-Q[Q}/krl,
respectively. The plot of SQlSo (slopes in the presence and absence or quencher) vs
[Ql is a straight line, and the slope or the straight line is (kq + kOX-Q)/kd (4,}9.30).
The rate constants (kd) of singlet oxygen decay in different solvents have been
repon ed (31). Therefore, the total singlet oxygen quenching rate constant (kq +
kOX_Q) or quencher can be determined rrom the slope of the plot of SQlSo vs lQl
[4].
The plots or(-d[021/dt)- L vs [soybean oil)-\ at di fferent levels orcapsanthin
are shown in Figure 4. The intercepts were the same at different levels or the
capsanthin, but the slopes or the plots increased as the concentration or capsanthin
increased from 0 to 1 x 10-
5
M. This resul t showed that capsanthin as well as other

'I
"
"
!
,
,
.,
,
" f
, ,
-1
: I
."
f
"
I
196
5
-

"
'-
c
3
Ji



c

2



-
1
5 1'l eKS: FlAVOR AND ANTIOXIDANT I' WOI'ERTn:s
no c:ap llml hin
0.25 1l105M
030 x IO5M
0.15 x IO5M
Q 1.00 x IO-5
M
10 20 30
1/011 in met hylene chloride ( 11M)
'0
Figure 4. Effect of capsanthin on the headspace oxygen depletion of soybean
oil in methylene chloride containing 4 ppm chlorophyll when exposed to light
for 2 hours at 25 0c.
7
..


y .: OM' + 5.7"'''
5

J!
- 3
<2
2
1
0
0.0 0.2 0.' 0.' 0.8 1.0 1.2
Capsanthin (x lO'5 M)
Figure S. Plot 0( 5'30 vs {capsanthin}.

IS. c m:N ET Effect oj CapSll rllhin 011 Photooxidatiotl ojSo),bean Oil 197
carotenoids quenched the singlet oxygen but did not quench the excited triplet state
of chlorophyll to reduce the photosensi tized oltidation of soybean oil (4,5, /5,/6).
To measure the 10lal singlel oxygen quenching rate constant (kq + kox-Q) of
capsanlhin, the regression line of setSo V5 (capsanthin] was plotted as shown in
Figure 5. So and So are the slope! of the plot of (-d[02ydl)"1 vs {soybean oi l]- I in
the presence and absence of capsanlhin, respcetively. As shown in Figure 5, the
slope of the plOI ofSQfSo vs {capsanthin1 is 5.746 x ,lOS M-l . Since the slope of the
plol ofSQfSo vs [quencher] is (k
q
+ kOX-Q)lkd (11.30) and the kd value of singlet
oxygen in methylene chloride is 1. 0 x 1()4 s-I (3/), Ihe total singlet oxygen quenching
rate constant (k
q
+ kox-Q) of capsanlhin is 5.746 x 10
9
M-Is-t in methylene
chloride.
CONCLUSI ON
II has been known that carotenoids are able to quench singlet oxygen to reduce the
photosensitized oxidation reaction (/6./7) . Our study showed that capsanthin, a
major carotenoid from paprika, as well as p-carotene and lutein exhibited the ability
to quench singl et oxygen to reduce the photosensitized oxidation of both soybean oil
and selected flavor compounds including 2-et hylfuran. 2,4.5-trimethyloxazole, and
DMHF. These results suggest that capsanthin as well as ot her carotenoids may be
applied to food system which contains food lipids or flavor compounds to minimize
food photodeterioration.
REFERENCES
I.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11
12.
13.
Rosenthal, I. in Singlet Vol IV Polymers alld Biorna/ecule.f. Frirner, A.A.,
Ed.; CRC Press: Florida, 1985, pp. 145-163.
Samuel, D.; Steckel. F. in Molecular Oxygen ;11 Biology, Hayaishi, 0 ., Ed.,
Elseviu: New York, 1974, I.
Chen, C-W.; Ho, C-T. J. Agric. FoodChem. in press.
Foote, C.S. in Singlet Orygell, Wasserman, H.H. and Murray, R.W., Eds.,
Academic Press: New York, J979, Pl'. 139-171.
Lee, S-.H., Min, D.B. l . Agric. FoodChem. 1991 ,39,642-646.
lung, MY. ; Choe, E.; Min, D.M. l . Food Sci. 1991,56,807.
Chou, P-.T.; Khan, A.U. Biochcm. Blophys. Res. Cornm. 1983, 1/5, 932-9]7.
Sherwin, E.R. in Food Aciditil'eS. Branen, P.M.; Davidson and Salminen,S.,
Eds., Marcel Dekker, Inc; New York. 1990, p. 139.
Clough, R.L.; Yee, B.G.; Foote, C.S. l. Amer. Chem. Soc. 1979,101,83.
Matsuura, T. ; Yoshimura, N.; Nishinaga, A.; Sai to, 1. Te/ra. Lett. 1969,2/,
1669-\671
Thomas, M.l .; Foote, C.S. Pno/ochclII. ,' n%hiol. 1978,27,683-693.
Tournaire, C.; Croux, S.; Maurette, M-.T. 1. Photochem. Ph%biol. 1993, 19,
205215.
Hal111I, c., Cuppett, S. l . Amer. Oil Chem. Soc. 1993, 70,477. 482.
I
,98 SI' ICF..S, FlAVOR CHEMI!o.'TRY AND ANTIOXmANT I' IW I'ERTIES
14. Hall III, C.; Cuppel!. S.; Wheeler, D., Fu, X. J. Amer. Oi/ Chem. Soc. 1994, 71,
533-535.
15. Lee, E.C.; Min, D.B. J. Food Sci. 1988,53, 18941895.
16. Lee. S-.W.; Min. O.B. 1. Agrie. FoodChem. 1990,38, 1630.1634.
17. Jung, M.Y.; Mjn, D.B. J. Amer. Oil Chem. Soc. 1991, 68, 653-658.
18. Mlnguu-Mosquera. M.I., Homero-Mindez, D. J. Agrie. FoodChem. J994, 42,
15551560.
19. Biaes, P.A. ; Daood, H.G.; Huszka, T.T.; Biaes, PK 1. Agrie. Food Chem.
1993, 41,1864-1867.
20. Chen, C.-. W.; Shu, C.-K.; Ho, C.-T. 1. Agrie. Food Chem. in press.
21. Ornata, T.; Murata, N. Pho/achem. Ph%biol. 1980, 31,18]-185.
22. Foote, C.S.; Denny, R.W. 1. Amer. Chem. Soc. 1968,90, 6233.6235.
23. Hirayama, 0 .; Nakamura,K. ; Hamada, S., Kobayasi, y , Lipids 1994, ]9,149-
ISO.
24. Chen, C.-W.; Ho, C.-T. 1. Food Lipids 1995, 2, 35-46.
25. Zhang., K.Q.; Bao, Y.; Wu, P., Rosen, R.T.; Ho, C.-T. 1. Agric. FoodChem.
1990,38, 1194- 1197.
26. Ho, C-.T.; Chen, Q.; Shi, H. ; Zhang, K.Q., Rosen, R.T. Prel'ellliVl! Med. 1992,
21,520-525 ,
27. Laubli, M.W.; Bruttel, P.A. 1. Amer. Oil Chem. Soc. 1986, 63, 792-795.
28. Gordon, M.H.; Mursi , E. 1. Amer. Oil Chern. Soc. 1994, 71, 649-65 1.
29. Foote, C.S.; Ching, Ta-Yen.; Geller, G.G. Pho/ochern. Ph% biol. 1974,20,
511-513.
)0. Yamauchi, R.; Matsushita, S. Agric. Bioi. Chern. 1977,41, 1425- 14) 0.
) I. Hurst, J.R. ; McDonald. J.D. ; Schuster, G.B. 1. Arner. Chem. Soc. 1982, 104,
2065-2067.

".'
Chap'er 16
Curcumin: An Ingredient that Reduces
Platelet Aggregation and Hyperlipidemia,
and Enhances Antioxidant and Immune
Functions
Vaguaog Uu
1.. Y. Resea rch Corporat ion, 67- 08 106A St reet, Flushing, NY 11365
The effects of curcumin on blood platelet aggregation,
hyperlipidemia, lipid peroxidation and immune function wcre
determined. Curcumin exhibi ted st rong inhibition of blood
platelet aggregat ion in humans and lipid peroxidation. by
adenosine diphosphate (ADP) of rau The marked nse til serum
Chole5tcrol levels in rats following 4 weeks of cholesterol fceding
was significantly reduced by curcumin. Addi tionally, curcumin
showed that it increased immune function, including macrophage
rate and Iymphoblastoid transformation in imlnunosuppression
rats induced by cycl ophosphane (CY). The above pharmaceutical
effects of curcumin with addition of flavone of Malricaria L.
were stronger than curcumin alollt. It can be seen that
effectively prevents cardiovascular disease, increases Immune
function and exhibits anti-oxidant activity.
Curcumin is isolated from the rhizomes of the plant , CllrCl/flla lal/ga L. which is a
medicinal plant widely used in China, India and Sout heast Asia. The rhi zomes of
the plant has been used in China as a reduction in serum hyperlipidemia li nd
stomachic drug (I) . . . .
Curcumi" has demonstrated anti_inflammatory and anll-oxldant properties and
it is essentially non-toxic (1-6). Recent studies have illdicated that a group of four
ingredients, which include cureumin, scoparone, ferulic acid and flavone . of
Malricaria L. has demonst rated marked hyperlipidemia and platelet-aggregatIOn
reducing propert ies (7). . .
This paper describes the effects of curcumm and fl avone of M. L.
on platelet-aggregation, hyperlipidemia, anti -oxidant and immune funcllons.
C 1997 American Chemical Society
, ,
"
I
200 SI'I Ct:s, HAVOK AND ANTIOXIDANT "KOI' EKTI ES
Materinls
Curcumin was extracted from C. 10llgel L. according to the following steps.
Rhizome of C IQ"ga L. o r C. aromaliea salish or C. zeJoaria RQ5CO was
driC<l and powered. One kilogram of the powder was soaked in 5 liters of 95%
ethanol, fOf 12 hoors at room temperature_ The resulting ethanol extract was
fiitered. The et hanol was then recovered under reduced pressure di stillation. A
was dissolved in 200 ml of I N NaOH. The resul ting solution was adjusted
to pH 7 wit h I N HCI and a light ycllow precipi tate was formed. The precipitate
was dissolved in 300 ml of 95% ethanol and the last procedure was repeated to
form a light yellow precipitate again. The light yellow preci pitate was washed with
acetone and ether successively and then dried under vacuum and was found to have
a melting point at 18) C.
Flavone was e:o: tracted fW1l1 M. chamQmilla L. according to the following
steps.
Plants of M. dutll/omillo L. were dried and powdered. One kg of the powder
was e:o:tracted in 2 liters of 95% ethanol for about 24 hours at room temperature.
The soluti on was fi lt ered afld the filtrate saved. Two lit ers of 950/. ethanol were
to the residue and renuxed in a water bath for 6 hours. The renuxed ethanol
was cooled and filtered and the filtrate combi ned wi th the ext ract fiitrate. Ethanol
then recovered by reduced pressure distillation and the residue was saved. One
thousand ml of acetic ether was added to the residue and rcnu:o:ed in a waler bath
for 6 hours. The renuxing procedure was repeated. Aceti c ether W3$ then
concentrated under reduced pressure distillation. Crystals were formed which were
then washed wilh waler. The final crystals were dried under vacuum and were found
to have a melting point of about 25f1 C (7).
C. 10llga I . aod M. chanlt)milla L. are recognized by Food and Drug
administration (FDA) as safe for human consumption.
Methods
mood Ilhutiet Aggregati oll . The blood platelet aggregat ion lest was performed
accordi ng to the method of 130m et al. (8). Blood was collected from vei ns of
humans using a needle attached to a plastic disposal syringe. The blood was
immediately transferred to a siliconized glass tube containing 0. 1 volume of 3.130/.
soohlm citrate. Platelet-rich plasma (PR!') was oblained by centrifugation of the
whole blood at 1000 rpm for 10 min at room temperature_ Pl atelet -poor plasma
(1'1'1') was prepared by centri fugation of the remaining blood at )000 rpm for 10
min. Pl atelet aggregation was performed at 37e. Human platelet studies were
carricd 0 11\ wil h a COnstant platelet number () x 10' Iml ). With regard to
detcflllillatior1 of plAtelet aggregation. maximum aggregation was induced by 0.2
I-\ M (ADI'). 0,4 Ill I PRP was added into each tube. Thirty six tubes were divided
into three groups. Fi fiy III of saline was added 10 each tube of control group, 50 1-\1
of curculllin (0.5 lIIg/ml) was added to each tube of the crucumin (t reatmE:nt I)
16. U U Curcumin: Platelet All1JT"f8atiQn d: flyperlipidemia
20'
group. and 50 of curcumi n (0.45 mg/ml) + 5 III flavone of M. chanwmilla I .
(0.05 mg/ml) were added to each tube of w rcumin plus flavone of M. chamomilla
L. (treatment 2) group. After incubation for 3 minutes at )1' C., 50 I.d of2 M ADP
was added 10 each tube. A five minute's aggregation GUrvc for eath lube was
plotted.
Percent inhibition of aggregat ion by PHP was calculated by:
% in control % IUn:gal'OQ ,..il ll tmtIUlC'"
%i nhibilionofagrrgalion - ---- ------------- 100 %
% aggregation in control
Determination of Lipid Per"Oxidation in Mitochondria and Mkr"050mt!J. Rat s
were killed by decapitation afier fasting for 24 hours and their liver tissue was
quickly removed. Microsomal and mitochondria fract ions were isolated from the
liver ti ssue by the method ofOda (9). The liver tissues wefe cut into small slices in
0.25 M sucrosc containing 3 mM Tris-HCI and 0. 1 mM EDTA (pH 7.4) at 4C, and
then small slices of liver ti ssues were homogenized with nine limes (Ihe amount by
weight ) 0.25 M sucrose solulion conlaining 3 mM Tri s- HCI and 0. 1 mM EDTA (pH
7.4) at 4 C. The homogenate solutions were adjusted to pH 7.4 by the addition or
o. t N KCI, and then the homogenate was centrifuged at 50 x g for 10 minutes al 4
D
C to remove the nuclear fractions and red blood cell s. The supernatant phase was
centri fuged al 700 :0: 8 for 10 minutes at 4C, and then further centri fuged at 5,000
x g for 45 minutes at 4
0
C to give mitochondrial fractions. The isolated
mitochondrial fract ions were washed twice wit h Krebs- Ringer phosphate buffer (pH
7.4). The supernatant phase centrifuged at 5,000 :0: 8 for 45 minutes was further
centrifuged at 24,000:0: g for 10 minutes 8t4 C. The precipitation was removed and
the supernatant phase was again cent rifuged at 54,000 x g for 60 minutes at 4' C to
obtain microsomal fractions. The isolated microsomal fractions were washed twice
with Krebs-Ringer phosphate buffer (pH 7.4). Mitochondria and microsomes were
suspended in Krebs-Ringer phosphate bulTer (pH 7.4). The final protein
concentrations of the mitochoodrial and microsomal suspensions were adjusted to
to mg protein/ml.
A mi:o:lure of mitochondrial suspension (0.5 ml, \0 mg protein/ml), Kerbs-
Ringer phosphate bulTer conlaining 40 mM ADP, 12 mM ascorbic acid and the
indicated amoullt of curcumin and flavone of M chamoll/illa L. were incubated at
31' C for 15 minutes in a final volume of I ml, In the treatment I group, buffer
contained 5 :0: 10-4 M curcumi n. In the treatment 2 group, buffer contained 0.45 x
10.
4
M curcumin and 0.05:0: 10-4 M flavone, Lipid peroxides of mitochondria were
determined by the method ofOhkawa ct al. (10).
Reducillg Hypcrlipidemi a Ted. Young male Wislar rats with an average weight
of 130 g were fed 1I0rmallaboratory dict and were housed individually in cages.
Room temperature was controlled at 25 l a C with 60"10 relati ve humidity. Li ghting
was controlled with cycl es of 12 hours light followed by 12 hours of dark.
202 SI'!CES: n,AVOR CIIEMISTRY ANIl ANTIOXIDANT PROl'ERTIFS
Fony male rats were separated into four groups. The normal gTOOp was given II
regular laboratory diet . Cont rol and treatment groups 1 and 2 were given II regular
diet supplemented with 1% cholesterol and 0.2% eholie acid. A restricted diet (8g)
was prescribed every morning. Curcumin (dose 100 mglkglday) was dissolved in
di stilled water (2 ml/rat) and was orally administered to rats of Irealmenl 1 group.
Curcumi n (dose 90 mg/kg/day) plus flavone (dose 10 mglkglday) were dissolved in
distilled water (2 O1ll(81) and was orally administered to rat s of treatment 2 group.
After 4 weeks of feeding. the ralS were starved J 8 hours, then sacrificed. Blood was
collected from Ihe jugular vein. The blood and liver were removed immediately and
portions of the tissue were exami ned to detcRTline cholesterol, triglyceride and total
lipid.
Lipid Conte'lI l . Total lipid contents of the' liver and serum were determined
by the methods of Sperry and Braoo (IJ). Total cholesterols were determined by
the methods of Carr and Drekter (11), frcc cholesterol by the method of Sperry
(13), and tri glyceddes by the method of Van handel -Zil versmit (14) .
Immune Flilielioli. Male mi ce weighting 18-20 g were used in the experiments and
were divided into treatment 1,2 and control groups. f or treatment I group, the
dosage or curcumin was 5.5 mglkg injected intrapctitoneally. For treatment 2
group, the dosage of curcumin was 5.0 mglkg and 0.5 mglkg flavone. The mice of
control group were injected with the same volume of normal saline. These
inj ections were repeated dail y for 7 days. On the last day, bot h trealed and control
groups were injected intraperitoneally with CY. The dosage ofCY was 4.5 mg/kg.
On the seventh day, the animals were killed by decapitation. The blood was
removed immediately and was examined to determine immune fimelion.
Lymphoblastoid Transrormlltion Test . Lymphoblastoid transformation test was
performed using the method of Hogan (15). Cells (J x 10'/ml) were maintained in
Eagles minimum essent ial medium containing 10% fetal bovine serum (FIlS) and 10
pglml polysaccahride-free purified phytohemagglutinin (PIIA.P). The cells were
incubated at J7oC/5% CO, 195o humidity and the medium changed every day.
I\-bcrophllge. The macrophage cells were cultured in PPM.I- I640 medium with
10% FBS. Aller overnight incubation at 31' C in an aunosphare of 50/. aoo
95% air, non-adherent cells were washed away. The adherent cell s were funher
incubated with 4 rul of the medium for 24 hours at 31' C. Antigens were added to
macrophage monolayers, After 6 hours incubation at 31' C, the monolayers were
washed four times to remove antigens and immediately used for experiments.
Macrophage cytotoxicity was measured as previously described (16). Results are
expressed as a specific percentage of !lCr release (percentage cylotoxicity) as
calculated by the following formula: percentage specific cylotoxieity - 100 x
(experimental C.p.m. - spontaneous c.p.m.)I(tolal c.p.m. -spontaneous c.p.m.) .
16. U U Curcumin: Platelet Aggregation & Hyperlipidemia 203
Result! Alld Discussion
The significant inhibitory effect of curcumin on plat elet -aggregation is summarized
in Table I below. Curcumin with flavone of M. dUll/lOll/ilia L. appeared to be
more effective than curcumi n alone. Platelet aggregation and subsequent adhesion
under the influence of local fl ow irregularities are primary and injury of the intimal
tissue is secondary (17) . Organ damage and dysfunction may be a consequence of
platelet awegati on and embolizat ion of plat elet thrombi in micro-circulation
Aggregation of platelets might cause damage result ing in obst ruction of vessels.
Intermittent infusions of adenosine diphosphat e int o the coronary circulation of pigs
caused circulatory collapse, electrocardiographic evidence of myocardial ischemia,
ventricular dysrhylhmias and formation of platelet aggregat es in the microcirculation
(18, 19).
Table I. Effects or CurCUlllin on Aggregation or Platelet
Rate of asgregati on of platelet
Control
Treatment 1
Treatment 2
67.6 5.0
26.S 0.8
24,7 0.7
Each value represents the mean of 10 animals S.E.
P<O.OI, significantly different from the normal group.
Treatment I: Curcumin only.
Treatmelll 2: Curcumin plus flavone of Matricaria c/llUIlomilla I ..
As shown in Tabl e II, the blood cholesterol level in rats fed I % cholesterol
was elevated 21:14% and supplementati on of curcumin and curcumin plus flavone of
M. chamomifla L. to the cholesterol diet depressed this elevation significn!l!l)'
Phospholipid levds showed no significant change. The addition of flavone or At!
challlomil/a L, appeared to be more effective than the curcumin-only group.
TAble II. ElTect OfCurc:ulllin 011 Cholesterol Levels in Blood
Diet Cholesterol
(mgldl) (mgldl) (%)
Normal 125 6 201 5 62
Cholest erol 4S0 40 230 9 209
Treatment I 2 10 10 20S 6 101
Treatmenl2 176tS' 195 5' 90
As shown in Table III , total lipid levels, free and total cholesterol, and triglyceride
levels of the liver decreased significantly when curcumin and navone of M,
chamolllilla L. wefe added to the hypercholesterolemic diet but free cholesterol in
the liver appeared no signi ficant change.
.,
i
,
!

,
I
\
i
r
I
'. ,
t.

i
!
.
. ,
,
I
,
f.
,.
,

r
I
,
;.
204 SI' IC,.:.s, FlAVOR CII EMJI'I"TRY AND ANTIOXIOANT
Tlll)lc lIJ. [ fTecl orCurcumin on Li pid uvd
Per 100 g li,"er li$SUe
lipid Total cholCSlcroi Flee cholesk:roI
III ( lng) (mg)
Norma! 7.!I.t I.O j I6 J:90 lJl.i21
Choleslerol 17.1:1:2. 1 2472:1:151 327.t26
Tlc-llmenl I 9.0 :to.S' 1101 :1:92 ' 291 :i2l'
Trc.,IIlICUll 8. 1 :t0.8' 921 :1:80' 280:1:24'
Each value represents the mean of 10 samples S.E.
Trigl)Wridc
(mg)
397:1: 2S
j20:t 27"
" !II :tlS"
"]].ill '
'I' < 0.05, signi ficantly different from the normal group; .. P < 0.01
Reccnt clinical studies have shown that the plant C. lUI/go L. significantly reduces
blood lipid (I). As shown in Table III , curcumin can also significantly reduce blood
and liver cholesterol levels. Therefore, these studies suggest that curcuntin may be
the active ingredient responsible for the lipid-lowering effects of the plant C. IVI/ga
L
Table IV li sts Ihe lipid pero:o.:idlltion in rat liver mitochondria. Treatment I and 2
significantly reduced the level of pero:o.:idation. Pero:o.:idation of li pids in the liver has
been shown to result in many kinds of to:o.:icity such as of membrane function and
damage to membrane bound enzymes (20).
Tallie IV, Effects or Cur<: umin on ADP plus Ascorbi c Acid-induced Lipid
PeroxidAl ion in Rllt Liver Mitochndria
Group
Normal
Treatment I
Inhibition
Treatment 2 69.1% 5.0
Each value represents the mean of I 0 samples S.E.
P<O.OI, signifi cantly different from the normal group.
CY is a very st rong immune inhibitor. It inhibited phagocytic rate to 77% and
lymphoblastoid to in rats. As shown in Table V and VI, curcumin and
curcumin with navone of M. chamvlI/if{a L. can significantly increase immune
function which has been markedly decreased by CY.
Tllble V. Effects orCurcumin on MacrOllhll ges
Group
Control
Cy
CY + Treatment I
CY + Treatment 2
Phagocytic rate SO (%)
35.10 2.01
8.00 0.36
21 .00 1. 0.20
23.20 0.20
16. Ll V
CllTCUmi,.: Platelet Aggregation & I/)perlipitlf'mia
Table VI. EIT ! or Cul"(: unlin 011 LY"l pholllasloili Trl'lliSrormRlion
Group
Control
Cy
CY + Treatment I
CY + Treatment 2
CPM SD

697 38
905 70
985 75
Each value represents the mean of 10 samplcs S.E.
' P<O.OI. $igniflcanlly different rrom cydophosph.unide group.
,.,
On the basis of the above results, we conclude that curcumin may protect
against cardiovascular disease, increase immune function and has anti-oxidant
activity.
Litenture Cited
1. Xue C.S. Pharmocology and application of Chinese Medicine; Wang Y. Ed.;
People Health Published House: Beijing, 1983; pp 849-850.
2. Shanna, C.P. Bioc. Phann. 1976,25,181 1-18 12.
3. TOOa, S., Miyase, T.; Anchi, H.; Tanizawa, H. ; Takino, Y. Chern. Pharm. Oul L
1985,33,lnS-I728.
4. Srimal R. C.; and Dhawan, B.N. J. Pharo Pharmacol. 1973,25,447-452.
5. Rao, T. S.; Basu, N.; Siddiqui, H.H. Indian J. Med. Res. 1982,75, 574-578.
6. Mukhopadhyay, A.; Basu, N.; Gujral, P.K. Agents and Act ions 1982,12,508-
SIS.
7. Yaguang Liu. U.S. patent 1#4,842,859, 1989, pp 2-3.
8. Born, F.V.R; Cross, MJ . J. Physiiol. 1963, 158, 178.
9. Ohkawa, H; Crush;, N; vagi , K. Biochcm. 1979,95,35 1.
10. Oda., T; Seki,S. J. Elec. Mi cro. 1965, 12,210.
II . Sperry, W.M.; Brand, F.C. J. Bi oi Chern. 1955, 213,69.
12. Carr, J.J.; Drekter,I.J. Clin. Chern. 1956,2,353.
l3. Sperry, W.M. ; Webb, M. J. Bioi Chern. 1959, 187,97.
14. Vanhandel, E.; Zilversmit, 0 .8 . 1. Lab Clin. Med. 1967, 50,152.
IS. Hogan, M.M.; Vogel, S.N. J. lmmuno. 1987, IJ 9, 3697-3703.
16. Schwartz, RH.L., Jackson 1..; Paul, W.E. J. Immuno!. 1975, 115, 1330.
17. Jorgensen, 1.. ; Haeren, lW.; Moe, N. Throm. Dial. Haem. 1973,29,470.
18. Jorgensen, L.; Rowosll, H.G.; Hovig, T. Lab Invest. 1967,17,616.
19. Moore, S.; Merserrcau, W.A. Art Pathol. 1968,85, 623.
20. Rice-Evans, C., Hochstein, P. Biochem. Biophys. Res. Comm. 1981, 100.15)7.

Chapter 17
Antioxidant Activity of Lavandin
(Lavandula x intermedia) Cell Cultures
in Relation to Their Rosmarinic Acid
Content
T. 16pcz-Arnaldos
l
, J. M. Zapata
1
, A. A. Calderon',
and A. Ros Barcelo!
lIkpartment of Plant Biology, Universi ty of Murcia,
E- JOI OO MUfcia Spoin
' D '
eparlment or Plant Biology, University IIf Al ca la de H
enn res,
E- 28871 Alcahi de Henares, Spain
AllIioxidants are onc oflhe principal ingredients thaI protect food quali t
oxidative deterioration ofbiomole<:ules. There is a
!nl,erest m. the appli,c8tion of antioxidants from plant material. herbs and
SpiCes. SPICes provide one oflhe most promisi ng sources of an Ii oxidant
compounds. Extracts from the Lamiaceae and Boraginaceae families have
be,en to be anlioxidams. The antioxidant activity is
pnm,anly to phenolic compounds that functio n as free radical
In some cases, also chelators. The present study
was u.ndenaken to assess the antiOXidant act ivity of extracts from
x illttrmedia) cell cult ures. We have compared the
supcroJOde arnon scavenger activity of rosmarinic acid (main constituent
of lavandin eJttracts) wi.th that di s.played by other structurally-related
compounds. Finally, the Iron-chelaung propenies ofrosmarinic acid are
reponed.
The food industry routinely anlioxidants such as DBA, DHT and TBHQ
to products from OXidative deteriorat ion. Although these antioxidants are
effechVC, sta.ble. and inexpensive, there is concem about potent ial adverse effects
antiOXidants (J). Toxicological data suppon the safety of the synthetic
antlOludants wh.en at use levels. Public perception that sYnlhetic compounds
are dangerous IS dnvmg the mdustry to search for "natural" antioxidants
Traditionall.y spices have been to improve flavor and of food. One
of preservation by plant extracts IS the antioxidant activity whlch protects foods
against. off fl avor related to lipid oxidat ion. The antioxidative activity found in many
plants IS due to the presence of compounds of a phenolic nature (I).
C 1997 American Chemicat Stlciely

17. UWE'l .... ARNALDOS .... Antioxidant Activity of LavQlldin Cell 207
Phenolic antioxidants function as free radical tenni nators and in some cases as
metal chelaton. Phenolic acids panicularly have been identified as potent antioxidants.
Caffeic acid and its esters are good examples of phenolic antioxidants (2).
Rosmarinic acid (a-O-caffeoyl-J,4-dihydroxyphenyllactic acid) (Figure I) is an
example of a caffei c acid ester occurring in plants. Rosmarinic acid is mainly found in
species of the Doraginaceae and Lamiaceae fami lies, but can also be detected in other
families (i .e. Apiaceae), ferns and homwons (Table I) (3,4). This suggests that the ability
10 synthesize this caffeoyl ester may actually be widespread as evidenced by rosmarinic
acid accuOlulation in a range of species (Tab!e I) (5).
Current research on rosmarinic acid centers on its physiological and pharmaco-
logical act ivities, and it is regarded as a potential pharmaceutical plant product (5). So,
rosmarinic acid has been shown to suppress the complement -dependent components of
endotoxin shocks in rabbits and t he oxidized compound displays antithyrotropic activity
on human thyroid membrane preparations (5). Funhermore, there is increasing concern
about the suitability of rosmarinic acid as a food additive.
One approach to recovery of uscful plant metabolites such as rosmarinic acid is
to extract plants as they are currently produced. Unfortunately the agronomic conditions
for production of such plants are extremely variable. Unreliable cultivation, htuvesting,
shipping techniques, infestat ion with pests and natural status of the plant can all impart
thC:' ]ost of the product produCi!d by the plant (6). Changes in any of these condi tions clln
adversely influence the production of compounds such as rosmarinic acid. The
application of modern culturing of plant cells offers an alternative means of production.
Condi tions of culturi ng can be optimized such that a cdl culture would produce
rosmarinic acid on a continuous basis at high yield and in a foml where it can b..: eusily
extracted (7).
Occurrence ofRos mari nic Acid in Lavandi n Cell Cullurn
Cultures have been successfully established from a number of plllnt speci..:s which
accumulate rosmarini c acid in their tissues. However, only a few of these culture.< hnve
demonstrated the ability to continuously produce the ester ill I'ilro (5). In 1994. 1_1' !'C7.-
Amaldos et al. (8) established cell cultures derived rrom leaves lind spikcs oflilviltltlin,
a sterile hybrid between i..,m'01xh,l" mlf:!lslijolia and La1't:mdll/(l illU!rmedia. Microscupic
analyses of the cells forming the callus showed the presence ofbille fluorescencc in
vacuoles when the preparations were observed under l lV light. This fluorescence is
indicative of the presence ofrosmarinic acid in the vacuoles. This result is in agreement
with those of Uiiusler et al. (9), who have recently the accumulation of
fOsmari nic acid in vacuoles of suspension-cul tured cells of COII:II$ by protoplast and
vacuole isolation (9).
To corroborat e the presence ofrosmarinic acid in lavandin cells cultured ill I'itro,
SO % (plv) methanolic extracts were obtained from the cell s. Tn Figure 2, the UV
spectrum of non-purified methanolic extracts is prnctically superimposible on that of pure
rosmarinic acid. This suggests that rosmari nic acid is the main phenolic present in the
methanolic extracts of the lavandin cells. Evidence in suppon of this suggestion arises
ITom the HPLC analyses of non-purified methanolic extracts obtained from !avandin cell
cultures (Figure 3).
Preparative low pressure chromatography on Sephadex LH-20 (0 26 x 420 mm;
methanol as eluent) ofmethanolic extracts of the cel ls and subsequent analysi s of the
i
,.
,
,.
'I ,
"
,I
' I
.,
1
I
1
,j
,
,1
I)
,
'I
"
!i
Ii
"
,
,.
j.'
Ii.
I"
If
I
r
;J
:1'

" I
" ,
::
,
,
!"
. ,
208 SI'lCES, FlAVOR CHEMISTRY AND ANTIOXIIJAI'i'T I'ROPERTIES
OOH
H
Figure I. SuuclIIre of rosmarinic acid.
a
0.8
b

v /
0
0.6 -
c
" 0
.0
L
0
-
0.4
.0
'"
\
0 .2
\
0.0
275 300 325 350
Wavelength (nm)
Figure 2. UV spectra af!,ure rosmarinic acid (a) and
melhol/olic extracts obtaine from cell cultures (b).
"
\
,
of non-purified
17. LOPEZ.ARNALOOS n AL. Antioxidant Activity 0/ Lavandin Cell Cultures 209
Table I. Occurence of rosmarinic acid in some plant families
Plant
Apiaceae
ryngium campeSlre
Sanicula europaea
Uoraginaccac
Alle/IIISG officinalis
Boraga officinalis
Uthospermum ojJicinaie
Pulmonaria o/ficinalis
Lamiaceae
Calamintha sylvatica S.l.
l.avandula angustifolia s.l.
Lovondu{o .f bumatii
Lavandula lo/ijolio
Lycopus europaeus
Melissa officilWlis s.l.
Memha x piperilo
Melllha pu/egium
Memha spicala
Manurda didylllll
Nepela cataria
Ocimum basilicuIII
Origanum vulgare
Orthosiphon ariS/alILS
Prunella vil/garis
RosmarinllS officinalis
Sa/via lavandillifolia s.l.
Salvia ojficinalis
Salureja //lontalla s.l.
Thy//l1lS vII/garis
Zosteraceac
Zooslera marina
Organ
leaf
flower
leaf
flower
'eaf
!lower
leaf
flower
m,,'
IRAJ (%, wfw)
0.8
3.0
0.4
0.7
0.3
0.8
0.4
1.2
0.5
0.7
3.7
4.7
2.8
2.7
3.9
1.8
0.2
1.9
5.0
0.8
6.1
2.5
0.6
3.3
1.2
2.6
1.9
21 .
SPICES: HAVOR CII EMISTRY .... NI> ANTIOXIDA..'tT " ROPERTIES
E E
c c
'"
'"
'"
'"
'" '"

c c

A

8
u u
c c
c c

c c
C 0


"'
"'
Time Time
Figure 3. HPLC chrofflllJograntS oj TIOIl -ruri/ied me/hanoUc extracts obtained
cell cu(/Ures derived lTOm (em'es (A and spikes (B) oj lovandin (Bar = 4
,mnllta).
17. ET Antioxidant Activity of f.Avandin Celt Cu/turf'.f 2 11
resultant fractions by TLC (silica gcl; acetic acid:mclhanol:dichloromcl hane, 4: 15'35
v/v/v) and UV-Vi sible spectroscopy revealed that, under the assay conditions. no
phenolic other Ihan rosmarinic acid could be detected aller examinat ion of TLC plates
under UV light (A"'36S run) and by spraying them with II solution of ferric chloride (2 %,
p/v) (datil nol shown). Analyses oflhe plat es showed Ihat only II few fractions were
enriched in rosmarinic acid, as judging by comparison orl he It, orlhe spots wilh Ihl\l of
pure rosmarinic acid. Furthermore, the UV spectra of these fractions were identical to
that showed by pure rosmarinic acid (data not shown).
The determination by HPLC ofrosn1alinic acid in methanolic extracts oflavandin
cell cultures reveal ed a content of about 3 micromoles per graAi of fresh weight (10)_
Thi s means a concentration or about 2-3 % in weight as regards the dry weight. Similar
concentrations have been reported working wi th other plant materials (Table I) (3)
Although rosmarinic acid has been previously detected and qUllntitied in ti eld-
grown plant material from the genus LawlI/dllla (3), LOpez- Amaldos e/ a/. (8) wer e the
first to demonstrate production of rosmarinic acid in cell cultures derived from plants
belongi ng to this genus. Funher studies are necessary in order to optimize production
and evaluate the potential for large-scale production of rosmarinic acid in lavumlin cell
cultures_
Superol ide alli on-scavenging activity of Cllrllcts obta ined from lavllndin cell
cult ures
The presence of rosmarinic acid in methallOlic extracts of lavandin cell cultures suggcstS
that these extracts may offer potential as antioxidant s. Despite the fact that lavandin
belongs to the Lamiaceae family, there is little information about its antioxidant
properties (J I).
In order to determi ne ant ioxidat ive activity of compounds or extracts obtained
from plant s, several methods have been developed (/2). Comparison of the results
obtained by ditfercnt tcchniques is complicated because antioxidative activi ty d('pcnds
on several fadars including the substrate used in the evaluation. However, for screening
a model system is much less time-consuming than traditional storage 51\1t1ies
(I),
In our study, the anlioxidative activity of methllnol ic extracts from lavandin eell
cultures was monitored by following the ability of these extracts to scavenge sup('roxide
anions generated in a phenazine methosulphate (PMS)-NADH system. Figure 4 shows
the absorbance at 270-280 nm (indicative or lOtal phenolics) and at 325 nm (indicat ive
of rosmarinic acid) as well as the inhibit ion or the superoxide dismutase-scnsitive
nitroblue tetrazolium chloride (J\'DT) reduction in fractions obtained after chrom:lIogra-
phy of lavandin methanolic extracts on Sephadex LH-20. From this figure, it can be
observed that maxi mum superoxide anion scavenging ability is associated wi th those
fractions containing rosmarinic acid. Thi s fact suggests that rosmarinic acid is the main
compound responsible for the superoxide anion-scavenging activity of met hallOlic
extract s obtained from lavamli n cell eultures_
,
i-
,
"
"
!I
P
,I
"
;'1
"
I
I
"
,
I
1
i
"
I i I
"
,
" ;i
i'
i
,
,
,
"
"
!!
i '


,
i
':
"
I
SPICES; HAVOR CII EMISTRY AND ANTIOXIDANT PROPERTIES

""
-

10,0 100
e
0 0


A
u
eX ,


,
,":
"'
,-- ,_ - I
"

,

'- -
, c
oe
,
f--
"
CD
u
"
Z
u
5,0 75
0 e

e 0
"
>



0.
0
" e
" .0
0
"
-

"

0
",
0
0,0 50 if>
30 40 50 60
Fraction number
FiSlire 4. Total phel/olic cOlllent, expressed a.r absorbance at 270-280 flm (-.),
rOSlIl{lI"inic acid COl1fent, expressed as absorbance at 325 nm (-), and
slIperoxide dismutase-sensitive NET reduction (---) of fractions obtained after
chromatographic separatioll all Sephadex LH-20 of methanolic extracts from
laV(mdill cell Cll/lares.
17. J,Ol' cZ-ARNALOOS lIT AI.- Antioxidant of Lamndin Cell Cultures 2 13
SlJpero}( ide 3nion-scavenging adivity of rosmari uic acid a nd otber related
compounds
To better understand how rosmarinic acid aclS as a superoxide anion scavenger, the
ability of olher structurally-related phenolics to inhibit the NBT reduction in a PMS-
NADH system was determillcd for comparison. The compounds tested were: (A)
Protocatechuic acid (a dihydroxybenroic acid), (8) Caffeic acid (a
dihydroxycillllamic acid), (C) Chlorogenic acid (an ester of caffeic acid and quinic
acid), and (D) Rosmarinic acid (which can De considered as a dimer of caffeic acid)
(Figure 5). lC.IO values for the scavenging of superOldde anions by phenolic
compounds wcre calculated from the inhibition's percentage- concentration
lines of the titrations represented in Figure 5.
Table II contains the 1C.IO values obtained from data in Figure 5. As can be
seen in this Table, of all the phenolics tested, rosmarinic acid was the most efficient
compound in deactivating the superoxide anion, inhibiting the supcroxide dismutase-
sensitive NBT reduction by 50 % with a concentrat ion of 33 p M.
The antioxidative potency of a compound is related to its structure, in
particular to electron delocaliz.ation of the aromatic nucleus and is also influenced by
the number of hydroxyl groups (13). In th is way, the higher antioxidative activity
of rosmarinic acid with rcspect to that of the other phenolics tested could De related
with the presence of two catechol groups on its molecule . However, since the lC.IO
value for rosmarinic acid is less than half of that for caffe ic acid (Table II), other
mechanisms for superoxide anion deactivation may be involved.
When othcr assays are used in order to compare antioxidant activity of
rosmarinic acid with that of the other mentioned phenolics, contradictory results are
found. Thus. in the methyl- linoleate test (14), the antioxidant efficiency increases in
the following order: chlorogenic aci d, protocatechuic acid, caffeic acid, rosmarinic
acid. However. using the method involving the free radical 2,2-Diphenyl-l-
picrylhydraz.yl (DPPH') (15) the increasing order of antioxidative efficiency is
rosmarinic acid, protocatechuic acid, caffeic acid. These data illustrate the above
mentioned difficulty of comparing rcsults obtained by application of different
experimental methods.
Iron-chclatill g act ivity or rnsmarinic acid
Superoxide allion is a relat ively nonreactive species in aqueous solution, but in the
presence of hydrogen peroxide and a transition metal such as iron, the extremely
reactive hydroxyl radical may be generated through a superoxidc anion-driven,
metal -cata)ysed Fenton reaction. The hydroxyl radical is able to initiate lipid
peroxidation directly through abstraction of hydrogen from fatty acids leading to the
formation of off-flavors and other undesirable compounds.
One p05sible strategy for minimizing pcroxidat ive damage consists in
removing traces of heavy metals by the use of chelators. In this way, rosmarinic acid
seems \0 be a good candidate for chel ating iron and other metals since it bears two
catechol and a carboxyl groups on its molecule.
When rosmar inic acid is incubated in the presence of iron ions. the formation
of a r05marinic acid-iron complex takes place. The formation of this complex can
214
SI'ICt:S: fLAVOR CHEMISTRY ANU ANTIOXIL>ANT PROI' ERTIES
A 8
50

t--
m
z
Q) loa
>
-

c
v
50
I
o
o
VI
c D

o 100 :500 500 100 300 500
Phenolic concentration Cu.M)
14 ft! 5. EffeCl of protocatechuic acid (A). caffeic acid (8), chlorogenic acid
( ,and rasmarinie acid (D) concelltrations on superoxide anions generated in
a MS/NADII system and dttermined by rhi! superoxide dismllfau-sensiti ve NBT
red/laiQ/!. Bars show standard errors.
17. I"OPEZ_ARNALDOS.:1' AI.. Antioxidant oj Lo.viJ/ldin Cell Cultures 2 15
Table n . Superoxide anion-scavenging efficieucy and str uct ure of the
phenolic acids tested
Compound
Structure
Protocatechuic acid 500
"
Caffeic acid 157
Chlorogenic acid 114

COOH
' 0
O.
Rosmarinic acid 33
.,,!\.CH.C.=</ Q
OH
HO HOOC
OH
"
" ,
"
Ii
I
'11
,J
.,;
0'
rl
,
"
I
;
,
I,

i
'I
.,
"
I,
" ;1
I
216 SP1Ct; ....: H.A VOR CIIY.MISTRY AND ANll OXIDANT 1'H. OPf.:lfrlt;S
r
E E
c C
N
..,
,... ..,
on
A
"
B
~ ~
0 0

u u
c c
0 0
-"
.0
" " 0 0

-" -"
., .,
Time Time
Figure 6. f/PLC cliromarogrmllT o/the rosmarillic acid-iroll complex mOllitored
at 572 nm (A) and at 333 nm (8) . Arrows indicate retention time/or rosmarinic
acid (Bar ... 4 mil/Utes) .
17. , .oPEZ-ARNALDOS "', A I ~ Antioxidant Adil'ity oj Lal'o.ndin C ~ 1 l Cu!tUrt., 217
he followed by VIS spectroscopy and it presents an absorption maximum at sn nm
(I"'.
Titrat ion curves for rosmarinic acid in the presence of different concentratiolls
of iron ions indicate a stoichiometry of 1 [0 1 for the complex. These dala were
confirmed by elemental analysis and inductively coupled plasma emission
spectrometry ( 10). This I to I complex seems to be lhe only absorbent species in the
reaction media. Support of this assertion come from graphical analysis of the
consecutive spectra according to Coleman a 0.1. (16) and from HPLC anal ysis of the
reaction media (Figure 6). From this figure. it can also be deduced thai , under the
assay conditions, almost all rosmarinic acid is forming part of the complex, since
absorbance at 333 nm for the expected retention lime of rosmarinic acid is very low.
Although chelating activity of compounds bearing catechol groups has been
largcly known, there is no report on complexing properties of rosmarinic aci d
towards iron ions. In 1985 Banthorpe et al. , working with callus cultures of
lAvandula angustifolia, demonstrated the secretion into the supporting agar of the
(2, E) -2-(3, 4-d ihydroxyphcny1)cthenyl ester of 3-0, 4-dih ydrox yphenyl)-2-propenoic
acid and of its (E,E)- isomer (17). These compounds, which are probably rcl ated
biosynt hetically with rosmarinie acid (18), are able to chelate Fe
H
ions forming
intensely blue pigments (IT} . Since )avandin cell cultures also display blue
pigmentation in the supporting agar and rosmarinic acid has been found to be the
main phenolic constituent present in extracts of the medium (19), a participation of
rosmari nic acid in the pigmentation of the culture media through the formation of
complexes with iron ions cannot be discarded (19).
Concl usions
Rosmari nic acid fulfils the requirements for being considered as a potent antioxidant
since it is not only capable of eHiciently scavenging superoxide anions, but is also
able to chelate iron ions. Studies on ()(hcr possible mechanisms for rosmarinic acid
acting as an antioxidant , such as hydrogen peroxide detox ification through
peroxidase-catalyzed reactions (8) , are now in progress. In the search for sources of
this comJ>Ound, lavandin cell cultures must be considered since they accumulate
relatively high amounlS of easy-to-purify rosmarinic acid.
Acknowledgments
This work has bcen partially supported by grants fron the CICYT (Spain), Projects
# ALI 93/573 and ALI 9511018.
Literature cited
I. Madsen. H. L.; Bertelsen, G. Trends Food Sci. Techno/. 1995,6,271.
2. Maruta, Y.; Kawabata, 1.; Nik i, R. 1. Agric. Food Chem. 1995,43,25<.)2.
3. Lamaison, 1. L. ; Petitjean-Freytet, C.; Carnal, A. Ann. PllOrm. Fr. 1990, {8.
103.
4. Petersen, M.; Hausler, E.; Meinhard, 1.; Karwatzki, S.; Gcrtlnw.\ ki. (. I'hllJ/
Cell TiS5. Org. Cull . 1994,38, 171.
I
218
S}, ICES: FL.\vOR CIIEMIS'J'RY AND ANTlOXJOANT " IWPF;JtTJES
S. De-Eknamkul , W. ; Ellis, 8. E. In Biouchnology in AgricuilUft' alld Fortslry;
Bajaj, Y. P. S., Ed.; Medicinal and Aromatic Plants I; Springer-Verlag:
Berlin, 1988, Vol. 4; pp 310-329.
6. Whitaker, R. J.; Hashimoto, T.; Evaru;, D. A. Ann. NY Acad. Sci. 1984,435,
364.
7. De-Eknamkul. W.; Ellis, B. E. Pfama Med. 1984.51,346.
8. L6pcz-Arnaldos, T. ; LOpez-Serrano, M.; Ros Barcel6, A.; Calder6n, A. A.;
Zapata. J. M. Biochem. Mof. Bioi. 1m. 1994,34,809.
9. HlI usler, E. ; Petersen, M.; Alfermann, W. Plant Cell Rtp, 1993,12,510.
10. L6pez-Arnaldos, T.; L6pez-Serrano, M.; Ros Barcel6, A.; Calder6n, A. A.;
Zapala, J. M. Frestnius J, Anal. Chtm. 1995, 351, 311.
11. Economou, K. D.; Oreopou\ou, V.; Thomopoulos, C. D. J, Am. Oil Chern. Soc.
1991 ,68, 109.
12. Frankel, E. N. Trends Food Sci. Technof. 1993,4,220.
13. Shahidi, F.; Wanasundara, P. K. J. P. D. Crit. Rev. Food Sci. Nutr. 1992, 32,
67.
14. Cuvelier, M. E.: Richard, H. ; Bersel , C. Biosci. Biotech. Biocl1em. 1992,56,
324.
15. Brand-Williams, W.: Cuvelier, M. E.; Bersel, C. Lebensm. WiJ"S. Technol.
1995, 28, 25.
16. Coleman. J. S.; Varga, L. P.; Maslin, S. H. lnorg. Chern. 1970,9,1015.
17. Banlhorpe, D. Y.: Bilyard, H. 1.; WalSOn, D. G. Phytochemistry 1985, 24.
2677.
18. Bamhorpe, D. V.; Bilyard, H. J.; Brown, O. D. Phytochemistry 1989, 28, 2109.
19. L6pel.-Arnaldos, T. ; L6pez-5errano, M.; Ros Barcel6. A.; Calder6n, A. A.;
Zapata, 1. M. Nat. Prod. Len. 1995, 7, 169.

Chapter 18
Anti-inflammatory Antioxidants from Tropical
Zingiberaceae Plan ts
Isolation and Synthesis of New Curcuminoids
Toshiyo Masuda
Faculty or Integrated Arts and Sciences, Unhersity of Toku shi ma,
Minamijosanjima i - I , Tokushima 770, Japan
Antiinflammatory active lUlt iol(idants were isolated from tropical
Zingiberaceae plant rhi zomes. Their chemical structures were
determi ned to be new curclll11ill oids. In part icu lar, Ihe compounds
from Zil1giber CUSSWIWll(lr (cassumunins and cassumunari ns) have
complex cLircuminoid structures and bOlh groups have sHonger
., ant iol<.idant and antii nfl ammatory activities than those of curCli lTli ll.
The total syttthesis of one of thcse complex curcuminoi ds.
cassumunin A, was accompli shed in 10 steps starting from 0 -
vanillin. The regio-specilic effcct of the hydroxyl and methoxyl
groups on the antioxidant activi ty of curcuminoid was invest igated
using synthctic curcumin analogs.
Natural antioxidants are important mat erial s for the oxidative deteri oration of
food, and some antioxidants also have potential prevention activity ill the initial
stages of oxidatioo-rdated di seases. I'lams used fOf food. includillg spice, are well
known to be ri ch in phenol s_ The phenol s have been targeted as natuml
antioxidants because most phenolic compounds have a radical trapping property
which produces an antioxidant activity ( I). There are about fourteen hundred
species of the Zingiberaceae plants in the world. Most of these grow or are
cuhivaled in tropi cal and subtropical Asia (2). The Zingiberaceae .plants typically
have large rhiwmes. Asian people have used these for and. spIce, and also for
traditi onal medicines to health (3). In conllect lon wll h our research 10
discover new natuml which are beneficial for human health, we have
been investi gating IropiClll Zingibcraceae plants as 11 source for new biologically
activc antioxidants.
Antioxidant ACl i,,] t)" of Tropicul Zingibcraccac Pl ant Rhizomes
Wc have examined Ihe alt,i oxidant activity of ninc cultivated species of
Zingiberaccae rhizomes which were collccted in Indonesia and Okinawa. The
ant ioxidant act ii ty was judged by the inhibit ion of linoleic aci d oxidation in an
C) 1997 American Oemiell l Society
I
220 SPI CES: HAVOR CIi EMI STRY ANTI OXIDANT PROPERTJt;S
et hanol-buffer system. The results obtained using the thiocyanate: and TBA
l11ethods (4) are shown in Figure I. The "ntioll idant activi ty of the acelone extract
of the rhi zomes incrcased in the order Curcuma heym!tlTUl < Phutomtda speciosa
< Curcuma (It Tuginosa < AmOnlum kepu/aga < Curcuma rna/rgga < Zingi/)er
CQS$umunar < Curcuma xanthorrhi<.J1 < Alpinia galanga < Curcuma domestira
using the thiocyanate method, and in the order Curcuma hqneana < Plllleomeria
SIH!ciosa < Curcuma aeruginosa "" Curcuma mangga < AnlOmum kepu/aga <
Zillgiber CaSSl/lllllnOr < Curcuma .mfl/horrhiw < Alpinia ga/ang(l < Curcuma
dUlllesrku usi ng the TBA method.
Curcumin is a typical constitue nt of tTopical Zingiberaceae plant (5) and has
strong antioxidant activi ty (6). We all alYl.ed the quantity of curcumin and two
known analogs in the acetone extract of the rhizomes. Although the extracts from
Curctlma domestica, Curcuma xa/lthorrlril.(l, and Zingiber cassumunar were
determined to contai n a relatively high amount of curcumin by HPLC anal ysis,
thei r strong amioxidant activity could nOI be explained only by the curcurninoid
contents (7). The results indicated the presence of new additional antioxidants ill
Ihe rhizomes.
New Ant ii nnammatory Anlioxidants from Tropical Zingiberareae Plall ts
Recently, furtumin has received much attelltion because of ils interesting
biological activity. whicb may be related to ils antioxidant property (8).
I\nliinnammatory activiTY is an important biological activiTy of curcomin (9).
because innammation is one of the peroxidation-related events in li ving
organisms and its inhibition suggests that eurcumi n works as an ant ioxidant even
in living cell s. Huang e/ al. reported antitUlllor promotion activi ty of curcumi n
and clarified that its activity is linked wit h the suppression of arachidonic acid
ruelabolislll ( 10). This metabolic pathway is one of the lipid peroxidation events
in living organisms. InnalllmHlion is al so well known to be closely relatcd to the
metabolism of arachidonic acid. We have been invest igating
in tropi cal plants and succeeded in
such ant ioxidants from three species of tropi cal Zi ngiberaccae plant s.
Curcuma domeslica. Half-processed rhizomes of Curcuma dOl/u!Stica are known
as turmeric and are used as a food coloring agent. The rhizomes are al so used as
ItlIditional mediciues. New (conlpounds 2 and 3) were isolal ed from
the dry rhizomes of the plant, the stmctures of which are shown in f-lgure 2.
,.\nlioxidant act ivity of the compounds are di splayed in !-- lgure 3 ( 11).
Curcuma XQ/ll llO"hiza. Curcuma .mntl/{)rrlliza is a medicinal gi nger in
Indonesia, however. thc rhilomeS are sometimes used for spice as a substit nt e for
CurCl/lllel dOlllc.l"licu. Wc isolated new ant ioxidants (compoull ds 4 and 5) from the
rhizomes. Their slructures and IIl1tio:>lid3nl activity are presented in !--lgures 2 and
4. respect ively ( 12).
Zingibu cassumunar. Zingiber Ca$$umwrar has large yellow rhizomes similar to
Curcullla dOlllcs/ica. The rhizomes are known to have anti innammatory activity
( 13) and are used as traditional in Thai land and Indonesia. We have
succeeded in isolating ne w antioxidants with antiinnammatory activity (rom the
rhizomes based on bot h the and antiinnammatory assays (14. IS).
Figure 2 shows the stroclllres of six new compounds Icassumunins A-C ( 16)
(compoullds 6, 7, and 8. respectively) and cassumunarins A-C (17) (compounds 9.
10. aud II , respectivelY)1 isolat ed along with curCllrnin aud S'-rnethollycurcurniu.
The antioxidant activity of the cassumunins and cassumunari ns is al so shown in
figures Sand 6. respectively.
18. MASUIJA
Arlii-inflammatory AI,tiQxidanls from Zingiberoceat Plants
r---------r"
"

I'" ...: , 01
u,
r-----------,'"
"I 0\
" .
"i 8
2
. . "
<i , .. V I
u,
m
SI' ICK'i, FlAVOK CHEMISTRY ANI> ANTIOXIDANT I'ROI'li K:nES
%
~ :
" 0 0
oc,
-
Figure 2. The Structures of Antioxidants from the Rhi zomes of Tropical
Zingibcraceae Plants

18. MASUDA kl/i. injlamtrw.lory Antioxidanl$ from Zingibtrtlct!at Plants
E
c
o
o
~
-


"
c

"'
-
o
~
"' -<
E
c
o
o
~
-


" c

"'
-
o
~
"' -<
0.5
~
no addi t ive
---
1
0.4
----
2
~ 3
0.3
0.2
0 .1
0 . 0 1 ~ ~ ~ ~
o 2
Rgure 3. Ant ioxidant Acti vity
domestico.
0 .5
4
Time ( days)
6
of the Isolated Compounds from CurCI/rna
~
no additive
---
1
0. 4
---
4
--0-
5
0 .3
0 .2
0.1
O . O ~ ~ ~ ~
o 2 4 6
T i me ( days)
Fi gure 4. Antioxidant Acti vity of the isolated Compounds from Curcum(1
Xllnrhorrliil.ll .
i
'"
SI'I(.:I\S, HAVOR CII E.'ttl,nRY AND ANT10XlOANT
E
=
e
Q
on
-


0
c
"
"'
0

.0
<:
-
=
Q
Q
on
o
O.S
-0-
n o addi ti ve

1
0.4
6
-0-
7

8
0.3
0.2
0 . 1
0 .0
0 4 6 8 1 0
T ime (d ays)
Fi gure 5. Antioxidarlt AClivily of Cas sum un ins A (6), B (7), and C (8) ,
0.2
o,f

o 2 4 6 8
T ime (d ays)
Figure 6. AntiOxi dllnt Acti vity of Cassumullarins A (9), B (1 0), and C
( I I) .
1 0
18. MASUDA Anti-inflammatory Anlioxifitlnt.\' jrom ZinKibtrllaar " lallU 225
Anli innammalory Acth'ily or the 150111100 Antioxidant" Newly isolated
compounds from the chree Zinl=iberaccac phill is previously rmnlioneti are
SlrllclUrally related to curcumin. We eXAmined their anliinnammil lory act ivi ty
( 18). Curcumi" (0.6 I.mol) showed inhibitory Rellvil), of edema formation, which
is caused by an innammalion inducer. TPA (l2-0 -ltlradeunoylphorbol-l3 .
aCel&le.2 IIg), on mouse car, The activity of Ihe isolated compound$ IIsinji! the
mouse ear method is summari7.ed in Table 1.
Synthe$is of Cassumuni n A, a Complex Cun: uminold from ZitfgiiJer
cossumutfar
We have begun the tOial synthesis of several new complu: curcuminoids and
sllcceeded in the synthesis of caSSUlllUnin A. Ihe moSI potent
from Zingiber cU$Yllmllnar. The synlhesis of curcumin was achieved
by many groups (19), howcver. thc most efficient mcthod was first reportcd by
Pabon (20), which involvcs thc Knocvenagel condensation of thc boron complex
of 2,4-pentadione wilh vanil1in. In Ihe synthesis strategy, we employed thi s
method for the coupling of two differcnt Iypes of benzaldehyde wilh 2,4-
pendadionc.
v-vanillin 12, which was prepared from and
was condensed wi th
giving the conjugated ketone t3 in 97% yield. One methyl group was introduccd
at the of the carbonyl group in 13 by treatment of (CIIJhCuLi in 94%
yield (compound 14). The methoxymet hyl group of 14 was removed by acid
treatmellt (I N HCI) in methanol to give 15 in 76% yield. The carbonyl group of
15 was reduced to a group by in 98% yield (compound
After diacdylation of 16, the diacetatc 17 was heated at 160 OC in DMSO to
introduce a double bond at the appropriate position in 90% overall yiel d
(compound 18) from 15. After hydrolysis of the acetyl group of 18 (compound
19), regiO$ekctiye introduction of aldehyde functio n at the 4-posi tion was
achieved by treatment with chloroform under alkaline conditions. giving aldehyde
20 in 40% yield, The coupling reaction of 20 with diketone 21 under Pabon's
condi tions (20) gave cassumunarin A in 41 % yield. The synthetic cassumunin A
was identical with the nalural compound based on spectroscopic methods (i-Igure
7,
Synthesis of New and Their Antioxida nt a nd
AnliiTinamma lory Acl il'ilies
Among planl consli tuents, there are many kinds of phenolil;: compounds, such as
navonoids. coumarins. and phenol carboxylic acids. It is well knC'wn that they
have various regioisomers in positions of the hydroxyl and group
substi tuents. The st ructure and efficiency for tocopherols
(21) and (22) have been well investigated. however. littl e is known
about curcurninoids. We have been int erested in the regio-effect of the hydroxyl
and methoxyl groups in the benzene ri ngs of cureumin on their ant iox idant and
antiinflammatory activiti es.
We synthesi7.ed cleven curcumin 311alogs as shown in Figure 8, and eMmined
their antioxidant activity using linoleic acid as a substrate and AlBN as
oxidation inducer. The data are shown in Figures 9- 12 (23). Figure 9 shows that
the hydroxyl group at 2 or 4-posilions is important for antioxidant activity.
Figures 10-12 show that a methoxyl group at the ortllo or pam-I)()sitions in(fC;iscS
the activity. The 4-hydroxy-3-methoxy compound 31 (curcumin), the 2-hydmAY
3-methoxy compound 26, and the coml)()und sh"w Ih"
I
iii
".
Sr IC.:S, FlAVOR CHEMISTRY AND ANTIOXIDANT I'ROrERTl fS
I'.
MASUDA Alltij"flamnwlory Alltio.rid(mls from ZingibuQcrtu Planls 227
Ir:

r
"
" " r
a
r
i 1
8 "
a
r
: i
"
8
r
fa
;ii
8
, I N
0
:
Tl$ble I. Inhibitory Act ivity of the Isolated Compounds (0.6 " nlOl)
"
on TPA Mouse Ear Innammation C


-
"
f
\
'j
Experimant Left Ear RighI Ear
""
Oh:t:SE(ms.!
Inhibil;OII( '1bJ Significance a ,i
I. f
,
i
TPA 10 16.MO.9
:l-
'l.
I+TPA TPA 10 9.7:t1.1 58 p<O.OI
r
9. 'l.
is
2+TPA TPA 5 8.7:t. 1.4 52 p<D.Ol
u u
a
'l.
r r
,
l+TPA TPA 5 5.310.4 32 p<O.OI
"
u u
f
,
r r
4+TPA TPA 6 r 1. 5:t:t.3 69 p41.01 u u

a
'I

eli

i
u
<
.1
2.
15 "
,
TPA

17.hO.9
Q
'\

l +TPA TPA 5 8,'h{),9 51 p<O.OI
"
"
"
r
r
,
6+TPA TPA 5 14.5%1.0 83 p<O.Ol
r ,
8
8
a
a
p<.O.O I
,
8 " 7+TPA TPA 5 13.3:t: 1.2 76
'"
d
8+TPA TPA 5 1l. 1%1.3 75 pdWI a <
'j
a
a .,

r
r
#

3.

f

,

TPA 6 16.5:tO. 6
0 0 '" a
;;
r "
"
I +TPA TPA 5 9.2%1.3 56 p<O.OI r
8 B 0
9+TPA TPA 5 12.h2. 1 75 p<Q.0I


I O+TPA TPA 5 1O.h1.4 62 p<O.Ol
II +TPA TPA 5 IO.hO.8 62 01<0.01
"
number of mice. bO: lnean of wei ght differnce betwee n right ear and left ear.
'l.
0
'l.
:l-

lin:
a a
"
r
"
r r
"
r u u
u u
u u


" ;1% f
,
,
0'
.
'"
e %
Table II. Inhibilory Atl ivity orCun:uminoids (0.6 I'fllol)


" r r
011 TPA (2 rS)-lndut'cd Mouse Ear Inna mmation
R
J
8
f
8
'l. 9.
r r
Trl.'almtfll
u u
uft Ear Righi Or
'"

Inhibjlion(%) Si gnificonu
" TPA 4 !7.4iO.9
a a
0
>
31( 1)+TPA TPA 4 9.7:t 1. 9 56 p<O.OI
r a
28+TPA TPA 4 10.0%1.2
5'
pd101
>
-
it

26+TPA TPA 4 4.4%1. 5 25 p<Q.OS


all : nu mber of mice, bO: mean of weight diITernce between right ear and left ear.
e
,


a R g
'l. 'l.
:l- :l- r r
u
'-'
u u
r
$I'ICt;S: CIi EMI.!.7K't' AND ANTIOXIDANT l'ItOPllTi FS
I) 13
1
0 y'
o 0
0
AA

R,
2) B(OBuh

Y J) ,,BuNH
1
R.
TBDM$ia
\",-,J-
H
U
OR
TBDM$i
R.
R,
0 0
'"
R, R,
&'
R,
220 11, .11, lI, gH. 11,_11. II, _U. "". H
2J: 112"11. IIj=lI . R..o-OII. lI,all.
U: 112"11. IIJ..otl. II.o-H. II".H. "". H
25.: II j2H. a..- Il. II,..H. l!"aH
U: II j-OM .. Jt..a H. II,. H.
21: 11:=011. R,a H. IIJ" H. l!"a H
lS: R,coOll. RJ211. ",,_H
:t9: 11,=011.
30: . R,=H. R,,-oll . R, . tt R,, - It
31: 1I1a H. R, =OM . Iyo.OU. RlaH, R,,: U
Jl: II,_U, 11.1 0011. R,e>(}M<. II,.H, R II
o 0
OR RO
Figure 8. Struclures and Symhesis or Curcuminoids (2 3-3 2) .
'.2 , _ ______ _ ______ -,
, .
E
0: 0.8

"


,!O.4

0.2
o
o nIIlnMbllor
" "
2
,

, ,
7

,
Time {hrJ
Fi gure 9. Inhibitory ACli vilY of CurCUlllinoids (22-25) against AIBN.
initiated Uooidc Acid Oxidation (Liool eic Acid: 54 mM. AIBN: 81 mM.
Curcumilloid: 170 1.M).
R,
R.
LS. MASUDA Anti. inflammatory Antioxidants from ZingiJnraCMt Plants

t:. U()'OIl)
E
o 111 ().OU.t.OM.)
0,"

"

c

"
0.4

0.7
0
0 2 , . ,
'rime (hr)
Fi gurelO. Inhibitory Activity of CUfCuminoids
initiated UlIOleic Acid Oxidation (LillOleic Aci d:
Curcuminoid: 170 JIM).
. ,
(2 5-29)
54 mM.
. ,
agai nst AIBN-
AIBN: 81 mM.
u.-- ---- --- ---------,
0 .... ...
"
E
13("011)
C
0.'

0 :Je("01l 100M.)

"

JI(4011 )-0.1. )
0.'

2
"
E
0.'

0.2
Time {hrJ
Figure ll. Inhibi tory ACli vity of Curcumiooids (23. 30. and 31) against
AIBNinitiated Uno/eic Acid Oxidation (Linoleic Acid: 54 mM. AIBN: 81
111 M, Curcuminoid: 170 t.M).
'"
230 SI' ICES: FUVOR CHEMISTRV AND ANTIOXIDANT l'ROI'ERTn ;s
..,,------------------,
o .. Iol>llli....
J.I (.l.ONi
!
~
0.'
r'
.B 0.4
~
0.'
Time (br)
Figure l2. InllibilOry Actiyity of Curcuminoids (24 and 32) againsl AISN-
initiated Linoleic Acid Oxidati on (Linoleic Acid: 54 mM, AIBN: 8 1 mM,
Curcuminoid: . 70 I'M).
~
- ~ 3 2
FigureD. Stabi lizati on Effect of Phenolic Radi cal by Mtthoxyt O)(ygen and
Olefin of Curcuminoids (2 6, 28, and J 1).
18. MASUDA Antiinflammatory Antioxidants from Zingiberaceae Plants
o
o
-
g
I
o
-
g
o
o
o
I
231
I
!.
I
I
i
!
1
i:
i
!
"
!
I
!
,
SI' ICt :.,,: .. LAVOl{ CHEMISTRY ANI) ANTIOXIDANT I'ROI'ERTI ES
'"
0
... Iwllli_
,
-
0 .)J

.)6CJ.OH 4,!-
_'bj .......... '))
0.'
-

J I (. OK ).()j.<<)
0.'
Time (hr)
Finllre 15. Inhibitory Activity of Curcuminoids ( 3 3 and 34) and Curcumin
(3i) against AIBN-initiated Lino[eic Acid Oxidation (Unoleic Acid: 54 mM,
AIBN: 81 mM, Cun:: umi lloid: 170mM).
strongest activity ill this assay sys.tem and Ihei r ?ffici ency is allTlO5t
These data indicate that the followlllg factors are Important for the curcum1l1old s
alllioxidant activity; I) the radical can.be to the bond
in the alkyl part at the I-posmon. 2) the phenoliC radical call be stabilized by the
lone pair of the mct hoxyl oxygen, as shown in 13. These pam.meters arc
si mi lar to those in tocopherol's case (24). We syntheSized two curcumll1 analogs,
the 4-hydroJly-2 3-dimet hoxy analog 33 and the
analog J4 14). In the former compound the phenolic radical can be
stabi lized by the two met hOJlyl oxygcns and the radical of the latter call
be stabilized by one of the mr:thylcndioxyl oxygens, whose lone pair well
overlapped wi th the phenolic mdical because the five membered nn.g rc:SIrIClS the
lone ()3lr obi tal of the oxygen in a good direction the del ocahl'.allo,:, the
phenolic radical to the oxygen (25).They showed slight ly stronger antlOJl.ldanl
activit y than that of cun::umin (Figure 15).
The allt iinflammatory activity of the two analogs (c?lllpounds 26
which showed equivalent activi ty to sllmmafl7.ed .In
Tllbl c II . These analogs have antllllflammatory activity SIIIII1 3(. to.CUrcUlmn,
indi cating the regio-specificity of sll bstituents on the aromatic rlllg.s t? the
1I111iinflammatory may not be so important only the antl.oxldant
effici ency of the substance may important FUl1hcr studIes along these 11IIe$ are
now in progress.
Acknowledgments
Thi s work was supponed in part by gronts from the Ue.hara M.emorial Foundation,
Takano Research Foundation, and Ministry of EducatIOn, SCience and CulllJre of
Japnn. The author thanks Dr. A. Jitoe, Ms. J. Isobe, and Ms. A. Kida of Osaka
Ci ty Uni versi ty ror their work in thi s project. The author also thanks Dr. Tom 1.
18. MASUDA Anli.injIammalllfY Antioxidants /I'llm ZinRiMrl1UrJe PlanL,
Mabry of the Univer.;ily of TCAaS at Austin for Ollf I;lIl1aborat itln in Tu u, and Ik
N. Nakatani of Osaka City Universily aod Dr. S. Yonemori uf the Univ. of
Ryukyus (Of" their suppor1 of plant collection.
Refercoces
1. Osawa, T.; Ramaralham, S.; Kawakishi. S.: !'la miki. M., "htnfl/i(" COntf/(J/lnli.f
in FQQd and Their 1:.1Jects on Hl!.alth II, ACS Symposium Series
American Chemical Society; WaShington. DC, 1992. Chapter [0, pp 122-134.
2. Hegnauer. R., Chem01axonQmie der PjlanUII, Birkhauser Verlag: Basel, 1963,
Vol. 2. : Chapter 48, pp 451-471 .
3. Corner. E. 1. H.; Watallabe, K . Illustrated Guidt to Tropical Pluws,
lIi rokawa Publishing; Tokyo, 1969, 1'1'1069- 1086.
4. O!;awa. T.; Namiki, M., Agric BioI. Chl!lIl., 1981 ,45,735-739.
5. Kuroyal1agi, M.; Natori, S., Yakuguku w sshi, 1970,90, [467- 1470.
6. TOOa, S.: Miyase, T.; Arichi, H., Tanizawa, H.; Takino, Y., Chem. Phl"flm.
8ull .. 1985, 33,1725- 1728.
7. Jitoe, A.: Masuda, T.; Tell gah, [. G. P.; Suprapata, D. N.; Gara. I. W.;
Nakatani. N., J. Agric, Food Chem., 1992,40, 1337- 1340.
8. Ammon, H. P. T.; Wahl, M. A., Planla Ml!d., 1991, 57,1 -7.
9. Rao, T. S.; Basu, N.; Siddiqui , H. H., Indian J. Ml!d. Res., 1982, 75,574-578.
10. Huang, M.-T.; Sman , R. c.; Wong, C. -Q.; Conney, A. H., Cancer Res .. 1991,
51.813-819.
I L Masuda, T.; litoe. A.; lsobe, 1.; Nakatani, N., PhYIOChtmislry, 1993, 32, [557.
1560.
12. Masuda. T., [robe, J.; J itoe. A.; Nakatani, N .. PhYlochl!mislry, 1992,31,3645-
3647.
13. Ponglux. D. ; Wongseripipatana, S.; I'tiadungcharoen, T.; Ruangrungsri, N.;
Likhiwitayawuid, K., Mtdicinal l'Ioms; VictOfY Poim; Bangkok, 1987. pp
275.
14. Masuda, T.: Ji toe, A., J. Agric. FoodChem., 1994, 42, 1850-1856.
IS. Masuda, T.; Jitoe, A.; Mabry, T. J., 1. Am. Oil Chl!m. Soc. , 1995, 72, 1053-
1057.
16. Masuda, T.; Ji toe, A.: Nakatani, N., ChI!III;s/ry l.ttl., 1993, 189- 192.
[7. Jitoe, A.: Masuda, T.; Mabry, T. J., TetrahedrQu Uti., 1994,35,981 -984.
18. Gshwendt, M.; Kittstein. K .. FUl"Stenberger, G.; Marks, F., Cancer utt., 1984,
25, 177-187.
19. Arri eta. A. E, Pharmaxie, 1993, 48, (f)(j-691.
20. Pabon, H. 1. J., Ru uil, 19\54. 83. 379-385.
21. Barclay. L. R. c.: Edwards, C. D.; Mukai, K.; Egawa. Y., Nishi, T., J. Org.
1995, 6{), 2139-2744.
22. Mora, A., Rios, P. 1. L: Alcara7., M. J.; 8iocltem. PIU/rnU/cal., 1990,40,79.\
797.
23. Masuda, T; Kida, A., Proceedi ng of Annual Meet ing of Japancse Society "f
Nutri tion and Food Science, 1995, pp 103.
24. Hughes, L.; Burt on, G. W.; Ingold, K. U.; Slaby, M.; Fosler, D.O., I'lun"Ii<
Compounds iu Food and Their Effects on Heal/h II, ACS Symposium Snies
507; American Chemi cal Society; WaShington, DC, 1992. Chaplcr 14. I'P
184-199.
25. Burton, G. W.; Ingold, K. U., Ace. Chem. Res., 1986, 19, 194201.
I
I
Curcumin:
of the
Chapter 19
A Pulse Radiolysis Investigation
Radical in Micellar Systems
A Model ror Behavior as a Biological Antioxidant
in Both Hydrophobic and Hydrophilic Environments
A. A. Gorman' , J. Oamblen' , T. J. Hill', H. Jooes' ,
V. S. Srlnivasa n
J
, and P. D. Wood!
t Department or Chemistry, University or Manchester,
Manchester M13 9PL, United Kingdom
ICenter ror I'hotocheml eal Sciences, Bowling Green Slate Univers ity,
Bowling Green, on 43403
Curcumin has been solubili sed in aqueous anionic, cationic and neutral
micellar systems and in aqueous buffer. The formation of its radical has been
ooSl.:rved in all situations. The radical is effectively inert to oxygen and reacts
with the natural vitamin C and vitamin E and Trolox 10 a degree
dependent on localion. II is concluded that the anlicarcinogenic activity of
Cllreutllin may welt be II. consequence of its ability (0 access both hydrophobic
nnd hydrophilic environmcrus, act as a mdi cal scavenger and be rcpaired by
both vitamin C and vitomin E,
Curcumin ( I ), the major coostituem of the spice turmeric, is a molecule of considerable
inte rest liS a consequence of ils known biological activity. This includes the light-
induced ollidative killing of thicleria (/.2) and anlicarcioogenesis related to inhibition of
lipid peroKidation (3-6). As part of a program aimed at undemanding the
anticarcillogcnic activi ty of thi s molecule (cj Ref. 7) we have examined the behavior of
curcumin in miccU<lr systems with cll1phasis on its location, the formation of the radical
and the reaction of that radical with and key 'tntioKidants. The SUrfll1;talllS
employed were sodium dodecyl sulfate (S])S), cetyl Iri mcthylammolliulII bromide
(crAll) and Tri ton X- HlO (TX) with allionic. cationic and neutral head groups
respecti vely. The cun:umin radica l was formed by the Slandard pul se radiolytic
sequence (8.9) summarised in etllt:ltiolls 15 where curcumin (Cur-OIl) ,III
electron 10 the azide radic:tl prior to Inss uf a I'f{)lnn.
1997 American Chemical Sudely
19. COllMAN At. Curtumin: A Pub."r Radiolysu' InvHigation
H20 + e H'

Oll'

e(solvated)
e(solvated)

N,o + H2O N,

OB:

OH'
OU'

N,. 01-1'

N,
N, + CurOH
N," + CurOI'I"
CurQH'-' Cur..o-
Experiment al
235
(I)
(2)
(')
(4)
(I)
Pllise radiolysis CKperimcnts WeT!! essentially as prtviously described (10). Digitized
data were Ir:lflsferred to the memory of a DAN 486 DX 33 PC for analysis. \Vmer was
doubly.distilled from potassium pemlanganate. CUTCunlin was as described (7). Cetyl
trimelhy!ammonium bromide (Fioons), sodium azide (BDH). sodium dodecyl sulf,ne
(Aldrich), Triton X- I(X) (Aldrich). Trolox (Aldrich). vitamin C (Sigma) and vit:lmin E
(Aldrich) were used as received_ Solutions were bubbled with either nitrous o:<ide
(British Oxygen Company) or nitrous oxide containing 10% oxygen (Air Products),
Rale const:lms delennim:d by lime-resolved lechniques are accurate to 5%.
Res ult s .
Curcumin was readily solubiliscd in aqueous surfactant systems in which the surfactant
concentr:l tion was above the critieal micelJe conccrur:uion (cme). In each Clse the
electronic :Ibsorption spectrum was independent of surfoctant concentration above the
cllle.
Absoq )\ iull Spectr a . The electronic absorptioll spectr3 of cl1rcl1min solubilised ill
:Jqueous SDS and TX (Figure la) systcms were essentially idcnlical - 417 nm)
and ahsohllely typical of the corresponding spectra in hydrocarbon solvent s such tiS
cyc1ohexane. a clear indication that, in such systems, a eurcutllin molec ul e on nvcrage
occupies a hydrophobic region of the micellar system, i.e. is "within" a mi celle. In
conlr:lst, the spectrum of curcumin in the aqueous crAB system (Figure lb) was
typical of Ihe or"nge curcumin anion ()"mv; - 480 nm) and it would appear tha!, in this
SYStCt ll, c.:UfCUllIi ll is predominantly On Ihe outer surface of the micelles, anionic, and
with the teoiary ammonium head groups of the CrAB component S, These
con<:imion., arc sllpponed by pnl se radiolysis experiments (vide infra).
I'Ub l' ]{ a diul yl ic fo r muti on of Ih e Radicu l in Micellar Systems. Pulse
radiolysis of N20-bubbled aqueous. pU 7 buffered, micellar solut ions of curcumin
(2.0 x !US moll
t
) comain ing sodium azide (2.0 x 10-
2
moll-I) led 10 fonm.tion of a
species with )"mu - 490 nnl, 111is is ellemplifi ed for the case of crAil in Figure 2
which shows the formation of the 490 nm species and the bleaching of the curcumin
ground stat e as a function of time. These processes take place on the same microsecond
(cf. Figures 2a and 2b) and the slow decDy of the 490 nm species on
milli scwnd timcscalcs (Figure 2c) was second-order, as cxpected for a mdical-radical
combin:l tion process. There were only minor variations in the kinetics of mdic:ll growth
and decay for the differem micellar systems for a given curcumin concemr:llion. Of kcy
impon,lIIce was the linding thai use of N:!O containing 10% oxygen led 10 absolutely no
change in the kinetics of decay of the radical absorption ;It 490 nm. This :llIuwS:III
upper limi t of - 1.0 K 1()2 1 mol
t
s l tO be placed on the rate COlISlant for r';:KtioliWith
oxygen. It is clear that in these systems the eUTCumin r.ldic:1I is indeed beill>! pn"lun'd
and is inert to oKygen on the timescales of its exiStence in our experiments. II III11SI I ...
recognised that in the real biologic:!l si atation the rJdical wuuld II Ot tli .. ""'''y vi"
236 SPICES: FlAVOR CII EM1STRY AND Ar-'TIOXII)ANT I' ROPERTlF..s
00
a
0.5
b
0.5
400 500 600
nm
Fieu r e I. Ekctronic absorpt ion of cun:umin (2.0 :t mol 1.
1
)
SO!libi!iscJ in iln iUluCOUS solution of (a) TX (5.0:t J()-3 mol I-I) and (b) crAB (5.0
.... 10-3 mol I-I).
19, GORMAN to" AI- Curcumin: A Rndiol)'$u Invtstigation 237
00
a
0.1


0.0
i j
1
, ,
; 1
t,,
, ,
-0.1
n
i
i" i ib r i
i i
c' ,
r I
,
-T' I
, I
r
,
300 400 500 nm
Fi gure 2. Transient absorptiOn spcctr:1 measured 6.4 jlS (Xl. 22 (D). 58 )1 S
(6) and 1261-15 (9) after pulse radiolysis of a solution of curcumi n (2.0)( 10-
5
mol
\-1). crAB (5.0)( 10-5 mol J.l) and sodium azide (2.0 K 10-
2
01011-
1
) i n N20-
saturated pH 7 bu ffer. Insets: lime dependence of (a) transient format ion 31 490 nm
with first-order fil, t' - 1.0)( lOS 5.
1
, 4.6 % absorption/division, 10 )ls/division.
(b) corresponding ground-stale bleac hing at 430 nm with first-order fi l, t' '" 0.71 x
10
5
s", 6.7 % absorptioll/division,](J !-I s/division alK! (el lransient decay a1490 nm
with second-order fil, 4.2 % absorplion/divisiol1, SO nls/division.
23'
SI'I C"::S: Jo"I-",VOR CIIEMI!oo1"RY ANI) ANTIOXlDANT PROPERTIES
''! I U
I I" :. I I
-.1. 1, Ll ,1 IJI

++ 1, j i b,
-! -j ! I
i

, I
.. -I ,
! j
Figure J. (a) and (b): Repe;u of c.xp<:tltnents com:sponding to Figure 2 showing
tntnsient dec:IY :1I 41)0 11m ill the present:e of (a) vi tamin C (2.0 x 10.
5
mol 1'\) wi th
first-order fit , k' .. 2.U x 10
1
S I, 4.9 % absorption/division, 300 !-,s/division and
(b) Trolox (2.0 x 10.
5
mol 1.
1
) with first -order fit, t'= 8.5 x 10
3
sl, 4.6 %
absorption I division, 50 !-Is/division. (e) and (d): Repeat of experiments
corresponding 10 t-igurc 2 using TX (5.0 x 10) mol I- I) in the presence of vitamin
E (2.0 x 10.5 mol 1.
1
) showing (c) u-,lIIsiem dccay at 490 nm with fit, /t.'
-= 1.1 x 103 S-l, 1.2 % :Ibsorption/division. 250 lis/division and (b) corresponding
growth of absorption at 430 nm wi th fil. t': 1.2 x 10
3
s", 0.6 %
absorption/division, 250 !-,s/division.
19. GORMAN ET Currumin: A Pum RadiolY5is
order combinat ion with itself, hut by a much slower pseudo first -order process
involving ambient scavengers. Inefficient reaction with oxygen might become morc
importpm in that Silu:lIion.
2

O_I/'OH
4
The nC)(1 (IUeslion co be addressed al Ihi s slUge was whether the location of
radical fomlP\ion would influence ils ability 10 react wilh additive mole!;ules expected to
be good radical scavengers. The potential scavengers chosen were vitamin E (2),
vilamin C (3) and the waler-soluble model of Ihe fonne r, Trolox (4). In the evenl of
efficient repair of the curcumin radical one would amicipll1e firstly a significam
shonening of the lauer's lifet ime and a change to pseudo first-order decay. if the
quencha concentration was significant ly hi gher than th<l t of the radical. Thus. the
above experiments were repeated in the presencc of 2.0 x 10.
5
mol l-I sc:lvenger. In all
of Ihese experiments the scavenger will of cOurse compete with CllrCl.lmin for the azide
radical. The cons<:quence of this is formation of both Ihe curcumin radical and thc
potential scavcngcr radical on microsecond limesc:lles (cf. FiguTC 2a). The subsequent
interacli ons can be monitored as a consequence of the hct thai 1111: absorption
of the cU!"Cllmin radical (490 nm) is signifkant ly shifted from those of vitamin C (- 3(iO
nm) ;"tnd vitamin E and Trolm; (- 430 nm).
Sc::.vengin g of Ihe Curcumin R::.di c::.1 Produced in CTAU Micelles. In lhe
presence of vitamin E the curcuntin radical decayed via clean second-order ki nctics on a
ti mescale identical 10 that of Figure 2c, i.e. in this situation vitamin E, which must be
solubili sed wi thin thc micelles, is unable to scavengc the curcumin radical. Since
subsequent experiments demonstrate repai r of the lall er by vi tamin E, these data clearly
agree wi th the propos;!1 that Ihe curcumin in this p;lT\ ieubr microhelcrogeneous
environment occu pies lin cKtramicellar location as Ihe (;orresponding mono-3nion.
Strong support for this conclusion C3me from the fillding 111:11 the w:lIersoluble vitamin
C and Trolox bolh quenched the cllTCumin r:ldic:ll cfficiellily as shown in FigllTCS 3a and
3b. The decay of the mdical absorption foll owed dc;m first-order kinetics,
10 rate for reaCtion betwcen the l:urcumin radical and the
scavenger of 1.0 x lOS J mol
t
s 1 for vit amin C and 4.3 x l(jll mol-I sol for Tro!lI.X.
Although the latt er is somewhat more efficient as a sC;lvenger, the c)( perimcm:d r:\to,:
conStantS are too close to allow significant concl usions to be
Scave nging of thl' CUfcunli n Il adkal in S US and TX \\I icelll' s. In SI:11k
COntr,tSIIO the situation for crAB above, the cu!"Cumin radical produced in eitl !<.: r SUS
or TX-bascd micellar systems was efficiently tluenched by vitamin . solubili sed within
the micelles as exemplified for TX in Figures 3c and 3d. These show the dec:ly of the
I
, I
,
I
I
: I
240
SI'I CF.S: FlAVOR CIIEMISTRY AND ANTIOXIDANT PROPERTI ES
OD
0.07
0.00
-0.07
400
500
nm 600
4. Transient absorptiOn spectr;1 Illc;lsurcd 5.2)ls (X), 12 Ils (0 ), 2S)ls
(6) and 58 llS ('i7) afler radiolysis of a sol uti on of cun.:umin (2.0 x 10.
5

, d 00' 'd (20' 10.2 mol \_1) in N, O-saturuted pH 7 buffer. Insets: lUoe
' )all , \U111a 2;1 C . h "_ Q4
dependence of (,I) Ir,lIl sient fornl;l\ion ,\[ 490 nm Wi th fi rst-o:der fit, k - 7.2 x 1
s.l , 1.0 % absorption/diviSion, 10 !-Is/division :md (0) tranSient decay at 49Q nm
wilil sccond()I"der fit. n.7 % absorplion/divisiOll. 25()
19. GOHMAN lIT AJ.. Curcllmin: A Puhe Radiolysi.f im'estigation 241
curcumi n radical al 490 nlll and the corresponding growth of the vitamin E radical at
430 nm on lhe snme limcscaleS. These dHta clearly indicate that the quenching of the
curcumin radical by vit:lTlIin E does indeed involve repair and supports thc cadler
conclusion that in the SDS and TX syStems the cli rcumi n is solubilised within the
hydrophobi c region of the micelle_ When Ihese experimenls were repeated in the
presence of vitamin C (20 x lOS moll -
t
), the decay or the euremnin radical was only
margi nally faster than the blank and still largely second-order. In the presence of the
same concent ration of Trolox quenching occurred lO' an extent which made the decay
predomi nantl y first-order but the Tate constant was an order of magnitude lower than in
the CTAB-based system (vide supra) . Clearly Trolox is more hydrocarbon-like than
vitamin C and has a higher probability of accessing the hydrophobic regions_
Formn(i oll of Ihe Curcll J11in Radical in Wa ler. It was possible to produce 2.0 x
10-
5
mol I-I eUfemnin in water by dropping a solution in 0.1 M NaOH into a large
excess of pH 7 buffer. Th:u this was a real homogeneuus solUlion, as opposed for
instance to a microc rystalline disperSion, was clearly demonstrated by the speclrum
which was essenti all y identicnl to thm in other hydroxylie solvents as et hanol and
in panicular the kinetics of its fCact ion with the azide radical, with and without addition
of vitamin C Trolox. Thu" pulse r'ldi olysis of such a solulion, N20-bubbled,
containing sodiutll azide led to formation of fhe cllrcumin radical (Figure 4a) on
timescales essentially idelllicalto those for the CrAB-based micellar sySTem for whieh
eurcumin is thoughl to be in the foml on the micellar surface, The spcctra!
changes with time (Figure 4) were vinuully identical to those for the micellar
pani cul arly the SOS and TX systems where the cUTCumin is in the neutral fornI , and the
radical was second-order as expected (Figure 4b).
Scave nging of the Cun: wni n Radi cal in Wuler _ Additi on of eit her vitamin C or
1'rolox (2.0 .x 10.
5
mol 1'\ in each case) led to efficient quenching of the curcumin
radic!l!. Rutecoll stallt s for scavenging were 1.2 x 1()81 mol-I s1 and 1.6 x 108j mol-t
s1 for vitamin C and Trolox respeCTively . The raw data are exemplified in fhe case of
Trolo:>:; in FigufC 5. Insets a and b show the decay of the cUfCumin radieal at 490 nm
and Ihe growth of the 1'rolox radical at 430 nm respectively. These occur on the same
ti mescales. Inset b shows the initial 'immediate' fOIUlation of the Trolox radical as a
consequence of competition for the ilzide radical. The spectra in FIgure 5 clearly show
the loss of the cmcumi n radical at 490 Hill with a concomitant increase in the intensity of
Ihe Trolox radical at -130 nm. An isosbestic poini close to 450 11111 is apparent. Again
these data confiml that the scavenging involves of the cUTCllmin radical.

It is clea r from the e.xperiments desl"fibcd that the molecule cllTCumin can access both
hydrophobic aud hydrophilic environments of a microhelerogeneous system. In
addition, the corresponding r'ldical is stable with respect 10 oxygen, aT least on our
experimental but can be repaired by both the naturalaJltiuxidanfs vitamins C
and E in hydroph il ic and hydrophobic environments respectively. It is currently our
view combi nation of character radical stability/repair behavior
referred to is the key to the importance of cure urn in as anticareinogenlc agent. One
can imagine a scenario where mole(;ules of thi s type as an interphasi c shunk for
radical repair.
Ack n owle dgme nt s
Experiments werc perfortlled at the Paterson lnst iwte for Cancer Research Free Radical
Research Facility, the Christ ie Hospital NBS Trust, Manchester. TIl e Fa(;ility is funded
under the European COlllmissiOl1 TMR PROGRAMME - ACCESS TO LARGE SCALE
242 SPI CJ.:S: Fl.AVOR CHEMISTRV ANI) ANTIOXIDANT !IWI fRTIE. ...
00
0.01
!
. ,
b
, ,

0.0
450 nm 550
350
Fillure S. Tr:msient absorptio n spe!; lr;t me:lsured 111 IlS (X), 171 }.Is (0 ),327 ).Is
(6) and 576 (V) after pul se radiolysis of;t soit llion of curcumin (2.() x 10,5 mol
\- 1). sodium (2.0 x 10-2111011,1) in N20-satur:lled pH 7 buffer conmining
Trolax (2.0 x \0-5 mo l J. I), ]11SCIS: 1in)( dependt':llCC: o f (a) transient decay at 490 nm
wit h rirsl -order fi l, k' = J.3 x 10
3
S l , 0.4 % absorplion/ division, 150 ).Is/division
and (b) cOrTcsponding growth of absorption :1I 430 mn with firs l-order 111,1:' '" 3.2
x loJ s I, OA % absorption/division. 150 ).ls/divi sion.
19. GORMAN":1' AI_ Curcumin: A Pulu Radiolysis Invutigation
f ACILITI ES. Gnln! ERBFMGECT 950084 Access 10 Pi e R FRR Facility. Thi s work
was supported by NIH Grant I RJ5GM4435-0 JAI (VSS) and the EPSRC (U. K.).
Literature Ci ted.
I.
2.
3.
4.
5
6.
7.
8 .
9.
10.
Tonnesen, H. ; De Vries, I I. ; Karlson. J .; V3n t Jenegouwen, G. G. J . Pharm.
Sci . 19K7, 76, pp. 371 -373.
Dahl, T. A.; McGow;m, M. A.; Shand, M. A.; Srinivasan, V. S. Arch.
Microbiol. 1989, /5/, pp. 183-185.
$h:lf111:I, O. P. 810(."111:111. Pharmacul. 1976,25, pp. 18111812.
Nag.lhh ush:ul, M.; Bhide , S. V. Nurr. CO/l cer 1986,8, pp. 201-210.
Harlfll:UI. P. E.; Sh;U1kel, D. M. Environ. Mol. MUloge/l. 1990, /5, pp. 145-
182.
Nagabhushan, M.; Bhide, S. V. J . Am. Coli. Nutr. 1992. II , pp. 192-198.
Gomlan, A. A.; Hamblet!, I.; Srinivasan, V. S.; Wood, P. O. Photochem.
Photoblol. 1994,59, pp. 389-398.
Hayon E.; Simic, M. J . Amer. Chem. Soc. 1970.92, pp. 7486 7487.
Butl er, 1.; Land, E. 1.; Swallow, A. 1.; Prutz, W. Radial. Phys. Chern. 1984,
23, pp. 265-27(1.
GOflll:lI1. A. A.; Hamblett, I. Chem. Phys. Lett. 1983,97, pp. 422-426.
,
,

,
,
I
,
I
!
I
.I
I
i INDEXES
I
,
I
,
,
I
01
Ii
.l
P
Ii
Ii
~
..
. ,
~ ;
,I
N
h'
o.
"
~ ; , '
[",
I'
I,
,.
f,
I ,
: ' , ~
:'
~ - :
ii
J
r
,
,

t"
I
Author Index
Dad:. Hyung Bee, 66
Barcel6, A. Ros, 206
Bene lscn. Grete, 176
Blank. Imre, 12
Block, Eric. 113
Cadwallader, Keith R., 66
Cai, Min, 66
Calderon, A. A., 206
Calvey, Elizabeth M., 113
Chen, C. C., 160
Chen, Chung-Wen. 188
Cheng, C. K., 160
Devaud, Stephanie, 12
Darn, Ruth, 29
Fay, Laurent B., 12
Feng. H. H., 160
Fujillori, E. M., 98
Fumeaux, Rene, 12
Gorman, A. A., 234
Hamblen, I .. 234
lIavldn- Frenkel, Daphna, 29
Hill. T.l., 234
Hi serodl, Richard D., 80
Ho, Chi-Tang, 80,188
Jones, H., 234
Lawrence, Brian M., 138
Lee, Tung Ching. 188
leistrilz, W", 7
Lin, Jianming. 12
Liu, Yaguang. 199
L6pez-Amaldos, T . 206
Madsen, Helle Lindberg, 176
Masuda, Toshiya, 219
Randle, Will iam M., 4 1
Risch, Sara J., 2
Rosen, Robert T., 80, 125
Shih, L. L., 160
Shu, Chi-Kuen, 138
Shu, W. Y. , 160
SkibSled. LeifH., 176
Srinivasan. V. S . 234
Wood, P. 0" 234
Yu, Tung-Hsi, 53
Zapata, J. M" 206
Zrybko, Carol L., 125
Affiliation Index
Antek Instruments, Inc" 98
Bowli ng Green Stale 234
Da-Yeh Institute of Technology, 53
David Michael & Company, Inc., 29
Food Industry Research and Development
Institute, 160
L. Y. Research Corporation. 199
Mi ssissippi State University. 66
Nabisco. Inc., 125
National Tsing-Hua University, 160
Nestec Limited, 12
R. J. Reynolds Tobacco Company, 138
RUl gers. The State University of New
Jersey, 80, I 25, 188
Science By Design, 2
SpiceTec Limited, 7
State University of New York-
Albany. 113
The Royal Veterinary and Agricultural
Universi ty, 176
U.S. Food and Drug Admini stration. 113
UniverSity of AJcall'i de Henares, 206
University of Georgia. 41
University of ManchesICr, 234
University of Murcia, 206
University of Tokushima. 2 I 9
246
INDEX
247
Subj ect Index
A
Aduileration. role in commercial essenti al
oil composition, 153,155- 159
Agronomic production of vanilla,
30-34,36/
S-A1k(en)y1-L-cysteine S-oxide flavor
precursors. role in onion flavor. 41-50
Allium
supercritical flui d extraction, 112- 123
volatile component fonnalion. 53-54
Allium (lsca/onicum aue!., Set Shallot
Allium ecpu L., Set Onion
Alliumjislu/osum L., See Welsh onion
Allium sat;vum L. , Set Gar lic
S-Ally1cyst eine sulfoxide, identification
of volatile compounds. 591,63
3-Alnino4,5-di methyl -3.4-di hydro-
2(5H)-furanone, formation of sotolone,
22-23
Anti -inflammatory antioxidants.
anti -inflammatory activity. 225,226r
Antioxidant activity
relationship to rosmarini c acid content
for iavandi n cell cultures, 206-2 17
spices, 6
Zingiberaceae plant rhizomes,
219,221/
Antioxidant(s)
delay of oxidation, 177,179
in spices and extracts
activity. 184- 185
evaluation of radical scavenging.
179-184
identification. 179- 181/
natural, 219
syntheti c. 206
Amiphotooxidation index, caps3l1thin,
188- 197
Aroma extract dilution analysis.
saffron flavor characterization. 66-78
B
Bacterial reduction in spices
dry steam procedure, 9-10
et hylene oxide gas procedure, 8
irradiation procedure, 8-9
methods, 3-4
wet steam procedures. 9
Benzoto:lpyrene. phase I oxidation. 81,83/
Biological antioxidant, curcumin radical
as model. 234-242
Biosynthctic pathways, production
of vanilla. 37
Biotechnological production of vanilla
biosynthetic pathways, 37
micropropagation, 35-37/
secondary products. 35,38-39/
Blood platelet aggregation. effect
ofcurcumin, 199- 205
Bra55ieo, See Mustard
c
a-O-Caffeoyl-3 ,4-dihydroxyphen y11actic
acid, See Rosmari nic acid
Capsaicin. analysis using chromatography
and chemiluminescent nitrogen
detection, 98- 111
Capsaicinoids, analyt ical methods.
98.101- 104
Capsanthin
antioxidative effect, 189
concentration in paprika. 189.190/
effect on chlorophyll-sensitized
photooxidation of soybean oil
and flavor compounds
2,5-di methy[-4-hydrox y-3(21f)-
furanone, 192.193/
2.ethylfuran, 192,193/
2.4.5-trimethyloxazole, 192.193/
experimental procedure, 189-192
I
..
24S SI'ICES: FL.,\VOR CIIEl'ollSTkY AND ANTIOXIDANT PROPERTI ES
Clpsanthin- COIrtilllled
induction ti me, 192,194- 195
kinetics, 195
quenching mechanism, 195-197
p. Carolcne, antioxidative effect,
188- 197
Cassumunin A, synthesis. 225,227[
Cclyllrimcthylammonium bromide
micelles, curcurnin radical formation,
238/.239
Chcmi luminesccnI nitrogen detection,
pungent flavor profiles and components
ofspiccs,98- 111
Chlorophyl l sensilized pholooxidalion
(1f soybean oil and navor compounds,
188- 197
Cholesterol levels in blood, effect
o/" curcumi n. 203
Chromatography with CLNO to analyze
pUllgenl navor profiles and components
of spices, 98- 111
Ci im'l!c, role in commercial essential oil
composition, 139- 146
Commercial essential oil composit ion
variation faclors
ildultcrmion, 153,155- 159
cl imate, 139- 146
experi mental descripl ion. 138
geographic origin, 139- 146
growth o f hnrvcslcd plant,
149.1 51- 1 52/,1541
infraspecifie differences. 149- 151 I, 1541
differences, 146- 150
processing parameter. 149, 152- 153,155/
Component analyses o f nuxed spices
accuracy, 171
experi mental procedure,
161,165- 166,167/
methods, 160-161
num..:rical :malysis, 165
simi lar ity index, 162- 164
vector representat ion o f
161-162
Crocus sflli v /Is L,. See Saffron
Cullivar, role in flavor intensity
of onions, 43,45,47
Curcuma domes/ic(/. anti-inflammatory
antioxidants, 220
C/lrcum(/ 10llga L., See Cureumin
Curcuma :calllhorrhiz(/, anti-inflammatory
antioxidants, 220,2231
Cureumin
antioxidant properties, 6
applications, 199
biological activity, 234
cholestcrol 1eve1s in blood, 203
description, 80--8]
experimental description, 200-202
lipid level . 203-204
lipid peroxidation, 204
Iymphoblastoid transformation,
204-205
macrophage, 204
piatelet aggregation, 203
properties, 199
radical formation, 234-242
reaction mechanism as free radical
scavenger, 81.821
resonance stabi lization of frce radical,
8 1,82/
structure, 234
Curcumin analogues
anti -inflammatory activities,
, .
226t,232
antioxidant act ivities, 225,228-232
synthesis, 225,228/
Curcumin radi cal in miceihrr systems
absorption spectra, 235.236/
experimental procedure, 235
pulse radiolytic formation of radical,
235,237/.239
radical formation in water, 240/.24 1
scavenging of radical
in cetyltrimethylammonium bromide
micelles, 238/,239
in sodi um dodecyl su lfat e and Triton
X-JOO micelles, 239,241
in water, 241,242/
Curcuminoids, isolation and synthesis,
2 19-232
Curing, steps, 34
Curlone, electron ionization- MS, 92/,96
INIlEX
D
Dihydrocapsaicin
analysis using chromatography and
chemiluminescent nitrogen detection,
98- 111
structure, 98,100/
2,5- Dirncthyl-4-hydroxy.3(2/ 1)-furanone,
role of capsanthin on chlorophyll -
sensit i7.cd photoox idation, 192,1931
Direct solvcnt extraction, saffron flavor
charactcrization, 66-78
Direct thennal extraction GC-MS
to characterize components of
lunncric, 80- 96
Dry stearn systcm, bacterial reduction in
spices, 9- 10
E
Electron spi n resonance spectroscopy
eval uat ion of antioxi dative activity
of spices and cxtracts, 176-185
eval uat ion of radical scavenging activity,
179,182-184
Electrospray-HPLC- MS. glucosinolate
determination in mustard, 125- 136
Esse ntial oils
factors affccting composit ion vari at ions,
138- 159
role in flavor, 160
See II/ SO Commercial essential oil
composit ion variat ion factors
Ethylene oxide gas procedure, bacterial
reduction in spices, 8
2.El hylfuran, role of capsanthin on
chlorophyll -sensitized photooxidation,
192,193/
r
Fats , scnsitivity to photooxidation, 188
Fenugreek
aroma composi tion, 17- 20
experimental descript ion, 13-16
flavor components, 13-26
24.

formation of sotolone from preeUfSOI' S,
22-26
quantilation of sotolone , 21
stereoisomeric characteriution of
sotolone, 20-21
volati le const ituents, 12
Flavone, role in pharmacelllicai effect of
curcumin, 199-205
Flavor compounds, role of capsanthin on
chlorophyll- sensitized photooxidation,
188- 197
Fl avor distribution, role in flavor
intensity of onions, 47
Fl avor of thcrmally processed Allium
vegetables, contribut ion of
nonvolatile sulfur-containing flavor
precursor of Allium, 53-63
G
Garlic
contribution of nonvolatil e sulfur-
containing flavor precursors to
flavor, 53-63
pungent flavor profiles and components
using chromatography and
ehcmil uminescent nitrogen detection,
106.108- 109/
supercriticaJ fluid extracti on, 112- 123
Gas chromatography-chcmiJuminescent
nitrogen detection, 102, I 03/
Gas chromatography-MS, saffron flavor
characterization, 66-78
Gas
description, 5,67
saffron flavor characterization, 66-78
Geographic origin, role in commercial
essential oil composit ion, 139-146
Glucosinolate determination in mustard
using HPLC-electrospray- MS
accuracy, 135
applications, 126
experimental procedure. 126-128
identifi cat ion methods, 126
nonvolatile compounds, 128
I
I
25. srlCI>S: .' LAVOR CII MISTRY AND ANTIOXIDANT PNON! RTI F.5
Glucosinol ate dct cnni nation in mustard
using HPLC-eJeclrospray-MS-
Continued
plllgoilri n. 130,131/
sinalbin, 130, 132/
sinapine. 133, 135,136/
sinigrin. 133.1341
Ii
Il igh-performance liquid chromatography-
c1ecl fospray-MS. gtucosinolate
determination in mustard, 125- 136
Hybrid peppers, degrees of hOlness, 98
Hydrophilic environment, curcumin
radical formation, 234-242
Hydrophobic environment, curcumin
rlldical formation, 234-242
). H ydrox y4 .5-dimclhyJ -2( 5H)-fu ranone,
See SOlolooc
4-Ii ydrox y. L. isoleucinc
formation of sotolone, 22
struclUrc, 13
2-Hydroxy-4,4,6-lrimclhyl -2,S-cycJo
hexadicn- I-one, Set Saffron
Hyperlipidemia. curcumin effect. 199-205
Immune func tion, effects of cun:umin,
J99- 205
Induction time, capsanthin, 188-197
Infraspeci fi ci ty, role in commercial
essential oil composi tion, 149-151/, 154/
hllcrspecifici ty, role in commercial
essenti al oil composition, 146-150
Iron-che/mi ng activity, rosmari ni c acid,
2 13,2 16-217
IlTlldi ation, bacterial reduction in spices, 8-9
L
Lavandin cell culture(s)
antiOJlOidant activity related 10 rosmarinic
acid content, 206-217
occurrence of rosmarinic acid,
207-208,210-21 1
Lavandin cell
superoxide anion scavcnging activity,
211-212
Lavandula x See Lavandin
cell eulture(s)
Lipid(s), generation of compounds, 54-56
Lipid level, effects of curcumin, 203-204
Lipid peroxidation, effects of curcumin,
199-205
Lutein, antioxidative effect on
photooxidation of 2-ethylfuTlln, 188-197
Lymphoblastoid transformation, effccts
of curcumi n, 204-205
M
Macrophage, curcumin effect, 204
S-Mcthylcysteine sul f01dde, id;:ntifi cati on
of volatil e compounds, 6Ot,63
Micellar systems, curcumin radi cal
formation, 234--242
Micropropagation, production of vanilla,
35-37/
Miled spices, componenl analyses,
160--173
Mass spectrometry_H PLC-elccttOspray,
glucosinolate determination in
mustard, 125- 136
Mustard, glucosinolate determination
using HPLC---electrospray-MS, 125-136
N
Natural antjolidants, phenols, 219
Nitrogcn-containing compouuds, source
of puugenl flavors in spi ces, 98
Nonvolatile sulfur-containing Oavor
precursor of AlliwlI, contribution to
flavor of thermally processed Allium
vegetables
volatile compounds generated from
thermal degl"Jdation or thermal
interw.:t ions or alk(en)yl cysteine
sul foxidcs, 58-63
volati le compounds gcnerJtcd ftOm
thermally treated blanched Allium
vegetables_ 54- 58
INllEX
Nordihydrocapsaicin, analysis using
chromatography and chemiluminescent
nitrogen d;:tec\ion, 98-1 11
o
Onion
contribution of nonvolatile sul fur-
contai ning flavor precursors to
flavor, 53--63
factors affecting flavor intensi ty
cult ivar, 43,45.47
environmental factors, 46/- 50/
flavor distribution, 47
flavor chemistry, 42-43,44/,1
pungent flavor profiles and components
using chromatography and chemi-
luminescent ni trogen detecti on,
106,108- 109/
supercrit ical fluid extracti on, 112- 123
Oxidation
delay using antioxidants, 177,179
frce radical generation, 177 ,178/
p
Paprika, pungent flavor profiles and
components using chromatography and
chemil uminescent nitrogen detect ion,
106,111/
Peppermint oils, composi tion variation
factors , 138- 159
pH, role in sotolone formation, 2S
Phenoli c compounds in plants, related
to ant ioxidative activity, 176-179
Photooxidation
sensit ivity of food constit uents, 188
soybean oi l and flavor compounds,
capsant hin effect , 188- 197
Photosensiti zed oxidat ion reaction, 189
Piperine
pungent fl avor profiles and compouents
using chromatography and chemi-
luminescent nitrogen detection,
106,110/
structure, 98,100/
-,- -
251
Platelet aggregat ion, effects of
curcumin, 203
Proct:ssi ng parameter, role in commercial
essential oil composition,
149, 152- 153,155t
Progoitrin, determination in mustard
using HPLC-electroSpray- MS. 125-136
S-( +)-cis- I-Propenyleysteine sulfoxide,
identification of volatile compounds,
621,63
S-Propylcystei ne sulfoxide,
ident ificat ion of volatile compounds,
61t,63
Pulse radiolysis, curcumi n radical
formation in micellar systems,
234-242
Pungent flavor profiles and components
of spices anllly7.cd by chromatography
and chcmiluminescent nitrogen dctecti on
advantages, 102, I 03/
detection mechanism for nitrogen
determination, 105-106
experimental procedure, 102,105
HPLC--chemilumi ncscent
nitrogen detection
garlic, 106, 108-109/
onion, 106,108-109/
paprika oleoresin, 106,111/
piperine, 106, 110/
red chit i powder, 106,107/
linear response, 101
R
Radi cal seavenging activity, evalUlltion
usi ng ESR spectroscopy, 179,182- 184
Ramp, supercri ti cal fluid extraction,
112- 123
Rancimat method, antiphotooxi (i;ltive
effect of carotenoids on induction time
of soybean oi\. 191-192, 194- 195
Red chili powder, pungent fl'lvor proJ'jks
and COmponents using c hwlllal",!!ral'hy
and chemiJuminescellt nilf11gl'U ,1<; 1",t i" n.
106,107/
I
I

,
r
i'
!
t
I
I
i

"
r
,
:1
:i
I'
I

,
t

i
I,
" ) .. ,
252 SPICES: I'I. AVOR CII I':MISTRY ANDANTIOXIIlANT I'ROI' I:: RTIF.5
Rosmarinic acid
iron-chelnting activit y. 2 13,2 16-2 17
occurrence, 207-2 ! 1
struct ure, 207,2081
supcroxi dc Dnion scavenging activity.
213,214j,215f
s
Saffron, protIucl ion, 66
Saffron flavor characterization using
aroma extract dilution analysis
aroma-acli vecomponcnts, 74-78
detection of aroma-active component
using GC-olfactomctry, 67
experimental procedure, 67-68
volatile components, 68-74
Sa[ranal, role in navor, 66-78
Semh'olatile component s in powdertd
turmeric, characterization usi ng direcl
thermal ext raction GC- MS. 80-96
ShailOl. contribut ion of nonvolatile
sulfur-containing flavor PfcClIrsors to
navor, 53-63
Sillalbin, determination in mustard
HPLC--clcctrospray- MS, 125-136
Sinapinc, determination in mustard using
HPLC--clectrospray- MS, 125- 136
Sini gri n, determination in mustard using
HPLC--clectrospray-MS, 125-136
Sodium dodccyl sulfate micelles,
cureumi n fadical format ion, 239,241
Soto!onc, 12- 26
Soybean oil. role of t;apsanthin on
chlorophyl l-sensitiud photoox idation,
188- 197
Spcurmint oi ls, composi tion variat ion
f3cto/'S,138- 159
Spi.;cs
,mlioxidati \'c activity, 6,176- 185,206
bacleri;11 reduction, 3-4,7- 10
component analyses, 160-173
factors affect ing consistency, 5-6
rcsearch into active component s, 5
role of oils in favor, J 60
Sterili7.ation techni ques. development
for b;lcterial reduction in spices, 7
Sugars, generation of compounds,
54-56
Sulfate fertility, role in flavor
int ensity of onions. 46/.47-49
Supcrcritical carbon dioxide, cxt raction
of Allium species. 1[3-123
SupcrcrilicaJ fluid chromalography-
chemi lumi neseent nitrogen detcction,
\02.104/
Supereriticai fluid extraction
Allium species
conditions 'Is. organosulfur compound
di stribution, 120/, 121.122/
cxperimcntal procedure, 121, 123
extraction condi tions. 115,1161
garlic. 117- 119/
onion, 117,121
ramp. J 17,119/
in food industry, 113
entrainer, I 13,115
patents. 11 3,1 141
system components, 11 5
use for analysis of spi ces, 5
Superoxide anion scavenging activity
lavandin cell culture extracts,
211- 212
rosmari nic acid, 213,214/,2 151
Synthetic antioxidants. 206
T
Thermosprny. description, 8 1
L , Ste
Fenugreek
2,6,6-Trimcthyl-I ,3-cyc1ohexadiene-l-
(;arboxaldchyde, Ste Safranal
2,4, 5-Trimethylollazole, role of
capsanthin on chlorophyll -sensiti zed
photOQx idati on, 192. 193/
Tropical Zingiberaceae plant s posscssing
ncw ant i-inflammatory ant ioxidants
ant i infl ammatory activity of isolated
antioxidant, 225,2261
Curcuma domestica, 220.222-223/
Curcuma :rMI1!orrhiZfI, 220,223/
ZillgibercclSsumar.220,224/
INDEX
Turmeric components characterized
by direct thcrmal extraction GC- MS
di rectt hemlal ext raction chromatogram,
87,88/
clectron ionization- MS for component s,
90-93,96
experi mental procedure, 84-87
methods, 81,84
semi quantitati ve values fOl' volatile
and semi volatile components.
87,891,96
structures. 94-95/,96
thcrmospray LC- MS chromatogram.
8 1,83/
ar-Turmcronc, electron ionization- MS,
90- 91/.96
Triton X- I 00 micelles, cureumin radical
formation, 239,241
v
Vanill a
agronomic production. 29- 34,36
biotechnology production, 34-39
secondary products, 35.38- 39/
Vanilla I' lani/ofia. Su Vilnillil
Volatile in powdered tU9
characterization lIsing direct thermal
extraction GC- MS. 80- 96
w
Welsh onion. contribution of nonvolatile'
sulfur-containing flavor precursors to
fla vor, 53-63
Wei steam system, bacterial reduction in
spices, 9
z
Zil1giber Ca,{SUmUllar
anti-inflammatory antioxidants,
220.224/
synthesis of cassumunin A, 225,227/
Zingiberaceae, members, 80
Zingiberaceae plant(s), source of
antioxidants. 219
Zingiberaceae plant rhizomes, antioxidant
activity, 219- 220,221/
Bestsellers rrom ACS Books
Tht ACS Styl' Guid,: A Ma,,,ud/Dr Author, and Edit""
Ediled by Janel S. Dodd
264 pp; clothbound ISBN 0 -8412-0917...{); paperback ISBN o-&412-0943-X
II'rinll, Iht u.boT"tJlory Nottbook
By Howard M. KllrWc
pp; elothbound ISBN 0-84 papc.back ISBN 0-S412...{)93l-2
CDr'" TT"tJ II Jitioll$/lJr Ch'mi5/J
By Dorothy P. Rodmann, Donald D. Bly, I'rederick H. Owens, and AnI\e- CI.i,e Anderson
240 pp; clothbound ISBN 0-84 12-3052-8; paperback ISBN 0-3412-303B-2
Ch,,,,ical Acti.iti,s (.,udenl.OO le:aeher edilion,)
By Cbri slie L. BOflford and Lee R. Summerl in
330 pp; spirulbound ISBN 0-84 12- 1417-4; cdilion, ISBN 0-8412-1416-6
Chtmical D' ", DItI/I"f1/ilJllS: A Sourt"tbook/or Tt at:htrs, Volltm" I and 2, Second Edition
Voh.mx: I by Lee R. Summerlin and bmes L. Ealy, Jr,
198 pp; spiralbound ISBN
Volume 2 by I...a: R. Summerlin, Christie L BorBfOld, and Julie B. r:.aJy
234 pp; spinlbound ISBN
FrDm C" .ttnllll /o Chemist
By Hugh W. SalzbcrB
300 pp; clolhbound ISBN 0-84 12- 11116-6; papc.back ISDN 0-&412- 1781-4
n., III/,m,,: A GltjJ'/"r Ch,,,,"u
Edited by Sleven M. Bachr:ll;h
360 pp; clothbound ISBN 0-8412-3223- 1; papc.back ISBN 0-8412- 3224_S
u.bortJlory 11'11$1' M"""I' ''''III: A Guid,book
ACS Task Force on Laboratory WIlSIe Mana,.,,,,,,nl
250 pp: clothbound ISBN 0-&412- 2135- 1; pape.back ISBN 0-8412- 2849-3
R'''t'", Ch, m;':,,/!, Eighlh FAli" on
100 pp: clo.hbound ISBN 0-S412- 2502-8
GiJod l.DbaT"tJlary PrtJt:tic, Siandanis: Applicari,,", /fJr H,Id and l.Dborotory Sludi,s
Edited by Willa Y. Garner, S, Barge. and James P. USW")'
pp; clOlhbound ISBN 0-1W12- 2 192-8
For fu,'her informatiOIl concoct:
American Chemical Socie. y
1155 Sixleench SItI, NW Washinlton. OC 20036
Telephone 800-221-9919 202- n6-8l00 (oulside: U.S.)
"The ACS Publicltion$ Cata10C i$ on the Inlernel a.