Said the researchers, “by measuring theabsorbency at 1 min, we have demonstrated that accuracy could besubstantially improved.” The results were published in “Underesti-mation of Pyruvic
Trang 2Industrial Applications of Selected JFS Articles
We are seeing just how important flavor compounds are—the
amount of pungency in onions, the characterization of flavor in
certain kinds of honey As standardization of flavor becomes key,
products can even be in trouble for tasting better than they should!
Instrumentation helps—the human tongue and nose is the final
judge, but numbers can help More papers are appearing with
work on nanotechnology, as well; an understanding of how this
new science works is imperative to the up-to-snuff technologist
(and engineer and marketer)
Learning a Little More about Nanoemulsion Phases
A paper titled “Optimization ofβ-Casein Stabilized
Nanoemul-sions Using Experimental Mixture Design” outlines the changes
in viscosity and glass transition temperature that exists in the
con-tinuous phase of nanoemulsion systems, and how those attributes
affected stability of a product Emulsions were made that included
β-casein at 4 levels, lactose and trehalose at several levels Higher
levels ofβ-casein content resulted in increased viscosity and
de-creases in melting temperatures A mixture design was used to
predict the optimum levels of lactose and trehalose needed to
reach minimum and maximum Tg and viscosity in solution at
fixed protein contents C1108–1117
Keeping Probiotics Useful
Freeze-drying probiotics and storing the dried powder, the usual
way to process the probiotics for commercial use, appears to reduce
the viability of the materials A group of researchers are looking
into the efficacy of using a glassy state for the little varmints, and
retaining the glassy state during drying In “Role of Glassy State
on Stabilities of Freeze-Dried Probiotics” the authors describe the
glassy state that should be retained during freezing, drying, and
storage of cells Insight into the role of glassy state has been largely
adopted from studies conducted with proteins and foods, especially
sugars Say the researchers, “Current understanding of the role of
the glassy state on viability of probiotics is not only valuable for
the production of fermented foods and nutraceuticals but also
for the development of nonfermented functional foods that use
the dried powder as an adjunct.” The paper, a review of recent
findings regarding glassy states and the effects on probiotic survival,
notes that 1000 articles and reviews were published in 2008, and
more information is appearing, supporting the introduction of $16
billion in the U.S alone in 2008 R152–156
How Pungent Is that Onion?
In determining how pungent an onion is, one standard reacts
onion juice with 2,4-dinitrophenylhydrazine (DNPH) It’s been a
standard test since 1961, measuring the pyruvic acid concentration
in onion juice However, researchers found that the absorbency ofthe color adduct of the reaction dropped fast, as compared to that
of the pyruvic acid standards, which meant that the pyruvic acidwas higher than estimated Said the researchers, “by measuring theabsorbency at 1 min, we have demonstrated that accuracy could besubstantially improved.” The results were published in “Underesti-mation of Pyruvic Acid Concentrations by Fructose and Cysteine
in 2,4-Dinitrophenylhydrazine-Mediated Onion Pungency Test.”Alliinase action in juice (fresh or cooked) and bulb colors did notaffect the degradation Some organic acids indigenously found inonion, such as ascorbic acid, proline, and glutamic acid, did not re-duce the absorbency However, fructose within the onion juice orsupplemented caused the degradation of the color adduct, whereassucrose and glucose also affected the test, but had a lesser effect.Degradation rates increased proportionally as fructose concentra-tions increased up to 70 mg/mL Cysteine was found to degradethe pyruvic acid itself before the pyruvic acid could react withDNPH This happened rapidly So if the degree of pungency ofonion ingredients is important to control, this study is important
C1136–1142
Films for Packaging Are Built with Agar, Special Clays
Water vapor adsorption behaviors are changed in nanofilmsmade from agar and nanoclay according to the type of nanoclaysused In the paper titled “Water Vapor Adsorption Isotherms ofAgar-Based Nanocomposite Films”, the author discusses methodsused to get an understanding of the effects of various nanoclays:The Guggenheim–Anderson–de Boer (GAB) isotherm model pa-rameters were estimated by using both polynomial regression andnonlinear regression methods, and the behavior of the nanocom-posite films was found to be greatly influenced by the type ofclay It was found that the GAB model fitted adequately for de-scribing experimental adsorption isotherm data for the film sam-
ples The monolayer moisture content (mo) of the film samples
was also greatly affected by the type of nanoclay used, that is,
mo of nanocomposite films was significantly lower than that of
the neat agar film For background information, for those notinto nanotechnology, bionanocomposites are hybrid materials ofbiopolymer with inorganic fillers which have at least one dimen-sion in the nanometer scale It is generally known that nanoscaledispersion of the filler phase in the polymer matrix leads to prop-erty enhancements such as decreased permeability to gases (O2,
CO2, and water vapor) and liquids, better resistance to solvents,increased thermal stability, and improved mechanical properties
N68–72
Trang 3Speaking of Word Count .
About a year ago, we informed the potential authors of future
manuscripts submitted to sections of the Journal of Food Science
that we would be limiting word count on manuscripts In this
action, we were trying to decrease the number of printed pages
particularly for JFS while enhancing the readability of the papers.
In a survey of printed papers at that time, Amanda Ferguson
discovered that most manuscripts to JFS were a little over 5,000
in word count, and manuscripts in Comprehensive Reviews in Food
Science and Food Safety were all over 10,000 words With those data
as the guideline, we set a JFS research manuscript word count limit
(excluding tables and figures) at 5,000 words and minimum word
count for CRFSFS at 10,000 words Starting November 2011,
au-thors were asked to voluntarily submit word count on manuscripts
and in January 2011, words count was required Needless to say,
we had some flexibility in applying the word count rule On the
other hand, I received not one letter to the editor decrying the
use of word count or complaining that it was too limiting Nor
did I receive any oral argument while I attended the IFT AMFE
in June I take this inaction as a sign of not only acceptance but of
support for word count limitation
With that in mind, I asked Amanda to again do a survey of word
count on our accepted manuscripts The data for a sampling of
papers published in 2011 are in and they tell the following story
JFS-Concise Reviews Avg. 7893.67 5266-12606
JFS-Research Sections Avg. 4544.28 1930-8454
Of the 141 papers in the JFS research sections, only 18 wereover the 5,000 word limit, and only 3 were more than 1000 words
over The mean of 4500 word count for the research sections of JFS
compares to the October 2010 reported mean of 5140 and range
of 2350-8560 Only 1 review paper in JFS was over the 10,000
word limit The effect of the decrease in word count was to duce the page count per paper from 7.3 to 7.0 In other words, thechange achieved the result we were hoping for and there were notuntoward consequences Now, for the kicker, should we be look-ing at a 4,500 word count limit for research papers submitted to
re-JFS for 2012? Send your thoughts to me at dlund@cals.wisc.edu.
Daryl LundEditor in Chief
Trang 4Chalat Santivarangkna, Mathias Aschenbrenner, Ulrich Kulozik, and Petra Foerst
Abstract: High viability of dried probiotics is of great importance for immediate recovery of activity in fermented foodsand for health-promoting effects of nutraceuticals The conventional process for the production of dried probiotics isfreeze-drying However, loss of viability occurs during the drying and storage of the dried powder It is believed thatachieving the “glassy state” is necessary for survival, and the glassy state should be retained during freezing, drying, andstorage of cells Insight into the role of glassy state has been largely adopted from studies conducted with proteins andfoods However, studies on the role of glassy state particularly with probiotic cells are on the increase, and both commonand explicit findings have been reported Current understanding of the role of the glassy state on viability of probiotics isnot only valuable for the production of fermented foods and nutraceuticals but also for the development of nonfermentedfunctional foods that use the dried powder as an adjunct Therefore, the aim of this review is to bring together recentfindings on the role of glassy state on survival of probiotics during each step of production and storage The prevailing state
of knowledge and recent finding are discussed The major gaps of knowledge have been identified and the perspective ofongoing and future research is addressed
Keywords: freeze drying, glass transition, lactic acid bacteria, probiotics
Introduction
Research and interest regarding probiotics is receiving more
focus than ever before For example, in 2008, more than
1000 articles and reviews were published on the subject and more
than 2000 probiotic products launched (Jankovic and others 2010)
The global market for probiotic products was worth US$16 billion
in 2008, and the estimated target is a total of US$19.6 billion in
sales in 2013 (Granato and others 2010) For food applications,
probiotics are mainly employed alone or together with starter
cultures in fermented dairy products (for example, yogurts and
cultured drinks) In addition, probiotics are also potentially used
as nutraceuticals and dry adjunct in nonfermented and nondairy
products, such as fruit juices, cereals, dried powder foods
(Rivera-Espinoza and Gallardo-Navarro 2010) In either case, high
vi-ability of probiotics is of great importance in order to ensure
immediate recovery of fermentation activity (especially Direct Vat
Inoculation-DVI cultures) and to meet the minimal requirements
for health-promoting effects
The standard process for the production of dry probiotics is
freeze-drying The typical freeze-drying process consists of 3 steps:
freezing, primary drying, and secondary drying During these
3 steps, cells are exposed to various stresses, especially
dehydra-tion, compromising cell survival A further survival reduction
oc-curs during storage of the dried powder, where the key storage
conditions such as relative humidity and temperature play a major
role In addition to physiological states of cells, the physical state
of the surrounding sample matrix is believed to be critical for
sur-vival of probiotics The high viscosity of a surrounding amorphous
MS 20110302 Submitted 3/9/2011, Accepted 6/21/2011 Authors are with Chair
for Food Process Engineering and Dairy Technology, Centre of Life and Food Sciences,
Technische Univ M¨unchen, 85354-Freising, Germany Direct inquiries to author
Foerst (E-mail: petra.foerst@wzw.tum.de).
glass can inhibit diffusion and slow down deleterious reactions orchanges in the structures and chemical composition The role
of glassy state on the functionalities, bioactivity, and stability ofenzymes, pharmaceutical proteins, and foods has been widely re-ported Nevertheless, current studies on the role of glassy state,particularly with probiotics, are on the increase and both com-mon and explicit findings have been found It is an aim of thisreview to bring together recent findings on the role of glassy state
on survival of probiotics during each step of the production and
of the storage period The prevailing state of knowledge related tothe role is revisited, and recent finding is discussed
Glassy State
The glassy state is referred to an amorphous metastable statethat resembles a solid but without any long-range lattice order,that is, the position of molecules relative to another is more ran-dom In other words, it has a solid characteristic/appearance but amolecular arrangement that is more typical for liquids A glass has
an extremely high viscosity (for example, typically≥ 1012 Pa s)and shows temperature-dependent transition (Slade and Levine1991) When a glass is heated above a certain temperature, themolecules gain translational mobility and enter a liquid-like state.The transformation of solid- to liquid-like state is known as glasstransition The most common parameter describing the glassy state
and its transition is the glass transition temperature (T g or T gfor
a maximally freeze-concentrated solution), below which
mate-rials exhibit the extremely high viscosity At T < T g, diffusionlimited deterioration reactions are inhibited because water in theamorphous glass is immobilized and unavailable
In various studies, sugars or strong glass-forming polymers have
been added in an effort to increase T g of dried probiotics This
is based on the observation that many anhydrobiotes accumulatelarge amount of sugars, especially trehalose and sucrose, inside thecells, for example, 25% of DW (Buitink and Leprince 2004) The
C
r Vol 76, Nr 8, 2011
Trang 5presence of sugars on both sides of the cells was considered being
necessary for protection against freezing (Duong and others 2006)
and dehydration (Tymczyszyn and others 2007) Crowe and others
(1987) found that intra- and extracellular sugars provide protection
of proteins and membranes by forming a glassy matrix that is able
to interact via hydrogen bonding This protective effect of sugars
on biomolecules was reported to appear only in case of
amor-phous carbohydrate glasses, but not in case of crystalline sugars
(Pikal and Rigsbee 1997; Izutsu and Kojima 2002) In contrast to
the drying of pharmaceutically relevant proteins, where the
sensi-tive biomolecules are directly embedded in the amorphous sugar
glass, several drying-sensitive components of bacterial cells are
sit-uated in the intracellular space Therefore, the protectant must be
transported through the cell membrane Thus, many high
molec-ular weight (MW) polymers that show limited transportation into
cells or in penetration and interaction with the phospholipid
head-groups of the cell membrane are not very effective in protecting
dried probiotics (Semyonov and others 2010)
According to Potts (1994), it is still not clear if bacterial glasses
exist The differential scanning calorimetry (DSC) technique
which is commonly used for the detection of glass transition,
is insensitive for cells, where changes in heat capacity are small and
or when many transitions occur in the same temperature range
For example, Fonseca and others (2001) showed that the T g of
washed cells of Lactobacillus bulgaricus cannot be detected
Further-more, Cerrutti and others (2000) and Fonseca and others (2001)
indicated that the reported T gis mainly obtained from the added
solutes or the drying matrix However, Hoekstra and others (2001)
and Sun and Leopold (1997) showed that the state of the cytoplasm
is considered to be of crucial importance for the dehydration
tol-erance in organisms Thus, the glassy state of dried cells (if there
is any) needs to be considered carefully
Role of Glassy State during Freezing
The typical freeze-drying or lyophilization process consists of
3 steps: freezing, primary drying or sublimation, and secondary
drying or desorption The ice crystals formed during freezing
determine the morphology and distribution of the pores
(macro-scopic cake structure) that are formed during the removal of ice
crystals Thus, the freezing process significantly determines the
drying behavior of the sample throughout the following process
steps In order to ensure rapid drying and to facilitate vapor
migra-tion during drying, ice crystals in the suspension should be large
and contiguous On the one hand, large ice crystals can only be
achieved by relatively slow freezing rates (not with liq N2) On
the other hand, these large ice crystals can induce cell damage
due to mechanical stress Generally, slow freezing may accompany
the eutectic crystallization of buffer salt components, which is also
suspected to cause membrane damage (Morgan and others 2006)
In other words, optimum freezing conditions often require a
com-promise between the requirements of the bacteria and the drying
performance For the production of probiotics, frozen granules are
commonly produced by distributing the cell suspension through
a droplet disc with pores or through a nozzle into liquid N2 In
addition to direct freezing in liquid N2, innovative rapid freezing
methods were explored by Volkert and others (2008) The frozen
granules are produced by the spray freezing, where cell suspension
is sprayed in an air blast freezer to get droplets (for example, size of
ca 5 to 30μm) For a frozen solution, at T g,the high viscosity of
the unfrozen phase will inhibit ice crystal formation; and therefore,
a maximally freeze-concentrated solution is formed (Roos 1997)
It has been suggested that the formation of a maximally
freeze-concentrated matrix with entrapped microbial cells is essential forthe survival during freezing (Pehkonen and others 2008) Given
T gof skim milk (–50◦C), sucrose (–46◦C), and trehalose (–40◦C),
which are the most common suspending medium used in many
studies, and T gof pure water (–135◦C), the maximally
freeze-concentrated matrix should be obtained by common industrialfreezing protocols, for example, immersion in liquid N2(–196◦C)
Otherwise, the freezing temperature should be selected with care
to ensure a maximally freeze-concentrated state and formation of
a glassy structure of the unfrozen continuous phase of the cell pension In freeze-drying formulations, cryoprotectants are used
sus-to forestall cell injury Some viscous cryo-protectants (for example,glycerol, sugars, and polymers) increase the viscosity of the freeze-concentrated solution or cytoplasm depending on their permeabil-ity Therefore, the glassy state can be reached at a lower coolingrate and a higher freezing temperature (Morris and others 2006)
The annealing step is commonly used for the freeze-drying
of proteins and pharmaceuticals However, it is not applied forthe production of freeze-dried probiotics Annealing is a pro-cess, where frozen samples are kept isothermally at a temperature
between T gand the onset of melting temperature During ing, freezable water entrapped in amorphous regions (due to rapidfreezing) is crystallized into larger and more uniform ice crystals
anneal-This is especially true for the frozen probiotics, which are monly produced by rapid freezing using liquid nitrogen A pellet
com-of 2 mm diameter is cooled from 0 to –50◦C in approximately
10 s or at a freezing rate of approximately 300◦C/min (Oetjen andHaseley 2003) The increased content and uniform ice formationimprove the efficacy of subsequent drying process The knowledge
of the influence of annealing on probiotics’ survival and stability isvery limited In a study by Ekdawi-Sever and others (2003), it wasreported that the annealing does not cause cell death, but improves
storage stability and reduces browning reaction of L acidophilus.
Role of Glassy State during Drying
It is the current opinion that the product temperature duringremoval of ice crystals should not be higher than the critical tem-
perature This critical temperature commonly refers to T g(or T
g
after sublimation) In this respect, the processing conditions, (forexample, pressure and shelf temperature during sublimation) must
be controlled so that the product temperature is not higher than
T gduring drying For this reason, solutes with high T
gare often
added to increase T g so that the drying can be carried out at a
higher temperature, and a high T gof final dried products is tained to provide a high storage stability Disaccharide sugars andoligomeric sugars are preferred as additives for freeze-drying not
ob-only because they exhibit a higher T g(Adams and Ramsay 1996),but also because they can be easily vitrified (Franks 1998; Ward andothers 1999) For sugar alcohols such as mannitol, caution must
be exercised Mannitol can easily separate from a frozen solution
in the form of a crystalline phase (Adams and Ramsay 1996; Kimand others 1998), resulting in a loss of the product stability afterfreeze-drying (Izutsu and others 1994; Izutsu and Kojima 2002)
This reason has been proposed for the little or no protection ferred by mannitol on freeze-dried malolactic cultures (Zhao andZhang 2005)
con-Recently, “collapse temperature” which is defined as the imum temperature preventing the structure of the dried product
max-from macroscopic collapse (T c) (Fonseca and others 2004a) wasproposed to use as the critical temperature for freeze-drying of
drying formulation with cells In comparison, T c and T gare sured with techniques based on different principles DSC has been
Trang 6Glassy state of dried probiotics
commonly used over decades to determine the T g as a mid- or
onset point of the temperature range that the endothermic shift in
heat capacity appears The measurement of T gis carried out with
a representative frozen sample ex situ at atmospheric pressure In
contrast, T cis commonly determined by a freeze-drying
micro-scope as the visual structural collapse during the simulated
subli-mation of ice crystals under a set vacuum level In other words, T c
reflects the physical state of frozen matrix during drying, whereas
T grather reflects the physical state regardless of the drying
condi-tions The state diagram showing glass transition temperature and
collapse temperature is depicted in Figure 1 (adapted and modified
from Roos 2010) In the absence of cells, the difference between
T gand T
cis very small Although the presence of cells does not
clearly influence T g(Fonseca and others 2001; Schoug and
oth-ers 2006), it significantly increases T c(Fonseca and others 2004a)
Bacteria can give some kind of structure and thus reduce or avoid
viscous flow when T g of pure sugar solution is reached The
in-crease depends on the cell types, that is, size, shapes, and cell chain
formation and concentration (Fonseca and others 2004b) As a
re-sult, a freeze-drying matrix with cells is more robust, and when T c
is taken as the critical temperature, it allows the drying stage with
a higher product temperature (or practically drying temperature)
This is of economical importance because it is estimated that the
increase in a degree of product temperature will decrease primary
drying time by about 13% (Tang and Pikal 2004)
Nevertheless, the opinion that cells should be retained in glassy
state during drying lacks clear empirical evidence The physical
state has been measured mostly by analyzing the frozen and dry
sample before and after drying, while the physical state during
drying is changed with drying time and conditions In our
re-cent study (Foerst and others 2010), several drying protocols were
carried out, and the viability was considered in relation to
physi-cal state, residence time in rubbery state and moisture content of
samples The study showed that glassy state may not play a
signif-icant role on viability of cells during drying step of freeze-drying
The survival decreased similarly for conditions where the samples
remained in glassy state for the whole part of drying, and for
con-dition where the difference between product and glass transitiontemperature was so high that the sample showed a collapsed struc-ture The results are in agreement with unexpected results from arecent study by Schersch and others (2010) with pharmaceuticallyrelevant proteins Loss of biological activities of the proteins wasnot observed in collapsed and noncollapsed cakes The collapsedlyophilizate, the appearance of which is unacceptable for pharma-ceutial products, is not critically important for probiotic products.The dried lyophilizate has to be milled into powder to be mixedwith other food components, mixed with other fillers for probi-otic tablets or capsules, as well as used as starter cultures Whenthe collapsed structure as a result of drying does not negatively af-fect the viability of cells, the proposed concept “collapse drying,”which is expected to substantially reduce drying time (Scherschand others 2010) can be possibly applied for the production offreeze-dried probiotics
In addition to free water, which is removed during primarydrying, cells contain ca 0.25 g water/g dry weight of stronglyassociated water that is unfreezable This strongly associated water
is removed by desorption The shelf temperature is elevated topromote desorption of water At the end of the secondary dry-ing, the sample should have a water content that is optimal forstorage It has been suggested that the moisture content of 1%
or less is required for a long-term shelf life (Nakamura 1996);however, Gardiner and others (2000) showed that moisture re-moval below 4% may be considered practically enough It wasalso reported in a study by Zayed and Roos (2004) that the opti-
mal moisture content for storage of freeze-dried L salivarius ssp salivarius is in the range of 2.8% to 5.6% Therefore, it is still
debatable whether cells should be dried to moisture content aslow as possible An argument is that lipid oxidation is enhanced
at very low water contents, and water acts as a protectant againstoxidation
Role of Glassy State during Storage
Low moisture contents of dried cells after drying do not tee high stability during storage The moisture is not constant, and
guaran-Figure 1–State diagram showing different
regions and state of a drying matrix: T m, onset
of ice melting; T m, onset of ice melting in a maximally freeze-concentrated solute matrix;
T g , glass transition temperature; T g, glass transition temperature of the maximally
freeze-concentrated solute matrix; T gw, glass
transition of water, and T c, collapsed
temperature T c approximate to T gand increases in the presence of cells The shaded area shows a temperature range for maximum
ice formation for the solute concentration C g The glass line indicates the glass transition temperature at various solute concentrations (adapted and modified from Roos 2010).
Trang 7desorption, adsorption, and the glass transition may occur during
storage depending on the specific combination of atmospheric
relative humidity and storage temperature Equilibrium of
ambi-ent conditions may only take a few hours or days (Kurtmann and
others 2009b) or as long as to 1 to 2 mo (Fonseca and others
2001) This equilibration period is often not taken into account
in studies with the storage stability and T g At a critical storage
temperature T g, the critical water content and storage relative
hu-midity (or a w) can be determined when information on sorption
isotherm of cells is available (Figure 2 adapted and modified from
Fonseca and others 2001) Information on the sorption isotherm
of dried probiotics is necessary to avoid inactivation due to
mois-ture gain during storage Furthermore, it can be used to
corre-late the storage relative humidity and the final moisture content
of dried cells to avoid unnecessary drying time from drying of
cells far beyond the moisture content corresponding to storage
a w However, in comparison to foods, very little information on
sorption isotherm of probiotics is available Notably, stability of
probiotics is commonly considered with respect to survival The
role of glassy state has never been studied in correlation to probiotic
properties, although the properties are vital for probiotic products
and there is a report about change in the properties during
stor-age The amount of bacteriocin production in freeze-dried vaginal
lactobacillus strains was significantly reduced after storage for 12 mo,
whereas the production of other antimicrobial substances (lactic
acid and hydrogenperoxide) and the auto-aggregation were still
retained (Juarez Tomas and others 2009)
Nevertheless, recent studies on storage of dried probiotics cast
doubt upon the hypothesis that chemical reactions are completely
halted in the glassy state It does not solely determine the storage
stability of dried cells, and inactivation still occurs during storage
at a temperature T – T g = 0 (Higl and others 2007; Kurtmann
and others 2009b) It was suggested that in order to achieve nearly
complete reduction of molecular movements, amorphous
phar-maceutical solids should be stored 50◦C below T g(Hancock and
others 1995) Similar findings were reported also in a study with
L coryniformis by Schoug and others (2010) and with L rhamnosus
by Miao and others (2008) where viability of cells clearly
de-creased at T – T ghigher than –50◦C and ca –30◦C, respectively
Furthermore, the glassy state does not play major protective roleagainst major storage deteriorative reactions such as lipid oxidation(Andersen and others 1999; Kurtmann and others 2009a), becausethe free radical reaction is not diffusion limited In addition, lipidcarbonyls as formed as secondary oxidation products from lipid ox-idation have been suggested to participate in browning reactionsthrough reactions with amino groups in Maillard-like reactions
Browning reaction is related to viability loss in freeze-dried cells(Kurtmann and others 2009c) It was suggested also that the na-ture of the sugar was more important for storage stability than thephysical state of the matrix with the nonreducing sucrose provid-ing better stability than the reducing lactose (Kurtmann and others2009b)
Alternatively, it was also proposed that a biomaterial is most
sta-ble at or below its monolayer moisture content or a w, which varies
in each biomaterial and with environmental conditions (Pitomboand others 1994) For example, the monolayer moisture con-
tent of L bulgaricus (Fonseca and others 2001) and encapsulated
L rhamnosus (Ying and others 2010) was 10% and 3%, respectively.
This concept cannot completely explain the storage stability ofdried cells as well, considering that an increase in temperature sig-nificantly affected the inactivation rate but minimally affected thewater activity Moreover, it provides only information on stronglyassociated and free water but does not indicate when the molecu-lar mobility starts to increase A strong combined effect of wateractivity and temperature was reported and it was suggested that
a w and T g are coupled (Kurtmann and others 2009b) Unlikefreeze-dried proteins, a bacterial cell is not a unique entity butcontains the outside compartment, which is mainly drying ma-trix and the inside compartment, which contains cell componentswith different affinities to water Hence, not all components orfunctions of cells are equally sensitive to moisture or temperature
In other words, there are systems where the inactivation is strongly
Figure 2–Relationships between glass transition temperature, water content, and water activity.
At critical storage temperature a (T g), the
according storage relative humidity (a w) and water content of a sample are b and c (adapted and modified from Fonseca and others 2001).
Trang 8Glassy state of dried probiotics
dependent on temperature and others that are strongly depend on
a w(Higl and others 2007)
Conclusion and Outlook
Despite the current opinion that a glassy state is highly
im-portant for the stability of freeze-dried biomaterial and thus has
to be maintained throughout the whole production process and
storage, no study could empirically confirm its decisive effects,
especially in case of probiotics This can be due to the fact that
unlike proteins and foods, dried probiotic cells are largely still
intact and actually separated from the drying matrices Thus,
in case of probiotics, common results relate to the physical state of
the surrounding drying matrices regardless of the actual physical
state of the cytoplasm For a better understanding of the role of the
glassy state on cells, and thus a systematic development of stable
freeze-dried probiotics, further studies are encouraged in order to
elucidate the physical states of cells during production and
stor-age Further studies with practical implication that should also be
addressed are, for example, the role of an annealing step during
freezing, optimal moisture content for stable storage of probiotics,
changes in probiotic properties due to freeze-drying and storage,
and sorption isotherm of dried probiotics
References
Adams GD, Ramsay JR 1996 Optimizing the lyophilization cycle and the consequences of
col-lapse on the pharmaceutical acceptability of Erwinia L-asparaginase J Pharm Sci 85:1301–5.
Andersen AB, Fog-Peterson MS, Larsen H, Skibsted LH 1999 Storage stability of freeze-dried
starter cultures (Streptococcus thermophilus) as related to physical state of freezing matrix Food
Sci Technol 32:540–7.
Buitink J, Leprince O 2004 Glass formation in plant anhydrobiotes: survival in the dry state.
Cryobiol 48:215–28.
Cerrutti P, Segovia de HM, Galvagno M, Schebor C, del Pilar BM 2000 Commercial baker’s
yeast stability as affected by intracellular content of trehalose, dehydration procedure and the
physical properties of external matrices Appl Microbiol Biotechnol 54:575–80.
Crowe JH, Crowe LM, Carpenter JF, Wistrom CA 1987 Stabilization of dry
phospholipid-bilayers and proteins by sugars Biochem J 242:1–10.
Duong T, Barrangou R, Russell WM, Klaenhammer TR 2006 Characterization of the tre
locus and analysis of trehalose cryoprotection in Lactobacillus acidophilus NCFM Appl Environ
Microbiol 72:1218–25.
Ekdawi-Sever N, Geontoro LA, de Pablo JJ 2003 Effects of annealing on freeze-dried
Lactobacillus acidophilus J Food Sci 68:2504–11.
Foerst P, Higl B, Santivarangkna C, Kulozik U 2010 Systematic investigation on the
in-fluence of glassy state on cell survival during the drying stage of freeze drying Cryobiol,
submitted.
Fonseca F, Obert JP, Beal C, Marin M 2001 State diagrams and sorption isotherms of bacterial
suspensions and fermented medium Thermochim Acta 366:167–82.
Fonseca F, Passot S, Cunin O, Marin M 2004a Collapse temperature of freeze-dried Lactobacillus
bulgaricus suspensions and protective media Biotechnol Prog 20:229–38.
Fonseca F, Passot S, Lieben P, Marin M 2004b Collapse temperature of bacterial suspensions:
the effect of cell type and concentration Cryo Lett 25:425–34.
Franks F 1998 Freeze-drying of bioproducts: putting principles into practice Eur J Pharm
Biopharm 45:221–9.
Gardiner GE, O’Sullivan E, Kelly J, Auty MA, Fitzgerald GF, Collins JK, Ross RP, Stanton
C 2000 Comparative survival rates of human-derived probiotic Lactobacillus paracasei and
L salivarius strains during heat treatment and spray drying Appl Environ Microbiol
66:2605–12.
Granato D, Branco GF, Nazzaro F, Cruz AG, Faria JAF 2010 Functional foods and nondairy
probiotic food development: trends, concepts, and products Comp Rev Food Sci Food Safety
9:292–302.
Hancock BC, Shamblin SL, Zografi G 1995 Molecular mobility of amorphous pharmaceutical
solids below their glass-transition temperatures Pharm Res 12:799–806.
Higl B, Kurtmann L, Carlsen CU, Ratjen J, Forst P, Skibsted LH, Kulozik U, Risbo J 2007.
Impact of water activity, temperature, and physical state on the storage stability of Lactobacillus
paracasei ssp paracasei freeze-dried in a lactose matrix Biotechnol Prog 23:794–800.
Hoekstra FA, Golovina EA, Buitink J 2001 Mechanisms of plant desiccation tolerance Trends
pro-Juarez Tomas MS, Bru E, Martos G, Nader-Macias ME 2009 Stability of freeze-dried vaginal
Lactobacillus strains in the presence of different lyoprotectors Can J Microbiol 55:544–52.
Kim AI, Akers MJ, Nail SL 1998 The physical state of mannitol after freeze-drying: effects of mannitol concentration, freezing rate, and a noncrystallizing cosolute J Pharm Sci 87:931–5.
Kurtmann L, Carlsen CU, Risbo J, Skibsted LH 2009a Storage stability of freeze-dried
Lacto-bacillus acidophilus (La-5) in relation to water activity and presence of oxygen and ascorbate.
Cryobiol 58:175–80.
Kurtmann L, Carlsen CU, Skibsted LH, Risbo J 2009b Water activity-temperature state
di-agrams of freeze-dried Lactobacillus acidophilus (La-5): influence of physical state on bacterial
survival during storage Biotechnol Prog 25:265–70.
Kurtmann L, Skibsted LH, Carlsen CU 2009c Browning of freeze-dried probiotic bacteria cultures in relation to loss of viability during storage J Agric Food Chem 57:6736–41 Miao S, Mills S, Stanton C, Fitzgerald GF, Roos Y, Ross RP 2008 Effect of disaccharides on survival during storage of freeze dried probiotics Dairy Sci Technol 88:19–30.
Morgan CA, Herman N, White PA, Vesey G 2006 Preservation of micro-organisms by drying:
a review J Microbiol Methods 66:183–93.
Morris GJ, Goodrich M, Acton E, Fonseca F 2006 The high viscosity encountered during freezing in glycerol solutions: effects on cryopreservation Cryobiol 52:323–34.
Nakamura LK 1996 Preservation and maintenance of Eubacteria 65–84 In: Hunter-Cevera
JC, Belt A, editors Maintaining cultures for biotechnology and industry Academic Press, Inc p 65–84.
Oetjen G, Haseley P 2003 Freeze-drying Weinheim, Germany: Wiley-VCH 395 p Pehkonen KS, Roos YH, Miao S, Ross RP, Stanton C 2008 State transitions and physico-
chemical aspects of cryoprotection and stabilization in freeze-drying of Lactobacillus rhamnosus
GG (LGG) J Appl Microbiol 104:1732–43.
Pikal MJ, Rigsbee DR 1997 The stability of insulin in crystalline and amorphous solids: observation of greater stability for the amorphous form Pharm Res 14:1379–87 Pitombo RNM, Spring C, Passos RF, Tonato M, Vitolo M 1994 Effect of moisture-content
on the invertase activity of freeze-dried Saccharomyces cerevisiae Cryobiol 31:383–92.
Potts M 1994 Desiccation tolerance of prokaryotes Microbiol Rev 58:755–805.
Rivera-Espinoza Y, Gallardo-Navarro Y 2010 Non-dairy probiotic products Food Microbiol 27:1–11.
Roos YH 1997 Frozen state transitions in relation to freeze drying J Thermal Anal 48: 535–44.
Roos YH 2010 Glass transition temperature and its relevance in food processing Annu Rev Food Sci Technol 1:469–96.
Schersch K, Betz O, Garidel P, Muehlau S, Bassarab S, Winter G 2010 Systematic investigation
of the effect of lyophilizate collapse on pharmaceutically relevant proteins I: stability after freeze-drying J Pharm Sci 99:2256–78.
Schoug A, Mahlin D, Jonson M, Hakansson S 2010 Differential effects of polymers PVP90
and Ficoll400 on storage stability and viability of Lactobacillus coryniformis Si3 freeze-dried in
sucrose J Appl Microbiol 108:1032–40.
Schoug A, Olsson J, Carlfors J, Schnurer J, Hakansson S 2006 Freeze-drying of Lactobacillus
coryniformis Si3-effects of sucrose concentration, cell density, and freezing rate on cell survival
and thermophysical properties Cryobiol 53:119–27.
Semyonov D, Ramon O, Kaplun Z, Levin-Brener L, Gurevich N, Shimoni E 2010
Microen-capsulation of Lactobacillus paracasei by spray freeze drying Food Res Int 43:193–202.
Slade L, Levine H 1991 Beyond water activity: recent advances based on an alternative approach
to the assessment of food quality and safety Crit Rev Food Sci Nutr 30:115–360 Sun WQ, Leopold AC 1997 Cytoplasmic vitrification acid survival of anhydrobiotic organisms Comp Biochem Physiol: A-Physiol 117:327–33.
Tang X, Pikal MJ 2004 Design of freeze-drying processes for pharmaceuticals: practical advice Pharm Res 21:191–200.
Tymczyszyn EE, Del Rosario DM, Gomez-Zavaglia A, Disalvo EA 2007 Volume recovery,
surface properties and membrane integrity of Lactobacillus delbrueckii subsp bulgaricus dehydrated
in the presence of trehalose or sucrose J Appl Microbiol 103:2410–9.
Volkert M, Ananta E, Luscher C, Knorr D 2008 Effect of air freezing, spray freezing, and
pressure shift freezing on membrane integrity and viability of Lactobacillus rhamnosus GG.
J Food Eng 87:532–40.
Ward KR, Adams GD, Alpar HO, Irwin WJ 1999 Protection of the enzyme L-asparaginase ing lyophilisation-a molecular modelling approach to predict required level of lyoprotectant Int J Pharm 187:153–62.
dur-Ying DY, Phoon MC, Sanguansri L, Weerakkody R, Burgar I, Augustin MA 2010
Microen-capsulated Lactobacillus rhamnosus GG powders: relationship of powder physical properties to
probiotic survival during storage J Food Sci 75:E588–95.
Zayed G, Roos YH 2004 Influence of trehalose and moisture content on survival of Lactobacillus
salivarius subjected to freeze-drying and storage Proc Biochem 39:1081–6.
Zhao G, Zhang G 2005 Effect of protective agents, freezing temperature, rehydration dia on viability of malolactic bacteria subjected to freeze-drying J Appl Microbiol 99: 333–8.
Trang 9Nanoemulsions Using Experimental
Mixture Design
Patrick G Maher, Mark A Fenelon, Yankun Zhou, Md Kamrul Haque, and Yrj¨o H Roos
Abstract: The objective of this study was to determine the effect of changing viscosity and glass transition temperature
in the continuous phase of nanoemulsion systems on subsequent stability Formulations comprising ofβ-casein (2.5%,
5%, 7.5%, and 10% w/w), lactose (0% to 20% w/w), and trehalose (0% to 20% w/w) were generated from Design ofExperiments (DOE) software and tested for glass transition temperature and onset of ice-melting temperature in maximallyfreeze-concentrated state (Tg & Tm), and viscosity (μ) Increasing β-casein content resulted in significant (P < 0.0001)increases in viscosity and Tm (P = 0.0003), and significant (P < 0.0001) decreases in Tg A mixture design was used
to predict the optimum levels of lactose and trehalose required to attain the minimum and maximum Tg and viscosity
in solution at fixed protein contents These mixtures were used to form the continuous phase ofβ-casein stabilized
nanoemulsions (10% w/w sunflower oil) prepared by microfluidization at 70 MPa Nanoemulsions were analyzed for Tg
& Tm , as well as viscosity, mean particle size, and stability Increasing levels ofβ-casein (2.5% to 10% w/w) resulted in
a significant (P < 0.0001) increase in viscosity (5 to 156 mPa.s), significant increase in particle size (P = 0.0115) from
186 to 199 nm, and significant decrease (P= 0.0001) in Tg (−45 to −50◦C) Increasing the protein content resulted
in a significant (P < 0.0001) increase in nanoemulsion stability A mixture DOE was successfully used to predict glass
transition and rheological properties for development of a continuous phase for use in nanoemulsions
Keywords: glass transition, nanoemulsion, optimization, stability, viscosity
Introduction
Food formulations are widely produced by combining
carbohy-drate, protein, and oil in emulsified systems The characteristics of
individual components can be used to control texture and shelf-life
(Dickinson and Pawlowsky 1997) Structure through glass
forma-tion is provided by carbohydrates Lactose is commonly used for
the continuous matrix-forming material during spray drying It
is a suitable carbohydrate due to its cheap availability, bland
fla-vor, low viscosity in concentrated solutions, and most importantly
its ability to produce a good glass Similarly, trehalose is also a
suitable carbohydrate as it has a wide pH-stability range, and is
nonreducing in the presence of protein during heat treatment
(Higashiyama 2002) Trehalose combined with lactose has been
found to delay lactose crystallization while keeping the glass
tran-sition temperature of powders (Tg) constant (Mazzobre and others
2001), due to modifications to the molecular environment by
ge-ometric, thermodynamic, or kinetic factors Carbohydrates are
often incorporated into dairy protein stabilized nutritional
bev-erages to meet nutritional and structural requirements Typically,
these emulsions are stabilized by amphiphilic moieties, such as
ca-seins The caseins are the most abundant proteins (approximately
Maher and Fenelon are with Teagasc Food Research Centre, Moorepark, Cork,
Ireland Authors Zhou, Haque, and Roos are with School of Food and
Nutritional Sciences, Univ College, Cork, Ireland Direct inquiries to Fenelon
(E-mail: mark.fenelon@teagasc.ie.).
80% w/w) in bovine milk and adsorb strongly at the oil-water terface during homogenization This layer provides stability to theemulsion during processing and storage and prevents coalescence(Dickinson and Stainsby 1982; Vega and Roos 2006) Caseins alsohave low heat sensitivity and better surface-active properties com-pared to whey proteins, which is beneficial for spray drying (Ped-ersen and others 1998; Hogan and others 2001; Dollo and others2003; Sliwinski and others 2003).β-casein is the most hydropho-
in-bic casein, and steric-stabilizing properties of the protein layeraround oil droplets in emulsions are provided by the phosphoser-ine residues ofβ-casein (Dickinson 1997a) β-casein is preferen-
tially adsorbed compared to other caseins due to lower interfacialenergies, and soβ-casein stabilized emulsions are less susceptible
to flocculation (Dickinson and others 1987, 1988; Dickinson andMatsumara 1994) Nanoemulsions are emulsions with a particlesize in the range of 20 to 200 nm that ensures stability as thediffusion rate, due to Brownian motion, is higher than that forsedimentation and creaming (Solans and others 2005)
The glass transition temperature (Tg) is the temperature atwhich, upon heating, an amorphous material changes from a glassy,solid-like state to a rubbery, fluid-like state (Langrish and Wang2006) Glass transition has an effect on many processes and proper-ties encountered in food science In powders, temperatures abovethe glass transition are found to increase the molecular mobilityand free volume, which can result in physico-chemical deteri-oration, such as crystallization of sugars, stickiness, and caking
of powders (White and Cakebread 1966; Slade and Levine 1991).During processing, stickiness may reduce product yields and retardflow of particles leading to operational problems for manufacturing
Trang 10equipment and increased down time (Bhandari and Howes 1999).
Also the storage stability and quality changes of food systems are
strongly affected by the glass transition (Schenz 1995)
Crystalliza-tion and caking are time-dependant and are both a funcCrystalliza-tion of
fluctuation in storage temperature and Tg The higher the
tem-perature above the Tg, the more viscosity decreases and molecular
mobility increases leading to crystallization and caking (Bhandari
and Howes 1999; Omar and Roos 2007) These phenomena may
also cause the continuous matrix to become unstable and release
entrapped components, resulting in impaired rehydration
prop-erties (Roos and Karel 1991a) It is therefore important that the
effects of Tgon food materials, such as emulsions in various states
(liquid or dehydrated systems), are taken into account in their
processing and storage
In this study, an experimental mixture design was generated
using Design of Experiments (DOE) software DOE was used
to statistically optimize the formulation of nanoemulsions using
viscosities and glass transition temperatures at fixed protein
con-centrations as response parameters to improve their stability andhence properties as encapsulants Data were analyzed statistically
to show by polynomial equations, adjusted R2and P-values which
variables in the formulation have the most significance on selectedresponses
The objective of this study was to produce nanoemulsions withvarying viscosities and glass transition temperatures at differentprotein levels, and observe the effect on mean particle size andstability through the application of a statistical mixture design
Materials and Methods
D-optimal experimental mixture designThe experimental mixture design was set up using the soft-ware Design Expert R Software version 7.1.3 (Stat-Ease Inc.,Minneapolis, Minn., U.S.A.) A D-optimal design was usedwith the constraints Lactose + Trehalose + β-casein = 20%
w/w with β-casein ≤10% w/w This was used to study the
Table 1– Mixture design formulations of carbohydrate-protein mixtures with experimental results of viscosity (μ at 400 s−1 ) and glass transition (T g , T m , T g ) values.
Trang 11continuous phase of carbohydrate-protein mixtures prior to
emulsification The model gave 12 different formulations plus
5 replicates leading to a total of 17 experimental
formula-tions (Table 1) These formulaformula-tions contained 20% w/w solids
and were analyzed for viscosity (μ), glass transition
tempera-ture of maximally freeze-concentrated liquid (Tg ), glass
tran-sition temperature of freeze-dried powder (Tg), and onset of
ice melting temperature of maximally freeze-concentrated
liq-uid (Tm ) The software was used to identify the formulations
that both minimized and maximized Tg and viscosity for the
carbohydrate-protein mixtures at fixed protein concentrations of
2.5%, 5%, 7.5%, and 10% w/w These optimized formulations
(Table 2) were input into the software (Historical Data Initial
Design) with the constraints; Lactose≤ 17.5% w/w, Trehalose
≤ 17.5% w/w, β-casein 2.5% to 10% w/w, where Lactose +
Trehalose+ β-casein = 20% w/w Sunflower oil (10% w/w) was
added to these formulations prior to emulsification
Nanoemul-sions were analyzed for viscosity, Tg , T
m , particle size, and aration rate The data were analyzed graphically and statistically
sep-using the software The statistical significance of the models was
given by the adjusted R2and P-values (Table 3), and the graphical
representations of models (Figure 1 to 8) along with the Scheffe
model polynomial equations for 3 components (Table 4) predict
effects of components in the formulation Models were deemed
significant at P < 0.05.
Carbohydrate-protein mixtures
The materials used in the experiments wereα-lactose
mono-hydrate (Sigma Aldrich, Ireland), trehalose (Cargill, U.K.), and
β-casein (Kerry Group, Ireland) β-casein was dispersed in
dis-tilled water at 45◦C for 1 h using mechanical agitation, adjusted
for water loss, and stored overnight (12 h) at 4◦C to ensure protein
hydration Lactose, trehalose, andβ-casein mixtures were prepared
according to Table 1, under mechanical agitation at 55 ◦C for
10 min Formulations were adjusted to pH 7.5 using 1N HCl at
25◦C
Nanoemulsion preparation
Oil in water nanoemulsions (10% w/w oil) withβ-casein as
emulsifier (2.5% to 10% w/w) were prepared using sunflower
oil (Trilby Trading, Ireland) Coarse emulsions were prepared by
blending 1 L of β-casein solutions with oil at 60 ◦C using a
high shear mixer (Silverson) at approximately 8000 rpm for 2
min Coarse emulsions were microfluidized (model M-110EH,
Microfluidics, Newton, Mass., U.S.A.) using a ceramic interaction
chamber (70 MPa at 60◦C) Specified quantities of carbohydrates(Table 2) were added to the nanoemulsions in a water bath at 55◦Cand mixed for 10 min Nanoemulsions were then cooled to 25◦Cand adjusted to pH 7.5 using 1N HCl The nanoemulsions con-taining lactose, trehalose,β-casein, and sunflower oil according to
the experimental design were heated to 60◦C and microfluidized
a second time (70 MPa) Approximately 500 mL of sample wascollected and stored at 4◦C prior to analysis
RheologyRheological analysis of carbohydrate-protein mixtures and na-noemulsions was carried out using an AR G2 rheometer (TAInstruments, Crawley, U.K.), equipped with a 60-mm aluminiumparallel plate (gap 800μm) Samples (approximately 2 mL) were
first presheared at 300 s−1for 1 min and equilibrated for 2 min.Samples were then sheared from 10 to 400 s−1over 3 mins, held
at 400 s−1 for 1 min, and then sheared from 400 to 10 s−1over
3 mins The apparent viscosity, taken at 400 s−1, was used in sequent statistical analysis A shear rate of 400 s−1was chosen tosimulate the conditions of high shear processing All measurementswere carried out at 25◦C
sub-Differential scanning calorimetry
A differential scanning calorimeter (DSC) equipped with uid N2 cooling (Mettler-Toledo DSC 821e, Schwerzenbach,Switzerland) was used to determine the glass transition temper-atures (onset glass transition temperature of maximally freeze-concentrated solids, Tg , for liquids and onset glass transitiontemeparture, Tg for anhydrous samples) and onset temperatures
liq-of ice melting in maximally freeze-concentrated solids (Tm ) incarbohydrate-protein mixtures and nanoemulsions Samples wereprepared in preweighed aluminium pans (40μL; Mettler Toledo-
27331) and scanned in hermetically sealed pans An empty panwas used as a reference The DSC was calibrated for temperatureusing n-hexane (melting point, 95.0◦C), mercury (melting point,
−38.8 ◦C), water (melting point, 0.0 ◦C), gallium (melting
point, 29.8◦C), and indium (melting point, 156.6◦C), and forheat flow using n-hexane (Hm, 151.8 J/g), mercury (Hm,11.4 J/g), water (Hm, 334.5 J/g), gallium (Hm, 80.0 J/g),and indium (Hm, 28.45 J/g) Liquid samples were hermeticallysealed in aluminium pans (40μL; Mettler Toledo) with an empty
pan as reference and scanned from−100◦C to 25◦C to
deter-mine Tg and Tm for nonannealed samples Samples were thenannealed for 15 min at approximately 1◦C below the initial onset
of Tm(measured during the 1st temperature cycle), followed by a
Table 2–Optimized formulations for minimum and maximum viscosity (μ at 400 s−1 ) and glass transition temperature (T g ) for sions at fixedβ-casein contents (2.5%, 5%, 7.5%, 10% w/w).
nanoemul-Lactose Trehalose β-casein Mean particle μEmul T g T m Separation rate Stokes’ creaming
Trang 12m Anhydrous samples were prepared by freeze drying
the carbohydrate-protein mixtures at their various concentration
levels They were 1st heated over the glass transition
tempera-ture region (60 to 140◦C), then cooled to below glass transition
temperature, and a 2nd heating scan was run over the glass
transi-tion region and the Tgwas measured Heating and cooling rates
for both liquid and anhydrous samples were 5 ◦C/min and 10
◦C/min, respectively T
g, Tg , and T
m were determined usingSTARethermal analysis software, version 8.10 (Mettler Toledo)
Measurements were carried out in triplicate
Particle size analysis
Particle size of the nanoemulsions was measured using a
Zetasizer Nano system (Malvern Instruments, Inc., Worcester,
U.K.) Measurements were carried out at 22 ◦C, at a
scatter-ing angle of 173◦ Samples were diluted using distilled water by a
factor of 20 prior to analysis to avoid the effects of multiple
scat-tering The mean average (z-average) droplet size was measured
using an intensity distribution
Stability analysis
A Lumifuge 116 stability analyzer (L.U.M GmbH, Berlin,
Germany) was used to measure the separation rate of
nanoemul-sions at 25◦C The Lumifuge is an analytical centrifuge that
con-tinually measures the light transmitted through a sample over the
Table 3–Statistical significance of models, both transformed and
un-transformed, for carbohydrate-protein mixtures and nanoemulsions.
Model significance Carbohydrate
a The natural log, (ln), of the viscosity results for the carbohydrate-protein mixtures was
modeled and the given R2 values reflect this transformed result.
b The inverse square root of the nanoemulsion viscosity results was modeled and the given
R2 values reflect this transformed result.
c Viscosity measurements are quoted at 400 s −1.
total length of the measurement cell The samples (0.4 mL aliquots)were placed in polycarbonate sample cells and centrifuged (1140
× g) for approximately 7.2 h, simulating 1 y of separation
un-der normal gravity Separation rates were determined using thesoftware package SepView 4.1 (L.U.M GmbH), which calculatedthe velocity of the separation of the individual particles from themeasurement results (μm/day)
Results and Discussion
Viscosity of carbohydrate-protein mixturesThe effect of increasingβ-casein content (0% to 10% w/w)
on viscosity is represented graphically in Figure 1 Viscosity creased exponentially with increasing protein content from 2 to
in-78 mPa.s, possibly as a result of increased interactions betweenhydrated protein molecules (Boye and others 1997) and the ability
of protein molecules to absorb water and swell (Damodaran 1997).Mora-Gutierrez and Farrell Jr (2000) suggested that sugarmolecules (lactose and sucrose) do not bind to proteins (casein),
or affect viscosity in protein-sugar systems However,
trehalose-β-casein mixtures had higher viscosity than lactose-trehalose-β-casein
mix-tures at equivalent carbohydrate contents (Table 1)
The data for viscosity analysis were fitted with a special cubicmodel, which was transformed using a natural log power law rec-ommended by the Box-Cox method This minimized the residualsum of squares in the model, thus improving the statistical analy-
sis The model was significant (P < 0.0001) with a high adjusted
R2(Table 3) The coefficients for this coded Scheffe polynomialmodel are shown in Table 4 and together with the ternary diagramand three-dimensional (3D) surface diagram in Figure 1 show that
β-casein (component C) had the greatest effect on viscosity Low
viscosity systems are the desired outcome The results show thatlow viscosity systems can be produced with lower quantities ofprotein (≤ 5% w/w) The negative coefficients (Table 4) and thedownward curves (Figure 1) for the nonlinear blending terms
AC and BC show that there is synergism between lactose with
β-casein, and between trehalose with β-casein, meaning that these
binary mixtures can result in low viscosity The antagonism ofadding a third component to the mixture causes an increase inviscosity (positive ABC)
Glass transition of carbohydrate-protein mixturesThe glass transition temperatures (Tg ) of the maximally freeze-concentrated carbohydrate-protein mixtures decreased with in-creasing protein content (Table 1; Figure 2) It is likely that
Table 4– Coefficients for each term in the coded polynomial equation, both transformed and untransformed, for carbohydrate-protein mixtures and nanoemulsions.
Coded polynomial equation coefficients for Scheffe model Carbohydrate
Trang 13increased viscosity retarded ice formation, thus decreasing the
so-lute concentration at maximum freeze concentration, Cg This
diluted the unfrozen matrix, and therefore the Tg decreased due
to the higher plasticizing effect of unfrozen water (Roos and Karel
1991b)
A special cubic model was used to fit the data for Tg ,
giv-ing significance at P < 0.0001 with a high adjusted R2 value
(Table 3).β-casein had the greatest effect on Tg , as demonstrated
by the negative 2nd-order nonlinear blending terms involving
β-casein (AC and BC) compared to the positive coefficient for model
term AB (Table 4; Figure 2) A high Tg is the desired outcome
for solutions Therefore, there is antagonism between lactose with
β-casein, and between trehalose with β-casein Varying the ratio
of lactose to trehalose had little effect on Tg at fixed protein levels.
The synergism of adding a 3rd component to the mixture causes
an increase in Tg , as indicated by the positive coefficient for the
three component nonlinear blending term from Table 4 (ABC) It
is suggested that increased viscosity is the primary reason for Tg
decrease with increasing protein content
The onset temperature of ice melting in the maximally
freeze-concentrated state, Tm , is an important property of a material
as it is the “mobility” point (Reid and others 1994) or point
at which the structure begins to collapse under its own weight
during freeze drying (Kaushik and Roos 2007) Tm increased
with increasing protein content (Table 1; Figure 3) As ice forms,
freeze-concentration of solids occurs simultaneously, decreaseing
the freezing/melting temperature of the unfrozen water (Roos
2007) Increasingβ-casein contents retarded ice formation with
a resultant lower Tg and higher T
m (ice formation ceased at ahigher temperature which corresponded to the higher Tm ) Roos
and Karel (1991b) found that in sucrose solutions ice formation
ceased at a viscosity of 107Pa s Although viscosities of the
freeze-concentrated carbohydrate-protein mixtures were not measured,
it is likely that the viscosities increased with the decreasing
temper-ature and increasing solids content of the unfrozen phase during
freeze concentration
The data for Tm analysis were fitted with a quadratic model
and had a high adjusted R2 with significance at P = 0.0003
(Table 3) β-casein had a greater effect on Tm than the vidual sugars (Table 1; Figure 3) Similar AC and BC values and
indi-a smindi-all coefficient for model term AB show thindi-at indi-at fixed proteinlevels, the ratio of lactose to treahlose is irrelevant It is suggestedthat increased viscosity is the biggest factor in increasing Tm ,therefore, it is beneficial to have highly viscous systems to produce
a high Tm .For freeze-dried powders, Tgincreased with increasing proteincontent (Table 1; Figure 4) The lowest protein content (2.5%w/w) system in the present study had a Tg of approximately
104◦C, whereas the highest protein content (10% w/w) systemhad a Tg of 124◦C Biliaderis and others (2002) reported thatthe addition of a high molecular weight carbohydrate (pullalan) tolactose increased Tg Similarly, Shamblin and others (1996) foundthat the Tg of sucrose solutions was higher in the presence ofpolymeric substances, such as starch and polyvinylpyrrolidone
A quadratic model was fitted to the data for Tgwith significance
at P < 0.0001 with a high adjusted R2(Table 3) Component C(β-casein) again had the most significant effect on Tg (Table 4;Figure 4) High Tgvalues are favorable in dairy powders, so highprotein systems give a beneficial outcome in this respect.The presence of protein in either liquid or dehydrated systems,affects the glass transition temperature of that system Typically, ahigh Tg in a liquid system should correspond to a high T
gpostdehydration Similarly, a low Tg should correspond to a low T
gpost dehydration In the current work, increasing protein contentincreased the Tg for dehydrated systems (Figure 4) while it de-creased Tg for liquid systems (Figure 2) As previously discussed,this may be due to increased solution viscosity at higher proteinconcentrations, which makes ice crystal formation more difficult,thus more unfrozen water remains in the continuous phase lower-ing Tg Therefore, the glass transition temperatures in both liquidand dehydrated forms are strongly influenced by protein content
Particle size of nanoemulsionsFor the nanoemulsions, mean particle size increased significantlywith increasing protein content (Table 2; Figure 5) Particle size
Figure 2 – Contour and response surface plots of onset glass transition temperature of maximally freeze-concentrated solids (T g) in protein mixtures comprising of lactose, trehalose, andβ-casein (0% to 20% w/w).
Trang 14increased from approximately 187 to 199 nm for increasing protein
contents up to 7.5% w/w β-casein, however mean particle size
decreased to 193 nm for samples containing 10% w/wβ-casein.
The general trend toward increased mean particle size may be due
to a layering affect of the protein around the small oil droplets
according to self-consistent field (SCF) theory This occurs when
the bulk protein concentration goes above a certain threshold value
(depending on ionic strength), causing multilayer condensation of
self-associating protein onto the oil droplet surface (Dickinson
1997a) In contrast with the current study, Qian and McClements
(2011) found that mean particle size decreased with increasing
protein content for sodium caseinate stabilized nanoemulsions,
however, they did obtain a similar narrow size distribution
The data for particle size were fitted with a quadratic model
and was significant (P < 0.0115) with a high adjusted R2
(Table 3) β-casein is the only component in the mixture that
affects particle size Lactose and trehalose are hydrophilic and main in the continuous phase, therefore having no direct effect
re-on particle size This is seen in Table 4 with the low negativecoefficient value for model term AB
Viscosities of nanoemulsionsThe viscosity of the nanoemulsions increased with increasingprotein content (Table 2, Figure 6) This was most likely due
to increased protein–protein interaction in the continuous phase
as the protein content increases The addition of sunflower oil
to the system increased the solids content by 10% (that is, 30%solids in total) and so the nanoemulsions are more viscous thanthe carbohydrate-protein mixtures
Figure 3–Contour and response surface plots of onset temperature of ice melting (T m) in carbohydrate-protein mixtures comprising of lactose, trehalose, andβ-casein (0% to 20% w/w).
Figure 4–Contour and response surface plots of glass transition temperature of freeze-dried powder (T g ) in carbohydrate-protein mixtures comprising
of lactose, trehalose, andβ-casein (0% to 20% w/w).
Trang 15A quadratic model was fitted to the viscosity data, which was
transformed using an inverse square-root power law recommended
by the Box-Cox method The model was significant (P < 0.0001)
with a high adjusted R2 (Table 3) The results again show that
low viscosity nanoemulsions (< 15 mPa.s) can be produced at
protein contents of≤ 5% w/w The negative coefficients (Table 4)
and the downward curves (Figure 6) for the nonlinear blending
terms AC and BC show that there is synergism between lactose
with β-casein, and between trehalose with β-casein, meaning
that these emulsions can keep a low viscosity The viscosities of
the nanoemulsions in Table 2 show that varying the level of lactose
to trehalose, at fixed protein levels, gave similar viscosities
Glass transitions of nanoemulsions
The Tg of the nanoemulsions decreased as protein content was
increased (Figure 7), in agreement with results for
carbohydrate-protein mixtures (Figure 2) The experimental Tg values for thenanoemulsions are similar to the Tg values predicted for the mix-tures by the optimization function in the DOE software Thisshows that sunflower oil did not affect Tg most likely becausethe oil was in the dispersed phase, while the continuous phase isprimarily responsible for changes in Tg .
The data for Tg were fitted with a quadratic model, and it was
significant (P= 0.0001) with a high adjusted R2(Table 3) Thelarge negative coefficients for the 2nd-order nonlinear blendingterms (Table 4) involvingβ-casein (AC and BC) compared to the
smaller AB value show that β-casein had the greatest effect on
Tg It was found that there is antagonism between lactose with
β-casein, and between trehalose with β-casein, as previously seen
for the carbohydrate-protein mixtures Therefore, low quantities
ofβ-casein combined with lactose or trehalose give the largest Tg .The ratio of lactose to trehalose in the emulsion, at fixed protein
Figure 5–Contour and response surface plots of mean particle size in nanoemulsions comprising of lactose, trehalose, andβ-casein contents.
Figure 6–Contour and response surface plots of viscosity (at 400 s−1) in nanoemulsions comprising of lactose, trehalose, andβ-casein contents.
Trang 16contents, is irrelevant as indicated by the relatively low coefficient
AB in Table 4 The decrease in Tg with increasing protein content
can be explained by a concomitant increase in viscosity
The Tm for nanoemulsions differed from those found for the
carbohydrate-protein mixtures In this case, the Tm decreased
with increasing protein content and the Tm values were lower in
each case for the nanoemulsions (Table 2), compared to that of
the carbohydrate-protein mixtures (Table 1) The overall model
generated was not significant at P = 0.0631 (Table 3), which
meant that the overall mean was a better predictor of the response
than the current model
Nanoemulsion stability
The nanoemulsion stability was measured by analyzing the
flota-tion/creaming rates of particles using an analytical centrifuge The
highest protein content nanoemulsions (7.5% and 10% w/w
pro-tein) were the most stable with separation rates of 7 and 4μm/d,
respectively, compared to separation rates of 41 and 24μm/day
for nanoemulsions with 2.5% and 5% w/w protein contents,
re-spectively (Table 2; Figure 8) Nanoemulsions are reported to
be more stable to creaming on storage compared to
microemul-sions due to the effects of Brownian motion being stronger than
gravitational forces They are also more stable to flocculation and
coalescence due to the lowering of the interfacial tension with
de-creasing particle size that decreases the stress required to break up
the droplets Typically, casein-coated emulsion droplets are prone
to depletion flocculation at protein concentrations higher than 2%
w/w (Dickinson and Golding 1997) This could be due to the
effect of unadsorbed casein in the continuous phase that passes
between oil droplets causing an osmotic pressure gradient, that
entropically force out the aggregated casein, causing the droplets
to flocculate (Dickinson 1997b, 1999) Increasing protein content
increases the occurrence of depletion flocculation In some cases,
it has been found that stability decreases with increasing sodium
caseinate content up to a certain concentration (6% w/w) Above
this concentration a more rigid structure is formed, preventing
rapid creaming (Dickinson and others 1997)
Experimental separation rates were compared to those lated using Stokes’ law equation (see Table 2):
calcu-v s = (ρf − ρ p)
where vs is the separation velocity of the particles, (ρf − ρp) isthe difference between the densities of the fluid in the continuousphase (1060 kg.m−3) and the particles of oil (917 kg.m−3), g is
gravitational acceleration (9.81 m.s−2), dpis the mean particle ameter (m), andμ is the viscosity of the nanoemulsion (kg/(m.s)).
di-The units are converted toμm/day and compared with the
exper-imental separation rates calculated from the Lumifuge in Table 2 Itwas also observed that as particle sizes of nanoemulsions were verysimilar overall that viscosity was the primary factor in determiningthe separation rate In general, the results compared well, howeverStokes’ law is applicable for ideal systems and this system is notideal due to hindrance caused by concentration effects (Robins2000)
A cubic model was fitted to the data for nanoemulsion stability
with significance at P < 0.0001 with a high adjusted R2 value(Table 3).β-casein had the most significant effect on nanoemulsion
separation rate due to increased viscosity at high protein levels Thenegative coefficients (Table 4) and the downward curves (Figure 8)for the nonlinear blending terms AC and BC show that there issynergism between lactose withβ-casein, and between trehalose
with β-casein, meaning that these mixtures can maintain low
separation rate The quantity of lactose or trehalose has no effect
on stability as indicated by the small coefficient for the modelterm AB
Overall, the desired result from an emulsion system is ideally toproduce a low viscosity, high Tg , that when dried corresponds
to a high Tg, low mean particle size and low separation rate
A fine balance can be achieved, whereby a low viscosity, highlystable emulsion system can be produced that is not susceptible
to physico-chemical deteriorative changes that are caused by lowglass transition temperature This study has demonstrated how tostatistically optimizeβ-casein stabilized nanoemulsion systems.
Figure 7–Contour and response surface plots of glass transition temperature of maximally freeze-concentrated solids in nanoemulsions (T g) comprising
of lactose, trehalose, andβ-casein.
Trang 17A statistical mixture design was successfully used for the
for-mulation of nanoemulsions based on optimization of
continu-ous phase parameters of viscosity and glass transition The study
showed that increasing β-casein content in mixtures containing
lactose and trehalose, significantly (P < 0.0001) increased
viscos-ity and decreased glass transition Tg A similar trend was observed
inβ-casein stabilized nanoemulsions containing lactose and
tre-halose This indicates that the physical properties of the continuous
phase are a primary factor influencing the stability and
function-ality of subsequent emulsions Nanoemulsion mean particle size
(P < 0.0115) and stability (P < 0.0001) were significantly
in-creased with increasingβ-casein content It was postulated that
the increased stability was due to the effect of increased viscosity
that reduced the separation rate of the dispersed phase according to
Stokes’ law Modulation of glass transition temperatures through
selection of ingredients, and their concentrations, may be
benefi-cial in powdered products to combat the phenomena of stickiness
and crystallization during storage
Statistical experimental mixture design can be a useful tool in
the formulation of nanoemulsions It reduces the workload by
giving a set number of experiments, and determines and analyzes
the impact of each component in the mixture
Acknowledgments
This research has been funded by the Irish Dept of Agriculture,
Fisheries & Food under the Food Institutional Research Measure
(FIRM)
References
Bhandari BR, Howes T 1999 Implication of glass transition for the drying and stability of dried
foods J Food Engr 40(1–2):71–9.
Biliaderis CG, Lazaridou A, Mavropoulos A, Barbayiannis N 2002 Water plasticizing effects on
crystallization behaviour of lactose in a co-lyophilized amorphous polysaccharide matrix and
its relevance to the glass transition Intl J Food Prop 5(2):463–82.
Boye JI, Ma CY, Harwalkar VR 1997 Thermal denaturation and coagulation of proteins.
In: Damodaran S, Paraf A, editors Food proteins and their applications New York: Marcel
Dekker Inc p 25–56.
Damodaran S 1997 Food proteins: an overview In: Damodaran S, Paraf A, editors Food
proteins and their applications New York: Marcel Dekker Inc p 1–24.
Dickinson E 1997a Properties of emulsions stabilized with milk proteins: overview of some
recent developments J Dairy Sci 80(10):2607–19.
Dickinson E 1997b Aggregation processes, particle interactions and colloidal structure In: Dickinson E, Bergenst˚ahl B, editors Food colloids: proteins, lipids and polysaccharides, Cambridge: Royal Society of Chemistry p 107–26.
Dickinson E 1999 Caseins in emulsions: interfacial properties and interactions Intl Dairy J 9(3–6):305–12.
Dickinson E, Golding M 1997 Depletion flocculation of emulsions containing unadsorbed sodium caseinate Food Hydrocolloids 11(1):13–8.
Dickinson E, Matsumara Y 1994 Proteins at liquid interfaces: role of the molten globule state Colloids Surf B 3(1–2):1–17.
Dickinson E, Pawlowsky K 1997 Influence of kappa-carrageenan on the properties of a stabilized emulsion Food Hydrocolloids 12(4):417–23.
protein-Dickinson E, Stainsby G 1982 Colloids in foods Applied Sci: 33–66.
Dickinson E, Whyman RH, Dalgleish DG 1987 Colloidal properties of model oil-in-water food emulsions stabilized separately byαs1 -casein,β-casein and κ-casein In: Dickinson E,
editor Food emulsions and foams London: Royal Society of Chemistry p 40–51 Dickinson E, Rolfe SE, Dalgleish DG 1988 Competitive adsorption ofαs1 -casein andβ-casein
in oil-in-water emulsions Food Hydrocolloids 2(5):397–405 Dickinson E, Golding M, Povey JW 1997 Creaming and flocculation of oil-in-water emulsions containing sodium caseinate J Colloid Interface Sci 185:515–29.
Dollo G, Le Corre P, Guerin A, Chevanne F, Burgot JL, Leverge R 2003 Spray-dried persable O/W emulsions to improve oral bioavailability of poorly soluble drugs Euro J Pharm Sci 19(4):273–80.
redis-Higashiyama T 2002 Novel functions and applications of trehalose Pure Appl Chem 74(7):1263–9.
Hogan SA, McNamee BF, O’Riordan ED, O’Sullivan M 2001 Emulsification and croencapsulation properties of sodium caseinate/carbohydrate blends Intl Dairy J 11(3): 137–44.
mi-Kaushik V, Roos YH 2007 Limonene encapsulation in freeze-drying of gum gelatin systems LWT- Food Sci and Technol 40(8):1381–91.
arabic-sucrose-Langrish T, Wang E 2006 Crystallisation of powders of spray-dried lactose, skim milk and lactose-salt mixtures Intl J Food Engr 2(4):1–15.
Mazzobre MF, Soto G, Aguilera JM, Buera MP 2001 Crystallization kinetics of lactose in sytems co-lyofilized with trehalose Analysis by differential scanning calorimetry Food Res Intl 34:903–11.
Mora-Gutierrez A, Farrell Jr HM 2000 Sugar-casein interaction in deuterated solutions of bovine and caprine casein as determined by oxygen-17 and carbon-13 nuclear magnetic resonance: a case of preferential interactions J Agric Food Chem 48:3245–55.
Omar EAM, Roos YH 2007 Glass transition and crystallization behaviour of freeze-dried lactose–salt mixtures LWT- Food Sci and Technol 40:536–43.
Pedersen GP, F¨aldt P, Bergens˚athl B, Kristensen HG 1998 Solid state characterization of a dry emulsion: a potential drug delivery system Intl J Pharmaceutics 171(2):257–70 Qian C, McClements DJ 2011 Formation of nanoemulsions stabilized by model food-grade emulsifiers using high pressure homogenization: factors affecting particle size Food Hydro- colloids 25(5):1000–8.
Reid DS, Kerr W, Hsu J 1994 The glass transition in the freezing process J Food Engr 22(1–4):483–94.
Robins MM 2000 Emulsions – creaming phenomena Curr Opin Colloid Interface Sci 5:265–72.
Roos YH 2007 Phase transitions and transformations in food systems In: Heldman DR, Lund DB, editors Handbook of food engineering 2nd ed Florida: Taylor & Francis Group.
Trang 18Shamblin SL, Huang EY, Zografi G 1996 The effects of co-lyophilized polymeric additives on
the glass transition temperature and crystallization of amorphous sucrose J Therm Analysis
and Calorim 47(5):1567–79.
Slade L, Levine H 1991 Beyond water activity: recent advances based on an alternative approach
to the assessment of food quality and safety Crit Rev Food Sci Nutr 30(2–3):115–360.
Sliwinski EL, Lavrijsen BWM, Vollenbroek JM, van der Stege HJ, Van Boekel MAJS, Wouters
JTM 2003 Effects of spray drying on physicochemical properties of milk protein-stabilised
emulsions Colloids and Surfaces B: Biointerfaces 31(1–4):219–29.
Solans C, Izquierdo P, Nolla J, Azemar N, Garcia-Celma MJ 2005 Nano-emulsions Curr Opinion Colloid Interface Sci 10:102–10.
Vega C, Roos YH 2006 Spray-dried dairy and dairy-like emulsions – compositional ations J Dairy Sci 89(2):383–401.
consider-White GW, Cakebread SH 1966 The glassy state in certain sugar-containing food products.
J Food Technol 1(1):73–82.
Trang 19Nutritional Characteristics of Seed
from Southern Spain
Elena Pastor-Cavada, Rocio Juan, Julio E Pastor, Manuel Alaiz, and Javier Vioque
Abstract: The study of local plants could provide useful data for the possible development of future crops with good
nutritional and functional properties The nutritional characteristics of the seed proteins of 28 Vicia species (31 taxa) (Fabaceae) from Spain were studied Protein contents in studied Vicia ranged from 20.1% in V articulata to 32% in
V pyrenaica V altissima, which is only deficient in Trp, showed the most balanced amino acid composition, while the remaining taxa were limiting in Met, Cys, and Trp In vitro protein digestibility (IVPD) ranged from 78% in V incana to 86.3% in V hirsuta In addition to the amino acid composition and IVPD, other nutritional parameters, such as amino acid
score, biological value, protein efficiency ratio, or protein digestibility corrected amino acid score, were studied Results
confirm the interest of studying wild populations of cultivated and non cultivated Vicia species as potentially interesting
sources of seeds with good nutritional characteristics
Keywords: amino acids, nutritional quality, seed proteins, Vicia
Introduction
Legume seeds may represent an important source of protein,
di-etary fibre, carbohydrates, several vitamins, and minerals (Granito
and others 2007; Mart´ın-Cabrejas and others 2008) Legumes can
be used for feeding livestock, for human food, or to increase the
ni-trogen levels in soil In several regions of the world, legumes play an
important role in human and animal nutrition as they are a cheap
source of protein and other nutrients (Cerletti and others 1978)
In recent years, several researchers, such as Rochfort and Panozzo
(2007), have shown that the consumption of proteins, certain
lipids, polyphenols, or bioactive peptides contained in legumes has
a beneficial effect on health According to Duranti (2006), legumes
may prevent diseases such as diabetes mellitus, colon cancer, and
coronary heart diseases Due to the presence of many beneficial
components in pulses, these positive effects can be related to more
than one compound For example, the cardio-protective effect of
pulses may be due to the synergistic effect of different compounds,
such as bioactive peptides or free fatty acids Although
consump-tion of legumes had been decreasing in many Western countries,
their high nutritional value has led to a recent increase in their
use in the human diet In addition, since the “green revolution,” a
large amount of the world’s phytodiversity has been lost because
lo-cal crops have been substituted by other genetilo-cally uniform crops,
which have higher yields However, the diversification of cultivars
is necessary to adapt cultivars to soil and water conditions, or to
limit phytosanitary risks Hence, knowledge of local plants could
provide useful data for the possible development of future crops
MS 20101443 Submitted 12/22/2010, Accepted 6/29/2011 Authors
Pastor-Cavada, Alaiz, and Vioque are with Inst de la Grasa (C.S.I.C.), Avda
Padre Garc´ıa Tejero 4, 41012-Sevilla, Spain Authors Juan and Pastor
are with Dept de Biolog´ıa Vegetal y Ecolog´ıa, Univ de Sevilla,
41012-Sevilla, Spain Direct inquiries to author Pastor-Cavada (E-mail: epastor@cica.es;
epastor@ig.csic.es).
The Vicia genus is a legume with more than 140 species Some Vicia species have been used as livestock feed or as a green manure
crop, and others have been used for local consumption For
example, the seeds of V articulata, V ervilia, V hirsuta, V monantha,
V narbonensis, or V sativa have been consumed in the past In other cases, the green parts are consumed, such as the leaves of V hirsuta,
V narbonensis, or V sativa In the latter species the leaves have been employed as a tea substitute Vicia species have also been inves-
tigated as sources of antioxidant compounds such as polyphenols(Pastor-Cavada and others 2010) The most commonly consumed
species is V faba, which is broadly cultivated in the Mediterranean
Region for human nutrition and livestock feed and studies on the
use of V faba for animal fodder have been carried out (Palander and others 2006) The protein fraction of V faba has also been
extensively characterized, including the amino acid composition(Lattanzio and others 1983), functional properties (Cepeda andothers 1998), the production of protein isolates (Macarulla andothers 2001), and protein hydrolysates (Chakraborty and others1979)
However, apart from V faba, there are very few studies on the
nutritional quality of seed proteins in wild populations of either
cultivated or noncultivated Vicia species.
In this study, the amino acid composition and nutritional
qual-ity of seed proteins were studied in 28 Vicia species (31 taxa)
distributed throughout Southern Spain Results provide useful formation concerning which of the species that are or were locallycultivated possess better nutritional properties in their seed pro-teins, and which wild taxa could be of interest from a nutritionalpoint of view
in-Material and Methods
MaterialsTrypsin, chymotrypsin and peptidase were from Sigma (TresCantos, Madrid, Spain) Diethyl ethoxymethylenemanolate waspurchased from Fluka (Buchs, Suiza) All other chemicals were of
Trang 20analytical grade Vicia seed samples were taken from wild
popu-lations except V faba, which was purchased from a local market.
Voucher specimens of studied populations are deposited in the
Herbarium of the Dept of Plant Biology and Ecology of the
Univ of Seville
Flour preparation
Select grains were milled using a grinder to obtain a uniform
flour
Protein estimation method
Total nitrogen was determined by the micro Kjeldahl method
according to AOAC (1999) 960.52 approved method Crude
pro-tein content was estimated using a conversion factor of 6.25 for
legumes
Amino acid analysis
Samples (10 mg) were hydrolyzed with 4 mL of 6 N HCl The
solutions were sealed in tubes under nitrogen and incubated in an
oven at 110 ◦C for 24 h Amino acids were determined after
derivatization with diethyl ethoxymethylenemalonate by
high-performance liquid chromatography (HPLC), according to the
method of Alaiz and others (1992), using D,L-α−aminobutyric
acid as an internal standard The HPLC system consisted of a
Model 600E multi-system with a 484 UV–vis detector (Waters,
Milford, Mass., U.S.A.; EEUU) equipped with a 300× 3.9 mm
i.d reversed-phase column (Novapack C18, 4μm; Waters) A
bi-nary gradient was used for elution with a flow of 0.9 mL/min
The solvents used were (A) sodium acetate (25 mM) containing
sodium azide (0.02%, w/v) pH 6 and (B) acetonitrile Elution was
as follows: time 0 to 3 min, linear gradient from A/B (91 : 9)
to A/B (86 : 14); 3 to 13 min, elution with A/B (86 : 14); 13 to
30 min, linear gradient from A/B (86 : 14) to A/B (69 : 31); 30 to
35 min, elution with A/B (69 : 31) The column was maintained
at 18◦C Tryptophan was analyzed by HPLC after basic hydrolysis
according to Yust and others (2004)
In vitro protein digestibility
In vitro protein digestibility (IVPD) was determined according
to the method of Hsu and others (1977) Samples containing
62.5 mg of protein were suspended in 10 mL of water and
the pH was adjusted to 8 An enzymatic solution containing
1.6 mg trypsin (17.7 BAEE U mg– ), 3.1 mg α-chymotrypsin
(43 U mg– ) and 1.3 mg peptidase (50 U g– ) per milliliter was
added to the protein suspension in a 1 : 10, v/v, ratio The pH of
the mixture was measured after 10 min, and the in vitro
digestibil-ity was calculated as a percentage of digestible protein using the
equation: % digestible protein= 210.464 – 18.103 × pH
Determination of nutritional parameters
The amino acid composition of studied Vicia was used for the
determination of several nutritional parameters of Vicia seed
pro-teins:
(1) Amino acid score (chemical score) was calculated as: %
sample essential amino acids contents /% recommended
essen-tial amino acids (FAO/WHO/UNU 1985)
(2) Protein efficiency ratio (PER) values were calculated from
the amino acid composition of Vicia seeds based on the following
3 equations (Alsmeyer and others 1974):
PER1= − 0.684 + 0.456 × Leu − 0.047 × Pro
PER2= − 0.468 + 0.454 × Leu − 0.105 × Tyr
PER3= − 1.816 + 0.435 × Met + 0.78 × Leu
± 211 × Hys − 0.944 × Tyr
(3) Protein digestibility corrected amino acid score (PDCAAS)(FAO/WHO 1989) was calculated as:
PDCAAS= Lowest uncorrected amino acid score × IVPD.
Uncorrected amino acid score= milligram of essential aminoacid in 1 g of test protein per milligram of essential amino acid in
1 g of reference protein (FAO/WHO/UNU 1985)
(4) Predicted biological value (BV) was calculated according toMorup and Olesen (1976) using the following equation:
BV= 102.15× Lys0.41× (Phe + Tyr)0.60× (Met + Cys)0.77
× Thr2.4× Trp0.21
where each amino acid symbol represents:
% amino acid/% amino acid FAO pattern(FAO/WHO/UNU
1985),
if % amino acid= % amino acid FAO pattern or
% amino acid FAO pattern (FAO/WHO/UNU 1985)/% aminoacid,
if % amino acid= % amino acid FAO pattern.
Cluster analysisResults are expressed as the mean values± SD of several sam-ples except for species with only 1 population Cluster analysis ofdifferent taxa was performed using PRIMER-pc program, em-ploying the Bray–Curtis index of dissimilarity (Bray and Curtis1957) The dissimilarity index was transformed to the index ofsimilarity (1 – dissimilarity index)× 100
Results and Discussion
Vicia seed protein contents ranged from 20.1% in V articulata to 32% in V pyrenaica (Table 1) These percentages observed in wild populations are similar to those reported for Vicia seeds belonging
to cultivated species such as V sativa In a screening of 840 V faba
lines, protein contents ranged from 18% to 31% with an averagecontent of 24% (Elsherbeeny and Robertson 1992) In addition,
the protein contents of several of the studied Vicia were also higher
than those observed in other cultivated legumes such as chickpeawith 24.7% seed protein ( S´anchez-Vioque and others 1999)
Amino acid composition of Vicia is shown in Table 1 Vicia species show an amino acid pattern characteristic of legumes
(Carbonaro and others 1997) According to FAO tions (FAO/WHO/UNU 1985), all taxa examined are deficient
recommenda-in sulfur amrecommenda-ino acids (Met and Cys) and Trp, except V altissima,
which was only deficient in Trp
In studied Vicia taxa, the most abundant essential amino
acids in decreasing order were Leu, Lys, Phe, Thr, and
Trang 22Val Leu values in all taxa are above FAO recommendations
(FAO/WHO/UNU 1985), with the highest contents observed in
V pubescens, which were significantly different (P < 0.05) from
those observed in V pseudocracca (Table 1) Lys, was not
limit-ing in any studied Vicia, with the highest percentages observed
in V ervilia, which were significantly different (P < 0.001) from
those observed in V eriocarpa and V vicioides Contents in aromatic
amino acids, Phe and Tyr, are also above FAO recommendations
(FAO/WHO/UNU 1985) These contents were similar in all
studied taxa, and only in V hybrida Phe percentages were
signifi-cantly different (P < 0.05) from those observed in V angustifolia,
V sativa subsp Sativa, and V tenuifolia His, Ile, and Val contents
are similar in all studied species with values in all cases above FAO
recommendations (FAO/WHO/UNU 1985)
Among the studied species, V altissima appears as the one with
the most balanced amino acid composition, because it was only
deficient in Trp Among the remaining taxa, V benghalensis and
V glauca subsp giennensis showed the best sulfur amino acids
con-tents (Table 1)
The importance of protein composition in the diet has been
well known for many years The nutritional quality of proteins
depends mainly on the amino acid composition To make use of
these amino acids, proteins must be digested to release them The
protein quality, also known as the nutritional or nutritive value of a
food, depends on its amino acid content and on the physiological
utilization of these amino acids after digestion and absorption
Many different factors may impair protein digestibility Hence, in
addition to amino acid composition, protein digestibility is another
protein characteristic which determines the protein quality
In vitro methods of protein evaluation are useful for determining
the nutritional value of new protein foods because they are faster
than in vivo methods (Monsoor and Yusuf 2002) Hsu and others
(1977) used a method that obtains in vitro digestibility value which
has high correlation with in vivo digestibility observed in rats.
IVPD in studied Vicia species ranged from 78% in V incana to
86.3% in V hirsuta (Table 2) Among the studied taxa, 14 of them
showed an IVPD higher (Table 2) than V faba (83.2%), which is
the most cultivated species Furthermore, most possessed an IVPD
higher than those observed in other legumes such as chickpea
(76.2%) (S´anchez-Vioque and others 1999), or other crops, such
as Amaranthus cruentus (77%) (Betschart 1982) or rapeseed (72.8%)
(Larbier and others 1991)
The highest Biological Value was observed in V benghalensis with
values of 67, whereas the lowest was observed in V hybrida with
18.7 (Table 2) The best biological values observed in this study
were similar to those reported in cereals such as wheat (61.6)
or triticale (65.3) (Friedman 1996), although lower than those
described for lactoalbumin (97%)
The PER is a good method for evaluating protein quality PER
values were calculated according to the formulae proposed by
Alsmeyer and others (1974), which have been shown to bear a
good relationship with real PER values The best PER values
were observed in V pubescens with PER1, PER2, and PER3 of
3.27, 3.29, and 3.77, respectively The lowest values were
calcu-lated for V tenuifolia with PER1 of 2.51 and PER2 of 2.58, and
V eriocarpa with PER3of 2.20 Similar values to those reported
here for Vicia have been described for other legumes (Alsmeyer and
others 1974) In this way, Sujak and others (2006) used PER2to
evaluate protein quality of Lupinus angustifolius (2.36) and Lupinus
albus (2.87).
The use of amino acid scores has been proposed as a more
ac-curate alternative to PER parameters (Sarwar and others 1984)
The chemical score is a simplified model for predicting dietaryprotein quality, but it does not take into account amino acid di-gestibility or availability (Moughan 2005) The amino acid score
was the highest in V hybrida with an average value of 129.5 (Table 2) The lowest value was observed in V benghalensis and
V eriocarpa with average values of 114.2 In a similar way, the % essential amino acids/total amino acids ratio was the highest in V hybrida with average values of 43.9, while the lowest was observed
in V benghalensis with average values of 38.8.
The PDCAAS is nowadays the most recommended theoreticalparameter for evaluating the nutritional quality of food proteins
It is based on FAO recommendations of amino acids requirements(FAO/WHO/UNU 1985) and IVPD It is superior to other meth-ods for evaluating the protein quality of food for humans because
it determines the quality of a protein based on the amino acidrequirements of a 2- to 5-year-old child adjusted for digestibility.The highest PDCAAS value for a given protein is 1.0, this proteinwill provide 100% of the indispensable amino acids recommended
by FAO/WHO (1989), however values≥ 0.5 are considered quite
good PDCAAS values in Vicia ranged from 0.32 in V faba to 0.55
in V lutea subsp vestita (Table 2) Most studied taxa have an
essen-tial amino acid composition similar or better in some indispensable
amino acid, like sulfur amino acids, than V faba Eleven of the
stud-ied taxa have a high nutritional quality proteins (PDCAAS≥ 0.5)(Table 2), showing similar PDCCAS values to those observed inother legumes, such as lentil (0.52) although lower than pea (0.69)(Henley and Kluster 1994)
Among mentioned taxa, V cordata, V lutea subsp vestita,
V monantha subsp calcarata, and V peregrina should be taken into
account because they have a good production of wild seeds withgood average size They have a balanced amino acid compositionand a high IVPD too Finally, all of them have a PDCAAS≥ 0.53,which is considered a good value and higher than value from
V faba (0.32), which is the most used as commercial legume.
The analysis of similarity of studied Vicia, based on seed protein
essential amino acid composition, shows that studied taxa have an
affinity higher than 93% (Figure 1) V vicioides (group A) clearly differs from all other Vicia taxa (group B), basically because of its high Thr contents Although V monardi also has also high percentages of Thr, it differs from V vicioides in its lower content
of Val and Ile (Table 1) Within group B another group (C) is
distinguished with a 94.6% affinity Group C includes V faba,
V pubescens, V hirsuta, V articulata, V ervilia, and V parviflora, which
are all annual species These species are characterized by higheramounts of Leu and lower amounts of Val in the seed proteins.Among the remaining taxa (group D), 3 groups (E, F, and G) weredistinguished Group E is formed by perennial species, except
V monardi, which is annual They tend to have higher contents
of Thr, ranging from 5.3 in V pyrenaica to 6.1 in V monardi.
Species in groups F and G are clearly distinguished due to thefact that those included in group F show higher contents of Phe,Ile, and Lys compared to those in group G In general, in bothgroups annual species predominate, although some species are
perennial such as V altissima in group F or V onobrychioides in group
G Group G is subdivided into 2 groups with all taxa belonging
to V lutea clustered in subgroup G2.
Species that are or have been used in crops in the past arelocated in groups C and F Although in other plant genera, such
as Teucrium, the amino acid composition has been useful from a
taxonomic point of view (Juan and others 2004), this was not so
in the case of Vicia Thus, clusters in Figure 1 do not correlate with recognized sections in the Vicia genus.
Trang 24Nutritional characterization of studied Vicia seed proteins shows
that these pulses have seed proteins of a high nutritional quality,
similar to those observed in other legumes and crops This
infor-mation may be useful for the revalorization of the Vicia species,
which were more broadly cultivated in the past It may also even
help in the domestication of new species or the use of wild
pop-ulations in breeding programmes, favoring the bioconservation of
Vicia.
References
Alaiz M, Navarro JL, Giron J, Vioque E 1992 Amino acid analysis by high-performance liquid
chromatography after derivatization with diethylethoxymethylenemalonate J Chromatogr
591:181–6.
Alsmeyer RH, Cunningham AE, Happich ML 1974 Equations predict PER from amino acid
analysis Food Technol Chicago 28(7):34–8.
AOAC 1999 Official methods of analysis (16th ed.) Arlington, Va.: Assoc of Official Analytical
Carbonaro M, Cappelloni M, Nicoli S, Lucarini M, Carnovale E 1997 Solubility–digestibility
relationship of legume proteins J Agric Food Chem 45:3387–94.
Cepeda E, Villaran MC, Aranguiz N 1998 Functional properties of faba bean (Vicia faba) protein
flour dried by spray drying and freeze drying J Food Eng 36:303–10.
Cerletti P, Fumagalli A, Venturin D 1978 Protein-composition of seeds of Lupinus-albus J Food
Sci 43:1409–12.
Chakraborty P, Bramsnaes F, Bose AN 1979 Characteristics of enzymatic hydrolysate of faba
bean (Vicia-faba minor) protein J Food Sci Technol Mysore 16:137–42.
Duranti M 2006 Grain legume proteins and nutraceutical properties Fitoterapia 77:67–82 Elsherbeeny MH, Robertson LD 1992 Protein content variation in a pure line faba bean
(Vicia-faba) collection J Sci Food Agric 58:193–6.
FAO/WHO 1989 Protein quality evaluation Report of a joint FAO/WHO expert consultation Food and nutrition paper nr 51 Rome, Italy: FAO/WHO.
FAO/WHO/UNU 1985 Energy and protein requirements Report of a joint FAO/WHO/ UNU expert consultation Technical report series nr 724 Geneva, Switzerland: FAO/WHO Friedman M 1996 Nutritional value of proteins from different food sources: a review J Agric Food Chem 44:6–29.
Granito M, Brito Y, Torres A 2007 Chemical composition, antioxidant capacity and
function-ality of raw and processed Phaseolus lunatus J Sci Food Agric 87:2801–9.
Henley EC, Kluster JM 1994 Protein quality evaluation by protein digestibility-corrected amino acid scoring Food Technol Chicago 48(4):74–7.
Hsu HW, Vavak D L, Satterlee LD, Miller GA 1977 A multienzyme technique for estimating protein digestibility J Food Sci 42:1269–73.
Juan R, Pastor J, Mill´an F, Vioque J 2004 Amino acids composition of Teucrium nutlet proteins.
Ann Bot London 94:615–21.
Larbier ZM, Chagneau AM, Lessire M 1991 Bioavailability of lysine in rapeseed and soybean meals determined by digestibility trial in cockerels and chick growth assay Anim Feed Sci Technol 35:237–46.
Lattanzio V, Bianco VV, Crivelli G, Miccolis V 1983 Variability of amino-acids, protein, vicine
and convicine in Vicia-faba (L) cultivars J Food Sci 48:992–3.
Macarulla MT, Medina C, De Diego MA, Chavarri M, Zulet MA, Martinez JA, Noel-Suberville
C, Higueret P, Portillo MP 2001 Effects of the whole seed and a protein isolate of faba bean
(Vicia faba) on the cholesterol metabolism of hypercholesterolaemic rats Brit J Nutr 85:607–14.
Mart´ın-Cabrejas M, D´ıaz MF, Aguilera Y, Ben´ıtez V, Moll´a E, Esteban RM 2008 Influence
of germination on the soluble carbohydrates and dietary fibre in non-conventional legumes Food Chem 107:1045–52.
Monsoor MA, Yusuf HKM 2002 In vitro protein digestibility of lathyrus pea (Lathyrus sativus), lentil (Lens culinaria) and chickpea (Cicer arietinum) Int J Food Sci Technol 37:97–9.
Morup IK, Olesen ES 1976 New method for prediction of protein value from essential amino acid pattern Nutr Rep Int 13:355–65.
Trang 25Palander S, Laurinen P, Perttila S, Valaja J, Partanen K 2006 Protein and amino acid digestibility
and metabolizable energy value of pea (Pisum sativum), faba bean (Vicia faba) and lupin (Lupinus
angustifolius) seeds for turkeys of different age Anim Feed Sci Technol 127:89–100.
Pastor-Cavada E, Juan R, Pastor J, Alaiz M, Vioque J 2011 Antioxidant activity in the seeds of
28 Vicia species from Southern Spain J Food Biochem Forthcoming.
Rochfort S, Panozzo J 2007 Phytochemicals for health, the role of pulses J Agric Food Chem
55:7981–94.
S´anchez-Vioque R, Clemente A, Vioque J, Bautista J, Mill´an F 1999 Protein isolate from
chickpea (Cicer arietinum L.): chemical composition, functional properties and protein
charac-terization Food Chem 64:237–43.
Sarwar G, Blair R, Friedman M, Gumbmann MR, Hackler LR, Pellet PL, Smith TK 1984 Inter- and intra-laboratory variability in rat growth assays for estimating protein quality in foods J Assoc Off An Chem 67:976–81.
Sujak A, Kotlarz A, Strobel W 2006 Compositional and nutritional evaluation of several lupin seeds Food Chem 98:711–9.
Yust MM, Pedroche J, Gir ´on-Calle J, Vioque J, Mill´an F, Alaiz M 2004 Determination of tryptophan by high-performance liquid chromatography of alkaline hydrolysates with spectrophotometric detection Food Chem 85:317– 20.
Trang 26Discrimination of Chinese Vinegars Based
on Headspace Solid-Phase Microextraction-Gas
Chromatography Mass Spectrometry of Volatile
Compounds and Multivariate Analysis
Zuobing Xiao, Shuiping Dai, Yunwei Niu, Haiyan Yu, Jiancai Zhu, Huaixiang Tian, and Yongbo Gu
Abstract: Headspace solid-phase microextraction (HS-SPME) coupled with gas chromatography mass spectrometry
(GC-MS) was applied for the determination of the characteristic volatile profiles of Chinese vinegars Multivariate
statistical techniques, such as principal component analysis (PCA) and cluster analysis (CA), were used to characterize
the different Chinese vinegars by types, fermentation method, and production area A total of 56 volatile compounds
were identified, including 15 esters, 10 aldehydes, 5 acids, 12 alcohols, 5 ketones, 4 volatile phenols, 2 pyrazines,
and 3 miscellaneous compounds The major compounds in Chinese vinegars were furfural, acetic acid, ethyl acetate,
3-hydroxy-2-butanone, 3-methyl-1-butanol, isopentyl acetate, benzaldehyde, phenylethyl alcohol The PCA results showed
that characterizing the Chinese vinegars by HS-SPME-GC-MS was highly related to their type, fermentation method, and
production area, and all these influencing factors were not independent The CA results indicated that the fermentation
method had a greater effect than vinegar type and production area The results showed that HS-SPME-GC-MS together
with chemometrics could provide practical reference for characterization of Chinese vinegars
Keywords: characterization, Chinese vinegar, cluster analysis, headspace solid-phase microextraction, principal component
analysis, volatiles
Practical Application: HS-SPME coupled with GC-MS was applied for the determination of the characteristic volatile
profiles of Chinese vinegars The major compounds in Chinese vinegars were furfural, acetic acid, ethyl acetate,
3-hydroxy-2-butanone, 3-methyl-1-butanol, isopentyl acetate, benzaldehyde, phenylethyl alcohol HS-SPME-GC-MS
together with chemometrics was an efficient tool for evaluating vinegar authenticity
Introduction
Vinegar is consumed worldwide as a food condiment and
preser-vative, especially in the Chinese diet Chinese vinegar has a history
of more than 3000 y (Shi 1999) Over 26 million hectoliters of
vinegar is produced every year, and more than 3.2 million liters
of vinegar is consumed every day in China Nowadays, Chinese
governments pay more and more attention on the quality control
of vinegar (Liu 2010) The aromatic vinegar, mature vinegar, rice
vinegar, and white vinegar are China’s 4 famous types of vinegar
The aromatic vinegar is dark gray or yellowish-brown, lustrous in
color and creamy, aromatic, mildly acidic, tasty, and slightly sweet
in flavor The mature vinegar ranges from brown to dark gray,
lustrous in color with a smoky smell, moderately acidic, and sweet
in flavor The rice vinegar is reddish-pink in color and exhibits a
pleasantly fresh and sweet smell, mildly acidic, slightly tasty, and
MS 20110094 Submitted 1/22/2011, Accepted 7/10/2011 Authors Xiao, Niu,
Yu, Zhu, Tian, and Gu are with School of Perfume and Aroma Technol., Shanghai
Inst of Technol., Shanghai 200233, China Author Dai is with School of Biological
Engineering, East China Univ of Science and Technol., Shanghai 200237, China.
Direct inquiries to author Xiao (E-mail: xzb@sit.edu.cn).
sweet in flavor The white vinegar is yellowish in color, aromatic,and strongly acidic in flavor
In Europe, the discrimination and characterization of gars have been widely investigated using gas chromatography massspectrometry (GC-MS) (Gerbi and others 1997; Tesfaye and oth-ers 2002; Callej ´on and others 2009) In China, several studieswere carried out to analyze Chinese vinegar using gas sensor ar-rays (Zhang and others 2006; Yin and others 2008) The maindifferences between Chinese vinegars and European vinegars areraw materials Chinese vinegars are usually produced from rice,sticky rice, sorghum, and wheat bran, while European vinegarsare mainly fermented from wine, cider, fruit juices, malted bar-ley, honey, and pure alcohol (Mej´ıas and others 2002) Chinesevinegars have a strong regional characteristic For example, Shanxiprovince is famous for mature vinegars and Jiangsu province isfamous for aromatic vinegars (Zhang and others 2006)
vine-Flavor is one of the important factors of vinegar character Theflavor of vinegar depends on the raw materials (rice, sticky rice,wheat bran, sugar, salt, and so on), the constituents formed dur-ing the fermentation and, in some cases, the substances formedduring aging (Cocchi and others 2008) Since the volatile profile
of vinegar represents a fingerprint of the sample, it is assumedthat a discrimination approach based on the analysis of volatile
C
r
Trang 27Headspace solid-phase microextraction (HS-SPME) is a fast,sensitive, and powerful method which offers large advantages overother sample preparation techniques as it eliminates the use ofextraction solvent and allows the extraction and the concentra-tion steps to be performed simultaneously (Guerrero and others2007) HS-SPME-GC-MS has been successfully used to analyzevolatile compounds from a large variety of food commodities forboth characterization and classification purposes (Wang and oth-ers 2009; Yang and others 2009; Zhao and others 2009; Callej ´onand others 2010; Garc´ıa-Mart´ın and others 2010) Pizarro andothers (2008) applied the HS-SPME-GC-MS and stepwise lineardiscriminant analysis (SLDA) method to classify different types ofEuropean vinegar Cirlini and others (2011) used the HS-SPME-GC-MS and principal component analysis (PCA) and classificationtree (CT) to classify the Balsamic vinegars of Modena (BVM) ofdifferent maturation and aging, and it indicated that this methodwas useful to discriminate BVM samples of different maturationand aging.
In this work, HS-SPME-GC-MS was applied to identify andquantify the volatile compounds in Chinese vinegars Using mul-tivariate analysis tools, such as PCA and cluster analysis (CA),the vinegar samples were characterized according to their type,fermentation method, and production area
Materials and Methods
Vinegar samplesEleven commercial Chinese vinegars were purchased from a lo-cal market in Shanghai, China The samples were kept refrigerated
at 4◦C Table 1 lists the name, type, raw materials, fermentationmethod, and production area which were copied from the vinegarbottle labels Since geographical discrimination was attempted, theproduction areas are shown in Figure 1
Trang 28be more efficient at extracting the volatile compounds of vinegar
than other fibers such as those coated with
polydimethylsilox-ane (PDMS), carbowax-divinylbenzene (CW-DVB), and
poly-dimethyl siloxane-divinylbenzene (PDMS-DVB) In this research,
a manual SPME holder together with one 75 μm
carboxyl-polydimethylsiloxane (CAR-PDMS) fiber, 20 mL vials, and Teflon
covers was used (Supelco Inc., Bellefonte, Pa., U.S.A.) Before
sorption, the fiber was preconditioned for 30 min on the Agilent
6890 gas chromatograph (Agilent Technologies, New York, N.Y.,
U.S.A.) with the injector temperature of 250◦C Between
injec-tions, the fiber was desorbed for 10 min at 250◦C in split mode
in order to prevent any contamination
The 20 mL vial with 5 mL Chinese vinegar sample and
20μL internal standard solution (2-octanol [40 mg/L in ethanol],
Sigma-Aldrich, St Louis, Mo., U.S.A.) was sealed up with one
Teflon cover The SPME fiber was exposed in the upper space of
the vial at 45◦C for 30 min while shaken at intervals, and then
withdrawn and directly introduced to the GC-MS injector for
desorption and analysis
GC-MS
GC-MS analysis was carried out on an Agilent 6890 gas
chro-matograph equipped with a mass selective detector (MSD) model
5973 N (Agilent Technologies) The capillary column used was
an HP-Innowax (60 m × 0.25 mm i.d × 0.25 μm film,
Agi-lent) from Agilent Technologies The injection was conducted in
splitless mode for 3 min at 220◦C The GC operation conditions
were as follows The injector temperature was 220◦C; the carrier
gas employed was helium (purity 99.999%) at a constant flow rate
of 1 mL/min The oven temperature was held at 40◦C for 2 min,
ramped to 150 ◦C at the rate of 3◦C/min, then again ramped
to 220◦C at the rate of 5 ◦C/min, and maintained for 5 min
The MSD was used for identification and electron impact energy
was 70 eV The ion source temperature was set at 230◦C The
quadrupole mass filter was operated at 150◦C The transfer line
temperature was at 250◦C The chromatograms were recorded by
monitoring the total ion currents in the 30 to 450 mass range In
terms of the number of peaks and peak area, several parameters in
the chromatogram were adjusted to increase reliability Peak width
was set to 0.04 and initial threshold was set to 13 All samples were
analyzed in triplicate and the average values were calculated
Compound identification and quantification
The volatile compounds were identified by comparing Kovats
retention indices (RI) (Kovats, 1965) and the MS fragmentation
patterns with those of reference compounds, or with mass spectra
in the Wiley7n.l and NIST05 Database (Agilent Technologies)and previously reported Kovats RI The Kovats RI of unknowncompounds were determined via sample injection with a homol-ogous series of alkanes (C6–C25) (Sigma-Aldrich) The GC-MSconditions were the same as described above To quantify thevolatiles, the integrated areas based on the total ion chromatogramswere normalized to the areas of the internal standard The rela-tive volatile concentrations in the 11 samples were determined
by comparison with the concentration of the internal standard(2-octanol)
Multivariate analysisPCA is essentially a descriptive method PCA was used to pro-vide a partial visualization of data in a reduced-dimension plot.The principal components or eigenvectors are orthogonal and theyare a linear combination of the original variables This method,normally the first step in data exploration, allows the main vari-ability aspects of a data set to be visualized without the constraint
of an initial hypothesis concerning the relationship within samplesand between samples and responses (variables) The main goals ofthis method are to find relationships between the different param-eters (objects and variables) and to detect possible clusters withinobjects or/and variables (Garc´ıa-Mart´ın and others 2010)
CA is an unsupervised chemometric technique that revealsthe natural groupings existing between samples characterized bythe values of measured variables A Euclidean metric is used in the
CA with an “average between groups” method of linkage
In this research, PCA was performed on the mean values of therelative volatile concentrations using the correlation matrix with
no rotation CA was employed to examine the relationships of theinfluencing factors (type, fermentation method, and productionarea) of various vinegars more intuitive All statistical analyses wereperformed using XLSTAT ver.7.5 (Addinsoft, New York, N.Y.,U.S.A.)
Results and Discussion
Volatile compounds in Chinese vinegar samplesThe assignments of the principal peaks of Chinese vinegar sam-ples’ volatile profile obtained from HS-SPME-GC-MS analysis areshown in Figure 2 Table 2 lists the identified volatile compounds
in the 11 samples by their chemical classes, relative tions, and RIs on the HP-Innowax column, respectively A total
concentra-of 56 volatile components, including 15 esters, 10 aldehydes, 5acids, 12 alcohols, 5 ketones, 4 volatile phenols, 2 pyrazines, and
3 miscellaneous compounds, were identified Among the fied volatiles, esters, acids, aldehydes, and alcohols were the largest
quanti-Figure 2–The representative total ion chromatogram for volatiles of vinegar sample.
Trang 30Esters. A total of 15 esters were detected and were among one
of the largest classes of detected compounds The ethyl acetate,isopentyl acetate, phenethyl acetate, and isobutyl acetate were themajor compounds in the 11 samples All the detected esters werepreviously found in various vinegars (Durante and others 2006;Garc´ıa-Mart´ın and others 2010) The content of ethyl acetate inthe type of rice vinegar and white vinegar (except Biaoding Baicu)was much higher than that in the other types of samples Themanufacturing process of rice vinegar and white vinegar employsliquid fermentation rather than solid fermentation The liquidfermentation may affect rice lipid utilization and formation of highmolecular weight esters (Ames and Macleod 1984) Esters werederived from the esterification of free fatty acids and alcohols.Thus, esters of the corresponding fatty acids were expected to bepresent However, the current results showed that some esters wereabsent even though their corresponding fatty acids were found.This could be an indication of either their presence in minuteamount, which was below the detection limit of the analyticalinstrument, or their absence in the process
Aldehydes. Ten aldehydes were identified in the ples, and 5 of them (furfural, benzaldehyde, acetaldehyde, 4-methoxybenzaldehyde, and cinnamaldehyde) were found in all the
sam-11 samples All the compounds were previously detected in variousvinegars (Del Signore 2001; Mar´ın and others 2002; Casale andothers 2006) The most abundant aldehydes in Chinese vinegarswere furfural, with the concentration of 35 mg/L The furfuralconcentrations in the 4 samples (Meigui Micu, Biaoding Baicu,Hengshun Baicu, and Qingfeng Baicu) were much lower thanthose in the other samples The reason could be owed to thedifferent fermentation methods Furans such as furfural and 5-methylfurfural can be formed by sugar dehydration or fragmenta-tion during the Maillard reaction (Fors 1983) Benzaldehyde wasdescribed as having sweet, fruity, nutty, and caramel-like odors
In addition, 4-methoxylbenzaldehyde (also named anisaldehyde)with sweet, powder, and spicy aroma, was the principal constituent
of anise oil (Rodrigues and others 2003) Due to low odor olds and distinctive odors, the above volatile aldehydes are potentialcontributors to Chinese vinegars (Xie and others 2008)
thresh-Acids. Five acids were identified in the samples, including lowcarbon chain saturated acid (C2–C10) The majority of the acids arereleased during extracellular digestion by lipases, while propanoicacid and 3-methylbutanoic acid (isovaleric acid) are consideredbyproducts of yeast protein metabolism (Han and others 2001).Among the 6 acids, the content of acetic acid was the highest.3-Methylbutanoic acid was a combination of milk aroma, souraroma, fruit aroma, and fat aroma It is a key compound in themajority of fermented products like cheese and yogurt (Ohatahand others 2009) Acids, such as hexanoic acid, were absent inthe type of Baicu samples (Biaoding, Hengshun, and Qingfeng).However, high concentrations of corresponding esters (ethyl hex-anoate) were found in these samples Because ethyl esters are
Trang 31derived from the esterification of free fatty acids and ethanol, their
corresponding fatty acids are expected to be present during the
fermentation process However, the current results indicated that
some fatty acids (lactic acid and octanoic acid) were absent even
though their corresponding esters (ethyl lactate and ethyl
oc-tanoate) were found This could be an indication of strong
es-terification in the processing of fermentation in samples
Alcohols. Alcohols were another large class, containing 12
compounds in the samples Among the alcohols,
3-methyl-1-butanol, 2,3-butanediol, and phenylethyl alcohol were the major
compounds These alcohols were also reported in balsamic vinegar
(Del Signore 2001) and wine vinegar (Morales and others 2004;Callej ´on and others 2010) as important contributors of flavor.Phenylethyl alcohol not only serves as a perfume formulation in-gredient in the flavor industry, but also acts as an aroma enhance-ment in the fermentation products such as bean paste (Lee andAhn 2009) The fusel alcohols, including 2-methylpropanol and3-methylbutanol, were mainly produced by the action of variousfungal enzymes during fermentation (Chung and others 2005).Hexanol, a common and important flavor component for somevegetables and plants, was the key component for compound-ing vegetable flavor It was derived from the bioremediation of
Variables (axes F1 and F2: 55.95 %)
AD1
AD2
AD6 AD7
AD8 AD9 AD10
ES1
ES2
ES3 ES4
ES5 ES6
ES7 ES8
ES9
ES10 ES11
KE4
KE5
AL1 AL2
AL3 AL4
AL5 AL6
PH2 PH3 PH4
by their chemical classes: , aldehyde (AD);, ester (ES);, ketone (KE); , alcohol (AL); ♦, acid (AC);◦, pyrazine (PY); +, phenol (PH);
–, miscellaneous (MI).
Observations (axes F1 and F2: 55.95 %)
aro_solid_shang
aro_solid_bei aro_solid_heng aro_solid_zheng
mat_solid_dong mat_solid_jin
ric_solid_jin ric_liquid_mei whi_liquid_biaowhi_liquid_heng
Trang 32unsaturated fatty acids, and was also the prerequisite chemical for
the formation of long-chain esters (Methven and others 2007)
Ketones. A total of 5 ketones were identified, including 4
aliphatic ketones (2-butanone, 2,3-butanedione, 2-heptanone,
3-hydroxy-2-butanone) and 1 cyclic ketone
(1-phenyl-1-propanone) The concentration of 3-hydroxy-2-butanone in the
vinegar samples, which accounted for approximately 7 mg/L, was
much higher than those other ketones 3-Hydroxy-2-butanone
was a characteristic compound of acetification (Morales and others
2001) 3-Hydroxy-2-butanone was produced during alcoholic
fer-mentation by the action of several microorganisms and was present
in fermented foods and beverages such as wine, vinegar, and
sev-eral dairy products (Caligiani and others 2007) In the QingfengBaicu sample, only 3-hydroxy-2-butanone was detected, lackingmajor ketones such as 2-butanone and 2,3-butanedione Somealiphatic ketones, such as 2-butanone and 2,3-butanedione, could
be products of lipid oxidation or degradation, assisted by enzymesfrom the inoculated mold (Wang and Hesseltine 1970)
Phenols. A total of 4 phenols (4-(1-methylpropyl)-phenol,2-methoxyphenol, 2-methoxy-4-methylphenol, phenol) werefound in the 11 vinegar samples As is well known, phenolswere responsible for smoky and phenolic odors Phenolic com-pounds were mainly produced by thermal degradation throughdepolymerization or oxidation of lignin, which was composed of
Observations (axes F1 and F2: 55.95 %)
whi_liquid_qing
whi_liquid_heng whi_liquid_biao
ric_liquid_mei ric_solid_jin
mat_solid_jin mat_solid_dong
aro_solid_zheng aro_solid_heng aro_solid_bei
Figure 5–PCA results of the vinegar samples;
making for vinegar type: , aromatic vinegar;+,
mature vinegar;, rice vinegar;, white vinegar.
Observations (axes F1 and F2: 55.95 %)
aro_solid_shang
aro_solid_bei aro_solid_heng aro_solid_zheng
mat_solid_dong mat_solid_jin
ric_solid_jin ric_liquid_mei
Figure 6–PCA results of the vinegar samples;
making for raw material: , sticky rice;+, neither
rice nor sticky rice;, rice.
Trang 33repeating phenol units having 3 carbon side chains (Natera and
others 2003) Because of their low threshold value, their flavor
is a characteristic aroma of the vinegars, although the content of
phenols was not very high
Pyrazines. Only 2 pyrazines (trimethylpyrazine and
3-ethyl-2, 5-dimethylpyrazine) were identified in the Chinese vinegars,
which were in relatively low concentration Pyrazine was not
found in the Meigui Micu, Biaoding Baicu, Hengshun Baicu,
and Qingfeng Baicu samples Pyrazines had a mixed aroma of
nut, peanut, and cocoa, particularly alkyl-pyrazines (Gilbertson
and Koenig 1979) In general, although pyrazines were present in
small quantities in natural samples, their contribution to the flavor
of these samples was considerable
Miscellaneous compounds. Three additional miscellaneouscompounds were identified in the 11 samples 2-Acetylfuran waspreviously found in balsamic vinegar (Cocchi and others 2004) and
coffee brews (Lopez-Galilea and others 2006) It was recognized as
having green rice and brown sugar aromas (Fors 1983) Furans can
be formed by sugar dehydration or fragmentation in the Maillardreaction (Fors 1983) Limonene considered to be one of the majorcontributors to orange flavor, had a weak, citrus-like aroma with
a high orthonasal odor threshold (13.7 ppm) in a deodorized
Observations (axes F1 and F2: 55.95 %)
aro_solid_shang
aro_solid_bei aro_solid_heng aro_solid_zheng
mat_solid_dong mat_solid_jin
ric_solid_jin ric_liquid_mei whi_liquid_biao
Observations (axes F1 and F2: 55.95 %)
aro_solid_shang
aro_solid_bei aro_solid_heng aro_solid_zheng
mat_solid_dong mat_solid_jin
ric_solid_jin ric_liquid_mei
Trang 34orange juice matrix (Plotto and others 2004) Anethole had an
aniselike odor and sweet taste It was a major component in anise
oil and is often used in the alcoholic beverage industry (Chung
and Cadwallader 1994)
Principal component analysis
PCA was used to provide a partial visualization of data in a
reduced-dimension plot Figure 3 and 4 show the first 2
prin-cipal components (PCs) accounted for 55.95% of the total
vari-ation across the samples (PC1 and PC2 accounted for 33.19%
and 18.77% of the variance, respectively) When examining the
volatile distributions (Figure 3), the major compounds positively
contributing to the PC1 dimension (factor loadings> 0.7) were
aldehydes such as acetaldehyde (AD1), 2-methylpropanal (AD2),
5-methylfurfural (AD8), 4-methoxybenzaldehyde (AD9), and
cinnamaldehyde (AD10); esters such as 2-butyl acetate (ES5) and
methyl cinnamate (ES15); ketones such as 2-butanone (KE1),
2,3-butanedione (KE2), and 3-hydroxy-2-butanone (KE4); alcohols
such as 2-butanol (AL1) and furfuryl alcohol (AL9); acids such as
propionic acid (AC2), hexanoic acid (AC5), and trimethylpyrazine
(PY1); phenols such as 2-methoxyphenol (PH2) and
2-methoxy-4-methylphenol (PH3); and miscellaneous compounds such as
limonene (MI1), 2-acetylfuran (MI2), and anethole (MI3) On
the negative side of the PC1 axis, ethyl acetate (ES2) was the main
compound (factor loadings< −0.7) In contrast, the important
compounds of the PC2 dimension were esters such as propyl
acetate (ES4), ethyl hexanoate (ES9), ethyl benzoate (ES12),
and diethyl succinate (ES13); ketones such as 2-heptanone
(KE3) and 1-phenyl-1-propanone (KE5); and alcohols such
as 1-propanol (AL2), 1-pentanol (AL5), and 1-hexanol
(AL6)
It was possible to determine the major volatile compoundscontributing to the differences in the samples by analyzing thecorrelation of each variable (Table 2) with the PC1 and PC2scores (Figure 4) In the upper side of PC1, the Shanghai Xi-angcu sample (aro_solid_shang) was located in the same direc-tion as the esters (ES4, ES9, ES13), ketones (KE3, KE5), andalcohols (AL2, AL5, AL6), which had the highest concentrations
of these compounds among the tested samples Similarly, in thelower right side of PC1, the Donghu Chencu (mat_solid_dong)and Jinbiao Chencu (mat_solid_jin) samples were highlyassociated with aldehydes (AD1, AD2, AD6, AD8, AD9, AD10),ketones (KE1, KE2, KE4), phenols (PH2, PH3), trimethylpyrazine(PY1), and miscellaneous compounds (MI1, MI2) On the nega-tive side of PC1 and on the positive side of PC2, the Jindan Micu(ric_solid_jin), Biaoding Baicu (whi_liquid_biao), and HengshunBaicu (whi_liquid_heng) samples were located in proximity withesters (ES2, ES6, ES8) and 2-methyl-1-propanol (AL3) On thenegative side of PC1 and PC2, the Meigui Micu (ric_liquid_mei)sample was highly associated with ethyl butanoate (ES7) and bu-tanoic acid (AC3)
To discriminate the samples of different types, raw terial, fermentation method, and production area, the samedata were plotted with different markings in Figure 5 to
ma-8, which showed the PCA plots of the vinegars based ontype, raw material, fermentation method, and production area,respectively
It can be observed that 4 distinct groups, which correspond toaromatic vinegar, mature vinegar, rice vinegar, and white vinegar,respectively, are well distinguished in Figure 5 It can be observedthat all the vinegars were well associated to their type, which is acritical factor to the character of Chinese vinegars The aromatic
Figure 9−CA dendrogram based on the
concentration of 56 volatiles to the 11 vinegar samples.
Trang 35vinegars and the mature vinegars are the most favorite vinegars
in China The aromatic vinegars are fermented from sticky rice
and mainly produced in Jiangsu province, the mature vinegars
are fermented from sorghum and produced in Shanxi province,
while the rice vinegars and white vinegars are fermented from
rice and mainly produced in Shanghai The difference between
the aromatic vinegars and the mature vinegars is big, while the
difference between the rice vinegars and the white vinegars is very
small in Figure 5 The difference between the Shanghai Xiangcu
and the other aromatic vinegars is also big, though they all belong
to the aromatic vinegar type The reason is owed to the different
production areas (Shanghai and Jiangsu province, respectively)
Comparing the relationships between the vinegars and their raw
materials was difficult in this study because each vinegar contained
more than 3 or 4 different materials As shown in Figure 6, the
vinegars can also be clearly divided into 3 groups according to their
main raw material, the samples aro_solid_zheng, aro_solid_bei,
aro_solid_heng, and aro_solid_shang including sticky rice in the
raw material, mat_solid_dong and mat_solid_jin without rice or
sticky rice, and ric_solid_jin, ric_liquid_mei, whi_liquid_biao,
whi_liquid_heng, and whi_liquid_qing made from rice So it is
clearly discernable that the main raw material is an important
factor in characterizing Chinese vinegars
As shown in Figure 7, the vinegars can be clearly divided into
2 groups according to their fermentation methods, solid
fermen-tation vinegar and liquid fermenfermen-tation vinegar The Jindan Micu
(ric_solid_jin), which belongs to the solid fermentation vinegar,
is very close to the group of the liquid fermentation vinegars in
Figure 7 The use of similar raw materials is the possible
rea-son for the small difference between Jindan Micu and the liquid
fermentation vinegars It was concluded that the fermentation
method was highly related to the vinegar raw materials and was
also an essential factor to the character of Chinese vinegars
As shown in Figure 8, the regional definition of the samples
is not that clear compared to type, raw material, or fermentation
method Traditionally, Chinese vinegars have a strong regional
characteristic For example, Shanxi province is famous for mature
vinegars and Jiangsu province is famous for aromatic vinegars The
difference between the Jiangsu vinegars and the Shanghai vinegars
is small, while the difference between the Jiangsu vinegars and the
Shanxi vinegars is very big This phenomenon can be explained
by Jiangsu being very close to Shanghai, but quite far from Shanxi,
indicating that the production area is also an important factor to
character the Chinese vinegars
From the above analysis, it can also be obtained that all the
influencing factors are not independent
Cluster analysis
In the present research, CA was applied to study the relationships
and the scale of the type, fermentation method, and production
area, when the vinegar samples were analyzed by
HS-SPME-GC-MS The similarities between samples were calculated on the basis
of the Euclidean distance and the CA dendrogram is shown in
Figure 9
Two different groups can be observed in the dendrogram
The 1st group included the samples ric_solid_jin, ric_liquid_mei,
whi_liquid_biao, whi_liquid_heng, and whi_liquid_qing, which
referred to rice vinegar and white vinegar except the sample
aro_solid_shang And the samples aro_solid_zheng, aro_solid_bei,
aro_solid_heng, mat_solid_dong, and mat_solid_jin formed the
2nd group, which referred to aromatic vinegar and mature vinegar
This shows that the rice vinegar and white vinegar are
similar, while the aromatic vinegar and the mature vinegar aresimilar The reason for this phenomenon is the same fermenta-tion method, the 1st group and the 2nd group belonged to theliquid fermentation and solid fermentation, respectively The Eu-clidean distance between the aromatic vinegars of aro_solid_zhengand aro_solid_heng is close However, the Shanghai Xiangcu(aro_solid_shang), which also belongs to the aromatic vinegars,
is close to the Jindan Micu (ric_solid_jin) This agrees with thePCA results and the reason has been discussed before Taken as
a whole, these results indicate that the fermentation method is amore important factor than the vinegar type and raw materials ininfluencing the characters of Chinese vinegars
Conclusions
This study represents one of the first approaches to the terization of Chinese vinegars by HS-SPME-GC-MS The resultsobtained from the analysis of the volatile profile of Chinese vine-gars coupled with statistical data analysis by means of PCA and
charac-CA have shown that this method can be useful to characterize ferent Chinese vinegar samples (types, fermentation method, andproduction area), to have a quality and authenticity evaluation ofthe product The volatile profile of Chinese vinegars belonged tothe following families: esters, aldehydes, acids, alcohols, ketones,volatile phenols, pyrazines, and miscellaneous compounds ThePCA results showed that a large number of volatiles corresponded
dif-to each of the Chinese vinegars Aromatic vinegar samples werelocated in the same direction as the alcohols and acids Similarly,mature vinegar samples were highly associated with aldehydes andketones on the positive side of the PC1 axis On the negative side
of PC1 and on the positive side of PC2, rice vinegar and whitevinegar samples were located in proximity with esters It was clearthat characterizing the Chinese vinegars by HS-SPME-GC-MSwas highly related to their type, fermentation method, and produc-tion area, and all these influencing factors were not independent.The CA results showed that the fermentation method was moreimportant than the type and production area Future studies mightexplore more vinegar samples of other varieties and build a morerobust model for the characterization and classification of vinegars
by HS-SPME-GC-MS
Acknowledgments
The authors gratefully acknowledge the financial supportprovided by National Natural Science Foundation of China(nr 20876097) and National 973 Foundation of China (nr2009CB226104)
Callej ´on RM, Torija MJ, Mas A, Morales ML, Troncoso AM 2010 Changes of volatile pounds in wine vinegars during their elaboration in barrels made from different woods Food Chem 120:561–71.
com-Casale M, Armanino C, Casolino C, Oliveros CC, Forina M 2006 A chemometrical approach for vinegar classification by headspace mass spectrometry of volatile compounds Food Sci Technol Res 12(3):223–30.
Chung HY, Cadwallader KR 1994 Aroma extract dilution analysis of blue crab claw meat volatiles J Agric Food Chem 42:2867–70.
Chung HY, Fung PK, Kim JS 2005 Aroma impact components in commercial plain sufu.
J Agric Food Chem 53:1684–91.
Cirlini M, Caligiani A, Palla L, Palla G 2011 HS-SPME/GC-MS and chemometrics for the classification of Balsamic Vinegars of Modena of different maturation and ageing Food Chem 124:1678–83.
Trang 36Cocchi M, Durante C, Foca G, Manzini D, Marchetti A, Ulrici A 2004 Application of a
wavelet-based algorithm on HS-SPME/GC signals for the classification of balsamic vinegars.
Chemom Intell Lab Syst 71(2):129–40.
Cocchi M, Durante C, Grandi M, Manzini D, Marchetti A 2008 Three-way principal
com-ponent analysis of the volatile fraction by HS-SPME/GC of aceto balsamico tradizionale of
modena Talanta 74(4):547–54.
Cocchi M, Durante C, Marchetti A, Armanino C, Casale M 2007 Characterization and
discrimination of different aged ‘Aceto Balsamico Tradizionale di Modena’ products by head
space mass spectrometry and chemometrics Anal Chim Acta 589(1):96–104.
Del Signore A 2001 Chemometric analysis and volatile compounds of traditional balsamic
vinegar from Modena J Food Eng 50:77–90.
Durante C, Cocchi M, Grandi M, Marchetti A, Bro R 2006 Application of N-PLS to gas
chromatographic and sensory data of traditional balsamic vinegars of modena Chemom Intell
Lab Syst 83(1):54–65.
Fors S 1983 Sensory properties of volatile Maillard reaction products and related compounds:
a literature review In: Waller GR Feather MS., editors The Maillard reaction in foods and
nutrition Washington, D.C.: American Chemical Society p 185–286.
Garc´ıa-Mart´ın S, Herrero C, Pe ˇna RM, Barciela J 2010 Solid-phase microextraction gas
chromatography-mass spectrometry (HS-SPME-GC-MS) determination of volatile
com-pounds in orujo spirits: multivariate chemometric characterization Food Chem 118:456–61.
Gerbi V, Zeppa G, Antonelli A, Carnacini A 1997 Sensory characterization of wine vinegars.
Food Qual Prefer 8:27–34.
Gilbertson G, Koenig RT 1979 Essential oils and related products Anal Chem 51(5):183–96.
Guerrero ED, Mar´ın RN, Mej´ıas RC, Barroso CG 2007 Stir bar sorptive extraction of volatile
compounds in vinegar: validation study and comparison with solid phase microextraction.
J Chromatogr A 1167(1):18–26.
Han BZ, Rombouts FM, Nou MJR 2001 A Chinese fermented soybean food Int J Food
Microbiol 65:1–10.
Lee SJ, Ahn B 2009 Comparison of volatile components in fermented soybean pastes using
simultaneous distillation and extraction (SDE) with sensory characterization Food Chem
114:600–9.
Liu K 2010 Discuss Chinese vinegar function and trend of development China Condiment
6:32–9.
Lopez-Galilea I, Fournier N, Cid C, Guichard E 2006 Changes in headspace volatile
concen-trations of coffee brews caused by the roasting process and the brewing procedure J Agric
Food Chem 54:8560–6.
Mar´ın RN, Mej´ıas RC, Moreno MVG, Rowe FG Barroso CG 2002 Headspace solid-phase
microextraction analysis of aroma compounds in vinegar Validation study J Chromatogr A
967:261–7.
Mej´ıas RC, Mar´ın RN, Moreno MVG, Barroso CG 2002 Optimisation of headspace
solid-phase microextration for analysis of aromatic compounds in vinegar J Chromatogr A
953:7–15.
Methven L, Tsoukka M, Oruna-concha MJ, Parker JK, Mottram DS 2007 Influence of sulfur
amino acids on the volatile and nonvolatile components of cooked salmon J Agric Food
Natera R, Castro R, De Valme Garc´ıa-Moreno M, Hern´andez MJ, Garc´ıa-Barroso C 2003 Chemometric studies of vinegars from different raw materials and processes of production.
J Agric Food Chem 51(11):3345–51.
Ohatah M, Tominaga T, Dubourdieu D, Kubota K, Sugawara E 2009 Quantification and odor contribution of 2-furanmethanethiol in different types of fermented soybean paste miso.
J Agric Food Chem 57(6):2481–5.
Pizarro C, Esteban-D´ıez I, S´aenz-Gonz´alez C, Gonz´alez-S´aiz JM 2008 Vinegar cation based on feature extraction and selection from headspace solid-phase microex- traction/gas chromatography volatile analyses: a feasibility study Anal Chim Acta 608: 38–47.
classifi-Plessi M, Bertelli D, Miglietta F 2006 Extraction and identification by GC-MS of nolic acids in traditional balsamic vinegar from Modena J Food Compos Anal 19: 49–54.
phe-Plotto A, Margar´ıa CA, Goodner KL, Goodrich R, Baldwin EA 2004 Odour and flavour thresholds for key aroma components in an orange juice matrix: terpenes and aldehydes Flavour Frag J 19:491–8.
Rodrigues V, Rosa PV, Marques MM, Petenate A, Meireles MA 2003 Super critical tion of essential oil from aniseed (pimpinella anisuml) using CO 2 : solubility, kinetics, and composition data J Agric Food Chem 51:1518–23.
extrac-Shi G 1999 Talk about Chinese vinegar China Brewing 6:39–40.
Tesfaye W, Morales ML, Garcia-Parrilla MC, Troncoso AM 2002 Wine vinegar: technology, authenticity and quality evaluation Trends Food Sci Tech 13:12–21.
Wang HL, Hesseltine CW 1970 Sufu and lao-chao J Agric Food Chem 18(4):572–5.
Wang YJ, Yang CX, Li SH, Yang L, Wang YN, Zhao JB, Jiang Q 2009 Volatile tics of 50 peaches and nectarines evaluated by HP-SPME with GC-MS Food Chem 116: 356–64.
characteris-Xie JC, Sun BG, Zheng FP, Wang SB 2008 Volatile flavor constituents in roasted pork of Mini-pig Food Chem 109:506–14.
Yang CX, Wang YJ, Liang ZC, Fan PG, Wu BH, Yang L, Wang YN, Li SH 2009 Volatiles
of grape berries evaluated at the germplasm level by headspace-SPME with GC-MS Food Chem 114:1106–14.
Yin Y, Yu HC, Zhang HS 2008 A feature extraction method based on wavelet packet analysis for discrimination of Chinese vinegars using a gas sensors array Sensors Actuators B 134: 1005–9.
Zhang QY, Zhang SP, Xie CS, Zeng DW, Fan CQ, Li DF, Bai ZK 2006 Characterization of Chinese vinegars by electronic nose Sensors Actuators B 119:538–46.
Zhao Y, Xu Y, Li J, Fan W, Jiang W 2009 Profile of volatile compounds in 11 brandies by headspace solid-phase microextraction followed by gas chromatography-mass spectrometry J Food Sci 74:90–9.
Trang 37Underestimation of Pyruvic Acid
Concentrations by Fructose and Cysteine
in 2,4-Dinitrophenylhydrazine-Mediated
Onion Pungency Test
Kil Sun Yoo, Eun Jin Lee, and Bhimanagouda S Patil
Abstract: Onion pungency has been routinely measured by determining pyruvic acid concentration in onion juice
by reacting with 2,4-dinitrophenylhydrazine (DNPH) since 1961 However, the absorbency of the color adduct of thereaction rapidly decreased in onion samples as compared to that of the pyruvic acid standards, resulting in underestimations
of the pyruvic acid concentrations By measuring the absorbency at 1 min, we have demonstrated that accuracy could besubstantially improved As a continuation, the causes of degradation of the color adduct after the reaction and pyruvic aciditself before the reaction were examined in this study Alliinase action in juice (fresh or cooked) and bulb colors did notinfluence the degradation Some organic acids indigenously found in onion, such as ascorbic acid, proline, and glutamicacid, did not reduce the absorbency However, fructose within the onion juice or supplemented caused the degradation
of the color adduct, whereas sucrose and glucose had a lesser effect Degradation rates increased proportionally as fructoseconcentrations increased up to 70 mg/mL Cysteine was found to degrade the pyruvic acid itself before the pyruvic acidcould react with DNPH Approximately 90% of the pyruvic acid was degraded after 60 min in samples of 7 mM pyruvicacid supplemented with 10 mg/mL cysteine Spectral comparisons of onion juice containing fructose naturally and pyruvicacid solution with supplemented fructose indicated identical patterns and confirmed that the color-adduct degradationwas caused by fructose Our study elucidated that fructose, a major sugar in onion juice, caused the degradation of coloradduct in the onion pungency test and resulted in underestimation of the pyruvic acid concentration
Keywords: cysteine, dinitrophenylhydrazine (DNPH), fructose, pungency, pyruvic acid
Introduction
Onion pungency has been widely measured by reacting
2,4-dinitrophenylhydrazine (DNPH) with the pyruvic acid found
in the onion juice Pyruvic acid is a byproduct of alliinase
hydrolysis of the flavor precursors in onions upon cell disruption
(Schwimmer and Western 1961; Randle and Bussard 1993; Yoo
and others 1995; Boyhan and others 1999; Yoo and Pike 1999;
Anthon and Barret 2003) Though pungency itself is caused by
many sulfur compounds from the alliinase reaction (Lancaster and
Boland 1990), pyruvic acid concentration is linearly proportional
to the levels of S-alk(en)yl-cysteine-sulfoxides, the flavor precursor
compounds (Lee and others 2009) The pungency level of onions
was largely determined by cultivar and secondarily modified by
environmental factors (Yoo and others 2006)
The pungency test method was first introduced by Schwimmer
and Western (1961) and has been modified for faster sample
processing and interference reduction (Randle and Bussard 1993;
Yoo and others 1995; Boyhan and others 1999; Yoo and Pike
1999; Anthon and Barret 2003) However, significant variation
MS 20110146 Submitted 2/2/2011, Accepted 7/7/2011 Authors are with
Veg-etable and Fruit Improvement Center, Dept of Horticultural Sciences, Texas A&M
Univ., College Station, TX 77843 Direct inquiries to author Yoo (E-mail: k-yoo@
tamu.edu).
existed among pyruvic acid test results from various laboratories
of the same source of onions (Havey and others 2002) Previously,our automated method measured approximately 17% higherlevels of pyruvic acid than did the manual spectrophotometricmethod when undiluted juice samples were used (Yoo and others2011) The absorbency of the color adduct rapidly decreased inundiluted onion juice samples as compared to that of the pyruvicacid standards, therefore resulting in underestimations of pyruvicacid concentrations (Yoo and others 2011) In that study, we havemodified the sample processing order and substantially improvedthe consistency by promptly measuring the absorbency at 1 minafter the color development
DNPH has long been used for detection of pyruvic acid andother carbonyl compounds and others by reacting nonselectivelywith carbonyl compounds to form the corresponding hydrazone( Jones and others 1961; Katsuki and others 1961; Schwimmer andWestern 1961; Critchfield 1963) Pyruvic acid andα-ketoglutaric
acid (Katsuki and others 1961), dicarbonyl compounds (Jones andothers 1961), and several aldehydes, such as n-octanal found incitrus (Peleg and Mannheim 1970), have been measured using thismethod In theory, slight overestimations of pyruvic acid contentsare likely due to the coexistence of related carbonyls in onions.Therefore, a high-performance liquid chromatography (HPLC)method was used to very accurately determine the pyruvic acidcontent in onions (Yoo and Pike 2001)
Trang 38The addition of NaOH or KOH to pyruvic acid-DNPH
reactions following heating at 37◦C results in a pink color adduct,
the absorbency of which can be measured spectrophotometrically
(Schwimmer and Western 1961) However, degradation of the
color adduct over time has not been well documented Measuring
the absorbency within 10 min of adding NaOH was
recom-mended and the degradation of DNPH by oxidizing agents was
indicated Ozone can destroy DNPH alone or following formation
of dinitrophenyl- hydrazones during determination of aldehydes
and ketones in the air (Achatz and others 1999) However, further
degradation of the pyruvic acid-DNPH adduct was not
exam-ined (Critchfield 1963) Anthon and Barret (2003) indicated a
2.5% reduction of the color adduct in 1 h and suggested prompt
absorbency measurements We have documented that about 20%
of absorbency was reduced in 60 min, regardless of bulb color
(Yoo and others 2011) However, the cause of the degradation
was not investigated or reported yet to our knowledge
Pyruvic acid is known to be degraded by polymerization and
decomposition in solution (The Merck Index 1991), but this slow
degradation was not noticeable within several hours and was not
regarded as a factor in this study Rather, we were interested in
other factors in onion juice that caused rapid degradation of
pyru-vic acid Only other factor causing degradation of pyrupyru-vic acid was
found to be cysteine that induced coupled oxidation of pyruvic
acid when cysteine was oxidized (Cavallini 1951)
As a continuation of our previous study of observing the rapid
degradation of the color adduct, we initiated this study to elucidate
the cause of color adduct degradation from the reaction of pyruvic
acid and DNPH used in onion pungency tests The reduction of
pyruvic acid concentrations through the addition of cysteine prior
to the reaction with DNPH was also investigated Spectral analysis
of these reactants was performed to confirm the changes in the
color adduct
Materials and Methods
Chemicals and plant materials
All chemicals, unless otherwise stated, were purchased from
Sigma (St Louis, Mo., U.S.A.) NaOH was purchased from Fisher
Scientific (Pittsburgh, Pa., U.S.A.) Red, yellow, and white onion
bulbs of unknown cultivars were purchased from a local store
Onion juice preparation
Fresh juice. The neck, base, and dry skin were removed from
the onion bulbs Bulbs were blended in a mixer for 2 min without
the addition of water The puree was incubated for 30 min at 24◦C
of room temperature to complete the alliinase reaction and filtered
through P-1 filters (Fisher Scientific) The juice was placed in a
50-mL tube and stored at−20◦C Frozen samples were defrosted
and centrifuged at 804 g for 30 min to remove particles Onion
juices, unless specified as cooked, refer to the fresh juice in this
study
Cooked juice. The white onion bulb were microwaved for
the rate of 1 min per 100 g tissue, blended, and filtered through a
cheese cloth Filtrates were centrifuged, filtered through P-1 filter
papers, and kept at−20◦C.
Production of color adduct in the pyruvic acid analysis
We used our previously published pyruvic acid analysis method
to examine undiluted onion juice (Yoo and others 1995) In
this study, the volume of each reactant was reduced to a third
of the originally published volumes to accommodate the use of
1 × 1 × 4.5 cm plastic cuvettes Twenty microliters of luted onion juice, pyruvic acid standard, or other mixtures werecombined in cuvettes with 1-mL DNPH (125 mg/L in 2N HCl)solution using an autodispenser (Hamilton Microlab 500 series,Reno, Nev., U.S.A.) The mixture was incubated in a 37◦C waterbath for 10 min, and 2-mL NaOH (1.0 N) was added with anEppendorf repetitive pipette Absorbencies at 485 nm weremeasured using a spectrophotometer (Model 601, Milton Roy,Rochester, N.Y., U.S.A.)
undi-Search for factors affecting color adduct degradation
Comparison of fresh and cooked onion juices. Since thealliinase reaction produces many changes in the composition ofonion juice, juice with or without the alliinase reaction was com-pared A white onion bulb (approximately 500 g) was cut in halflongitudinally Half was blended and processed as described in thefresh juice section The other half was microwaved for 2.5 minand processed as described in the cooked juice section
Because the cooked juice contained only approximately0.2 mM pyruvic acid (Yoo and Pike 2001), 38.5 mg sodiumpyruvic acid was dissolved in 50-mL cooked juice in order toobtain pyruvic acid concentrations similar to the 7 mM pyruvicacid standard and fresh juice The final pyruvic acid concentration
in the cooked juice was not measured, but was estimated to bearound 7.2 mM Fresh onion juice had pyruvic acid levels ofapproximately 6 to 7 mM as determined in a separate test Thecooked- and fresh-juice samples and the 7 mM pyruvic acid stan-dard were treated with DNPH as described above Absorbencieswere monitored for 120 min, and relative absorbencies (%, 100×absorbency at a given time/initial absorbency) were calculated
Comparison between bulb colors. Differences in the rates
of color adduct degradation were examined between red, low, and white onions because yellow and red onions containantioxidant compounds, such as quercetin and anthocyanins (Kimand others 2004) Fresh red, yellow, and white onion juices wereassayed as above, and absorbencies were monitored for 60 min.Relative absorbencies were calculated The pyruvic acid concen-trations of the red, yellow, and white onions were 11.8, 3.7, and6.4 mM, respectively
yel-Screening of candidate compounds. Several compoundsthat naturally exist in onion tissues were evaluated for their role
in color adduct degradation A 7 mM pyruvic acid solution wasprepared and 50 mg/mL of ascorbic acid, amino acids (proline,cysteine, and glutamic acid), or sugars (glucose, fructose, andsucrose) were individually dissolved in 10 mL of the pyruvic acidsolution The duplicated samples were analyzed as above, and ab-sorbencies were monitored for 60 min Actual absorbencies arepresented to more clearly represent the reduction of color adductdue to cysteine
The effect of fructose on color adduct degradation
Addition of fructose to various juice mixtures. Solutions
of 7 mM pyruvic acid, fresh or cooked white onion juice, andfresh or cooked white onion juice supplemented with 50 mg/mLfructose were prepared Samples were treated with DNPH, ab-sorbencies were measured for 60 min, and relative absorbencieswere calculated
Fructose concentrations. Solutions of 7 mM pyruvic acidsupplemented with 0 (control), 10, 30, 50, or 70 mg/mL fructosewere prepared Samples were treated with DNPH, absorbencieswere measured over 60 min, and relative absorbencies werecalculated
Trang 39Comparison of different sugars. The 7 mM pyruvic acid
solution was supplemented with 50 mg/mL fructose, glucose, or
sucrose The control had no sugar added (pyruvic acid alone)
Samples were treated with DNPH, absorbencies were measured
for 60 min, and relative absorbencies were calculated Samples
were prepared in duplicate
The effect of cysteine on absorbencies
Color adduct degradation and cysteine concentrations.
A 7 mM pyruvic acid solution was supplemented with 0, 5, 10,
or 15 mg/mL cysteine Samples were treated with DNPH,
ab-sorbencies were measured for 60 min, and relative abab-sorbencies
were calculated Samples were prepared in duplicate
HPLC measurement of pyruvic acid. The HPLC system
included a binary pump, an autosampler, and a UV/Vis
detec-tor (Perkin Elmer LC 200 Series, Norwalk, Conn., U.S.A.) An
Alltech IOA 100 organic acid column (Deerfield, Ill., U.S.A.,
7.8× 300 mm2) with a guard cartridge was used to separate
pyru-vic acid (Yoo and Pike 2001) The solvent was water containing
0.3-mL sulfuric acid/L at a rate of 0.5-mL/min and a 20-μL
sample was injected The column temperature was 45◦C, and
de-tection was made at 210 nm External standard of sodium pyruvic
acid ranging from 0 to 5 mM was prepared for a standard curve
It took 30 min to run a sample
A 7 mM pyruvic acid solution was supplemented with
10 mg/mL cysteine and the changes in pyruvic
acidconcentra-tions were measured at 0, 1, 31, 61, and 720 min The peak area at
0 min was measured prior to the addition of cysteine The pyruvic
acid+ cysteine samples were injected into the HPLC system and
peak areas were recorded Relative pyruvic acid concentrations
(%, 100× peak area at a given time/initial peak area) and
regres-sion coefficients were calculated
Spectral analysis of the color adduct
The color adduct developed during the pyruvic acid test
method was placed in a plastic cuvette and the spectrum between
340 and 700 nm was recorded by a spectrophotometer (Model
U-2900, Hitach High-Tech, Schaumburg, Ill., U.S.A.) The
ref-erence control, containing water rather than pyruvic acid, was
used as the background measurement
Comparison of onion juice and the pyruvic acid
standard. Fresh yellow onion juice and 7 mM pyruvic acid
stan-dards were analyzed as described above Their spectra were plotted
on the actual scale and were normalized to overlap the spectra The
λmax was determined and the reference wavelengths used in the
previous papers are indicated
Changes of spectra in the standard and
fructose-supplemented samples. The 7 mM pyruvic acid standard,
cooked white onion juice fortified with 7 mM pyruvic acid
and 50 mg/mL fructose, and 7 mM pyruvic acid fortified with
70 mg/mL fructose were treated with DNPH and the color adduct
developed The spectral data from each sample were collected at
0, 10, and 20 min
Results and Discussion
Effects of alliinase activity and onion color on the
reduction of the color adduct
The color adduct resulting from the pyruvic acid standard was
fairly stable and degraded little over 120 min (Figure 1, left)
How-ever, the color adducts that developed in the reactions containing
fresh juice and cooked juice supplemented with 7 mM pyruvic
acid decreased rapidly The cooked juice samples lost slightly morecolor than the fresh juice samples after 2 h There were no differ-ences in the rates of color reduction in the reactions containingjuice from the different colored onions (Figure 1, right).These results indicated that there was no difference in the coloradduct produced by samples with active or denatured alliinase.Therefore, another factor caused the color adduct degradation inthe onion juice itself Since cooking denatured the alliinase, whichcatalyzes the 1-propenyl-cysteine-sulfoxide and methyl-cysteine-sulfoxide reactions that produce numerous sulfur volatiles and re-lated compounds (Lancaster and Boland 1990), any product fromthose hydrolysis reactions were excluded from the candidate com-pounds There was no retardation or acceleration of the coloradduct degradation by quercetin and anthocyanin compounds,compounds that determine onion bulb colors (Kim and others2004) Red onions contain anthocyanins and quercetin, yellowonions contain quercetin, and white onions do not contain eithertype of compound The onions examined had different colors andpungency levels (3.7 to 11.8 mM), but showed very similar patterns
of reduction Therefore, neither bulb color nor pungency leveleffected the reduction rate of the color adduct Additionally, thio-sulfinates and other sulfur compounds were excluded by this study.Effect of amino acids, sugars, and ascorbic acid on thereduction of the color adduct
The effect of amino acids, sugars, and ascorbic acid on therates of color adduct degradation was examined A 7 mM pyru-vic acid sample was supplemented with 50 mg/mL of proline,cysteine, glutamic acid, ascorbic acid, or fructose and subjected
to the DNPH reaction All samples, including the pyruvic acidcontrol, showed rapid reduction of absorbency between 0 and 5min following the addition of NaOH This reduction was possiblydue to a mixing problem or to an unknown mechanism of initialdegradation (Figure 2) However, there was little linear reduction
of absorbencies between 5 and 60 min, with a total reduction ofless than 6% in 30 min, in all samples except for the fructose-supplemented sample The sample supplemented with fructoseshowed a more rapid reduction in absorbencies over time There-fore, we further studied the effects of fructose on the degradation
of the color adduct Interestingly, cysteine showed a much-reducedinitial absorbency, as compared with the other treatments How-ever, the rate of degradation over time was similar to the othertreatments We suspect that cysteine has some reaction with pyru-vic acid even before the initiation of the DNPH reaction Ascorbicacid at a concentration of 50 mg/mL induced a reduction by 5min, but was not further studied because of its naturally low con-centration in onion (approximately 50μg/mL).
Effect of fructose on the reduction of the color adduct
Addition of fructose to fresh or cooked onion juice. Theabsorbencies of the cooked juice supplemented with 7 mM pyru-vic acid and the fresh juice were reduced by about 20% and 30%after 60 min, respectively (Figure 3) The fresh and cooked juicessupplemented with 50 mg/mL fructose lost about 60% of theirinitial absorbencies in 60 min However, the absorbency of thepyruvic acid alone did not change much There was slightly morereduction of the color adduct (10%) in the cooked juice supple-mented with pyruvic acid than in the fresh juice sample, thoughfructose content was expected to be the same in both samples.The samples supplemented with fructose lost the color adduct
at equal rates These results indicated that higher concentrations
of fructose correlated with higher rates of absorbency reduction
Trang 40It was estimated that the shortday onion “Texas Grano 1015Y”
contains approximately 20 mg/mL fructose, 23.6 mg/mL glucose,
and 3 mg/mL sucrose (Hamilton and others 1997) Therefore, the
fresh and cooked onion juices used in this study likely contained
about 20 mg/mL fructose and the fructose-fortified juice samples
contained about 70 mg/mL fructose The rates of reduction of
the color adduct absorbencies positively correlated with the
con-centration of fructose in the test samples, thereby demonstrating
that fructose caused the reductions of the color adduct As
com-pared with the Schwimmer and Western method (using 1 mL of
200-fold diluted juice), our method (using 80μL of undiluted
juice) was calculated to contain 16 times more juice (or fructose)
As a result, more rapid absorbency reduction was thought to be
possible
Comparison of fructose concentrations and other
sugars. Pyruvic acid supplemented with 30 or 70 mg/mL
fruc-tose showed 10% and 60% reductions in absorbency, respectively
(Figure 4, left) Color adduct degradation was proportional to
the fructose concentrations As compared to sucrose and glucose,
fructose showed the greatest effect on the absorbencies of the
color adduct (Figure 4, right) There was a 4% to 6% and 7%
to 12% reduction at 30 and 60 min in pyruvic acid samples
supplemented with 50 mg/mL glucose and fructose, respectively
Fructose showed very rapid color reduction of 12% and 38% by
30 and 60 min, respectively
The addition of fructose up to 70 mg/mL to the 7 mM pyruvic
acid sample did not increase the actual absorbency at 485 nm, or
the addition of 50 mg/mL sucrose or glucose, or fructose increased
the absorbency by only 0.010, 0.019, or 0.034 units, respectively
(data not shown) The absorbency of the 7 mM pyruvic acid
was about 0.480 Therefore, the increases in absorbency caused
by sucrose, glucose, and fructose at 50 mg/mL (278 mM) were
insignificant or negligible
Fructose and glucose have a ketone and an aldehyde group,
respectively, and are reducing sugars (The Merck Index 1991)
However, fructose caused an approximately 3-fold greater
degra-dation of the color adduct than did glucose The ketone or
alde-hyde groups in fructose and glucose were believed not to react
with DNPH, because no significant increase in absorbency was
observed Fructose and glucose in solution are mostly in the cyclic
forms of pyranose and furanose (Hyv ¨onen and others 1977; Flood
and others 1996) Therefore, the keto and aldehyde groups in
the acyclic forms were incorporated into the ring structures and
Time after adding NaOH (min)
Control Proline Ascorbic Cysteine Glutamic Fructose
Figure 2–Changes in absorbency measurements of the DNPH-pyruvic acid color adduct in samples of 7 mM pyruvic acid supplemented with
50 mg/mL proline, ascorbic acid, cysteine, glutamic acid, or fructose over
7 mM pyruvate Fresh juice Fresh juice + fructose Cooked juice + pyruvate Cooked juice + pyruvate + fructose
Figure 3–Changes in the relative absorbencies of the DNPH-pyruvic acid color adduct in 7 mM pyruvic acid, fresh or cooked onion juice, and fresh or cooked onion juice supplemented with 50 mg/mL fructose over 60 min.
Time after adding NaOH (min)
Time after adding NaOH (min)
84 86 88 90 92 94 96 98 100
7 mM pyruvate Yellow White Red
Figure 1–Reduction of the relative absorbencies of the DNPH-pyruvic acid color adduct following the addition of NaOH in a 7 mM pyruvic acid standard, fresh onion juice, and cooked onion juice supplemented with 7 mM pyruvate over 120 min (left), and juice samples from yellow, white, or red onions over 60 min (right).