LWT - Food Science and Technology 82 (2017) 184e191

Contents lists available at ScienceDirect

LWT - Food Science and Technology

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

Determination of potential off-flavour in yeast extract

Ya Zhang a, Huanlu Song a, *, Pei Li b, Juan Yao b, Jian Xiong b
a
Beijing Innovation Centre of Food Nutrition and Human Health, Laboratory of Molecular Sensory Science, Beijing Technology and Business University, No.
11, Fucheng Road, Haidian District, 100048, Beijing, China
b
Yeast Extract Seasoning Division, Angel Yeast Co. Ltd, Yichang, Hubei Province, China

a r t i c l e i n f o a b s t r a c t

Article history: Yeast extracts is a kind of food ingredient for enhancing flavor, but some yeast extract samples had yeasty
Received 14 September 2016 odour (off-flavour). The aim of this study was to detect key aroma compounds leading to the yeasty
Received in revised form odour. Volatile compounds of 3 yeast extracts (with yeasty odour), FA and FB, LC were extracted by solid
21 February 2017
phase micro-extraction (SPME) and solvent assisted flavour evaporation (SAFE) respectively, and ana-
Accepted 10 April 2017
Available online 15 April 2017
lysed by gas chromatography-olfactrometry-mass spectrometry (GC-O-MS) and aroma extract dilution
analysis (AEDA) techniques. Thirty aroma-active compounds were detected, among them 4-
methylphenol, 3-methylpyridine, 3-methylbutanoic acid, propanoic acid with high FD factors
Keywords:
Off-flavour
(FD ¼ 125, 125, 625, 125, 25), were identified as the key off-flavours responsible for yeasty and their
Gas chromatography-olfactrometry-mass accurate quantification were done by external standard method in SIM mode. Calculation of the OAVs
spectrometry (GC-O-MS) revealed 3-methylbutanoic acid, 3-methylpyridine, 4-methylphenol as the key odourants in the extract.
Aroma extract dilution analysis (AEDA) For the verification of analyzed off-flavours, sensory evaluations of aroma model were conducted by
Odour activity value (OAV) eight panelists, and the result showed that aroma recombinate had the yeasty odour profile similar with
Aroma model the original yeast extracts.
The methods were successful for the detection of off-flavors in yeast extract and the four off-
compounds analyzed were mainly responsible for the yeasty flavour.
© 2017 Elsevier Ltd. All rights reserved.

1. Introduction (which normally had basic and characteristic meaty or caramel

Yeast extracts were made from baker's yeast (Verduyn,
Suksomcheep & Suphantharika, 1999), brewer's yeast, torula
yeast or kluyveromyces by autolysis process produced under
controlled conditions (Festring & Hofmann, 2010; Moresi, Orban,
Quaglia, & Casini, 1995), the degradation of intracellular nucleic
acid, protein and other macromolecular compounds produced the
tasty substances such as amino acids and nucleotides (Chae, Joo, &
In, 2001), the yeast extract products were usually in the forms of
liquid, paste or powder. Yeast extracts were excellent and natural
seasoning, widely used as ingredients for the production of savoury
foods (Ikeda, 2002).
The ‘yeast flavour’, which was formed during products pro-
cessing, affected the overall aroma quality of yeast extract products
Abbreviations: YE, yeast extract; SPME, solid phase micro-extraction; SAFE,
solvent assisted flavour evaporation; GC-O-MS, gas chromatograph-olfactometry-
mass spectrometry; AEDA, aroma extract dilution analysis; FD, flavour dilution;
SIM, selected ion monitoring; OAV, odour activity value.
* Corresponding author.
E-mail address: songhl@th.btbu.edu.cn (H. Song).
notes), and was therefore sometimes the focus of consumers
complaints. Although a study conducted by Lin et al. had reported
aroma components that included off-flavour of yeast extracts (Lin
et al., 2014), the specific compounds that produce yeasty odour
are rarely the focus of studies. It is particularly important and ur-
gent to identify the cause of the off-flavour that is always described
as yeasty in yeast extract products. However, the intensity of the
off-flavour and the overall aroma profile of yeast extract samples
varied with different processing techniques.
Until now, only a range of studies on yeast extracts are available
that identify with different aroma characteristics that enhance the
food flavour, but the yeasty odour in samples has barely been
investigated. Münch, Hofmann and Schieberle et al. (Münch,
Hofmann & Schieberle, 1997) identified 2-furanmethanethiol, 4-
hydroxy-2,5-dimethyl-3(2H)-furanone as the key odorants from
the intensely roasted, sweet smelling solution of yeast extract.
Mahadevan and Farmer (2006) compared three types of yeast
extract pastes from two different suppliers and the key odorants
such as 2-methyl-3-furanthiol, 2-methyl-3-methyldithiofuran,
methional, 1-octen-3-one were identified as the cause for the
meaty and roasted note of the samples. Werkhoff et al. reported

http://dx.doi.org/10.1016/j.lwt.2017.04.030
0023-6438/© 2017 Elsevier Ltd. All rights reserved.
 Y. Zhang et al. / LWT - Food Science and Technology 82 (2017) 184e191
sulfur-containing volatiles generated by thermal treatment of yeast
extract (Werkhoff, Bruening, Emberger, Guentert, Koepsel & Kuhn,
1990). Ames and MacLeod identified some aroma compounds with
meaty odour properties from yeast extract (Ames & MacLeod,
1984), and Grosch (2001) investigated amino-carbonyl in-
teractions and their possible contribution to the overall odor profile
of yeast extract.
The aim of this study was to identify the key aroma compounds
responsible for the ‘yeasty’ odour in yeast extract samples then
verified the accuracy of detected compounds by the sensory eval-
uation of an aroma model. The aroma model was made by
combining authentic standard compounds according to their nat-
ural concentrations present in the yeast extracts (Lawless &
Heymann, 2010). The quantitative analysis of key off flavours
compounds was completed using an external standard method in
SIM (selected ion monitoring) mode (Steingass, Langen, Carle &
Schmarr, 2015; Hoscheid, Reinas, Cortez, Costa, & Cardoso, 2012).
2. Materials and methods
2.1. Samples
Three kinds of yeast extracts were provided by Angel Yeast Co.
Ltd. (Yichang, Hubei Province, China) and each extract had been the
focus of complaints about off-flavour (yeasty) properties, in com-
parison with other yeast extract products. The yeast extract powers
FA and FB were prepared by the enzymatic hydrolysis of yeast cells
with protease/RNase, thermal-inactivation of enzymes, centrifugal
separation, vacuum concentration and spray drying. The yeast
extract LC was prepared by the same process except no vacuum
concentration was used.
2.2. Chemicals
Diethyl ether (AR), n-pentane (AR), n-hexane (AR), anhydrous
sodium sulfate (AR), sodium chloride (AR) were purchased from
Yifengtiancheng Scientific Instruments Co. Ltd. (Beijing, China). n-
Alkanes (C7eC22) (chromatographic reagent) and 2-methyl-3-
heptanone (chromatographic reagent) were purchased from Sig-
maeAldrich (St Louis, MO, USA). Nitrogen (99.995% purity) was
supplied by Beijing Haipu Beifen Gas Industry Co. Ltd. (Beijing,
China).
Distilled water was purified by the Millipore Simplicity system
(Millipore Corp., Saint-Quentin, France) to obtain the ultrapure
water. Standards of volatile compounds such as 3-methylbutanal,
3-methylthiophene, 2-methylthiazole, methylpyrazine, octanal,
2,5-dimethylpyrazine, 2,6-dimethylpyrazine, ethylpyrazine, 2,3-
dimethylpyrazine, dimethyltrisulfide, 2-methyl-3-furanthiol, 3-
(methylthio)propanal, nonanal, trimethylpyrazine, benzeneace-
taldehyde, 2-methyl-5-(1-methylethyl)pyrazine, acetic acid,
furfural, 2,6-diethylpyrazine, benzaldehyde, propanoic acid, 2,3-
butanediol, 2-methylpropanoic acid, acetophenone, 3-
methylbutanoic acid, 3-methylpyridine, phenylethyl alcohol, 4-
methylphenol, and indole were purchased from SigmaeAldrich
Co. Ltd (Santa Clara, CA, USA).
2.3. Quantification of volatile compounds
The extraction/isolation of volatile substances was carried out
by solid phase micro-extraction (SPME) and solvent assisted flavour
evaporation (SAFE), respectively. 2-Methyl-3-heptanone was used
as an internal standard because it eluted quickly (Song &
Cadwallader, 2008), without interfering with the target com-
pounds. For SPME and SAFE analysis, the content of internal stan-
dards as follows.
185
In this study, two quantitative methods were used, including an
internal standard method for total aroma compounds and an
external standard method in SIM mode for the four off-flavour
compounds detected. For the internal standard method, the
authentic standards listed in Table 1 were mixed well, then 2-
methyl-3-heptanone (0.816 mg/mL in hexane) was added as the
internal standard.
Response factor ¼ (concentration of authentic standard/peak area
of standard)/(concentration of internal standard/peak area of the
internal standard)
Concentration of an odorant ¼ concentration of internal
standard  peak area of compound/(peak area of the internal
standard/response factor).
For the external standard method in SIM mode, four standard
solutions of 3-methylbutanoic acid, 3-methylpyridine, 4-
methylphenol, and propanoic acid compounds were prepared at
five levels: 50, 100, 200, 500 and 1000 mg/L in selected ion moni-
toring mode according to the mass spectra SIM modes of m/z 60, 41,
93, 66, 107, 108 and 74, 73, respectively.
2.4. Solid phase micro-extraction (SPME)
SPME analytical method was applied in previous research (Lin,
Wong, & Kao, 2002). Yeast extract powder (4 g) was placed in a
40 mL headspace vial, then 1 mL internal standard 2-methyl-3-
heptanone at the concentration of 0.816 mg/mL was added, and
thermostated at 61 Cin a water bath for 20 min. The SPME needle
coating divinylbenzene/carboxen/polydimethylsiloxane (75 mm)
was inserted into the vial and adsorbed for 44 min, then desorbed
at the gas chromatography inlet for 5 min before the GC-O-MS
analysis.
2.5. Solvent assisted flavor evaporation (SAFE)
Five grams of the each of yeast extracts was weighed accurately,
dissolved in 100 mL of ultrapure water, after which the internal
standard compound 2-methyl-3-heptanone was added at a con-
centration of 0.816 g/mL. The YE sample was mixed with 120 mL
combined solvent of diethyl ether and pentane (v:v ¼ 2:1) and was
extracted repeatedly three times, producing a total of 220 mL
mixture. It was put on shaking plate, rotating at a speed of 180r/min
at 4  C for 8 h. The organic phase of 100 mL was obtained using a
separatory funnel after static layering. The installation of the SAFE
apparatus (Deutsche Forschungsanstalt für Lebensmittelchemie,
Freising, Germany) and operation procedure were as follows: a
500 mL round bottom flask was put on a thermostatic water bath
pot of 40  C, another 250 mL round bottom flask used as a receiver
was put into a liquid nitrogen environment; the cold trap was also
filled with liquid nitrogen, the distilled water temperature was held
at 50  C, and the pressure under vacuum was 104Pa by a combi-
nation of a high vacuum pump and a turbo pump (Edwards, UK).
The 100 mL volume of organic phase extracted by ether/n-pentane
was dropped into the distillation flask slowly and uniformly. The
extracts were separated from organic solvent, then dried by
anhydrous sodium sulfate two times, and concentrated to 2 mL by
Vigreux column under 40  C, and finally 0.5 mL concentrate was
obtained by a gentle purge of nitrogen. All extracts being free of
water and impurities were subsequently stored at 20  C in a 2 mL
glass vial equipped with Teflon-lined cap before undergoing GC-O-
MS analysis.
186 Y. Zhang et al. / LWT - Food Science and Technology 82 (2017) 184e191

Table 1
Aroma-active compounds of three yeasty yeast extracts samples by AEDA.

Nr Compound namea Odor propertyb Retention Retention index Relative Method of FD factord
time(min)e concentration(mg g1)c identification

DB-WAX DB-5 LC FA FB

1 3-Methylbutanoic acid Cheesy, rancid 24.993 1573 755 3.65 1.45 1.32 LRI,O,MS,STD 625
2 2-Methyl-3-furanthiol Meaty 17.552 1211 870 1.37 1.97 LRI,O,MS,STD 625
3 4-Methylphenol Medicine 33.408 1986 1075 1.21 1.33 1.24 LRI,O,MS,STD 125
4 3-Methylpyridine Medicine, bitter 16.856 1236 723 3.94 9.55 2.76 LRI,O,MS,STD 125
5 Propanoic acid Rancid, pungent 21.899 1441 660 3.59 2.67 6.81 LRI,O,MS,STD 125
6 3-Methylbutanal Dark chocolate 5.815 921 <500 67.77 28.58 31.32 LRI,O,MS,STD 125
7 3-(Methylthio)propanal Cooked potato 19.87 1354 902 3.43 LRI,O,MS,STD 125
8 Benzeneacetaldehyde Honey, sweet 25.586 1600 1043 4.14 14.55 LRI,O,MS,STD 125
9 Ethylpyrazine Popcorn, peanut butter 16.89 1241 820 1.13 0.89 1.93 LRI,O,MS,STD 125
10 3-Methylthiophene Metal 9.953 1044 1.84 1.75 0.26 LRI,O,MS,STD 25
11 2-Methylthiazole Roast, smoky 14.25 1144 1.11 1.34 LRI,O,MS,STD 5
12 Methylpyrazine Popcorn 15.099 1174 707 1.85 1.46 4.27 LRI,O,MS,STD 25
13 Octanal Green, fruity 15.566 1191 1001 1.10 4.11 LRI,O,MS,STD 25
14 2,5-Dimethylpyrazine Roasted nut 16.553 1230 809 5.16 3.96 12.92 LRI,O,MS,STD 25
15 2,6-Dimethylpyrazine Roasted nut, cocoa 16.724 1237 812 3.54 2.20 4.82 LRI,O,MS,STD 25
16 2,3-Dimethylpyrazine Nutty, peanut butter 17.202 1255 836 0.39 0.52 0.95 LRI,O,MS,STD 25
17 Dimethyl trisulfide Cooked cabbage, garlic 17.994 1283 976 0.48 LRI,O,MS,STD 25
18 Nonanal Green 18.27 1292 1103 3.23 LRI,O,MS,STD 5
19 Trimethylpyrazine Roast, nutty 18.771 1313 904 4.30 2.61 5.87 LRI,O,MS,STD 25
20 2-Methyl-5-(1-methylethyl)pyrazine Sweat 18.912 1318 1053 5.09 0.64 LRI,O,MS,STD 5
21 Acetic acid Sour 19.727 1342 <500 4.98 8.77 LRI,O,MS,STD 25
22 Furfural Bread, sweet 19.472 1366 829 0.55 LRI,O,MS,STD 25
23 2,6-Diethylpyrazine Roast, buttery 20.094 1370 1.81 2.20 1.72 LRI,MS,O,STD 5
24 Benzaldehyde Almond 21.645 1428 960 20.15 53.33 91.17 LRI,O,MS,STD 25
25 2,3-Butanediol Buttery, cheesy 21.985 1444 4.35 2.63 LRI,MS,STD 25
26 2-Methylpropanoic acid Butter, cheese 22.648 1471 1106 1.80 3.29 LRI,O,MS,STD 5
27 Acetophenone Floral 24.647 1558 1041 6.99 13.33 LRI,MS,STD 5
28 Phenylethyl Alcohol Rosy, floral 30.191 1819 1113 1.29 3.53 8.06 LRI,O,MS,STD 1
29 1-(1H-pyrrol-2-yl)ethanone Nutty, popcorn 31.36 1879 961 1.32 0.80 1.39 LRI,O,MS 25
30 Indole Camphor 39.058 2357 1294 2.62 9.45 17.49 LRI,O,MS,STD 5
a
Every compound was identified using the two methods meanwhile of SPME and SAFE by comparing it with an authentic standard except 1-(1H-pyrrol-2-yl)ethanone
based on the following criteria: (1) having same retention time on the same column under the same conditions; (2) descriptions of its aroma property; (3) matching mass
spectrum.
b
Odor description at the olfactory detection port (Sniffer 9000; Brechbühler, Switzerland).
c
Results are the means of three repetitions as mg g1 using internal standard quantitative method.
d
FD: Flavor dilution factor. The concentrated yeast extracts by SAFE was diluted by ethyl ether. The ratios were 1:5, 1:25, 1:125, and 1:625, corresponding to FD factors of 5,
25, 125, and 625, respectively.
e
Retention time: The time that different volatile compounds elute from the chromatography column.

2.6. Gas chromatography-olfactometry-mass spectrometry (GC-O-
MS) analysis
The detection of yeast extract volatiles was performed on a
7890A gas chromatograph coupled to a Triple Quad 7000B (both
Agilent, Palo Alto, CA, USA), and connected a Sniffer 9000 Olfac-
tometer (Brechbühler, Schlieren, Switzerland). The aroma compo-
nents were separated on two different polar columns of DB-WAX
and DB-5 (both were 30 m  0.25 mm  0.25 mm thickness; J & W
Scientific, Folsom, CA, USA). GC-O-MS was conducted on both col-
umns. The flow rate of ultra-high purity helium used as carrier gas
was 1.2 mL min1 with. The oven temperature was held at 40  C for
3 min, which was then increased to 200  C at a rate of 5  C min1
and then increased to 230  C at 10  C min1 with a final hold at
230  C for 3 min. The concentrated aroma components were des-
orbed in the GC injector at 250  C for 5 min in splitless mode. The
effluent of GC was split 1:1 between MSD and olfactometer though
a triple valve. Electron-impact mass spectra were generated at
70 eV, with an m/z scan range from 45 to 450 amu. The ion source
and interface temperature were 230  C and 280  C, respectively.
The separated volatile compounds all run though Sniffer 9000
coupled with wet gas (high purity nitrogen and distilled water) to
reach the nose, to improve the sensitivity and comfort of the pan-
elists during the olfactometry process (sniffing).
2.7. Aroma extract dilution analysis (AEDA)
The AEDA method was applied to compare the flavour dilution
(FD) factors which determined the key aroma compounds (Guth &
Grosch, 1999). As values of FD factors increased, the more critical
the aroma compounds became. For AEDA, the concentrated yeast
extract volatiles were stepwise diluted with diethyl ether to 1:5,
1:25, 1:125, and 1:625 of the original extracts (Fickert & Schieberle,
1998). The original and diluted extracts were detected by GC-O-MS
until no odour could be detected on both the DB-5 and DB-Wax
columns. The maximum dilution factor of aroma compounds was
the last diluted multiple that could be detected with GC-O-MS.
2.8. Compounds identification
The volatile compounds were identified by mass spectrometry
compared with NIST Mass Spectral Library 05 Vision and linear
retention indices (LRI) with reference compounds. n-Alkanes (C7-
C30) were analysed under the same conditions on two different
polar columns, DB-WAX and DB-5. All aroma compounds except 1-
(1H-pyrrol-2-yl)ethanone were confirmed by the injection of
standard compounds into the GC-MS detector. LRI was calculated
using the following equation: LRI ¼ 100N þ 100n{[tRa  tRN]/
[(tR(Nþn)  tRN]}, where N represents the number of carbon atoms
with peaks in front of the identified compounds, and n represents
 Y. Zhang et al. / LWT - Food Science and Technology 82 (2017) 184e191
the numerical difference of upper and lower n-alkanes in gas
chromatography. The variables tRa, tRN, tR (Nþn) denote the retention
time of the identified compounds, and the upper alkane and lower
alkanes, respectively (Lin et al., 2014).
2.9. Sensory evaluation
Sensory evaluation was carried out by eight panelists (three men
and five women, 23e50 years of age) from Laboratory of Molecular
Sensory Science of Beijing Technology and Business University that
have over 400 h of sensory experience and were experienced with
yeast extracts. A specific sensory evaluation of 10-point universal
intensity scale was performed by the panelists (0e2, weaker; 3e4,
weak; 5e6, middle; 7e8, strong; 9e10, stronger). Yeast extracts
evaluations were conducted in triplicate by each panelist. The
odour property of a yeasty note was described as something rancid,
vitamin/medicine-like.
2.10. Aroma model
In order to prove that the detection results of off-flavour sub-
stances were correct, aroma model of synthetic blends was pre-
pared according to the analytical data quantitatively determined by
external standard methods in SIM mode, and their aroma profiles
were compared with those of the original yeast extracts by sensory
evaluation (Grosch, 2001).
3. Results and discussion
3.1. Sensory analysis
The characteristic aroma intensity ratings for the three extracts
samples were drawn on a spider graph using 8 descriptors
(Narasimhan, Chand, & Rajalakshmi, 1992). Lin et al. (2014) re-
ported that aroma intensities of yeast extract pastes were divided
into basic meaty/characteristic meaty, roasted, burnt, sweet, sour,
off-flavour. According to the panelists, the three yeast extracts (FA,
FB, and LC) had different odour characteristics through sensory
evaluations, but all presented with the overall notes of yeasty,
meaty, roasted, sweet, sour, burnt, popcorn and caramel. As shown
in Fig. 1, all of the three samples tested were identified with a
distinct odour of yeasty. The FA sample tended to have more intense
off flavor but with higher scores of meaty and roasted notes. Among
Fig. 1. Aroma profile by sensory evaluations of 3 kinds of yeast extracts with characteristic yeasty odor.
187
the no-yeasty descriptions, notes of caramel, meaty, and roasted
had the highest scores, while sourness and sweetness gave the
lowest scores. Interestingly, the sourness of FB was higher than
those of the other two yeast extracts. The LC sample (yeast extract
paste) had basic meaty and caramel notes, with lower scores of
yeasty odour, owing to the precursor difference and processing
method. The results of instrumental analysis (GC-O-MS) of the
samples were similar with those of the sensory evaluations.
3.2. Identification of the aroma-active compounds
3.2.1. GC split mode
The mode of GC injector was selected for split or splitless, which
showed obvious differences in olfactometer intensity and detection
response. In the split mode, aroma compounds were partially
transferred into the GC column partial according to a certain ratio
(1:1, 1:5, et al.) and were vaporized under high temperature. In the
splitless mode, samples were all transferred into the column before
the first 0.5 min. Fig. 2 shows the TIC diagram by split (1:1) or
splitless mode; the red represents the splitless results, which were
much higher than the split ones (see Figs. 3 and 4).
3.2.2. Determination of key aroma-active compounds
Volatile compounds were identified and quantified in three
yeast extracts (FA, FB, and LC) by linear retention index values on
DB-Wax and DB-5 columns. Although the detection with GC-MS
was coupled with an olfactometer (GC-O-MS), a total of 30 com-
pounds were analysed covering a variety of odor properties, such as
the basic aroma of meaty, roasted, popcorn-like, cooked potato-
like, nutty, buttery, cocoa-like, cheesy, flowery, and green. Most of
these aromas had already been identified by previous studies
(Münch & Schieberle, 1998). Mahadevan and Farmer (2006) re-
ported some aroma-active compounds as key odorants in three
yeast extract pastes, including 2-methyl-3-furanthiol, 2-methyl-3-
methyldithiofuran, methional, 1-octen-3-one, dimethyl trisul-
phide together with a number of pyrazines, thiophenes, and
aliphatic compounds. More importantly, besides the identification
of key odorants, at the same time, in this study the characteristic
off-flavor of medicine, rancid, bitter medicine and pungent notes
were also detected, leading to the yeasty odor of YE samples. The
three yeast extracts (FA, FB, and LC), which had similar odor
properties had the same key aroma-active compounds, though the
detection of all the compounds had same FD factors. The volatile
188 Y. Zhang et al. / LWT - Food Science and Technology 82 (2017) 184e191

Fig. 2. Comparison the abundance of GC splitless (red line) and split (1:1) (green line) mode.

Fig. 3. Mass spectra and structures of the analyzed four key compounds in samples.

compounds with FD factors 25 were 3-methylbutanal (dark
chocolate), methylpyrazine (popcorn), octanal (green, fruity),
2,5(6)-dimethylpyrazine (roasted nut), dimethyl trisulfide (cooked
cabbage, garlic), 3-(methylthio)propanal (cooked potato), benzene
acetaldehyde (honey, sweet), furfural (bread, sweet), 1-(1H-pyrrol-
2-yl)ethanone (nutty, popcorn), 3-methylbutanoic acid (cheesy,
rancid), 4-methylphenol (medicine), 3-methylpyridine (medicine,
bitter), 2,3-butanediol (buttery, cheesy), benzaldehyde (almond),
and acetic acid (sour). Among these compounds, characteristic
meaty and roasted sweet flavour compounds as well as some off-
 Y. Zhang et al. / LWT - Food Science and Technology 82 (2017) 184e191
Fig. 4. Chromatograph peaks of major volatile compounds listed in Table 1.
flavour compounds were included.
The 30 aroma-active compounds listed in Table 1 were identi-
fied from all three YE samples (FA, FB, LC) by the methods of GC-O-
MS and AEDA. Compounds with FD factors 5 had contributed
majorly to the aroma profile of samples. In contrast, compounds
with FD factors <5 were considered to make only minor contri-
butions to the overall aroma. Lin et al. (2014) reported 3-
methylbutanal, 3-butanedione, 2,3,5-trimethyl-pyrazine, acetic
acid ethenyl ester, 2-methyl-5-(methio)-furan, 2-furan meth-
anethiol, 1-octen-3-one, 2-acetyl-1-pyrroline and (E,E)-2,4-
decadienal as key aroma-active compounds, for they were
partially responsible for meaty, roasted and sweet notes in the yeast
extract pastes. The result of Münch et al. (Münch et al., 1997)
showed that methional, 2-acetyl-2-thiazoline, phenylacetic acid,
and 2,3-diethyl-5-methylpyrazine had higher flavor dilution (FD)
factors from a thermally-treated yeast extract autolysate. Consid-
ering of FD factors listed in Table 1, the following compounds with
higher values might have made greater contribution to the whole
aroma profile of the samples, 3-methylbutanoic acid (cheesy,
rancid, FD ¼ 625), 2-methyl-3-furanthiol (meaty, FD ¼ 625), 4-
methylphenol (medicine, FD ¼ 125), 3-methylpyridine (medicine,
bitter, FD ¼ 125), propanoic acid (rancid, pungent, FD ¼ 125), 3-
methylbutanal (dark chocolate, FD ¼ 125), 3-(methylthio)propa-
nal (cooked potato, FD ¼ 125), benzene acetaldehyde (honey,
sweet, FD ¼ 125). It is worth noting that, the key aroma-active
compounds mentioned above can be divided into two types: (1)
The first type contributes the off-flavour, which includes 3-
methylbutanoic acid, 4-methylphenol, 3-methylpyridine, and
propanoic acid which had FD factor of 625 or 125. These four
compounds were identified unequivocally as responsible for the
medicine/rancid/cheesy/pungent properties of samples; (2) The
189
other type of aroma-active compounds were desirable, with meaty,
sweet, nutty and roast odour properties.
The four aroma-active compounds with the highest FD factors
that were responsible for the off-flavour of yeast extracts were
accurately quantified by an external standard in selected ion
monitoring (SIM) mode according to the mass spectra SIM mode
being m/z 60, 41, 93, 66, 107, 108, and 74, 73. The concentration of
the three samples differed from each other. Comparing the con-
centrations obtained from both external and internal standard
methods, some differences were examined. The differences may be
due to influence on concentration when all authentic standards
mixed together when internal standard method was used. Because
of these possible influences, the external standard method was
chosen for the calculation of OAVs (odour activity value). FD-factor
and OAV values were listed in Table 2, including 3-methylbutanoic
acid (cheesy, rancid), 3-methylpyridine (medicine, bitter), 4-
methylphenol (medicine), and propanoic acid (rancid, pungent).
To make a further comparison with the other compounds, olfactory
thresholds in water and in air were also taken as a reference to
calculate the OAVs. In general, odor compound with OAV1
contributed to the overall aroma profile and those with OAV<1
made no contribution. Usually FD-factor and OAV values were fairly
well aligned, but there were some exceptions. For example, in this
study, calculation of the OAVs revealed 3-methylbutanoic acid, 3-
methylpyridine, 4-methylphenol as the key odorants in the yeast
extracts, however, even though FD factor was 125, the OAVs of
propanoic acid were less than 1 with all samples.
The reason for the differences in results mentioned above was
probably that, for the OAVs, the antagonistic and synergistic
interaction effect between the aroma compounds and odour
threshold was not taken into consideration, that is, the threshold of
190 Y. Zhang et al. / LWT - Food Science and Technology 82 (2017) 184e191

each compound in water matrix is not equal to that in true sample

Every compound was identified by comparing it with an authentic standard based on the following criteria: (1) having same retention time on the same column under the same conditions; (2) descriptions of its aroma
factore
matrix. (Rothe, 1976). Thus, if the OAVs of some odorants are less

625
125
125
125
FD
than 1, it is possible that here is a synergistic effect between the
aroma and the experimental compounds. On the other side, in the

107(100), 108(84), 77(29), 79(18), 51(12), 39(11), 53(10), 80(10), 27(8), 90(8)
74(100), 28(99), 29(88), 45(62), 73(61), 27(58), 57(28), 26(17), 30(14), 55(13)
OAV approach, it is assumed that the aroma intensity of each

60(100), 41(33), 43(32), 87(17), 45(16), 27(16), 39(13), 42(7), 29(7), 69(5)
93(100), 66(37), 92(29), 39(28), 65(22), 40(13), 38(9), 63(8), 67(8), 94(8)
compound increases linearly with the increase of its concentration,
and every aroma compound has the same psychometric functions.

FD: Flavor dilution factor. The concentrated yeast extracts was diluted by ethyl ether. The ratios were 1:5, 1:25, 1:125, and 1:625, corresponding to FD factors of 5, 25, 125, and 625, respectively.
However, the relationship between log concentration and resulting
aroma does not appear to be linear but, instead, is an S-shaped
curve, so the values of OAVs are not equal to the same odour in-
tensity (Pang, Hu, Liao, Sun, Zhang & Wu, 2012).

3.2.3. Aroma model

The odor properties of the four off-flavour compounds were
evaluated through an aroma model, in which the concentrations of
four compounds were in accordance with those naturally occurring
in three yeast extracts (LC, FA, and FB) by the external standard
method in SIM mode. The four standard compounds were added to
the solution of water matrix for sensory verification by the panel
members. The odor characters of aqueous solutions of four stan-
dard compounds were always described as rancid and medicine-
like, and the overall aroma profile resembled a yeasty odour.
Mass spectra

The odour characterizations of the three aroma models were

described via sensory evaluation from all eight panelists. Overall,
the odor properties were defined as a combination of rancid,
pungent and medicine-like note, similar to the yeasty extract
EI

Results are the means of three repetitions as mg/L use the external standard method in SIM mode according to the Mass spectra.

samples. Furthermore, the off-flavour intensities were obviously

(mg L1)

higher than those in the actual yeast extracts, but the three aroma
Water
Odor thresholdd

2190
250

model had different predominant odour profiles. Model LC gave the

30
55

high score impression of rancid note. Owing to higher concentra-

(ng L1)

tions of 3-methylpyridine and 4-methylphenol, model FA had a

The threshold values are expressed as ng/L or mg/L and were measured as reported by the reference, respectively.
Sensory and analytical properties of key odor compounds responsible for yeasty-flavor in three yeast extract samples.

higher score of medicine-like odour; model FB showed evidence of

1.5

0.5
Air

some rancid property combined with sourness. However, the most

<1
FB

predominant compounds of the four detected need to be verified by

1
8
2

reconstitution and omission experiments and studies of possible

<1
FA

34
2

synergistic effects (Strube, Buettner & Groetzinger, 2009).

OAV

<1
11
LC

4. Conclusion
Odor description at the olfactory detection port (Sniffer 9000; Brechbühler, Switzerland).
1075 ± 2.08
253 ± 0.11
249 ± 0.02

789 ± 2.31

It was shwon that yeasty flavours were always found in some

yeast extracts in the presence of 3-methylbutanoic acid, 3-
FB
Relative concentration(mg/L)c

methylpyridine, 4-methylphenol and propanoic acid, which were

1020 ± 1.52
1986 ± 1.53

detected by the GC-O-MS technique, and AEDA results showed that

498 ± 0.06

290 ± 0.01

these compounds had high FD factors, 625, 125, 125, 125, respec-
tively. Using an external standard method in SIM mode for the
FA

quantitative analysis of aroma-active compounds, their odor ac-

tivity values (OAVs) were calculated. It was revealed that odorants
0.02f
0.12

0.57
0.2

with higher OAVs were essential for the overall aroma profile, but
±
±
±
±

some compounds with lower OAVs might also be important con-

370
338
140
400
LC

tributors. An aroma model was constructed by blending synthetic

odourants according to their real concentrations in the original
Rancid, pungent
Medicine, bitter
Odor propertyb

Cheesy, rancid

samples. Sensory experiments indicated that the overall odour

properties were the combination of rancid, pungent and medicine-
Medicine

property; (3) matching mass spectrum.

like characteristics, similar to the yeasty extract samples. However,

the more predominant compounds out of four standards analysed
still need to be verified by aroma reconstitution and omission ex-
3-Methylbutanoic acid

periments in the future study.

Values are mean ± SD.
Compound namea

3-Methylpyridine
4-Methylphenol
Propanoic acid

References

Ames, J. M., & MacLeod, G. (1984). Volatile components of a yeast extract compo-
sition. Journal of Food Science, 50, 125e131.
Chae, H. J., Joo, H., & In, M. J. (2001). Utilization of brewer's yeast cells for the
Table 2

production of food-grade yeast extract. Part 1. Effects of different enzymatic

treatments on solid and protein recovery and flavor characteristics. Bioresource
Nr

1
2
3
4
a

e
b

f
c
d

Technology, 76, 253e258.

 Y. Zhang et al. / LWT - Food Science and Technology 82 (2017) 184e191
Festring, D., & Hofmann, T. (2010). Discovery of N2-(1-carboxyethyl)guanosine 5'-
monophosphate as an umami-enhancing Maillard-modified nucleotide in yeast
extracts. Journal of Agricultural and Food Chemistry, 58, 14e22.
Fickert, B., & Schieberle, P. (1998). Identification of the key odorants in barley malt
(caramal) using GC/MS techniques and odor dilution analyses. Nahrung (Food),
42, 371e375.
Grosch, W. (2001). Evaluation of the key odorants of foods by dilution experiments,
aroma models and omission. Chemical Senses, 26, 533e545.
Guth, H., & Grosch, W. (1999). Evaluation of important odorants in foods by dilution
techniques. In R. Teranishi, L. Wick Emily, & I. Hornstein (Eds.), Flavor Chemistry:
30 Years of progress (pp. 377e386). Kluwer Academic/Plenum Publishers.
Hoscheid, J., Reinas, A., Cortez, D. A. G., Costa, W. F., & Cardoso, M. L. C. (2012).
Determination by GC-MS-SIM of furanoditerpenes in Pterodon pubescens
benth.: Development and validation. Talanta, 100, 372e376.
Ikeda, K. (2002). New seasonings. Chemical Senses, 27, 847e849.
Lawless, H. T., & Heymann, H. (2010). Sensory evaluation of food: Principles & prac-
tices (2nd ed.). Springer.
Lin, M. L., Liu, X. S., Xu, Q. Q., Song, H. L., Li, P., & Yao, J. (2014). Aroma-active
components of yeast extract pastes with a basic and characteristic meaty
flavour. Journal of the Science of Food and Agriculture, 94, 882e889.
Lin, T. F., Wong, J. Y., & Kao, H. P. (2002). Correlation of musty odor and 2-MIB in two
drinking water treatment plants in South Taiwan. Science of the Total Environ-
ment, 289, 225e235.
Mahadevan, K., & Farmer, L. (2006). Key odor impact compounds in three yeast
extract pastes. Journal of Agricultural and Food Chemistry, 54, 7242e7250.
Moresi, M., Orban, E., Quaglia, G. B., & Casini, I. (1995). Effect of some physic-
chemical treatments on the kinetics of autolysed-yeast extract production
from whey. Journal of the Science of Food and Agriculture, 67, 347e357.
Münch, P., Hofmann, T., & Schieberle, P. (1997). Comparison of key odorants
generated by thermal treatment of commercial and self-prepared yeast
191
extracts: influence of the amino acid composition on odorant formation.
Journal of Agricultural and Food Chemistry, 45, 1338e1344.
Münch, P., & Schieberle, P. (1998). Quantitative studies on the formation of key
odorants in thermally treated yeast extracts using stable isotope dilution assays.
Journal of Agricultural and Food Chemistry, 46, 4695e4701.
Narasimhan, S., Chand, N., & Rajalakshmi, D. (1992). Saffron: Quality evaluation by
sensory profile and gas chromatography. Journal of Food Quality, 15, 303e314.
Pang, X. L., Hu, X. S., Liao, X. J., Sun, Z. J., Zhang, M. W., & Wu, J. H. (2012). Study on
Two Evaluation methods of odor-active compounds in Hami melon: Frequency
detection-gas chromatography-olfactometry method and odor activity value
analysis. Journal of Chinese Institute of Food Science and Technology, 12, 174e182
(in Chinese).
Rothe, M. (1976). Aroma values-a useful concept? Nahrung, 20, 259e266.
Song, H. L., & Cadwallader, K. R. (2007). Aroma components of American country
ham. Journal of Food Science, 73, C29eC35.
Steingass, C. B., Langen, J., Carle, R., & Schmarr, H. G. (2015). Authentication of
pineapple (Ananas comosus [L.] merr.) fruit maturity stages by quantitative
analysis of g- and d-lactones using headspace solid-phase microextraction and
chirospecific gas chromatographyeselected ion monitoring mass spectrometry
(HS-SPME-GC-SIM-MS). Food Chemistry, 168, 496e503.
Strube, A., Buettner, A., & Groetzinger, C. (2009). Characterization and identification
of a plastic-like off-odor in mineral water. Water Science and Technology. Water
Supply, 9, 299e309.
Verduyn, C., Suksomcheep, A., & Suphantharika, M. (1999). Effect of high pressure
homogenization and papain on the preparation of autolysed yeast extract.
World Journal of Microbiology and Biotechnology, 15, 57e63.
Werkhoff, P., Bruening, J., Emberger, R., Guentert, M., Koepsel, M., & Kuhn, W.
(1990). Isolation and characterization of volatile sulfur-containing meat flavor
components in model systems. Journal of Agricultural and Food Chemistry, 38,
777e791.