Analysis of Volatile Flavor Compounds of Sardine (Sardinops

S:Se

nsor

y&Fo

odQu

ality

JFS S: Sensory and Food Quality

Analysis of Volatile Flavor Compoundsof Sardine (Sardinops melanostica) bySolid Phase MicroextractionN. GANEKO, M. SHODA, I. HIROHARA, A. BHADRA, T. ISHIDA, H. MATSUDA, H. TAKAMURA, AND T. MATOBA

ABSTRACT: Generally the main component of fishy flavor is considered to be trimethylamine. On the otherhand, carbonyl compounds, produced from oxidation of polyunsaturated fatty acid by lipoxygenase or by autox-idation, might have some contribution to the fishy flavor. Since sardine skin contains high levels of polyunsatu-rated fatty acids and lipoxygenase, carbonyl compounds may be generated more easily than trimethylamine. Inthis study, volatile flavor compounds of sardine were analyzed by gas chromatograph-mass spectrometry and gaschromatograph-olfactometry combined with solid phase microextraction. Then, the flavor components that con-tribute to fishy flavor were identified. At normal pH (6.2), trimethylamine was not detected or sensed from thefresh sardines. When the pH was raised, the amount of trimethylamine became higher. Trimethylamine flavorwas weak at pH 9 and strongly sensed at pH 11 or higher. On the other hand, 33 other compounds were posi-tively or tentatively identified, including 8 hydrocarbons, 5 ketones, 1 furan, 1 sulfur compound, 12 aldehydes, and6 alcohols in fresh sardines. Among them, 2,3-pentanedione, hexanal, and 1-penten-3-ol were the main compo-nents. Forty-seven flavors were detected by gas chromatograph-olfactometry. Among them, paint-like (1-penten-3-one), caramel-like (2,3-pentanedione), green-like (hexanal), shore-like ((Z)-4-heptenal), citrus note (octanal),mushroom-like (1-octen-3-one), potato-like (methional), insect-like ((E,Z)-2,6-nonadienal), and bloody note (notidentified) were strongly sensed. From the aforementioned results, it can be concluded that these compounds ratherthan trimethylamine contributed to fresh sardine flavor.

Keywords: fishy flavor, sardine, volatile flavor compounds

Introduction

Flavor is one of the most important factors for consumer accep-tance of products. Fish is widely consumed as a source of pro-

tein all over the world. However, the characteristic fishy flavor oftenmakes fish less acceptable to some extent. Though the fishy flavorof fresh fish is hardly present, low molecular volatile compoundsare generated by the actions of microbes or by self-digestion dur-ing storage and transportation (Prost and others 2004). As a conse-quence of these actions, fishy flavor is generated.

Sardine (Sardinops melanostica) is one of the most popular fishin the world. The content of polyunsaturated fatty acids such as do-cosahexaenoic and eicosapentaenoic acids in sardine is very high(Gamez-Meza and others 1999). However, sardine is easy to deteri-orate and form a characteristic fishy flavor.

Generally, volatile amines such as trimethylamine are consid-ered as the main components of fishy flavor (Koizumi and oth-

MS 20070544 Submitted 7/14/2007, Accepted 10/13/2007. Authors Ganeko,Shoda, and Bhadra are with Graduate School of Humanities and Sciences,Nara Women’s Univ., Kitauoya-Nishimachi, Nara 630-8506, Japan. AuthorsHirohara, Takamura, and Matoba are with Dept. of Food Science and Nu-trition, Nara Women’s Univ., Kitauoya-Nishimachi, Nara 630-8506, Japan.Authors Ishida and Matsuda are with Takara Shuzo Co., Ltd., 1 Shimotoba-yoshidencho, Fushimi, Kyoto 612-8381, Japan. Author Matsuda is also withYaizu Suisankagaku Industry Co., Ltd., 8–13, Kogawashinmachi 5-chome,Yaizu, Shizuoka 425-8570, Japan. Author Takamura is also with KYOU-SEI Science Center for Life and Nature, Nara Women’s Univ., Kitauoya-Nishimachi, Nara 630-8506, Japan. Direct inquiries to author Takamura(E-mail: [email protected]).

A part of the study was presented at the 2004 Annual Meeting of theInstitute of Food Technologists, Las Vegas, NV, U.S.A., in July 2004.

ers 1979). Trimethylamine is generated from trimethylamine oxideby microorganisms (Baixas-Nogueras and others 2001). It was de-tected in cooked fishes (Koizumi and others 1979; Kasahara andNishibori 1985; Horiuchi and others 1998; Prost and others 1998)and in cod and trout (Milo and Grosch 1995). However, the charac-teristic of fishy odor from sardine and other fish is somewhat differ-ent from that of trimethylamine. Therefore, other compounds mayalso contribute fishy flavor.

On the other hand, carbonyl compounds, which are producedfrom oxidation of polyunsaturated fatty acids by lipoxygenase or byautoxidation, might also have some contribution to the fishy fla-vor (Josephson and others 1984). Since sardine skin contains highlevels of polyunsaturated fatty acids and lipoxygenase (Mohri andothers 1992), carbonyl compounds may be generated more easilythan trimethylamine.

In this study, we evaluate the fishy flavor components in sar-dine by gas chromatograph-mass spectrometry (GC-MS) and gaschromatograph-olfactometry (GC-O) combined with solid phasemicroextraction (SPME) in order to indicate that not trimethy-lamine but carbonyl compounds contribute to fishy odor in sar-dine.

Materials and Methods

ReagentsAuthentic standards and solvents were obtained from Nacalai

Tesque Inc. (Kyoto, Japan) and Wako Pure Chemical Industries (Os-aka, Japan). The water used in this experiment was purified withMilli-Q Labo equipment (Millipore Japan, Tokyo, Japan).

C© 2007 Institute of Food Technologists Vol. 73, Nr. 1, 2008—JOURNAL OF FOOD SCIENCE S83doi: 10.1111/j.1750-3841.2007.00608.xFurther reproduction without permission is prohibited

S:Sensory&Food

Quality

Volatile flavor compounds of sardine. . .

Fish samplesFresh chilled sardines were purchased from a fish shop in Nara,

Japan. Sardines were used in analyses immediately or after 24- or48-h refrigeration at 4 ◦C.

Sample preparationOnly edible parts (that is, muscle and skin) of sardines were used.

Sardine flesh was homogenized in ultrapure water (5 times to fishweight). Three grams of sodium chloride were added to 10 g of

Figure 1 --- Standard curve for quantification of trimethylamine.

Figure 2 --- Comparison of SPME fibers.Fresh sardine homogenate was analyzedby GC with SPME using A, polyacrylate;B, polydimethylsiloxane (PDMS); C,divinylbenzene/polydimethylsiloxane(DVB/PDMS); D, divinylbenzene/carbowax(DVB/CW); E, carboxen/polydimethylsilox-ane (CAR/PDMS); and F, divinylben-zene/carboxen/polydimethylsiloxane(DVB/CAR/PDMS) fibers.

sample homogenate in a 20-mL sealed glass vial. For detection ofvolatile amines, sample pH was adjusted to 6.2 to 12.

SPME devicesThe SPME fibers and the manual holder were purchased from

Supelco Co. (Bellefonte, Pa., U.S.A.). The SPME fibers used werepolyacrylate, polydimethylsiloxane (PDMS), divinylbenzene/

Table 1 --- Changes in trimethylamine content and flavor intensityat different pH.

pH Trimethylamine content (ppm) Flavor intensity

6.2 ND ND7 0.21 ± 0.17 ND8 0.86 ± 0.55 ND9 2.42 ± 1.50 ±10 6.37 ± 2.67 +11 23.95 ± 10.14 ++12 29.55 ± 8.75 ++Data are means ± SD of 3 determinations.ND, not detected; ±, very weak; +, strong; ++, very strong.

Table 2 --- Changes in trimethylamine content and flavor intensityin refrigerated sardine.

pH 6.2 pH 11

Trimethylamine Flavor Trimethylamine FlavorTime (h) content (ppm) intensity content (ppm) intensity

0 ND ND 20.38 ± 9.27 +24 0.04 ± 0.04 ND 30.52 ± 8.04 ++48 0.13 ± 0.16 ND 67.42 ± 20.66 ++Data are means ± SD of 3 determinations.ND, not detected; +, strong; ++, very strong.

S84 JOURNAL OF FOOD SCIENCE—Vol. 73, Nr. 1, 2008

S:Se

nsor

y&Fo

odQu

ality


polydimethylsiloxane (DVB/PDMS), divinylbenzene/carbowax(DVB/CW), carboxen/polydimethylsiloxane (CAR/PDMS), anddivinylbenzene/carboxen/polydimethylsiloxane (DVB/CAR/PD-MS).

Extraction of volatile compounds in sardineThe SPME fiber length was set to 2 cm for regular analysis and

changed to 0.25 to 2 cm for SPME dilution analysis. The vial con-taining homogenized sample was incubated at 40 ◦C for 1 h. ThenSPME fiber was inserted into the headspace above the solution andexposed for 30 min at 40 ◦C. The fiber was retrieved and injectedinto GC and GC-MS. Volatile compounds were desorbed within10 s.

GC analysis of trimethylamineGC analysis of trimethylamine was carried out on a GC-17A (Shi-

madzu, Kyoto, Japan) equipped with a flame ionization detector(FID) and sniffing port. CP-Volamine column (60 m × 0.32 mmi.d., Varian, Walnut Creek, Calif., U.S.A.) was used for separation.The column was preheated to 40 ◦C. The temperature was raised to60 ◦C at 1 ◦C/min and then to 200 ◦C at 5 ◦C/min. The effluent fromthe capillary column was split 1:1 between the FID and the sniffingport using “Y” splitter. Sniffing was carried out using OSS-2 sniffer(SGE Intl., Melbourne, Australia).

In this condition, trimethylamine was eluted at 16.3 min andthe retention index was 495. Quantification of trimethylamine wascarried out with standard curve calculated from FID peak areaand amount of trimethylamine (Figure 1). The detection limit was0.025 ppm.

GC analysis of volatile compounds other thantrimethylamine

GC analysis of volatile compounds other than trimethylaminewas carried out on a GC-17A equipped with FID and sniffing port.DB-WAX column (60 m × 0.32 mm i.d., J&W Scientific, Folsom,Calif., U.S.A.) was used for separation. The column was preheatedto 40 ◦C. Then the temperature was raised to 200 ◦C at 4 ◦C/min.The injector temperature was 250 ◦C, and the detector temperaturewas 260 ◦C. The effluent from the capillary column was split 1:1 be-tween the FID and the sniffing port using “Y” splitter. Sniffing wascarried out using OSS-2 sniffer.

Figure 3 --- Gas chromatogram of volatilecompounds in fresh sardine. Peak numberscorrespond to Table 3 to 5.

GC-MS analysisGC-MS analysis was carried out on a GCMS-QP5050 (Shimadzu,

Kyoto, Japan). The oven condition and column were the sameas described previously. The mass spectrometer was operated at

Table 3 --- Identification of volatile compounds in fresh sardine byGC-MS.

Peak nr Retention index Identification

Hydrocarbons (8)1 500 Pentane2 522 2-pentene4 600 Hexane6 700 Heptane11 860 2-octene20 937 Benzene65 1484 Pentadecane76 1695 HeptadecaneKetones (5)10 821 Acetone25 1022 1-penten-3-one27 1056 2,3-pentanedione48 1309 1-octen-3-one51 1325 2,3-octanedioneAldehydes (12)9 783 Propanal13 880 Butanal22 977 Pentanal29 1084 Hexanal34 1138 (E)-2-pentenal37 1185 Heptanal39 1228 (E)-2-hexenal41 1250 (Z)-4-heptenal46 1295 Octanal59 1315 (E)-2-octenal68 1515 (E,E)-2,4-heptadienal73 1612 (E,Z)-2,6-nonadienalAlcohols (6)36 1157 1-penten-3-ol49 1228 (E)-2-penten-1-ol50 1250 (Z)-2-penten-1-ol54 1353 Hexanol60 1446 1-octen-3-ol61 1454 HeptanolFuran (1)21 944 2-ethylfuranSulfur compound (1)7 710 Carbon disulfide

Peak numbers correspond to Figure 3.

Vol. 73, Nr. 1, 2008—JOURNAL OF FOOD SCIENCE S85

S:Sensory&Food

Quality


an ionization voltage of 70 eV and an ion source temperature of290 ◦C. The injection port temperature was set at 250 ◦C. Heliumwas used as a carrier gas at a column flow rate of 1.8 mL/min.

Identification of volatile compoundsFlavor compounds were identified by comparison of the reten-

tion time and mass spectrum with those of authentic standards.Characteristic flavor compounds were specified by GC-O.

Results and Discussion

Selection of fiberSix types of SPME fiber were compared for their adsorption capa-

bilities (Figure 2). DVB/CAR/PDMS (50/30 μm layer) was selectedfor the following experiments, since the highest level of fish flavorwas obtained.

Table 4 --- Compounds and their characteristic flavor identified through GC-O.

Fiber length

Peak nr RI Characteristic flavor Compounds 2.0 cm 1.0 cm 0.5 cm 0.25 cm

2 522 Gas, chemical, paint-like 2-pentene ± − − −--- 675 Enteruria-like, sulfide-like NI ++ ++ ++ ±9 783 Alcoholic-like Propanal ++ ++ + ±--- 924 Paint-like, chemical-like NI ± ± − −--- 967 Caramel-like, rotten (2,3-butanedione) + + − −22 977 Green pentanal ± − − −--- 982 Gas-like NI ++ ++ ± −--- 1007 Paint-like, chemical-like NI ++ ++ ++ ++25 1022 Paint-like, chemical-like 1-penten-3-one +++ +++ +++ +++27 1056 Caramel-like, rotten 2,3-pentanedione +++ +++ +++ +++29 1084 Green hexanal +++ +++ +++ +++--- 1087 Coffee-like, allium-like NI ++ ++ ++ ++--- 1106 Paint-like, chemical-like NI ++ ++ ++ ++34 1138 Green (E)-2-pentenal ± ± − −--- 1143 Green NI ++ ++ ++ ±--- 1153 Potato-like NI + ± − −37 1185 Citrus-like heptanal + + ± −--- 1196 Bloody NI + + ± −41 1250 Shore-like (Z)-4-heptenal +++ +++ +++ +++--- 1279 Shiitake mushroom-like NI ++ ++ + +46 1295 Citrus-like octanal ++ ++ ++ ++48 1309 Shiitake mushroom-like 1-octen-3-one ++ ++ ++ ++--- 1316 Melon-like, sweet NI + − − −--- 1323 sweet NI ± − − −--- 1340 Paper-like NI ++ ++ ++ +--- 1353 Roasted NI + − − −--- 1359 Sweet NI ++ ++ ++ +--- 1372 Metalic NI ++ ++ + ±--- 1380 Bloody NI +++ ++ ++ ++--- 1385 Paper-like NI ++ + + ±--- 1440 Sweet NI ++ ++ + +60 1443 Aromatic, oxidized oil-like (E)-2-octenal ++ + + ±--- 1447 Paper-like NI ++ ++ ++ +--- 1453 Burdock-like NI ± ± ± −--- 1474 Potato-like (methional) +++ +++ +++ +++--- 1485 Metalic NI + + + ±68 1515 Aromatic, oxidized oil-like (E,E)-2,4-heptadienal ++ ++ + ±--- 1555 Watermelon-like NI ++ ++ ++ +73 1612 Insect-like (E,Z)-2,6-nonadienal +++ +++ ++ ++--- 1618 Melon-like, sweet NI ++ ++ ++ +--- 1733 Aromatic, oxidized oil-like NI + + − −--- 1747 Dried fish-like, dustcloth-like NI ++ ++ + +--- 1822 Roasted NI ++ ++ ++ ++--- 1862 Aromatic, oxidized oil-like NI + + + ±Peak numbers correspond to Figure 3.Compound names in parentheses are estimated by the aroma and retention time of authentic samplesRI = retention index; NI = not identified.−, none; ±, very weak; +, weak; ++, strong; +++, very strong.

Trimethylamine in the fresh sardine at different pHThe effect of pH on the volatility of trimethylamine and flavor

intensity is shown in Table 1. At normal pH (6.2), trimethylaminewas not detected from the fresh sample. Triqui and Bouchriti (2003)reported a similar result. When the pH was raised, the level oftrimethylamine became higher. The smell of trimethylamine wasnot detected at normal PH (6.2), and it was weak at pH 9 butwas strongly sensed at pH 11 or higher. These results suggestthat trimethylamine was hardly volatilized to be detected belowpH 8.

Trimethylamine in refrigerated sardineTable 2 shows the changes in trimethylamine content and the

flavor intensity in sardine refrigerated at 4 ◦C for 24 and 48 h.Trimethylamine was analyzed at pH 6 and pH 11. At pH 6.2,trimethylamine was not sensed even if stored for 48 h. When the pHwas raised to 11, trimethylamine increased with the storage time.


S:Se

nsor

y&Fo

odQu

ality


Its fishy flavor was sensed more strongly as the period of storage be-came longer. Therefore, trimethylamine increased with long-timestorage, but it was scarcely volatilized and sensed at normal pH. Al-though it is often thought that trimethylamine is the main contrib-utor of fishy flavor, trimethylamine did not contribute to sardineflavor at normal pH.

Volatile compounds other than trimethylamine infresh sardine

Gas chromatograms of volatile compounds in fresh sardinesare shown in Figure 3. The identified volatile compounds arelisted in Table 3. In fresh sardines, 33 compounds were def-initely or tentatively identified, including 8 hydrocarbons, 5ketones, 1 furan, 1 sulfur compound, 12 aldehydes, and 6alcohols. 2,3-Pentanedione, hexanal, and 1-penten-3-ol weremainly detected. Contents of propanal, 2-ethylfuran, pentanal,

Table 5 --- Changes in flavor intensity during refrigeration.

Odor strength

Peak nr RI Characteristic flavor Compounds Fresh 24 h 48 h

2 522 Gas, chemical-like 2-pentene ± ++ ++--- 675 Enteruria-like NI ++ ++ ++9 783 Alcoholic-like Propanal ++ ++ ++13 880 Grassy Butanal − − ±--- 924 Paint-like, chemical-like NI ± + ++--- 967 Caramel-like, rotten (2,3-butanedione) + ++ ++22 977 Green Pentanal − ± ±--- 982 Gas-like NI ++ ++ ++--- 1007 Paint-like, chemical-like NI ++ ++ ++25 1022 Paint-like, chemical-like 1-penten-3-one +++ ++++ ++++27 1056 Caramel-like, rotten 2,3-pentanedione +++ ++++ ++++29 1084 Green Hexanal +++ ++++ ++++--- 1087 Coffee-like, allium-like NI ++ +++ +++--- 1106 Paint-like, chemical-like NI ++ ++ +++34 1138 Green (E)-2-pentenal ± + +--- 1143 Green NI ++ ++ ++--- 1153 Potato-like NI + + ++36 1157 Paint-like, chemical-like 1-penten-3-ol − − ±37 1185 Citrus-like Heptanal + ++ ++--- 1196 Bloody NI + ++ ++41 1250 Shore-like (Z)-4-heptenal +++ ++++ ++++--- 1279 Shiitake mushroom-like NI ++ ++ ++46 1295 Citrus-like Octanal ++ +++ ++++48 1309 Shiitake mushroom-like 1-octen-3-one ++ +++ ++++--- 1316 Melon-like, sweet NI + + ++--- 1323 Sweet NI − ± +--- 1340 Paper-like NI ++ +++ +++--- 1353 Roasted NI + + +++--- 1359 Sweet NI ++ ++ +++--- 1372 Metalic NI ++ ++ +++--- 1380 Bloody NI +++ ++++ ++++--- 1385 Paper-like NI ++ ++ ++--- 1440 Sweet NI ++ ++ ++60 1443 Aromatic, oxidized oil-like (E)-2-octenal ++ ++ ++--- 1447 Paper-like NI ++ ++ ++--- 1453 Burdock-like NI ± ++ ++--- 1474 Potato-like (Methional) +++ ++++ ++++--- 1485 Metalic NI + ++ ++68 1515 Aromatic, oxidized oil-like (E,E)-2,4-heptadienal ++ ++ +++--- 1555 Watermelon-like NI ++ ++ ++73 1612 Insect-like (E,Z)-2,6-nonadienal +++ ++++ ++++--- 1618 Melon-like, sweet NI ++ ++ ++--- 1733 Aromatic, oxidized oil-like NI + + ++--- 1747 Dried fish-like, dustcloth-like NI ++ ++ ++--- 1822 Roasted NI ++ +++ ++++--- 1862 Aromatic, oxidized oil-like NI + + ++Peak numbers correspond to Figure 3.Compound names in parentheses are estimated by the aroma and retention time of authentic samples.RI = retention index; NI, not identified.−, none; ±, very weak; +, weak; ++, strong; +++, very strong; ++++, strongest.

1-penten-3-one, 2,3-pentanedione, hexanal, 1-penten-3-ol, 2,3-octanedione, heptanol, and pentadecane were high in sardine fla-vor. Yoshiwa and others (1992) detected pentane, hexane, heptane,propanal, (E)-2-octene, butanal, benzene, pentanal, 1-penten-3-one, 2,3-butanedione, hexanal, (E)-2-pentenal, 1-penten-3-ol, (Z)-2-penten-1-ol, 1-octen-3-ol, and (E,E)-2,4-heptadienal from freshsardine by purge and trap extraction method. Volatile compoundsdetected in this study are similar to their results. Yoshiwa and oth-ers (1997) identified 2,3-pentanedione, hexanal, (E)-2-pentenal, 1-penten-3-ol, (E)-2-hexenal, (Z)-4-heptenal, octanal, (E)-2-penten-1-ol, (Z)-2-penten-1-ol, hexanol, (E)-2-octenal, 1-octen-3-ol, (E,E)-2,4-heptadienal, (E,Z)-2,6-nonadienal, and heptadecane from freshsardines by simultaneous distillation and extraction under re-duced pressure. They detected heptanal in sardines stored at10 ◦C for 24 h. Similar compounds were also identified in ourstudy.

Vol. 73, Nr. 1, 2008—JOURNAL OF FOOD SCIENCE S87

S:Sensory&Food

Quality


Flavor analysis of fresh and refrigerated sardines byGC-O

In GC-O analysis, 47 flavors were sensed (Table 4). Amongthem, 14 flavor compounds were identified by GC-MS. Paint-like (1-penten-3-one), caramel-like (2,3-pentanedione), green-like(hexanal), shore-like ((Z)-4-heptenal), and potato-like (methional)odors were sensed most strongly in SPME dilution analysis.

1-Penten-3-ol was detected by purge and trap extraction method(Yoshiwa and others 1992), which is a responsible flavor of freshmarine products and generated from polyunsaturated fatty acid bythe influence of an enzyme such as lipoxygenase or hydroperoxi-dase. We detected the same compound by GC, but its flavor wassensed by GC-O occasionally. Consequently, hexanol, heptanol,and 1-octen-3-ol were detected in this study, but were not sensedin GC-O analysis. Contribution of these compounds to the flavor offresh sardines seems quite low in the initial stage of freshness de-cline because their sensory threshold might be too high.

Among aldehydes, propanal (alcohol-like), pentanal (green),hexanal (green), (E)-2-pentenal (grass-like), heptenal (citrus-like), (E)-2-octenal (cinnamon-like), methional (potato-like), (E,E)-2,4-heptadienal (cinnamon-like/oxidized oil-like), and (E,Z)-2,6-nonadienal (cucumber-like/insect-like) were sensed by GC-O. Es-pecially, hexanal, (Z)-4-heptenal, octanal, methional, and (2,6)-nonadienal were sensed strongly. (E,Z)-2,6-nonadienal is thoughtto be an important flavor of ayu fish (Hirano and others 1992), andprecursor of (Z)-4-heptenal (McGill and others 1974, 1977) which isthought to be the main compound leading to off-flavor in refriger-ated cod. Methional, a sulfur-containing aldehyde, is a compoundfound in soy sauce and potato chips. Although it was not detectedby GC-FID or GC-MS, it was estimated by its aroma and retentiontime of an authentic sample.

Among ketones, 2,3-butanedione (caramel-like), 1-penten-3-one (paint-like), 2,3-pentanedione (caramel-like), and 1-octen-3-one (mushroom-like) were sensed. Although 2,3-butanedione wasnot detected by GC-FID or GC-MS, it was estimated by its aromaand retention time of an authentic sample.

Changes in flavor intensity during refrigeration are presentedin Table 5. Aldehydes and ketones tended to increase dur-ing refrigeration. 1-Penten-3-one (paint-like), hexanal (grassy),(Z)-4-heptenal (shore-like), octanal (citrus-like), 1-octen-3-one(mushroom-like), methional (potato-like), (E,Z)-2,6-nonadienal(cucumber-like/insect-like), and some unidentified odors (bloody,aromatic) were sensed strongly in samples preserved for 48 h.

Conclusions

Trimethylamine, which is usually considered as the main com-ponent of fishy flavor generation, was detected at 24 h but not

sensed from fresh sardine even after 48-h refrigeration. At pH 9 orhigher, trimethylamine was detected and sensed. Trimethylaminewas increased at pH 11 with refrigeration time and was also sensedstrongly.

Besides trimethylamine, 33 other compounds were definitelyor tentatively identified, including 8 hydrocarbons, 5 ketones,1 furanoid, 1 sulfur-containing compound, 12 aldehydes, and6 alcohols. 2,3-Pentanedione, hexanal, and 1-penten-3-ol weredetected mostly. Among the compounds, hexanal (grassy),(Z)-4-heptenal (shore-like), (E,Z)-2,6-nonadienal (cucumber-like/insect-like), 1-penten-3-one (paint-like), 2,3-pentanedione(caramelized), and methional (potato-like) were detected andsensed strongly. Therefore, it can be concluded that these carbonylcompounds rather than trimethylamine contribute to sardine fishyodor.

ReferencesBaixas-Nogueras S, Bover-Cid S, Vidal-Carou MC, Veciana-Nogues MT, Marine-Font

A. 2001. Trimethylamine and total volatile basic nitrogen determination by flow in-jection/gas diffusion in Mediterranean hake (Merluccius merluccius). J Agric FoodChem 49:1681–6.

Gamez-Meza N, Higuera-Ciapara I, Calderon de la Barca AM, Vazquez-Moreno L,Noriega-Rodrıguez J, Angulo-Guerrero O. 1999. Seasonal variation in the fatty acidcomposition and quality of sardine oil from Sardinops sagax caeruleus of the Gulfof California. Lipids 34:639–42.

Hirano T, Zhang CH, Morishita A, Suzuki T, Shirai T. 1992. Identification of volatilecompounds in ayu fish and its feeds. Nippon Suisan Gakkaishi 58:547–57.

Horiuchi M, Umano K, Shibamoto T. 1998. Analysis of volatile compounds formedfrom fish oil heated with cysteine and trimethylamine oxide. J Agric Food Chem46:5232–7.

Josephson DB, Lindsay RC, Stuiber DA. 1984. Biogenesis of lipid-derived volatilearoma compounds in the emerald shiner (Notropis atherinoides). J Agric FoodChem 32:1347–52.

Kasahara K, Nishibori K. 1985. Volatile compounds of roasted fishes. Nippon SuisanGakkaishi 51:489–92.

Koizumi C, Keiu-Thu C, Nonaka J. 1979. Undesirable odor of cooked sardine meat.Nippon Suisan Gakkaishi 45:1307–12.

McGill AS, Hardy R, Burt JR, Gunstone FD. 1974. Hept-cis-4-enal and its contributionto the off-flavour in cold stored cod. J Sci Food Agric 25:1477–89.

McGill AS, Hardy R, Gunstone FD. 1977. Further analysis of the volatile componentsof frozen cold stored cod and the influence of these on flavour. J Sci Food Agric28:200–5.

Milo C, Grosch W. 1995. Detection of odor defects in boiled cod and trout by gaschromatography-olfactometry of headspace samples. J Agric Food Chem 43:459–62.

Mohri S, Cho S-Y, Endo Y, Fujimoto K. 1992. Linoleate 13(S)-lipoxygenase in sardineskin. J Agric Food Chem 40:573–6.

Prost C, Serot T, Demaimay M. 1998. Identification of the most potent odorants inwild and farmed cooked turbot (Scophtalamus maximus L.). J Agric Food Chem46:3214–9.

Prost C, Hallier A, Cardinal M, Serot T, Courcoux P. 2004. Effect of storage time on rawsardine (Sardina pilchardus) flavor and aroma quality. J Food Sci 69:198–204.

Triqui R, Bouchriti N. 2003. Freshness assessments of moroccan sardine (Sardinapilchardus): comparison of overall sensory changes to instrumentally determinedvolatiles. J Agric Food Chem 51:7540–6.

Yoshiwa T, Morimoto K, Sakamoto K, Ishikawa Y. 1992. Analysis of volatile compo-nents in sardine by purge-and-trap method. Nippon Suisan Gakkaishi 58:2105–10.

Yoshiwa T, Morimoto K, Sakamoto K, Ishikawa Y, Tokita M, Morita M. 1997. Volatilecompounds of fishy odor in sardine by simultaneous distillation and extraction un-der reduced pressure. Nippon Suisan Gakkaishi 63:222–30.


Documents

Analysis of Volatile Flavor Compounds of Sardine (Sardinops