Identification and Characterization of the Aroma-Impact Components

Mar 2, 2015 - evaporation (DSE−SAFE) and aroma extract dilution analysis (AEDA). Meanwhile, static headspace dilution analysis (SHDA) and headspace ...
0 downloads 0 Views 1020KB Size
Download PDF
Article pubs.acs.org/JAFC

Identification and Characterization of the Aroma-Impact Components of Thai Fish Sauce Nawaporn Lapsongphon,† Jirawat Yongsawatdigul,† and Keith R. Cadwallader*,‡ †

School of Food Technology, Suranaree University of Technology, Nakhon Ratchasima 30000, Thailand Department of Food Science and Human Nutrition, University of Illinois, 1302 West Pennsylvania Avenue, Urbana, Illinois 61801, United States

‡

ABSTRACT: Comprehensive analysis of the potent odorants in Thai premium fish sauce samples was accomplished by use of complementary volatile isolation methods combined with gas chromatography−olfactometry (GC−O) and GC−mass spectrometry. Odorants of intermediate and low volatility were determined by direct solvent extraction/solvent-assisted flavor evaporation (DSE−SAFE) and aroma extract dilution analysis (AEDA). Meanwhile, static headspace dilution analysis (SHDA) and headspace solid-phase microextraction (H-SPME) were used to determine the highly volatile odorants. Results of AEDA indicated the importance (log3FD factor ≥6) of five acidic odorants (butanoic acid, 3-methylbutanoic acid, 4-hydroxy-2,5dimethyl-3(2H)-furanone, 4-hydroxy-2-ethyl-5-methyl-3(2H)-furanone, and 2-phenylacetic acid) and four neutral/basic odorants (3-methylbutanal, (Z)-1,5-octadien-3-one, phenylacetaldehyde, and o-aminoacetophone). Results of SHDA indicated the predominant (log3FD factors ≥5) headspace odorants were methanethiol, 2-methylpropanal, 2-methylbutanal, 3-methylbutanal, dimethyl trisulfide, 3-(methylthio)propanal, and butanoic acid. Concentrations for 21 odorants were determined by stable isotope dilution analysis (SIDA), and their odor-activity values (OAVs) were calculated. Among these, methanethiol, 2methylpropanal, 3-methylbutanal, dimethyl trisulfide, 3-(methylthio)propanal, and butanoic acid had the highest OAVs (>500). Results of aroma recombination and omission studies revealed the importance of acids, aldehydes, and sulfur-containing compounds to the overall aroma of the Thai fish sauce. KEYWORDS: fish sauce, aroma extract dilution analysis, odorant, odor-activity value, recombination study, omission study

■

INTRODUCTION Thai fish sauce or nam pla is widely consumed throughout Southeast Asia, and demand for this product is growing worldwide due the increased popularity of Southeast Asian cuisine. The typical production of Thai fish sauce involves combining two to three parts of anchovy (Stolephourus spp.) with one part of solar salt.1 The mixture is allowed to undergo a natural fermentation for 12−18 months, after which the liquid is drained off and ripened for an additional 2−12 weeks before bottling.2 Although the product is a rich source of protein, it contains a high level of salt and is, therefore, mainly used as a condiment or seasoning agent. Flavor, and to a lesser extent appearance attributes, are the main determinants of fish sauce quality. The flavor components of fish sauce are formed primarily by enzymatic and microbial degradation of proteins and lipids. The earliest detailed study of the volatile composition of Thai fish sauce was conducted by MacIver et al.3 They identified 52 volatile compounds, of which 42 had not been previously reported in fish sauce. Later, Sanceda et al.4 compared the volatile compositions of Thai, Japanese, and Vietnamese fish sauces and reported that the three fish sauces differed mainly in their contents of volatile short-chain fatty acids. Several studies have made use of gas chromatography−olfactometry and GC− mass spectrometry combined with either dynamic headspace analysis (DHA)5,6 or solid phase microextraction (SPME)7 to identify the aroma-active components of Thai fish sauce. In particular, the study by Giri et al.5 compared the odor-activity values (OAVs) of premium Thai fish sauce and reported that © 2015 American Chemical Society

butanoic acid had the highest odor-activity value (OAV), followed by 2-methylbutanoic acid, dimethyl trisulfide, 3(methylthio)propanal, acetic acid, and trimethylamine. The accuracy of an OAV depends on the accuracy of both the odor detection threshold and the quantification results for the target odorant. Quantification of the volatile compounds in fish sauce has, thus far, been carried out using semiquantitative analysis methods involving the use of internal standards.5,6 Currently, stable isotope dilution assay (SIDA) is the preferred method for quantitative analysis due to its great precision and accuracy. SIDA makes use of isotopically labeled analogues as internal standards. The physical and chemical properties of the labeled internal standards are nearly identical to those of the unlabeled analytes. In combination with GC−MS, the analyte and labeled internal standard can be differentiated according to their different molecular masses or electron-impact mass spectra.8 Although potent odorants can be ranked by their relative odor potencies based on results of GC−O or by determination of their OAVs, the actual contribution of these compounds to the overall aroma of the product should be verified by construction and subsequent sensory analysis of aroma recombination models. To formulate an accurate recombination model, it is critical to know with some certainty the Received: Revised: Accepted: Published: 2628

December 17, 2014 February 24, 2015 March 2, 2015 March 2, 2015 DOI: 10.1021/jf5061248 J. Agric. Food Chem. 2015, 63, 2628−2638

Article

Journal of Agricultural and Food Chemistry

Figure 1. Structures for labeled standards used in stable isotope dilution assays. (Firmenich, Princeton, NJ); 10 (Bedoukian Research Inc., Danbury, CT); 31 and 32 (TCI America, Portland, OR); 11 (Lancaster, Windham, NH). 2-Acetyl-1-pyrroline (no. 12) was synthesized using the procedure described by Fuganti et al.,9 and (Z)-1,5-octadien-3-one (no. 14) was synthesized according to Lin et al.10 Odorless deionized− distilled water was prepared by boiling glass-distilled water in an open flask until its volume was reduced by one-fourth of the original volume. Isotopically Labeled Compounds. Structures for labeled standards used for quantification by stable isotope dilutions assays are provided in Figure 1. The following labeled compounds were obtained from commercial sources: [2H9]-trimethylamine (I-1), [2H2]3-methylbutanal (I-7), [2H3]-acetic acid (I-15) (CDN, Pointe-Claire, Quebec, Canada); [2H6]-dimethyl sulfide (I-4̅) (Sigma-Aldrich); [2H5]-propanoic acid (I-18) (Cambridge Isotope Laboratories, Inc., Tewksbury MA); and [13C2]-2-phenylacetic acid (I-37) (Isotec, Miamisburg, OH). The following labeled compounds were synthesized using previously published methods: [2H3]-methanethiol (I-2),11 [2H6]dimethyltrisulfide (I-13),12 [2H3]-3-(methylthio)propanal (I-16),12 [2H5]-3,6-dimethyl-2-ethylpyrazine (I-17),13 [2H2]-butanoic acid (I21),14 [13C2]-phenylacetaldehyde (I-22),15 [2H2]-3-methylbutanoic acid (I-23),14 [13C2]-4-hydroxy-2,5-dimethyl-3(2H)-furanone (I30),16 [2H3]-o-aminoacetophenone (I-34),17 and [2H3]-skatole (I36).18 Synthesis. [2,3-2H2]-2-Methylpropanal (I-5) and [2,3-2H2]-2methylpropanoic acid were synthesized from the corresponding labeled alcohol. The uniformly deuterium labeled alcohol was synthesized following a published procedure described for the synthesis of 5,6-[2H2]-hexan-1-ol19 with some modifications.

concentration of each odorant as well as the matrix composition the product. This approach can be used to verify the results of instrumental analyses and reveal which odorants make the greatest contribution to the overall aroma of the product. The aim of this study was to identify the key aroma components of premium commercial Thai fish sauce by application of the complementary GC−O techniques, aroma extract dilution analysis (AEDA), static headspace dilution analysis (SHDA), and by calculation of OAVs for predominant odorants based on accurate concentrations determined by SIDA. In addition, the impact of selected predominant odorants on the overall aroma of fish sauce was elucidated by aroma recombination and omission studies.

■

MATERIALS AND METHODS

Fish Sauce Samples. Two leading brands of commercial premium fish sauce (nam pla) were obtained from a local market (Bangkok, Thailand). They are referred to as fish sauce A (Rayong Fish Sauce Industry Co., Ltd., Rayong, Thailand) and fish sauce B (Marine Resources and Development Co., Ltd., Chantaburi, Thailand). Fish sauce A contained 23.99% (w/w) NaCl and 2.58% (w/w) total nitrogen, while NaCl and total nitrogen contents for fish sauce B were 24.85% (w/w) and 2.18% (w/w), respectively. Reference Standard Compounds. The following chemicals (listed in Tables 2−7) were obtained from commercial sources given in parentheses: nos. 1−8, 13, 16−25, 27, 29, 30, and 33−38 (SigmaAldrich, St. Louis, MO); 15 (Fisher Scientific, Pittsburgh, PA); 28 2629

DOI: 10.1021/jf5061248 J. Agric. Food Chem. 2015, 63, 2628−2638

Article

Journal of Agricultural and Food Chemistry [2,3-2H2]-2-Methylpropan-1-ol. Chlorotri(triphenylphosphine)rhodium(I) (Wilkinson’s catalyst; Sigma-Aldrich) (15 wt % of the alkenol; 0.45 g) and 2-methyl-2-propen-1-ol (Sigma-Aldrich; 2.5 g, 34.7 mmol) were placed in a pressure reactor equipped with stir bar and rubber septum. The reactor was sealed, then flushed for 5 min with deuterium gas (UHP grade 99.995%; isotopic enrichment 99.7%; Matheson Tri-Gas, Parsippany, NJ) using a needle placed below the surface of the solution. Reaction was carried out at 40 psi until the reaction was complete. The spent catalyst was removed by centrifugation. The target compound was purified by vacuum distillation. Yield: 1.83 g (99% purity). MS-EI, m/z (%): 76 (11) (M+), 58 (5), 57 (7), 46 (5), 45 (100), 44 (64), 43 (59), 42 (42), 41 (13), 40 (13), 39 (8). [2,3-2H2]-2-Methylpropanal (I-5). [2,3-2H2]-2-Methylpropan-1ol (0.38 g; 5.0 mmol) in 2 mL of dichloromethane was added to a 10 mL suspension of pyridinium chlorochromate (Sigma-Aldrich) (1.2 g; 5.5 mmol) in dichloromethane. The mixture was stirred for 1.5 h at room temperature, and 20 mL of diethyl ether was added. The supernatant was decanted, and the residue was extracted with diethyl ether (3 × 5 mL) until the black gum became granular in consistency. The combined organic phase was filtered through a 10 g bed of Florisil (Sigma-Aldrich). The target compound (I-5) was purified by vacuum distillation prior to removal of solvent. Yield: 0.18 g (92.2% purity). MS-EI, m/z (%): 75 (8) (M + 1), 74 (54) (M+), 73 (21), 58 (6), 59 (7), 45 (100), 44 (67), 43 (60), 42 (65), 40 (45), 39 (21). [2,3-2H2]-2-Methylpropanoic Acid (I-20). [2,3-2H2]-2-Methylpropan-1-ol (1.0 g; 13 mmol) was added to a vigorously stirred solution of potassium permanganate (2.5 g, 16 mmol) dissolved in 50 mL of aqueous 0.15 M NaOH in a 200 mL flask. After 2 h, the mixture was acidified with 1 M HCl and then extracted with diethyl ether (3 × 20 mL). The ether layer was extracted with aqueous 0.5 M NaOH (3 × 20 mL). The NaOH extract was acidified with aqueous 4 M HCl and then extracted with ether (3 × 20 mL). The ether layer was washed with aqueous saturated NaCl (2 × 20 mL) and dried over anhydrous Na2SO4 (5 g). The target compound (I-20) was obtained after removal of the solvent. Yield: 0.2552 g (98% purity). MS-EI m/z (%): 90 (8) (M+), 75 (15), 74 (19), 46 (6), 45 (100), 44 (11), 43 (29), 42 (20), 41 (6), 40 (9). [2,3-2H2]-3-Phenylpropanoic Acid (I-38). The deuterated acid was synthesized from cinnamic acid using a modification of the procedure described above for synthesis of [2,3-2H2]-2-methylpropan1-ol. Chlorotri(triphenylphosphine)rhodium(I) (Wilkinson’s catalyst) (15 wt % of the alkenoic acid; 0.15 g) plus (E)-cinnamic acid (Eastman Kodak Co., Rochester, NY; 1.0 g; 34.7 mmol) and 5.0 mL of methanol-d (Sigma-Aldrich) were placed in a pressure reactor equipped with stir bar and rubber septum. The reactor was sealed then flushed for 5 min with deuterium gas (UHP grade 99.995%; isotopic enrichment 99.7%; Matheson Tri-Gas) using a needle placed below the surface of the solution. Reaction was carried out at 40 psi until the reaction was complete. The spent catalyst was removed by centrifugation and the solvent evaporated to yield the target compound (I-38). Yield: 0.92 g (95% purity). MS-EI, m/z (%): 152 (42) (M+), 107 (15), 106 (43), 105 (22), 93 (9), 92 (100), 91 (16), 80 (5), 79 (10), 78 (13), 77 (6), 66 (8), 52 (6), 51 (10). Sample Preparation. Direct Solvent Extraction (DSE). Fish sauce (20 g) plus 100 mL of odorless-distilled water was mixed well with 25 μL of an internal standard stock solution (1.15 μg/μL of 2-ethyl butyric acid in methanol as acidic fraction internal standard, 1.32 μg/ μL of 2-methyl-3-heptanone, and 1.13 μg/μL of 2,4,6-trimethylpyridine in methanol as neutral/basic fraction internal standard). The pH was adjusted to ∼2.0 using 2 M HCl and extracted three times with diethyl ether (100 mL total volume). After the third extraction, the pH of sample was raised to ∼9.0 using 2 M NaOH and extracted three times with diethyl ether (100 mL total volume). The pooled solvent extract (200 mL) was concentrated to 50 mL using a Vigreux column in a 43 °C water bath.

Solvent-Assisted Flavor Evaporation (SAFE). Extracts prepared by DSE were subjected to SAFE to remove any nonvolatile materials, as previously described.20 The SAFE extract was concentrated to 30 mL using a Vigreux column in a 43 °C water bath and separated into acidic (AF) and neutral/basic (NBF) fractions. The aroma extract (30 mL) was extracted with aqueous 0.1 M NaOH (3 × 20 mL), and the upper layer (diethyl ether) containing the neutral/basic volatiles fraction (NBF) was collected. The aqueous layers were pooled, acidified to pH ∼2.0 with 4 M HCl, saturated with NaCl, and then extracted with diethyl ether (3 × 20 mL). The upper layer containing the acidic volatiles (AF) was collected. The two fractions were washed twice with 15 mL of saturated NaCl solution to 10 mL using a Vigreux column and then dried over 2 g of anhydrous Na2SO4. Final volumes of extracts were concentrated using a gentle stream of nitrogen gas to 900 μL for AF and 100 μL for NBF. Extracts were stored in 2 mL glass vials at −70 °C until analysis. Aroma Extract Dilution Analysis (AEDA). AF and NBF extracts were diluted by volume (1 part aroma extract to 2 parts solvent) in diethyl ether. Each dilution was analyzed by gas chromatography− olfactometry (GC−O) using a 6890 GC (Agilent Technologies, Inc., Palo Alto, CA) equipped with a flame ionization detector (FID) and olfactory detection port (DATU Technology Transfer, Geneva, NY). The sample (2 μL) was injected using direct cool-on-column mode (+3 °C temperature tracking). Separations were performed on polar (Rtx-Wax, 15 m × 0.32 mm i.d.; 0.5 μm film; Restek, Bellefonate, PA) and nonpolar (Rtx-5 SLIMS, 15 m × 0.32 mm i.d.; 0.5 μm film; Restek) columns. The GC oven temperature was programmed from 40 to 225 °C at a rate of 10 °C/min, with initial and final holding times of 5 and 20 min, respectively. Helium was used as the carrier gas at a constant flow rate of 2.2 mL/min. Column effluent was split 1:1 between the FID and the sniffing port using deactivated fused silica tubing (each 1 m × 0.25 mm i.d.; Restek). The FID and sniffing port were maintained at a temperature of 250 °C. GC−O was conducted by two experienced panelists. The flavor dilution factor (FD factor) for a compound was the highest dilution at which it was detected by GC− O. Other details of the procedure have been previously described.21 Static Headspace Dilution Analysis (SHDA). GC−O was conducted using a 6890 GC (Agilent Technologies, Inc.) equipped with an FID and olfactory detector port (ODP2; Gerstel, Inc., Germany). Fish sauce sample (5 g) was placed in a 250 mL roundbottom flask and sealed with a PTFE-lined septum. The flask was equilibrated at 40 °C for 30 min with agitation provided by a magnetic stir bar. Each headspace volume (25, 5, 1, 0.2, or 0.04 mL) was injected via a heated (50 °C) gastight syringe into a cooled injection system (CIS-4, Gerstel, Inc.) operating in the solvent vent mode [−120 °C (0.10 min hold) ramped at 12 °C/s to 260 °C (10 min hold); 50 mL/min initial vent flow (0.10 min hold), 1.10 min splitless time, 50 mL/min final split flow]. Separations were performed on RtxWax (15 m × 0.53 mm i.d.; 1 μm film; Restek) and Rtx-5 columns (15 m × 0.53 mm i.d.; 1 μm film; Restek). The GC oven temperature was programmed from 35 to 225 °C at a rate of 6 °C/min, with initial and final holding times of 5 and 10 min, respectively. The FID and ODP2 temperatures were maintained at 250 °C. Column effluent was split 1:1 between the FID and ODP2 using deactivated fused silica tubing as described earlier. Other details of the procedure have been previously described.21 Gas Chromatography−Mass Spectrometry (GC−MS). The GC−MS analyses were performed using a 6890 GC/5973 mass selective detector (MSD; Agilent Technologies, Inc.) system. Each aroma extract (1 μL) was injected into a cooled injection system (CIS4, Gerstel, Inc.) operating in the cold splitless-mode [−50 °C initial temperature (0.10 min hold); ramp rate 12 °C/s; 260 °C final temperature (10 min hold); 1.10 min valve-delay, the 50 mL/min vent flow]. Separations were performed on Stabilwax (30 m × 0.25 mm i.d.; 0.25 μm film; Restek) and Sac-5 columns (30 m × 0.25 mm i.d.; 0.25 μm film; Supleco, Bellefonte, PA). Helium was used as carrier gas at a constant rate of 1.0 mL/min. The GC oven temperature was programmed from 35 to 225 °C at a rate of 4 °C/min with initial and final holding times of 5 and 20 min, respectively. MSD conditions were: capillary direct interface temperature, 280 °C; ionization energy, 2630

DOI: 10.1021/jf5061248 J. Agric. Food Chem. 2015, 63, 2628−2638

Article

Journal of Agricultural and Food Chemistry

fishy/briny, malty, and sulfur (Table 1). Intensity of each reference material was rated on 15 cm universal scales, with intensity ratings of 0

70 eV; mass range, 35−350 amu; electron multiplier voltage (Autotune +200 V); scan rate, 5.2 scans/s. Headspace-Solid Phase Microextraction (H-SPME). For identification of highly volatile odorants, fish sauce sample (1 g) was placed in a 22 mL headspace vial and spiked with 5 μL of internal standards (1.15 μg/μL of 2-ethyl butyric acid, 1.32 μg/μL of 2-methyl3-heptanone, and 1.13 μg/μL of 2,4,6-trimethylpyridine in methanol). The sample was sealed immediately with a PTFE-lined septum and preincubated at 40 °C for 10 min with agitation (500 rpm, 60 s on, 10 s off) using a MPS2 autosampler (Gerstel, Inc.). Then, a SPME fiber (50/30 μm DVB/carboxen/polydimethylsiloxane fiber; Supelco) was exposed to the vial headspace for 20 min. Then fiber was desorbed by splitless injection (injector temperature 260 °C; splitless time 4 min; vent flow 50 mL/min) into GC−MS system with the same settings as described above. The GC oven temperature was programmed from 35 to 225 °C at a rate of 6 °C/min with an initial and final holding time of 5 and 20 min, respectively. Compound Identification. Compounds were tentatively identified by GC−O by comparing their retention indices (RIs) and odor properties with those of authentic reference compounds. Compounds were positively identified by comparing their RIs, odor properties, and mass spectra with authentic reference compounds. A homologous series of n-alkanes was used for determination of RIs according to the method of van den Dool and Kratz.22 Compound Quantification. For quantification of methanethiol (no. 2), the concentration of the synthesized [2H3]-methanethiol was first determined by analysis using a 6890 GC−flame photometric detector (GC−FPD) system (Agilent Technologies, Inc.). Headspace volume of methanethiol (50−300 μL) or [2H3]-methanethiol (5 mL) were drawn by a gastight syringe and injected into 5 mL of methanol in 22 mL sealed headspace vials. Each methanolic sample (1 μL) was injected using hot split mode into an Rtx-Wax column (15 m × 0.53 mm i.d.; 1 um film; Restek). The concentration of [2H3]-methanethiol (13.7 μg/mL) in the stock solution was determined by comparing its peak area with a standard curve generated from known concentrations of methanethiol (seven levels, from 7.4 to 48.9 μg/mL). For analysis, fish sauce sample was spiked with an appropriate headspace volume of [2H3]-methanethiol and quantified by H-SPME as described above. Other compounds in low abundance were quantified by H-SPME and DSE-SAFE as described above, respectively. Highly abundant compounds were analyzed by DSE as described above, except that a 5 g sample was first mixed with isotopically labeled compounds. Then sample was diluted and mixed well with 25 mL of odorless-distilled water. The sample was extracted three times with diethyl ether (60 mL total volume). The pooled solvent extract (60 mL) was evaporated to 10 mL using a Vigreux column in a 43 °C water bath, dried over 2 g of anhydrous Na2SO4, and further concentrated to 200 μL under nitrogen gas stream. The selected ions and response factors of the target (unlabeled) and labeled compounds were determined by H-SPME-GC−MS and GC− MS as described above (Table 2). Response factor for each compound was determined by measuring five levels of defined mixtures of the respective labeled and unlabeled compounds. Concentration for each compound in fish sauce was calculated using the response factor and area of the selected ion and expressed as micrograms per kilogram (ppb). Odor-activity values (OAVs) were calculated by dividing the concentrations by the respective odor thresholds in water, obtained from literature. Sensory Analysis. Comparison of Fish Sauce and the Complete Fish Sauce Aroma Model. Aroma profiling was done by descriptive sensory analysis as previously described.21 One gram of fish sauce was dissolved in the matrix (10 mL of 0.1 M phosphate buffer, pH 5.6, 25% NaCl) and presented to panelists in 125 mL Nalgene PTFE wash bottles (Fisher, Pittsburgh, PA, USA) with siphon tubes removed from the caps. Bottles were covered with aluminum foil to prevent any visual bias and labeled with random three-digit codes. The 11-member panel (five males, six females, 20−48 years old) were asked to identify and define descriptive terms of fish sauce sample and to determine appropriate aroma references. Odor notes representing fish sauce samples were selected and included cheesy, potatoey, brothy/meaty,

Table 1. Attributes, References, and Rating for Sensory Descriptive Analysis of Fish Sauce attribute cheesy potatoey brothy/meaty fishy/briny malty sulfur

referencea 0.1 g of parmesan cheese 10 g of mashed potato 10 mL of a 1% yeast extract aqueous solution 10 mL of a 100 ppb trimethylamine aqueous solution 0.1 g of dark chocolate 2 g of boiled Brussels sprouts

ratingb 9 8 5 8 8 12

a

Samples were presented in 125 mL PTFE bottles. bAroma intensity was rated on a 15 cm universal scale, with intensity ratings of 0 and 15 corresponding to none and very, respectively.

and 15 corresponding to none and very intense, respectively (Table 1). Intensity ratings for the standards were used as calibration references for rating the intensities of the fish sauce samples and the aroma reconstitution models. The complete recombination models, consisting of 21 odorants, were prepared by mixing 1 mL of each stock solution (Table 3) in 10 mL of the aforementioned matrix based on the determined concentrations (Table 3). The panelists were asked to evaluate the intensities of each odor attribute in the fish sauce samples and models. Samples were evaluated in duplicate. Differences among samples were evaluated by analysis of variance with means separation using SPSS version 11 software (SPSS Inc., Chicago, IL). Omission Experiments. Omission models were prepared in the same matrix described for the recombination models, except that various groups of compounds, namely acids, aldehydes, sulfur compounds, and low OVA compounds, were omitted from each model (Table 3). Sensory comparison of the omission models to the recombination (complete) models was performed using the R-index by ranking method.23 Panelists were asked to rank the omission models on how different they were from the complete model, from most similar to least similar (n = 19; 5 males, 14 females). The degree of difference of each model compared against complete model was calculated using John Brown computations.24

■

RESULTS AND DISCUSSION Identification of Odor-Active Compounds. Odor-active compounds from two brands of premium Thai fish sauce (A and B) were isolated by DSE/SAFE. To avoid interferences during GC−O, the aroma extracts were first separated into acidic (AF) and neutral/basic (NBF) fractions. The 29 compounds with log3FD factor of ≥1 were detected by AEDA, consisting of 11 acidic compounds (Table 4) and 18 neutral/basic compounds (Table 5). The AEDA results for the two fish sauce samples were similar. Compounds in AF with high log3FD values included acetic acid (vinegar), butanoic acid (cheesy), 3-methyl butanoic acid (cheesy, sweaty), 4-hydroxy2,5-dimethyl-3(2H)-furanone (burnt sugar), 4-hydroxy-2-ethyl5-methyl-3(2H)-furanone (burnt sugar), and 2-phenylacetic acid (rosy, plastic). These were regarded as the predominant acidic odorants based on their high log3FD factor (≥6). Likewise, predominant odorants in the NBF (log3FD factor ≥3) were 3-methylbutanal (malty, dark chocolate), 3(methylthio)propanal (cooked potato), phenylacetaldehyde (rosy, plastic), and o-aminoacetophenone (musky, grape, stale). In a previous study, 2-methylpropanal, 2-methylbutanal, 2pentanone, 2-ethylpyridine, dimethyl trisulfide, 3-(methylthio)propanal, and 3-methylbutanoic acid were reported with high FD factors and were regarded as principal contributors to the aroma of Thai fish sauce.5 The results of the present study 2631

DOI: 10.1021/jf5061248 J. Agric. Food Chem. 2015, 63, 2628−2638

Article

Journal of Agricultural and Food Chemistry Table 2. Selected Ions (m/z) and Response Factors Used in Stable Isotope Dilution Assays no.a

ionb

compd trimethylamine methanethiol dimethyl sulfide 2-methylpropanal 2-methylbutanal 3-methylbutanal dimethyl trisulfide acetic acid 3-(methylthio)propanal 3,6-dimethyl-2-ethylpyrazine propanoic acid 2-methylpropanoic acid butanoic acid phenylacetaldehyde 3-methylbutanoic acid 3-(methylthio)-1-propanol 4-hydroxy-2,5-dimethyl-3(2H)-furanone o-aminoacetophenone skatole 2-phenylacetic acid 3-phenylpropionic acid

1 2 4 5 6 7 13 15 16 17 18 20 21 22 23 25 30 34 36 37 38

58 48 62 72 86 86 126 60 104 135 74 73 88 120 87 106 128 135 130 136 150

no.c I-1 I-2 I-3 I-4 I-7e I-7 I-13 I-15 I-16 I-17 I-18 I-20 I-21 I-22 I-23 I-25 I-30 I-34 I-36 I-37 I-38

labeled compd

ionb

Rf d

[ H9]-trimethylamine [2H3]-methanethiol [2H6]-dimethyl sulfide [2H2]-2-methylpropanal [2H2]-3-methylbutanal [2H2]-3-methylbutanal [2H6]-dimethyl trisulfide [2H3]-acetic acid [2H3]-3-(methylthio)propanal [2H5]-3,6-dimethyl-2-ethylpyrazine [2H5]-propanoic acid [2H2]-2-methylpropanoic acid [2H2]-butanoic acid [13C2]-phenylacetaldehyde [2H2]-3-methylbutanoic acid [2H3]-3-(methylthio) propanol [13C2]-4-hydroxy-2,5-dimethyl-3(2H)-furanone [2H3]-o-aminoacetophenone [2H3]-skatole [13C2]-2-phenylacetic acid [2H2]-3-phenylpropionic acid

66 51 68 74 88 88 132 63 107 141 79 75 90 122 89 109 130 138 133 138 152

0.57 0.96 4.27 2.13 0.96 1.45 1.14 0.99 1.70 0.72 1.28 0.35 1.20 2.03 0.50 1.07 0.95 1.65 0.37 0.76 0.96

2

a Numbers correspond to those in Tables 3−7 and Figures 1 and 2. bSelected ion. cThe letter “I” indicates isotopically labeled compound. dResponse factor. eA structurally similar compound was selected as the internal standard.

Table 3. Concentrations of High Purity Standard Compounds Used to Prepare Fish Sauce Models concn (μg/L) a

no.

15 18 20 21 23 37 38 5 6 7 22 2 4 13 16 25 1 17 30 34 36 a

compd

% purity

b

Group 1, Acidsc d acetic acid 99.9 propanoic acidd 99.4 2-methylpropanoic acid 99.5 butanoic acid 99.8 3-methylbutanoic acid 99.7 2-phenylacetic acid 99.3 3-phenylpropanoic acid 99.9 Group 2, Aldehydese 2-methylpropanal 98.0 2-methylbutanal 95.8 3-methylbutanal 96.7 phenylacetaldehyde 99.3 Group 3, Sulfur-Containing Compoundse methanethiol 99.5 dimethyl sulfide 99.6 dimethyl trisulfide 94.6 3-(methylthio)propanal 97.4 3-(methylthio)-1-propanol 99.6 Group 4, Low OAV Compoundse trimethylamine 99.9 3,6-dimethyl-2-ethylpyrazine 99.4 4-hydroxy-2,5-dimethyl-3(2H)-furanone 94.4 o-aminoacetophenone 99.8 skatole 99.6

A model

B model

4960000 99850 6450 131500 20050 2700 3600

4510000 56100 22400 308500 95000 20600 1800

654 376 1360 177

930 906 1720 184

62.5 41.1 18.4 388 710 0.860 1.32 2600 4.02 0.474

62.5 11.4 27.2 1138 648 1.34 5.54 2500 3.00 0.482

Numbers corresponded to those in Tables 2, 4−7. bPurity determined by GC−O and GC−FID analysis. cStock solution prepared in water. Compound was used neat. eStock solution prepared in methanol.

d

methyl-2(5H) furanone were identified for the first time as odor-active compounds in fish sauce. 4-Hydroxy-2,5-dimethyl-3(2H)-furanone has been found in many foods, including pineapple,25 strawberry,26 beef broth,27

confirm the importance of most of the above compounds, except for 2-pentanone and 2-ethylpyridine. Furthermore, 4hydroxy-2,5-dimethyl-3(2H)-furanone, 4-hydroxy-2(or 5)ethyl-5(or 2)-methyl-3(2H)-furanone, and 3-hydroxy-4,5-di2632

DOI: 10.1021/jf5061248 J. Agric. Food Chem. 2015, 63, 2628−2638

Article

Journal of Agricultural and Food Chemistry

Table 4. Predominant Acidic Odorants (Log3FD factor ≥1) in Fish Sauce Determined by Aroma Extract Dilution Analysis RIc no.

a

15 18 20 21 23 30 31 32 33 37 38

compd e

acetic acid propanoic acide 2-methylpropanoic acide butanoic acide 3-methylbutanoic acide 4-hydroxy-2,5-dimethyl-3(2H)-furanonee 4-hydroxy-2-ethyl-5-methyl-3(2H)-furanoneg 4-hydroxy-5-ethyl-2-methyl-3(2H)-furanoneg 3-hydroxy-4,5-dimethyl-2(5H)-furanoneg 2-phenylacetic acide 3-phenylpropanoic acide

odor

b

Wax

sour cheesy, fecal cheesy, Swiss cheese cheesy cheesy, sweaty burnt sugar burnt sugar burnt sugar curry, spicy rosy, plastic rosy

log3 FD factord Rtx-5 f

1433 1514 1550 1611 1654 2005 2056 2077 2173 2515 2590

nd nd 804 829 893 1078 1149 nd 1115 1272 1366

A

B

6 4 4 8 7 6 7 2 3 8 2

6 5 6 9 9 6 6 nd 5 8 1

a

Numbers correspond to those in Tables 2, 3, 5−7. bOdor quality as perceived during GC−O. cRetention indices were calculated from GC−O data; Wax, Rtx-Wax. dFlavor dilution (FD) factor determined on Rtx-Wax column. eCompound positively identified by comparing RIs, mass spectrum and odor properties with those of reference compound. fnd, not detected. gCompound tentatively identified by comparing RIs and odor properties with those of reference compound.

Table 5. Predominant Neutral/Basic Odorants (Log3FD Factor ≥1) in Fish Sauce Determined by Aroma Extract Dilution Analysis RIc no.a 5 6 7 10 11 12 14 16 17 19 22 25 27 28 29 34 35 36

compd g

2-methylpropanal 2-methylbutanale 3-methylbutanale (Z)-4-heptenalg 1-octen-3-oneg 2-acetyl-1-pyrrolineg (Z)-1,5-octadien-3-oneg 3-(methylthio)propanale 3,6-dimethyl-2-ethylpyrazinee (E)-2-nonenalg phenylacetaldehydee 3-(methylthio)-1-propanole 2-acetyl-2-thiazolineg (E)-β-damascenoneg 2-phenylethanole o-aminoacetophenonee indolee skatolee

Log3FD factord

odorb

Wax

Rtx-5

A

B

malty, dark chocolate malty, dark chocolate malty, dark chocolate crabby, stale, fishy mushroom popcorn metallic cooked potato roasted potato stale, hay rosy, plastic potato popcorn floral, apple sauce wine-like, rosy musky, grape, stale fecal fecal