Odorant Synergy Effects as the Cause of Fishy Malodors in Algal

Article pubs.acs.org/JAFC

Odorant Synergy Effects as the Cause of Fishy Malodors in Algal Marine Oils Raymond T. Marsili* and Charles R. Laskonis Marsili Consulting Group, Rockford University, Starr Science Building, Room 120, 5050 East State Street, Rockford, Illinois 61108, United States ABSTRACT: As unsaturated lipids oxidize, they form hydroperoxides, which are susceptible to further oxidation or decomposition to secondary reaction products including aldehydes, ketones, acids, and alcohols. While oxidation reactions of docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA) are responsible for fishy off-flavors in marine oils, gas chromatography-olfactometry (GC-O) and other types of analytical studies have failed to reveal which specific oxidation products are involved. Previous research (Marsili, R.T.; Laskonis, C. The importance of odourant synergy effects in understanding malodour problems in DHA and EPA products. Lipid Technol. 2014, 26 (2), 31−34) has indicated that fishy malodor may be caused by the presence of two lipid oxidation products, heptanal and (E,Z)-3,5-octadien-2-one. The aims of the present study are to provide experimental method details and offer further evidence that these two oxidation products are indeed the cause of fishy malodors. Initial GC-MS-O studies of marine oils with fishy malodors revealed numerous oxidation products, but none were characterized as fishy. However, when all sample volatiles were captured together and then desorbed simultaneously in GC-O experiments, the fishy malodor was evident, indicating odorant synergy effects were responsible. A simple, novel method was developed using an olfactometry detector as a fraction collector to trap various peaks in marine oil chromatograms. The nose cone of the olfactometry detector was replaced with a PDMS foam absorption tube at various times during GC analysis. Combinations of GC peaks were trapped on PDMS tubes, desorbed in a Gerstel thermal extractor (off-line), and sniffed. The combination of two analytes was found to cause fishy malodors: heptanal and (E,Z)-3,5-octadien-2-one. Purgeand-trap, solid phase microextraction (SPME), and headspace stir bar sorptive extraction (HSSE) sample preparation methods prior to GC-MS were investigated. All methods confirmed the combination of heptanal and (E,Z)-3,5-octadien-2-one as the cause of fishy odor. KEYWORDS: fishy, olfactometry, synergy, lipid oxidation products, algal marine oils, DHA, EPA

■

in fishy malodor development to establish improved stability tests and end points of quality deterioration. Several volatile components have been characterized in fish oil, in fish itself, and in fish oil enriched foods such as mayonnaise and milk.1 Sixty different volatiles comprising alkenals, alkadienals, alkatrienals, and vinyl ketones have been identified in fish oil enriched milk.2 The most potent odorants identified in this system by GC-O were 1-penten-3-one; (Z)-4heptenal; 1-octen-3-one; (Z)-1,5-octadien-3-one; (E,E)-2,4heptadienal; and (E,Z)-2,6-nonadienal. However, despite their low odor thresholds, none of the separated individual volatiles produced a fishy odor. Therefore, the researchers hypothesized that the fishy and metallic off-flavors were due to a combination of some of these potent odorants. 1-Penten-3-one may contribute to unpleasant off-flavors described as sharp-fishy in fish oil3 or rancid and plastic in fish oil enriched mayonnaise4 and fish oil enriched milk.3 The increase in fishy, metallic, and rancid off-flavors has been correlated to high concentrations of (E,E)-2,4-heptadienal in fish oil enriched mayonnaise.5 The perception of off-flavors has been correlated to the development of 1-penten-3-one; (E,E)-2,4-heptadienal; and (E,Z)-2,6-

INTRODUCTION Increasing evidence compiled over the past 30 years supports the nutritional benefits of dietary long-chain n−3 polyunsaturated fatty acids (n−3 PUFA). The main focus has been on eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA). Due to the high degree of unsaturation of EPA and DHA, triglycerides rich in these fatty acids undergo rapid oxidative deterioration, which seems to be a particularly pronounced problem in emulsions and complex food systems. As lipids oxidize, they form hydroperoxides that are susceptible to further oxidation or decomposition to secondary reaction products including aldehydes, ketones, acids, and alcohols. The oxidative stability of long-chain PUFA and DHA containing fish and algal oils varies widely according to their fatty acid composition, the physical and colloidal states of the lipids, the contents of tocopherols and other antioxidants, and the presence and activity of transition metals. Chemicals that form from DHA and EPA oxidation are largely responsible for malodors ranging from metallic, green, painty, burnt, glue-like, earthy, pungent, deep-fried, and fishy. The fishy malodors may be the most offensive off-flavors in marine oils. Gas chromatography−olfactometry (GC-O) and other types of analytical studies have failed to reveal which specific oxidation products are responsible for fishy malodors. The current high interest in using long-chain PUFA oils warrants an improved understanding of the chemicals involved © 2014 American Chemical Society

Received: Revised: Accepted: Published: 9676

March 19, 2013 September 16, 2014 September 18, 2014 September 18, 2014 dx.doi.org/10.1021/jf502252q | J. Agric. Food Chem. 2014, 62, 9676−9682

Journal of Agricultural and Food Chemistry

Article

nonadienal in fish oil enriched milk.6 The volatiles (Z)-4heptenal and (E,Z)-2,6-nonadienal have been associated with fishy off-flavors in oxidized fish oil.7 One study concluded that the volatile secondary oxidation products (E,Z)-2,6-nonadienal, 1-penten-3-one, (Z)-4-heptenal, and (E,E)-2,4-heptadienal, while individually not responsible for fishy malodors, were the cause of fishy malodors as a result of a combinatory effect.1 They did not include 1,5-octadien-3-one in their sensory descriptive experiments even though it had been previously reported as a potent odorant in fish oil because it was not detected in their fish oil enriched milk having a strong fishy offflavor. Researchers studying fish oil enriched mayonnaise concluded that the following ketones probably contribute to fishy malodors: 1-penten-3-one, (E)-3-penten-2-one, 3-heptanone, (Z)-1,5-octadien-3-one, 1-nonen-3-one, 2-nonanone, (E,Z)-3,5octadien-2-one, and (E,E)-3,5-octadien-2-one.4 Furthermore, they surmise that (Z)-1,5-octadien-3-one is probably the most potent fishy odorant due its extremely low threshold value (0.00003 mg/kg in oil) and characteristic geranium-like odor.2 While the most potent odorants associated with DHA and EPA oxidation have been identified, demonstration of correlation between their presence and concentration levels with fishy malodor development is lacking. Assuming that the most potent odorants are the cause of fishy malodor could be erroneous. Few if any GC-O supportive studies have investigated odorant synergy effects as the cause of fishy malodors. GC-O is traditionally used to identify individual odor-active GC fractions. Potential synergy effects cannot be evaluated when single compounds are evaluated over time. Many synergistic effects are known to occur between single compounds in complex food matrices.8 Combinations of single substances can produce enhancing or masking interactive effects. For example, researchers investigating the aging of beer reported a change in sensory perception of components when evaluated singularly versus when evaluated in combination. Specifically, (E)-2-nonenal was evaluated by panelists as cardboard-like and (E,Z)-2,6-nonadienal was perceived as cucumber-like when the single chemicals were subjected to GC-O. However, the combined effect of the two compounds generated a sweet fruity flavor perception distinctly different from the single substances alone.8 The researchers concluded that the new sensory perception of the combination appeared to be related to both the absolute and relative concentrations of the two compounds. Numerous sample preparation analytical techniques prior to GC-O have been applied to the analysis of fishy off-flavors, including dynamic headspace with Tenax GR trapping,4,9 simultaneous steam-distillation-solvent extraction (SDE using a Likens−Nickerson apparatus),10 solid-phase microextraction (SPME) using a Carboxen/PDMS fiber,11 and static headspace.12 Headspace sorptive extraction (HSSE) with 2 cm PDMS Twisters was used in the present study.13 Naude and Rohwer developed a novel GC-O approach to studying odorant synergy effects. In their method, sorptive extraction of aroma volatiles is accomplished using multichannel open tubular silicone (polydimethylsiloxane [PDMS]) rubber traps (MCTs) for fraction collection of aroma compounds.14 GC peaks from milk samples were collected on MCTs that were fitted into the GC FID (FID and flame gases were turned off) for fraction collection purposes. Selection of peaks to trap was made on a carefully timed

basis after preanalyzing samples to determine significant odorants. The captured chemicals on the MCT were transferred to an off-line portable heating device (with a nitrogen flow of 20 mL/min and a temperature of 130 or 160 °C).The aroma of the thermally released compounds could then be evaluated by panelists. As aroma eluted from the MCT, it was sniffed by a team of six evaluators. Using this approach, researchers determined that a synergistic combination of 2heptanone and 2-nonanone was responsible for a pungent cheese, sour milk-like aroma detected in UHT milk. Interestingly, when smelled individually, 2-heptanone has a fruity-floral-soapy aroma and 2-nonanone has a fruity-sweet aroma. In the present study, an olfactometry detector was used as a fraction collector to trap various peaks in marine oil chromatograms. The nose cone of the olfactometry detector was replaced with a Gerstel PDMS foam absorption tube at various precise times during GC analysis. Combinations of GC peaks were trapped on PDMS tubes, desorbed in a Gerstel thermal extractor (off-line) at 160 °C, and sniffed by aroma evaluators to determine the combined synergistic perception of mixtures of aroma compounds present in marine oil contaminated with fishy malodors. This study was conducted to (1) determine which chemicals cause fishy malodors and (2) provide an analytical methodology for studying odor synergy effects.

■

MATERIALS AND METHODS

Experimental Design. This study was divided into several experiments. Algal oils with varying degrees of fishy malodor were tested by GC-MS. Sample preparation techniques investigated included purge-and-trap, SPME, and HSSE GC-MS. HSSE was selected as the primary analytical technique because it had the highest sensitivity and provided excellent replicate determinations. In the next step, GC-O experiments were performed on samples with strong fishy malodors. When no single peak was found to contribute fishy malodors, synergy effects between two or more odorants were investigated. First, all sample analytes were trapped on a PDMS foam trap, desorbed simultaneously with the Gerstel Thermal Desorber, and sniffed by panelists. A strong fishy odor was detected by the panel of sniffers, suggesting that odorant synergy effects could be responsible for the fishy malodor. Attempts were made to isolate the odorants responsible for the fishy malodor by trapping odorants in various regions of the chromatogram and then evaluating odor characteristics of the trapped volatiles after simultaneous desorption in the Thermal Desorber. Using this process, only two chemicals heptanal and (E,Z)-3,5-octadien-2-onewhen present in the trapped volatiles were found essential for generating fishy malodor. To confirm that these two odorants were all that was necessary to generate fishy malodor, only these two chemicals were trapped and subjected to sniffing with the thermal extractor. Omission experiments were also conducted. In omission experiments all volatiles in the chromatogram except heptanal and (E,Z)-3,5-octadien-2-one were trapped and then desorbed in the Thermal Desorber to see if the fishy odor was present. Experiments were then conducted to confirm the identity of these two chemicals and to check for the possibility of coelution of contaminating peaks. Identification of (E,Z)-3,5-octadien-2-one is based on retention index values using a DB-5 column and olfactometry odor characteristics compared to literature results, as well as mass spectra (EI and CI). Accurate mass measurements for the (E,Z)-3,5octadien-2-one peak were made using a high resolution time-of-flight mass spectrometer (Leco Pegasus GC-HRT). Marine Oil Samples. Fermentor grown microalgae samples were processed to extract the DHA-rich oil. The finished commercial product is a clear, amber-colored oil rich in DHA. Six algal oil samples were analyzed (two with no fishy malodor and four with fishy malodor). A commercial microencapsulated powder version of the 9677

dx.doi.org/10.1021/jf502252q | J. Agric. Food Chem. 2014, 62, 9676−9682


Article

Figure 1. 2.00 g of marine oil extracted with 20 mm × df = 0.5 mm (47 μL) Twister stir bar (PDMS) in headspace for 2 h at 50 °C while stirring with Teflon-coated micro stir bar. The same sample was analyzed six times. Figure shows how peaks contributing to fishy malodor were isolated and identified. Off-line thermal desorption with a Gerstel thermal extractor was used for odor characterization. Peaks that are not numbered are background siloxane impurities in PDMS. Identity of numbered peaks and retention indices [in brackets] are as follows: (1) 1-penten-3-ol [682]; (2) 2-pentenal [750]; (3) 1-pentanol [767]; (4) hexanal [800]; (5) 2-hexenal [854]; (6) 1-hexanol [877]; (7) heptanal [901]; (8) 2,4-hexadienal [912]; (9) methylcyclohexane [933]; (10) 1,1-diethoxy-3-methylbutane [954]; (11) 2-heptenal [958]; (12) benzaldehyde [964] and 1-heptanol [967] coeluting; (13) 1-octen-3-one [970]; (14) 3,5,5,-trimethyl-1-hexene [975]; (15) (E,Z)-2,4-heptadienal [996]; (16) octanal [1004]; (17) (E,E)-2,4heptadienal [1013]; (18) 2-ethyl-1-hexanol [1028]; (19) limonene [1030]; (20) eucalyptol [1032]; (21) 2-octenal [1063]; (22) 1-octanol [1074] and (E,Z)-3,5-octadien-2-one [1081] coeluting; (23) 2-nonanone [1090]; (24) (E,E)-3,5-octadien-2-one [1098]; (25) nonanal [1104]; (26) (E,Z)2,6-nonadienal [1146]; (27) 2-nonenal [1150]; (28) decanal [1207]; (29) 2-ethylhexyl acrylate [1235]; (30) 2-decenal [1253]; (31) 2-undecanone [1291]; (32) 2-dodecenal [1469]. Reprinted with permission from ref 19. Copyright 2014 John Wiley & Sons Ltd. DHA oil sample with a strong fishy malodor was also analyzed. The samples were sourced from a major supplier of DHA oils. Algal oil is composed mainly of triglycerides. The composition is approximately 35−45% DHA and other triglycerides, including myristic acid (13−20%), palmitic acid (12−25%), oleic acid (10− 25%), lauric acid (2−6%), and capric acid (1%). Heptanal, octanal, nonanal, decanal, 2-decenal, and 2-undecenal are formed from autoxidation of oleic acid, an n−9 monounsaturated fatty acid. The (E,Z)-3,5-octadien-2-one observed in the samples is derived from DHA oxidation. Chemicals. Standards were obtained to assist in the identification of (E,Z)-3,5-octadien-2-one. A standard of (E,E)-3,5-octadien-2-one (CAS No. 30086-02-3) was purchased from Penta Manufacturing Company (Fairfield, NJ). A standard consisting of a mixture of (E,E)3,5-octadien-2-one and (E,Z)-3,5-octadien-2-one (90% v/v and 10% v/v, respectively) was obtained as a gift from Allard Cossé, Crop

Bioprotection Research Unit, USDA−ARS, Nat. Center for Agricultural Utilization Research (NCAUR), Peoria, IL. While this standard assisted in the identification of (E,Z)-3,5-octadien-2-one, it was not useful for conducting recombination/model system studies. Recombination studies, which involve spiking odorless control oils with odorant suspects, are frequently used for sensorial confirmation of which odorants are responsible for the odor of a product. Since (E,Z)3,5-octadien-2-one was present in minor amounts compared to the major component of (E,E)-3,5-octadien-2-one, which has a strong geranium-like odor, the standard from Allard Cossé could not be used for recombination studies. A pure standard of (E,Z)-3,5-octadien-2one was not commercially available. HSSE GC-MS Analysis. Samples were analyzed by headspace sorptive extraction (HSSE) using the Gerstel Twister, length = 20 mm, df = 0.5 mm, 47 μL PDMS coating (Part No. 011444-001-00). Two grams of algal oil sample was placed in a 20 mL glass GC vial with a 10 9678



Article

Table 1. Olfactometry Results for Typical Algal Oil with Fishy Malodora trial

region of chromatogram trapped (min)

odor characteristic

comment

T1 T2 T3 T4 T5 T6 T7 T8

7.62−16.60 9.30−15.00 10.30−12.70 9.30−12.70 9.30−12.30 9.30−9.75 and 12.35−12.50 9.75−12.50 9.30−12.35

fishy fishy not fishy fishy not fishy fishy not fishy not fishy

fishy odorants trapped fishy odorants trapped at least one contributor to fishy malodor excluded from trapping heptanal suspected as contributor (E,Z)-3,5-octadien-2-one suspected as contributor heptanal and (E,Z)-3,5-octadien-2-one confirmed contributors heptanal excluded; confirms heptanal as contributor (E,Z)-3,5-octadien-2-one excluded; confirms its contribution to fishy odor

a

Odor of combinations of chemicals evaluated off-line after trapping on PDMS foam followed by thermal desorption. Six evaluators used per test; all samples tested in triplicate. mm Teflon coated micro stir bar. A paperclip was attached to the septum, and the Gerstel Twister was suspended from the paperclip in the headspace of the vial. The sample was heated to 50 °C. Extraction was continued for 2 h at 50 °C while the Teflon micro stir bar was stirred at 350 rpm; the Twister attached to the paperclip remained stationary. The Twister was then removed from the vial and subjected to thermal desorption using the Gerstel TDU. Samples were tested in triplicate. Concentrations of the two oxidation productsheptanal and (E,Z)3,5-octadien-2-onewere measured in samples using a method-ofadditions HSSE technique. A nonfishy algal oil sample was spiked with varying levels of heptanal and (E,Z)-3,5-octadien-2-one to develop a standard calibration curve. Since a pure standard of (E,Z)-3,5octadien-2-one was not available, the standard containing a mixture of 10% (v/v) (E,Z)-3,5-octadien-2-one and 90% (v/v) (E,E)-3,5octadien-2-one was used for the calibration. Calibration curves were prepared by adding 0, 40, 100, 200, and 500 ppb heptanal (R2 = 0.979) and 0, 4.8, 12, 23.9, and 60 ppb (E,Z)-3,5-octadien-2-one (R2 = 0.969). Thermal Desorption Parameters. The Gerstel cooled injection system (CIS), a programmed temperature vaporization (PTV) type inlet, was used in the solvent vent mode at 50 mL/min; cryofocusing was at −100 °C followed by heating at 12.0 °C/s to 280 °C and then holding at 280 °C for 3.00 min. The Gerstel cooled injection system (CIS) liner (Gerstel part number 012442-010-00) used was quartz wool. Gerstel TDU parameters were as follows: initial temperature was 30 °C, followed by heating at 60.0 °C/min to 280 °C and holding at 280 °C for 4.00 min. GC Parameters. The column used was DB-5MS, 30 m × 0.25 mm × 0.25 μm (Agilent Technologies, catalog number 122-5532).The initial column temperature was 40 °C for 3 min, followed by heating at 10 °C/min to 270 °C for 3 min. Carrier gas flow was 1.0 mL/min through the column; the inlet was set to splitless mode. The flow was split 1:1 between the MS detector and the olfactometry detector. To check for possible coeluters, samples were reanalyzed on a Leco Pegasus HT GC-TOFMS instrument (sampling rate at 40 spectra/ second) using the same method and GC conditions. The peak deconvolution capability of this instrument showed that 1-octanol tended to coelute on the front leading edge of the (E,Z)-3,5-octadien2-one peak. Subsequent experiments showed that 1-octanol was not a contributor to the fishy odor. No other coelution of chemicals with heptanal or (E,Z)-3,5-octadien-2-one peaks was observed. Additional Dynamic Headspace Analyses. Several samples were also analyzed by dynamic headspace GC-MS using either PDMS foam traps (Gerstel catalog number 013758-005-00) or Tenax TA traps (Gerstel catalog number 013741-005-00), both purchased from Gerstel. The PDMS foam traps, which can be heated up to 300 °C, were identical to those used to collect volatiles with the olfactometry detector (see Olfactometry Experiments). Two grams of algal oil and a Teflon stir bar (stirring at 350 rpm) were placed in a desorption tube with the sparging needle just above the surface of the oil. The dynamic headspace glassware was purchased from Scientific Instrument Services (Ringoes, NJ). The sample was heated to 50 °C for 20 min with a nitrogen purge gas flow rate of 50 mL/min. Results were not significantly different from HSSE results and are not reported here.

The same TDU program as used for HSSE was used to desorb volatiles from the traps. Olfactometry Experiments. All samples were subjected to GC-O experiments to assess odor quality and intensity of each individual peak. Four trained panelists, three females ages 24−35 years old and one male age 23, were used for GC-O experiments. The Gerstel ODP3 was used at 160 °C. Panelists sniffed chromatogram effluents for the entire GC run time of the sample (29.0 min). Even though samples had a fishy malodor, none of the individual peaks shown in Figure 1 had a fishy smell. Studying synergistic olfactometry responses from combinations of odorants was more challenging. For example, the strategy used for collecting specific regions of the chromatogram is illustrated in Figure 1 and Table 1 (column titled “region of chromatogram trapped”). Initially (trial 1), all volatiles in the sample were trapped on a polydimethylsiloxane (PDMS) foam sorbent tube and desorbed off-line to characterize the odor of the entire sample. In trial 2, a wide swath of the chromatogram was trapped onto a PDMS tube. The PDMS foam trap is an open porous foam packing used for analyte trapping. PDMS strongly retains nonpolar analytes. The trap is then inserted into a Gerstel thermal extractor (off-line), which is preheated to 160 °C with a nitrogen flow gas at 50 mL/min. Evaluators (four) consecutively sniffed the effluent from the PDMS foam tube for 30 s after tube was placed in the thermal desorber to assess the overall odor impact (fishy or not fishy). If fishy aroma is not detected in the desorbed volatiles mixture but the algal oil sample has a fishy malodor, then the analytical method may not be trapping the analytes responsible for the fishy malodor and alternate sample preparation methods or traps should be considered. However, this was not the case. The “fishy” aroma was noted in all initial trials of complaint samples. The sample was reanalyzed (trial 3), but this time fewer chromatographic peaks were collected. If this sample fraction demonstrated a fishy aroma, then the sample was analyzed again with even fewer peaks collected. In this process, either the front or back end of the chromatogram was sequentially omitted from collection. This process continues until the sample sniffed at the thermal extractor shows no fishy odor, indicating that an important fishy odor contributor has not been captured. Monitoring the MS TIC trace on the computer screen during GC-MS analysis allows for accurate determination of times to attach and detach the PDMS tubes for trapping. To test the accuracy of the trapping technique to collect the desired peaks of interest, one PDMS tube was thermally desorbed after trapping only heptanal and (E,Z)-3,5-octadien-2-one and analyzed by GC-MS. Its chromatogram showed that the two major contributors to fishy malodor were indeed trapped with no significant background contamination from other peaks. When potential fishy odor causing candidates have been identified, only they are collected on the trap and desorbed off-line for aroma evaluation (trial 6). If the fishy odor is detected, these volatiles are confirmed as the cause of fishy odor development in the oil. Results are then confirmed in omission experiments: by reanalyzing the sample after all peaks are trapped except the ones suspected of causing the fishy odor. 9679



Article

Figure 2. EI and CI spectra for (E,Z)-3,5-octadien-2-one on Leco Pegasus GC-HRT.

■

RESULTS AND DISCUSSION A typical HSSE chromatogram of algal oil with a strong fishy aroma is shown in Figure 1 (trial 1). Approximately 32 significant odor-active peaks were detected. None of the chemicals in singularity imparts a fishy malodor to the sample, even though many have high odor activity values (OAVs), the ratio of the chemical concentration in the food to its threshold concentration in that food. No unusual chemicals were detected; most are expected lipid oxidation products that have been previously reported in algal oils. Recent research has shown that OAV measurements for predicting the perceived intensity of odorants and their contribution to their overall aroma are not good indicators of the actual sensory experience.15

As indicated in Figure 1 (trial 6) and Table 1, the combination of heptanal and (E,Z)-3,5-octadien-2-one when smelled together has been shown to cause fishy malodors. The results from olfactometry assessment show that the aromas of the individual oxidation products heptanal and (E,Z)-3,5octadien-2-one are distinctly different from the aroma of the two compounds when evaluated in combination. With the offline sniffing capability of the thermal extractor it is possible to use a team of several aroma evaluators for each sample that is thermally desorbed. Using PDMS foam as the trapping medium is important because, like the MCTs used by Naude and Rohwer,14 it has low in-line backpressure and does not cause peak shifting in 9680



Article

diminish a food’s perceived aroma. Even more remarkably, it can result in the emergence of a strikingly different sensory perception, completely unrelated to that of the individual compounds alone. While GC-O experiments have resolved numerous flavor and off-flavor issues in the past, the limitations of methods that only assess odor contribution from individual components need to be reconsidered. Odorant behavior in mixtures demands more studying. Studying synergy effects in olfactometry experiments should be a significant future trend. GC-O experiments are useful for determining which aroma compounds most likely make a contribution to the odor of a food. However, these experiments should be supplemented with sensory work and olfactometric synergy studies. When traditional GC-O experiments fail to detect the malodor present in the food sample, the researcher should attempt to determine if (1) the analytical sample preparation method is not collecting/trapping the offending odorant; (2) the offending odorant is being trapped efficiently but its concentration level is below the detection threshold of the analytical (MS) detector; (3) the offending odorant is being extracted from the sample but is thermally labile and is decomposing in the injector; or (4) the perceived malodor of the sample is due to odorant synergy effects related to smelling two or more odorants in combination. An analytical technique for studying odorant synergy effects has been presented as well as an example of how it can be applied. Heptanal and (E,Z)-3,5-octadien-2-one appear to be significant contributors to fishy malodor in algal oils. It is possible that some additional higher boiling point compounds, not isolated in our methods, could contribute to fishy malodor as has been recently reported.18 However, according to our present study, these higher molecular weight compounds do not appear to be as significant to fishy off-flavor as heptanal and (E,Z)-3,5-octadien-2-one, since simple headspace extractions by HSSE and SPME at relatively low temperature (50 °C) were shown to extract strong fishy odors from all samples. This article provides more details of the analytical methods and techniques used to identify odorants in our previous algal oil research.19 Confirmation of the roles of heptanal and (E,Z)-3,5octadien-2-one in causing fishy odors should be conducted using sensory paneling of odorless oil samples spiked with pure standards of these compounds.

chromatograms. Peak retention times when the PDMS trap is attached and when it is not attached are identical. Identification of heptanal in samples is based on mass spectra (EI), rentention index (compared to pure heptanal standard using DB-5 column), and olfactometry odor characteristics compared to pure standard. Identification of (E,Z)-3,5octadien-2-one is based on retention index values using a DB-5 column and olfactometry odor characteristics compared to literature results, as well as mass spectra (EI and CI). Accurate mass measurements for the (E,Z)-3,5-octadien-2-one peak were made using a high resolution time-of-flight mass spectrometer (Leco Pegasus GC-HRT). The EI spectrum for (E,Z)-3,5-octadien-2-one indicates the molecular ion at m/z 124.08831; the formula search indicated C8H12O with a mass accuracy of 0.38 ppm. The CI spectrum for (E,Z)-3,5-octadien2-one shows a strong protonated molecular ion at m/z 125.09609 that confirmed the molecular weight of the compound. The accurate mass measurement of the CI molecular ion also suggested C8H12O with a mass accuracy of −0.007 ppm. Figure 2 shows accurate mass results and spectra. Instrument parameters for the Pegasus GC-HRT were as follows: mode, high resolution (R = 25,000); mass range, 40− 400 m/z (EI), 60−510 m/z (CI); acquisition rate, 6 spectra/s; electron energy: −70 eV (EI), −140 eV (CI); and reagent gas for CI, 5% ammonia in methane at 0.8 mL/min. Table 2 shows that concentrations of both heptanal and (E,Z)-3,5-octadien-2-one are higher in complaint fishy malodor Table 2. Concentrations of Heptanal and (E,Z)-3,5Octadien-2-one in Algal Oil Samples with and without Fishy Malodor and DHA Powder with Strong Fishy Malodora sample description oil #1a: not fishy oil #2a: not fishy oil #3a: medium fishy odor oil #3b: medium fishy odor oil #4a: medium fishy odor oil #4b: strong fishy odor DHA microencapsulated powder: strongest fishy odor

heptanal (μg/kg)

(E,Z)-3,5-octadien-2one (μg/kg)

1,715 1,175 1,695 2,025 5,141 3,260 10,226

6 9 49 115 58 246 3,127

a

Samples analyzed by HSSE GC-MS. Method-of-additions calibration used.

■

samples compared to samples lacking the fishy aroma defect and tend to be highest in the most offensive sample (the DHA microencapsulated powder sample). Octanol, which tended to coelute on the front end of the (E,Z)-3,5-octadien-2-one peak with the chromatographic conditions used, was investigated as a potential contributor to fishy malodor; it did not make a significant impact on fishy malodor. (E,E)-3,5-Octadien-2-one and chemicals other than heptanal and (E,Z)-3,5-octadien-2-one did not contribute significantly to the fishy malodor. Future studies will involve synthesizing (E,Z)-3,5-octadien-2one in order to use it along with heptanal in recombination/ model system studies to assess odor threshold levels required to generate fishy malodors in marine oils. Model system studies have been shown to be highly beneficial for studying combinations of odorants in food products.16,17 Synergy effects can impart unexpected and potent aroma changes to foods. For example, synergy can either enhance or

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

REFERENCES

(1) Venkateshwarlu, G.; Let, M. B.; Meyer, A. S.; Jocobsen, C. Modeling the sensory impact of defined combination of volatile lipid oxidation products on fishy and metallic off-flavors. J. Agric. Food. Chem. 2004, 52, 1635−1641. (2) Venkateshwarlu, G.; Let, M. B.; Meyer, A. S.; Jacobsen, C. Chemical and olfactometric characterization of volatile flavor compounds in a fish oil enriched milk emulsion. J. Agric. Food Chem. 2004, 52, 311−317. (3) Grosch, W. Low-MW products of hydroperoxide reactions. 1987. In Autoxidation of Unsaturated Lipids; Chan, H. W. S., Ed.; Academic Press: London, U.K., 1987; pp 95−139. (4) Hartvigsen, K.; Lund, P.; Hansen, L. F.; Hølmer, G. Dynamic headspace gas chromatography/mass spectrometry characterization of

9681



Article

volatiles produced in fish oil enriched mayonnaise during storage. J. Agric. Food Chem. 2000, 48, 4858−4867. (5) Jacobsen, C. Sensory impact of lipid oxidation in complex food systems. Fett/Lipid 1999, 101, 484−492. (6) Let, M. B.; Jacobsen, C.; Frankel, E. N.; Meyer, A. S. Oxidative flavour deterioration of fish oil enriched milk. Eur. J. Lipid Sci. Technol. 2003, 105, 518−528. (7) Karahadian, C.; Lindsay, R. C. Evaluation of compounds contributing characterizing fishy flavors in fish oils. J. Am. Oil. Chem. Soc. 1989, 66, 953−960. (8) Hermann, M. B.; Klotzbucher, B.; Wurzbzcher, M. A new validation of relevant substances for the evaluation of beer aging depending on the employed boiling system. J. Inst. Brewing 2010, 116 (1), 41−48. (9) Refsgaard, H. H. F.; Haahr, A. M.; Jensen, B. Isolation and quantification of volatiles in fish by dynamic headspace sampling and mass spectrometry. J. Agric. Food. Chem. 1999, 47, 1114−1118. (10) Le Guen, S.; Prost, C.; Demaimay, M. Critical comparison of three olfactometric methods for the identification of the most potent odorants in cooked mussels (Mytilus edulis). J. Agric. Food Chem. 2000, 48, 1307−1314. (11) Shen, Z.; Mann, M. A.; Sanguansri, L.; Cheng, L. J. Oxidative stability of microencapsulated fish oil powders stabilized by blends of chitosan, modified starch, and glucose. J. Agric. Food Chem. 2010, 58, 4487−4493. (12) Milo, C.; Grosch, W. Detection of odor defects in boiled cod and trout by gas chromatography-olfactometry of headspace samples. J. Agric. Food Chem. 1995, 43, 459−462. (13) Marsili, R. Analysis of musty microbial metabolites by stir bar sorptive extraction. In Flavor, Fragrance and Odor Analysis, 2nd ed.; Marsili, R., Ed.; CRC Press: Boca Raton, FL, 2012; pp 63−91. (14) Naude, Y.; Rohwer, E. R. The olfactometric analysis of milk volatiles using a novel gas-chromatography-based method: A case study in synergistic perception of aroma compounds. In Flavor, Fragrance and Odor Analysis, 2nd ed.; Marsili, R., Ed.; CRC Press: Boca Raton, FL, 2012; pp 93−110. (15) Audouin, V.; Florence, B.; Vickers, Z. M.; Reineccius, G. A. Limitations in the use of odor activity values to determine important odorants in foods. In Gas Chromatography-Olfactometry: The State of the Art; Leland, J. V., Schieberle, P., Buettner, A., Acree, T. E., Eds.; American Chemical Society: Washington, DC, 2001; pp 156−171. (16) Marsili, R.; Miller, N.; Kilmer, G. J.; Simmons, R.E. Identification and quantitation of the primary chemicals responsible for the characteristic malodor of beet sugar by purge and trap GC-MSOD techniques. J. Chromatogr. Sci. 1994, 32, 165−171. (17) Marsili, R. T.; Miller, N. Determination of major aroma impact compounds in fermented cucumbers by solid-phase microextractiongas chromatography-mass spectrometry-olfactometry detection. J. Chromatogr. Sci. 2000, 18, 307−314. (18) Hammer, M.; Schieberle, P. Model studies on the key aroma compounds formed by an oxidative degradation of omega-3 fatty acids initiated by either copper(II) Ions or lipoxygenase. J. Agric. Food Chem. 2013, 61 (46), 10891−10900. (19) Marsili, R. T.; Laskonis, C. The importance of odourant synergy effects in understanding malodour problems in DHA and EPA products. Lipid Technology 2014, 26 (2), 31−34.

9682


Odorant Synergy Effects as the Cause of Fishy Malodors in Algal

Recommend Documents