Characterization of the Key Aroma Constituents in Fry Breads by Means of the Sensomics Concept

The key aroma constituents in the volatile fractions isolated FROM two differently processed fry breads by solvent-assisted flavor evaporation were characterized by an aroma extract dilution analysis (AEDA). Twenty-two compounds were identified with flavor dilution (FD) factor ranges of 2–516. Among them, 13 compounds (FD ≥ 16) were quantified by stable isotope dilution assays and analyzed by odor activity values (OAVs). Of these, 11 compounds had OAVs ≥ 1, and the highest concentrations were determined for δ-decalactone and 2,3-butanedione. Two recombination models of the fry breads showed similarity to the corresponding fry breads. Omission tests confirmed that aroma-active constituents, such as δ-decalactone (oily/peach), 2-acetyl-1-pyrroline (roasty/popcorn-like), 3-methylbutanal (malty), methional (baked potato-like), 2,3-butanedione (buttery), phenyl acetaldehyde (flowery), (E,E)-2,4-decadienal (deep-fried), butanoic acid, and 3-methylbutanoic acid, were the key aroma constituents of fry bread. In addition, 3-methoxy-4-vinylphenol (smoky) and 4-hydroxy-2,5-dimethyl-3(2H)-furanone were also identified as important aroma constituents of fry bread.


Introduction
Fry bread or scone is Native American bread widely consumed in the United States. Fry bread is produced from frozen or unfrozen flat dough which is fried or deep-fried in oil, shortening or lard. Fry breads' formulation is quite similar to that of bread rolls: flour, sugar, powdered milk, salt, and water. While bread is simply a mixture of flour, water, yeast, and salt with or without butter in the right proportions, kneaded, fermented, and baked in an oven [1]; fry bread formula is devoid of yeast, and the dough is deep-fried at a temperature of 176.7 • C. The quality of fry bread is normally defined by its texture, color and flavor (i.e., the sum of the gustative and olfactory responses observed during food consumption) [2]. Among these qualities, the flavor of fry bread, as in other bread types, is one of the most important factors that determine its acceptance by consumers [3].
The flavor of bread as well as the key aroma compounds responsible for its characteristic flavor has been well documented [4][5][6]. Oftentimes the flavor of bread is brought about by the interaction of a large number of compounds which exhibit different olfactive characteristics. Some of these compounds include pyrazines, aldehydes, esters, ketones, acids, alcohols, hydrocarbons, pyrrolines, furans, etc. [2,3]. Other determinant factors influencing the odor quality of bread are the type of flour, type of fermentation [7], dough improvers [8], and production process [9]. While more than 540 volatiles have been identified in bread [10], only a small fraction of these volatile compounds plays a significant role in the bread's aroma [1]; interestingly, these small fractions are the ones detectable by the human olfactory receptors. In recent times, the development of extraction methods and analytical procedures has been employed to identify these volatile fractions [4,11,12]. However, most of Foods 2020, 9,1129 2 of 12 the methods employed for the quantitation of these volatile fractions are usually carried out with external calibration without necessarily addressing losses during the workup procedure. To address this issue, stable isotope dilution assays (SIDAs) in combination with gas chromatography-mass spectrometry-olfactometry as well as the calculation of odor activity values (OAVs) are now being employed [13,14]. Currently there are many documented reports on the volatile constituents of bread. However, for Native Americans, with an estimated population of 42 million [15], fry bread is primarily a food for special occasion, similar to cake [16]. It can be served as a savory meal topped with cheese or meat. However, the flavor characteristic of this product, which is eaten across the United States, has not been reported to the best of our knowledge.
Therefore, there is a need for a concerted effort to elucidate the flavor compounds in fry bread. This study was therefore aimed at characterizing the key aroma constituents of fry bread using the sensomics approach. The sensomics approach is the best method to date for identifying compounds which play an active role in aroma perception. The approach can also be used to isolate taste components in food along with aroma compounds [17]. In addition, while other flavor analytical techniques rely on separation-based chromatographic methods to quantify the aroma strength of individual compounds in a food matrix, the sensomics approach combines the aforementioned techniques with reconstitution and omission experiments to evaluate the role of specific compounds in the perceived aroma of a mixture. The practical fallout of this approach is the so-called flavor blueprint or flavor signature of a food, i.e., the combinatorial code of the entire set of odor-and taste-active food components in their natural concentrations in food [18]. Furthermore, the sensomics approach has been employed in the characterization of aroma compounds of yeast dough dumpling [14] and the crust of soft pretzels [13].

Fry Bread Production
Fry breads were produced by employing the two commonly used methods by Native North Americans. Fry bread can be made from either frozen (−30 • C) or freshly made doughs. The dough recipes contained high protein (13%) (enriched bakers patent flour from Pastry Product, Sdn., Malaysia) (1000 g); warm water (550 g, 50 • C); salt (15 g); granulated sugar (30 g); powdered milk (15 g); and baking powder (30 g). The flour, salt, granulated sugar, and baking powder were mixed separately in a large bowl. The powdered milk was dissolved in the warm water and subsequently added slowly to the dry mixture. The ingredients were introduced into a mixer (Stephan, Hameln, Germany) and mixed for 3 min to produce fluffy dough. The dough mass was divided into 8 golf ball-sized pieces. Four of the golf ball-sized pieces (A) were kept frozen (−30 • C) for 24 h. The other four pieces (i.e., unfrozen) (B) were allowed to rise in a warm spot (29 ± 2 • C) for approximately 20 min. Each of the dough pieces was flattened with a rolling pin to approximately 0.102-0.127 M circular discs and fried in 0.051 to 0.076 M sunflower oil maintained at 180 • C in a mini electric frying pan (Model/SKU 745409792, Helenite). The flattened dough pieces were fried on each side for about 15 s. The golden-brown fry breads (UFBs) (approximately 0.153 M) were removed and drained on a paper towel. The frozen dough pieces (A) were removed after 24 h and kept in a zippered plastic bag inside a fridge overnight to defrost. The thawed dough pieces were flattened into circular discs (0.102-0.127 M) and fried (FBs) as described earlier.

Isolation of Volatiles from Pulverized Fry Bread for Aroma Extracts Dilution Analysis (AEDA)
Fry breads (i.e., UFBs and FBs) were each (150 g) sliced into pieces and frozen separately in liquid nitrogen. The frozen bread pieces were pulverized in a warring blender. The volatiles from the pulverized crumb (150 g) were extracted at room temperature (30 ± 2 • C) with dichloromethane (350 mL, 1 h), and the extract was subsequently distilled at 40 • C using the solvent-assisted flavor evaporation (SAFE) distillation protocol [28]. The distillate was treated with aqueous sodium carbonate solution (0.5 mol/L, 50 mL × 3) to remove the acidic volatiles [29]. Furthermore, the aqueous solution was washed with dichloromethane (50 mL), and the organic phases were combined and dried over anhydrous sodium sulphate. It was filtered and concentrated to 1 mL with a small size Vigreux column (40 cm × 1 cm). All analyses were repeated in triplicate.

Gas Chromatography-Mass Spectrometry Analysis
The volatile constituents of the fry bread extracts were analyzed by the GC-MS system (GC-MS, QP-5050A, Shimadzu, Kyoto, Japan) equipped with a (30 m × 0.32 mm, 0.5 µm film thickness) DB-WAX (J & W Scientific, Folsom, USA) column. The extracts (2 µL) were applied by the on-column injection technique at 220 • C. The temperature of the oven was raised from 40 • C. Min −1 to 50 • C, held for 2 min isothermally, and then raised from 4 • C. Min −1 to 250 • C. The flow rate of the carrier helium was 2.0 mL Min −1 . The retention indices (RIs) of the compounds were calculated as described previously [29].
Mass spectra were recorded in the electron impact positive mode (EI) over a scan range of m/z 40-270 (scan frequency 5.8 Hz) applying an electron energy of 70 eV. The total run time was 45 min. The source and transfer line temperatures were 200 and 240 • C, respectively. Mass spectra were evaluated using the Xcalibur software (Thermos Scientific, Dreieich, Germany).

Gas Chromatography-Olfactometry (GC-O)
In order to identify the aroma-active constituents in the fry bread extracts, an olfactory detection port ODP-3 (Gestalt, Mulheim, Germany) which was connected to a Trace Ultra 1300 gas chromatograph (Thermos Scientific, Waltham, MA, USA) was used to conduct the sniffing test. The GC-O system was fitted with a DB-Wax column (30 m × 0.32 mm and 0.5 µm film thickness, J & W Scientific, Folsom, CA, USA). Sniffing was conducted as described previously [30]. Three experienced panelists (two females and a male) with strong gustative and olfactory responses in earlier sessions were used for the sniffing test. The sniffing analysis was divided into three sessions of 20 min, and each assessor participated in the exercise. All analyses were repeated in triplicate by each assessor.

Aroma Extracts Dilution Analysis (AEDA)
The flavor dilution (FD) factors of the aroma constituents were determined by GC-O as reported by Lasekan and Yap [30]. The original extracts (200 µL) containing the neutral/basic volatile constituents obtained from the pulverized fry bread (150 g) were diluted in a stepwise fashion by the addition of dichloromethane as described earlier [30]. Three panelists evaluated all dilutions in triplicate. Only the aroma compounds detected by more than two panelists were recorded.

Aroma Constituents' Quantification by Stable Isotope Dilution Assays (ACQSIDAs)
Labelled standards (10-60 µg) which were earlier dissolved in dichloromethane (5 mL) were added to each pulverized fry bread sample (100 g). The extract was distilled using the SAFE distillation method described earlier in Section 2.3. Aliquots (0.5 µL) of the concentrates were analyzed by means of two-dimensional GC-MS as described previously [31]. The calibration factor for each compound was determined by analyzing mixtures of the defined quantity of the labelled compounds in five different mass ratios (1:5, 1:3, 1:1, 3:1, and 5:1) using the GC-MS. The obtained response factors from the peak area and the amounts of labelled compound are shown in Table 1.

Aroma Profile Determination
An hour after frying, the pulverized fry breads (i.e., UFBs and FBs) (approximately 8 g with similar crust covering) were placed inside glass beakers (height 7 cm, volume 45 mL) with three random digitals and orthonasally analyzed by panel members at room temperature (29 ± 2 • C). In addition, samples were rotated among panelists to avoid a carry-over effect. The panel was made up of 10 members, aged 24 to 35 years, and composed of seven women and three men. These panelists participated in weekly sensory training sessions for at least a year to be able to recognize and describe different aroma qualities. The sensory analyses were conducted in a sensory room according to international standards (ISO 8589, 2007) [33] with individual booths equipped with uniforms and glare-free white light (D65). Descriptors used were determined in preliminary sensory experiments as described by Steinhause et al. [34]. The reference compounds used as stimuli were: 10 µg L −1 of 2-acetyl-1-pyrroline (roasty); 100 µg L −1 of 3-methylbutanal (malty); 70 µg L −1 of 2,3-butanedione (buttery); 50 µg L −1 of 4-vinyl-2-methoxyphenol (smoky); 100 µg L −1 of methional (baked potato-like); and 50 µg L −1 of δ-decalactone (oily/peach). During evaluation, the panelists had 5 min to rest after each set of samples was tested. All samples were repeated in triplicate. The intensities of the attributes were rated on a 7 point linear scale from 0 (not perceivable) to 3 (strongly perceivable) in steps of 0.5 by the panelists. The sensory data were analyzed in triplicate using the Student's t-test, and statistical analyses were performed using SPSS 20.0 (SPSS Inc., Chicago, IL, USA). In addition, all procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. The study protocol and consent procedure received ethical approval from the Institutional Review Board (IRB) of the University Putra Malaysia. Informed consent was obtained from all individual participants included in the study.

Aroma Model Recombinants and Omission Tests of the Fry Breads (UFBs and FBs)
Reference standards of aroma constituents with OAVs above 1 ( Table 2) were prepared in ethanolic solution [35]. The combined ethanolic stock solutions of the 11 aroma compounds (500 µL) with OAVs ≥ 1 was mixed with 30 mL of citrate buffer (pH 5.6; 0.1 mol L −1 ) and 30 g of free corn starch in a closed Teflon cup. The Teflon cup was stirred continuously for 15 min at room temperature (i.e., 29 • C). A triangle test was performed to determine the significance of one odorant on the aroma recombination models (UFBs and FBs) reported in (Table 4). For each of the models, a glass of the mixture (20 mL) was prepared by omitting one or a group of selected odorants from the complete recombination model (Table 5). This mixture and two other glasses containing the complete recombination models were presented to the sensory panel in a triangle test [36]. The results of the triangle tests were analyzed by comparing the total number of correct responses with the minimum number of responses required for statistical significance (ISO 4120, 2004) [37]. Panel performance was obtained by applying analysis of variance (ANOVA) to the sensory profile data. The data were analyzed using SAS statistical software (SAS Institute, Inc., Cary, NC, USA, 1996). The significance α was calculated according to the method of Callejo et al. [36]. In addition, each of the aroma models was evaluated orthonasally in comparison with the corresponding fry bread (i.e., UFB or FB) as described above (Section 2.6.). a The compounds were identified by comparing the mass spectra through the mass spectra/electron ionization (MS/EI), the retention indices was detected on bonded low bleed wax capillary column (DB-Wax), and the odor note perceived at the sniffing port. * Flavor dilution (FD) determined by aroma extract dilution analysis (AEDA) for fry breads produced from unfrozen (UFB) and frozen (FB) doughs. NB: Neutral-basic fraction, A: acidic fraction.

The Aroma-Active Constituents of Fry Breads
A total of 22 aroma-active constituents were identified in the fry breads (i.e., fry breads produced from frozen dough, (FB) and those made from unfrozen dough (UFB)). Among these compounds, seven acids, eight aldehydes, four ketones, one heterocyclic compound, one furan, and a phenol were positively identified (Table 2). These aroma constituents exhibited an array of odor nuances such as malty, buttery, roast-like, baked potato-like, sweaty, deep-fried, caramel, smoky, and oily/peach-like. To establish differences among the flavors of the fry bread samples, the volatile fractions of the fry bread extracts were subjected to AEDA. The results of the AEDA and the identification experiments carried out are shown in Table 2. In the neutral-basic fraction (Table 2)
In addition, nearly all of the aroma-active constituents, disregarding their boiling point, polarity or functional group, decreased slightly (i.e., <7%) in the fry breads produced from the frozen doughs (FB) ( Table 3). The only exception to this were 2,3-butanedione, 2-acetyl-1-pyrroline, and 2-methoxy-4-vinylphenol. The slight decreases (i.e., <7%) obtained between the FB and UFB were probably due to the frozen dough used in the production of FB. Studies have shown that freezing often results in the disruption of the gluten structure as a result of water migration brought about by crystal formation. This phenomenon results in the dissociation of starch granules from the gluten network [41]. The disruption of the starch granules is known to influence the interaction of starch with the volatile compounds as well as the retention of volatile compounds [42] Overall, the aldehydes and acids constituted the largest number of aroma constituents (FD ≥ 16) (Table 3) in the fry breads. Aldehydes are typical lipid oxidation products that are associated with characteristic aroma of whole wheat bread [43]. In addition, aldehydes, such as 3-methylbutanal (malty) and phenyl acetaldehyde, identified in the fry breads can also be formed during the Maillard reaction [44]. Another route for the formation of the aldehydes could be through the amino acid biosynthetic pathway in which the aldehydes formed during the Ehrlich pathway is oxidized to their corresponding acids such as 3-methylbutanoic acid by aldehyde dehydrogenase [45]. It should be mentioned that acetic acid is also a well-known product of the Maillard reaction. Butanoic acid 350 ± 8.8 a 348 ± 5.5 a 100 3.5 3.5 7 Phenyl acetaldehyde 107 ± 5.0 a 102 ± 7.1 b ND ND ND * Thresholds in starch (Rychlik, Schieberle, and Grosch (1998) [46], OAV odor activity values on the basis of threshold in starch. UFB, fry breads produced from unfrozen doughs; FB, fry breads produced from frozen doughs; ND not detectable. Different letters within the same row represent significant differences.

Aroma Profile Analysis and Aroma Simulation Models of the Fry Breads (UFBs and FBs)
The results of the sensory evaluation of the different fry breads (UFBs and FBs) are shown in ( Figure 1A, Table 4). The aroma profiles of the fry breads were characterized as roasty/popcorn-like, malty, buttery, baked potato-like, smoky, sweaty, deep-fried, and oily/peach-like. However, with the exception of the malty and buttery notes, the aroma profiles of the UFBs and FBs were quite similar. The malty and buttery nuances in the FB samples were less intense as compared to that of the UFB samples (Table 4). The Duncan's multiple comparison test results (Table 4) revealed that the eight attributes (roasty/popcorn-like, malty, buttery, baked potato-like, smoky, sweaty, deep-fried, and oily/peach-like) with different superscripts seemed to well explain their aroma characteristics. To validate this observation, recombination experiments were performed by mixing solutions of the pure reference compounds in the same amounts as obtained for both UFB and FB ( Table 5). The recombination models were evaluated in parallel with the UFB and FB samples. Sensory results revealed that the recombinant models imitated well the flavor of the fry breads ( Figure 1B,C). In addition, the roasty/popcorn aroma note was perceived as equally intense in the aroma models as well as in the fry breads.

Omission Tests
To assess the contribution of individual compound to the overall aroma of the fry bread (UFB and FB), omission tests were conducted on the fry bread aroma models (Table 6) [47]. In this study, 7 aroma omission models (M1-M7) comprised of either a single or group of compounds were prepared. Each of the omission models was evaluated in triangular experiments with two complete recombination models ( Table 6). The results revealed that, when the entire group of aldehydes (M1) was omitted, their omission from the complete recombination model could be detected by 9 out of the 10 assessors. This is an indication of the importance of these aldehydes (i.e., 3-methylbutanal, methional, phenyl acetaldehyde and (E,E)-2,4-decadienal) in the overall aroma of the fry breads.

Conclusions
In conclusion, differences in the aroma-active constituents of fry breads produced from frozen doughs (FBs) and freshly made doughs (UFBs) were characterized for the first time. A total of twenty-two aroma constituents were identified in the fry breads. The results of the OAVs and sensory studies showed that the aroma profiles of the fry breads were characterized as roasty/popcorn-like, malty, buttery, baked potato-like, smoky, sweaty, deep-fried, and oily/peach-like. However, with the exception of the malty and buttery notes, the aroma profiles of UFBs and FBs were quite similar, and the malty and buttery nuances in the FB samples were less intense as compared to that of the UFB samples. Aroma-active constituents, such as δ-decalactone (oily/peach), 2-acetyl-1-pyrroline (roasty/popcorn-like), 3-methylbutanal (malty), methional (baked potato-like), 2,3-butanedione (buttery), phenyl acetaldehyde (flowery), (E,E)-2,4-decadienal (deep-fried), butanoic acid, and 3-methylbutanoic acid, were identified as the key aroma constituents of fry bread. In addition, 3-methoxy-4-vinylphenol (smoky) and 4-hydroxy-2,5-dimethyl-3(2H)-furanone (caramel-like) were identified as important aroma constituents of fry bread. Finally, these findings establish a basis for further work on the identification of the taste-active food components in fry breads as well as consumers' preferences for the differently produced fry breads. Funding: Financial support for this research was provided by the University Putra Malaysia research scheme (Grant No. 9478500).

Conflicts of Interest:
The authors declare they do not have any conflict of interest.