Identification of Thiols in Yellow Onion (Allium

Identification of Thiols in Yellow Onion (Allium cepa L.) Using Solvent Vented Large Volume Injection GC-MS , Robert Cannon, Sytze Haasnoot, Hans Colstee, Cor Niedeveld, Gijs Koopmanschap, and Neil C. Da Costa

Abstract: Thiols are often highly odor active molecules and as such can significantly contribute to aroma while being present at extremely low concentrations. This paper details the identification of thiols in yellow onion juice by solvent extraction followed by thiol enrichment using a mercuric agarose gel column. Due to the inherent thermal instability and low concentrations of thiols in onion, chromatographic analysis utilized larger volume solvent elimination injections. New sulfur compounds in onion included 1,1-propanedithiol, bis-(1-sulfanylpropyl)-sulfide, 1-methylsulfanyl-1-propanethiol, 1-propylsulfanyl-1-propanethiol, and 1-allylsulfanyl-1-propanethiol. A discussion on the potential route of formation for each compound is included along with the orthonasal and retronasal evaluations of the synthesized molecules. This work investigated and identified 5 newly identified compounds present in onions that can impart onion character at low concentrations levels. Keywords: Allium cepa, 1-alkylsulfanyl-1-alkanethiols, multimode inlet, onion juice, sulfur compounds

Introduction For thousands of years onions have been cultivated for human consumption. The desired, intense flavor of onion, and other Allium species, develops enzymatically. Within the intact plant, the enzyme allinase is stored in the cell’s vacuole, which is isolated from the S-alk(en)yl-L-Cysteine sulfoxide substrates found in the cytoplasm (Block and Calvey 1994). When a stressor disrupts the plant tissue, the flavor precursor molecules are made available to the enzymes, which produces a cascade of chemical reactions, including the lachrymatory effect of onions and its characteristic aroma. S-1-propenyl cysteine sulfoxide (isoalliin) is the main cysteine sulfoxide present in onions (Storsberg and others 2004). The action of allinase on isoalliin generates 1-propenyl sulfenic acid, which is transformed to propanethial-S-oxide, the lachrymatory compound in onions, by the action of the lachrymatory factor synthase (LFS) enzyme. In the presence of water, propanethialS-oxide decomposes to propanal and hydrogen sulfide (Poll and others 2003; Feiberg and others 2012). This basic pathway is shown in Figure 1. The products generated from this decomposition are inherently reactive and can lead to the formation of further highly odorous molecules (Boelens and others 1974). Given the presence of propanal and hydrogen sulfide in onions the occurrence of 1,1-propandithiol was postulated, however until now has not been reported in nature. The powerful onion character of this molecule and the relative stability of terminal gem-dithiols, make it a desirable choice for use in formulating savory flavors (Cairns and others 1952; Timokhina and others 2004). The combination of reactive, enzymatically generated volatile compounds in Allium species creates a challenging matrix for analytical investigations. Hot injection conditions have been criticized in the literature, as providing an erroneous picture JFDS-2017-0994 Submitted 6/21/2017, Accepted 9/14/2017. Authors are with International Flavors & Fragrances Inc., Research & Development, 1515 State Highway 36, Union Beach, N.J. 07735, U.S.A. Direct inquiries to author Wermes (E-mail: [email protected]).

R C 2017 Institute of Food Technologists

doi: 10.1111/1750-3841.13943 Further reproduction without permission is prohibited

of the composition of these plants especially when describing the thiosulfinate composition (Block and others 1992; Block and others 1993; Block and Calvey 1994). Elevated injection temperatures have been implicated in the formation of 3vinyl-4H-1,2-dithiin, 2-vinyl-4H-1,3-dithiin, 3,4-dimethyl-2,3dihydrothiophene-2-thiol, 3,4-dimethylthiophene, and many diand trisulfides, most notably diallyl disulfide and allicin, in onion samples (Block and others 1993; Block and Calvey 1994; Iranshahi 2012) Because of these concerns, cool injection techniques have become the favored technique for Allium analysis by gas chromatography (Block and others 1993). Traditional on-column injection however limit the amount of sample which can be loaded, and also deposits any non-volatile material present in the extract onto the column. In an attempt to overcome these limitations, this research used an Agilent multimode inlet (MMI) operating in the solvent vent mode to study the volatile thiols in onion, a few of which are reported for the first time in onion and natural products. A discussion on the potential mechanism of formation for each compound is included along with the orthonasal and retronasal evaluations of the synthesized ingredients.

Materials and Methods Materials Yellow onions were purchased at a local supermarket for immediate preparation. Chemicals The following reagents were purchased from Sigma-Aldrich (St. Louis, Mo., U.S.A.) unless otherwise noted: pentane, diethyl ether, dichloromethane, deuterated chloroform, tetramethylsilane, and anhydrous sodium sulfate. Extraction of yellow onion juice Yellow onions (2 kg) were peeled and quartered prior to processing with a commercial juicer (Breville Juice Fountain, Breville Vol. 00, Nr. 0, 2017 r Journal of Food Science 1

Food Chemistry

Clint Wermes

Identification of thiols in yellow onion . . .

Figure 1–Formation of propanal and hydrogen sulfide in Allium cepa.3

Food Chemistry

USA Inc., Torrance, Calif., U.S.A.). The juicing was carried out in a fume hood to minimize exposure to the lachrymatory compounds. The expressed juice (1 L) was diluted with an equal volume of 8% salt water and allowed to stand at room temperature for 30 min. The mixture was then passed through a cheesecloth filter to remove any plant solids followed by extraction with pentanediethyl ether (2:1; 3 × 200 mL). The extract was dried over anhydrous sodium sulfate, filtered and reduced to approximately 10 mL using a Zymark Turbovap (Biotage, Uppsala, Sweden).

and 250 amu. Mass spectral identification was achieved using inhouse and commercial libraries, as well as synthesized or purchased referenced standards when available.

Nuclear magnetic resonance NMR spectra were recorded at 26.8 °C in deuterated chloroform (containing 0.05% v/v tetramethylsilane) on a Bruker Avance III 400 MHZ or a Bruker Avance 500 MHz spectrometer (Billerica, Mass., U.S.A.), with 5 mm BBO probes. 1 H chemical shifts are expressed as parts per million (ppm) with residual chloroform (δ 7.26) or tetramethylsilane (δ 0.00) as a reference and are reported as chemical shift (δH), relative integral, multiplicity (s = singlet, br = broad, d = doublet, t = triplet, higher multiplicities as for example, dd = doublet of doublets, m = multiplet); and coupling constants (J) reported in Hz.

Thiol enrichment The onion juice extract was further enriched by affinity chromatography. The extract (10 mL) was added to a mercuric agarose gel column, washed with dichloromethane (DCM; 3 × 10 mL) and the collected effluent discarded. The thiol enriched fraction was then eluted with 10 mL of a 10 mM dithiothreitol (SigmaAldrich, Saint Louis, Mo., U.S.A.) solution in DCM. This collec- Synthesis of 1,1-propanedithiol 9 tion was concentrated to 1 mL under a gentle stream of nitrogen. The synthesis of compound 9 involved a 2-step pathway via Preparation of the mercuric agarose gel and the thiol enrichment 1,1-propanedithioacetate as a stable intermediate. process followed in this study have been described in detail by 1,1-Propanedithioacetate. A 2 L, 3-necked flask was Schreier and Full (1994). charged with acetic anhydride (Acros Organics, Geel, Belgium; 300.0 g, 2.9 mol) and thioacetic acid (Acros Organics, Geel, Gas chromatography-mass spectrometry Belgium; 450.0 g, 5.9 mol) with mechanical stirring and cooled The extract was analyzed on apolar and mid-polar phase in an ice bath. At 0 °C, catalyst boron trifluoride dimethyl ethercolumns using an Agilent 7890A gas chromatograph (Santa Clara, ate (Sigma-Aldrich, Schnelldorf, Germany; 75.0 g, 660 mmol) Calif., U.S.A.) 5977 mass selective detector. The apolar capillary was added dropwise over 15 min. Once the flask temperature recolumn had dimensions of 50 m x 320 μm x 0.4 um (Agilent VF- turned to 0 °C, propanal (Acros Organics, Geel, Belgium; 175.0 g, 1) and the midpolar capillary column was 30 m x 320 μm x 1.5 3.0 mol) was added dropwise over 2 h. The reaction mixture was μm RTX-1701 (Restek Bellefonte, Pa., U.S.A.). Samples were allowed to warm to ambient temperature and stirring was conintroduced to the gas chromatograph (GC) using a multimode tinued for an additional 16 h. The mixture was cooled in an ice inlet configured for a large volume, solvent elimination injection bath and neutralized to pH 7 by dropwise addition of a 33% aque(5 μl injection volume @ 0.3 μl/sec) at an initial inlet temperature ous sodium hydroxide (Brenntag, Dordrecht, The Netherlands) of 20 °C for 0.2 min. An inlet liner prepacked with deactivated solution. The neutralized mixture was twice extracted with glass wool positioned near the middle of the liner was used to dichloromehane (Brenntag, Dordrecht, The Netherlands) and the allow the sample to be deposited directly onto the wool by the organic layer was dried over magnesium sulfate (Acros Organautosampler. The split vent during the injection was set to a flow ics, Geel, Belgium) and filtered. The solvent was evaporated in of 30 mL/min and a pressure of 1.0 psi until 0.2 min, after which vacuo and 1,1-propanedithioacetate (461.0 g, 2.4 mol) was obthe vent was then closed. The inlet temperature was ramped at tained by fractionated vacuum distillation (1 mbar, 115 °C) with a 720 °C/min to 150 °C and held for 1.6 min, followed by a purity of 99% (80% yield based on propanal). 1 H NMR (CDCl3 , 720 °C/min ramp to 250 °C for 60 min. The purge to split vent 500 MHz): δ 5.03 to 5.07 (m, 1H), 2.33 (s, 6H), 1.90 to 2.01 (m, was reopened at 1.55 min after the injection, with a purge flow 2H), 1.01 (t, J = 7.1 Hz, 3H). EI-MS: 192 (4, M+ ), 43 (100), 148 of 60 mL/min. For the apolar column, the helium carrier gas (31), 149 (23), 75 (18), 117 (17), 107 (9), 74 (8), 73 (7), 106 (4). flow rate was held constant at 1.5 mL/min, and the temperature 1,1-Propanedithiol 9. A 6 L, 3-necked flask was charged program of the oven started at an initial temperature of 0 °C for with dry methanol (Acros Organics, Geel, Belgium; 3.0 kg). 1 min, then increased 40 °C/min to 40 °C followed by 2 °C/min A second 2 L, 3-necked flask was charged with sodium chloto 250 °C with an 8 min hold. The same method was used for ride (Akzo Nobel, Hengelo, The Netherlands; 1.0 kg), equipped the mid-polar column except for a helium carrier gas flow rate with a dropping funnel and connected to the first flask by means of 2 mL/min. The instrument was calibrated using a homologous of a Liebig condenser. Concentrated sulfuric acid (Brenntag, series of alkanes in order to generate retention index values for the Dordrecht, The Netherlands; 0.2 kg) was added dropwise into observed peaks. the second flask. Hydrogen chloride gas was generated upon conThe data was acquired using an Agilent 5977 single-quadrupole tact of sulfuric acid with the salt. The hydrogen chloride gas was GC/MS system in electron ionization (EI) mode. The ion source condensed into the dry methanol until it reached a concentration was set at 230 °C with an electron energy of 70 eV. The transfer line of 5% by weight. The flasks were disconnected and the first flask temperature was set to 100 °C. Spectra were acquired between 33 was equipped with a heating mantle, a reflux condenser and a 2 Journal of Food Science r Vol. 00, Nr. 0, 2017

dropping funnel. The methanol/dry hydrogen chloride mixture was heated to 50 °C. While maintaining this temperature, 1,1propanedithioacetate (461.0 g, 2.4 mol) was slowly added to the flask with stirring. After 2 h at 50 °C, the hydrolysis was completed. The solvent was evaporated in vacuo. 1,1-Propanedithiol 9 (180.0 g, 1.7 mol) was obtained by fractionated vacuum distillation (1 mbar, 50 °C) with a purity of 99% (69% yield based on 1,1-propanedithioacetate). The total yield based on the 2-step synthesis was 55%. 1 H NMR (CDCl3 , 500 MHz): δ 4.00 to 4.07 ppm (m, 1H), 2.38 ppm (d, J = 6.3 Hz, 2H), 1.86 to 1.94 ppm (m, 2H), 1.07 ppm (t, J = 7.3 Hz, 3H). EI-MS: 108 (48, M+ ), 75 (100), 41 (44), 47 (25), 74 (24), 45 (22), 79 (15), 39 (8), 59 (7), 76 (7).

Synthesis of 1-methylsulfanyl-1-propanethiol 11 In a 2 L, 3-necked flask, propanal (70.3 g, 1.2 mol) was added dropwise to a mixture of a 7.2% solution of methanethiol (SigmaAldrich, Schnelldorf, Germany; 69.8 g, 1.5 mol) in water at 0 °C. The resulting solution was transferred to a 5 L autoclave. Hydrogen sulfide (Air Liquide Industrie B.V., Rotterdam, The Netherlands) pressure was applied and maintained at 6.75 bar over 2 h. During the reaction, the temperature was kept below 23 °C by wall cooling. After ceasing the supply of hydrogen sulfide gas, the mixture was reacted overnight on its residual hydrogen sulfide pressure. Remaining hydrogen sulfide was removed through a lye scrubber. Sodium chloride (400.0 g) was added and the oil layer separated from the aqueous solution, dried over anhydrous magnesium sulfate and filtered. The crude product was purified 3 times by distillation and yeilded 1-methylsulfanyl-1-propanethiol 11 (11.0 g, 90.0 mmol) with a purity of 99% (7.44% yield based on propanal). 1 H NMR (CDCl3 , 400 MHz): δ 3.77 to 3.83 (m, 1H), 2.20 (s, 3H), 1.99 (d, J = 7.1 Hz, 1H), 1.75 to 1.96 (m, 2H), 1.06 (t, J = 7.3 Hz, 3H). EI-MS: 122 (45, M+ ), 89 (100), 41 (47), 75 (24), 45 (23), 47 (21), 74 (21), 61 (14), 93 (11), 73 (9).

Synthesis of 1-allylsulfanyl-1-propanethiol 17 An autoclave was charged with ethanol (Tereos SA, Lille, France; 1.7 kg), propanal (132.0 g, 2.3 mol), and freshly distilled allylthiol (TCI Europe N.V., Zwijndrecht, Belgium; 164.0 g, 2.2 mol). The mixture was stirred and 6 bar of hydrogen sulfide gas was applied over 2 h. The temperature rose to 60 °C due to reaction. After ceasing the supply of hydrogen sulfide gas, the mixture was reacted overnight on its residual hydrogen sulfide pressure. The mixture was allowed to cool to ambient temperature. The remaining hydrogen sulfide was removed through a lye scrubber. The reaction mixture was transferred to a distillation setup and the solvent was evaporated at 175 mbar at 40 to 50 °C. The residue was distilled using a small Vigreux at 1 mbar, 60 to 95 °C. Subsequently, the obtained distillate (72.0 g) was redistilled repeatedly in fractions at 1 mbar, 40 to 50 °C using a 3 plate column, thereby keeping the fractions richest in desired product for the next distillation. The pure product distilled at 45 °C at 1 mbar. A fraction of 1allylsulfanyl-1-propanethiol 17 (20.0 g, 135.0 mmol) with a purity of 98% (5.95% yield based on propanal) was obtained. 1 H NMR (CDCl3 , 400 MHz), δ:3.77 to 3.83 (m, 1H), 2.20 (s, 3H), 1.99 (d, J = 7.1 Hz, 1H), 1.75 to 1.96 (m, 2H), 1.06 (t, J = 7.3 Hz, 3H). EI-MS: 148 (13, M+ ), 74 (100), 115 (61), 41 (68), 73 (59), 75 (49), 45 (33), 119 (25), 39 (22), 47 (22), 59 (14), 114 (14), 71 (12), 81 (12).

Synthesis of 1-propylsulfanyl-1-propanethiol 18 An autoclave was charged with ethanol (1.5 kg), 1-propanethiol (Sigma-Aldrich, Schnelldorf, Germany; 150.0 g, 2.0 mol), and propanal (120.0 g, 2.1 mol). Hydrogen sulfide pressure was applied and maintained at 6.75 bar over 2 h. During the reaction, the temperature rose spontaneously to 42 °C. After ceasing the supply of hydrogen sulfide gas, the mixture was reacted overnight on its residual hydrogen sulfide pressure, after which the mixture was allowed to cool to ambient temperature. Remaining hydrogen sulfide was removed through a lye scrubber. The reaction mixture was collected in a single-necked flask, ethanol was evaporated at 40°C and 175 mbar. The residue was distilled at 1 mbar to isolate the product. Fractions with purity > 99 % were combined to yield 1-propylsulfanyl-1-propanethiol 18 (33.8 g, 225.0 mmol; 10.8% yield based on propanal). 1 H NMR (CDCl3 , 400 MHz): δ 3.83 to 3.88 (m, 1H), 2.71 to 2.77 (m, 1H), 2.60 to 2.68 (m, 1H), 1.90 to 1.99 (m, 2H), 1.70 to 1.90 (m, 1H), 1.57 to 1.70 (m, 2H), 1.09 (t, J = 7.3 Hz, 3H), 1.01 (t, J = 7.4 Hz, 3H). EI-MS: 150 (28, M+ ), 117 (100), 75 (69), 41 (43), 74 (37), 47 (27), 45 (22), 73 (14), 39 (11), 76 (10), 118 (9), 59 (5). Synthesis of bis-(1-sulfanylpropyl)-sulfide isomer 25 and 26 A 1 L, 3-necked flask was charged with water (200.0 g) and a 85% solution of phosphoric acid (Univar N.V., Zwijndrecht, The Netherlands; 14.0 g). Potassium hydroxide (45% solution in water; 22.5 g) was used to adjust the solution to a pH of 7. The created phosphate buffer was cooled down to ambient temperature. 1,1-Propanedithiol (65.0 g, 600.0 mmol) was added and the mixture was stirred for 6 h. The mixture was extracted twice with ethyl acetate (Acros Organics, Geel, Belgium; 200.0 g), and the organic layer dried over anhydrous magnesium sulfate and filtered. Solvent was evaporated in vacuo. The residue was distilled at 1 mbar. Bis-(1-sulfanylpropyl)-sulfide isomers 25 and 26 were obtained by fractionated vacuum distillation (56.0 g, 311.0 mmol; purity 98 %). 1 H NMR (CDCl3 , 500 MHz): δ 4.16 to 4.24 (m, 1H), 4.05 to 4.11 (m, 1H), 1.75 to 2.10 (m, 6H), 1.05 to 1.14 (m, 6H). EI-MS: 182 (7, M+ ), 75 (100), 41 (51), 47 (27), 74 (24), 45 (23), 73 (17), 149 (13), 39 (11), 148 (8), 108 (7), 59 (5). Odor and taste evaluation Six expert panelists conducted orthonasal evaluations of each of the synthesized molecules at 0.01% in ethanol using paper smelling strips and retronasal evaluations at 250 μg/L in water. Tasting of new chemicals is based on established safe levels using industry standard toxicological principles which include structure evaluation, comparison of structure and metabolism to known molecules used in flavorings and applying the concept of Threshold of Toxicological Concern (TTC).

Results and Discussion Thiols in fresh onion are of interest since they are often olfactively potent molecules that can contribute significantly to the aroma even when present at extremely low concentrations (Iranshahi 2012). Due to the presence of propanal and hydrogen sulfide from the enzymatic degradation of isoalliin, gem-dithiol compounds were suspected to be present in onions. To investigate this hypothesis, the expressed juice from 2 kg of yellow onions was allowed to stand for 30 min prior to extraction. This allowed time for the enzymatic and chemical reactions which generate the desired aroma compounds to form (Block and others 1992). A low density extraction solvent (pentane/diethyl ether) was selected to reduce complications with emulsion formation and eliminate Vol. 00, Nr. 0, 2017 r Journal of Food Science 3

Food Chemistry


Identification of thiols in yellow onion . . . Table 1–Relative concentration of thiols in raw onion. Kovats No.a

Food Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

Compound

RI-VFIb

RI-1701c

%d

Ide

Methanethiol Propanethiol Allylthiol Methanedithiol 2-Mercaptoethanol 1-Sulfanylpropan-2-one 1-Sulfanylpropan-2-ol 1,3-Dithiabutane 1,1-Propanedithiol 1,2-Propanedithiol 1-Methylsulfanyl-1-propanethiol Unidentified 3-Sulfanyl-2-methylvaleraldehyde isomer 1 3-Sulfanyl-2-methylvaleraldehyde isomer 2 Unidentified Unidentified 1-Allylsulfanyl-1-propanethiol 1-Propylsulfanyl-1-propanethiol Unidentified 3,4-Dimethyl-2,3-dihydrothiophene-2-thiol isomer 1 3-Sulfanyl-2-methylpentan-1-ol Unidentified Unidentified 3,4-Dimethyl-2,3-dihydrothiophene-2-thiol isomer 2 Bis-(1-sulfanylpropyl)-sulfide isomer 1 Bis-(1-sulfanylpropyl)-sulfide isomer 2

521 600 604 642 709 741 773 780 823 848 920 976 992 997 1027 1033 1064 1089 1103 1108 1085 1138 1166 1151 1282 1311

533 710 706 801 909 925 917 885 921 959 885 1017 1155 1164 1159 1203 1162 1183 1214 1239 1271 1279 1309 1289 1415 1466

15.34 21.50 3.60 0.03 0.02 0.43 0.43 0.04 7.05 10.14 0.04 0.56 0.72 0.24 0.18 0.92 0.80 15.70 9.85 1.32 2.18 2.45 5.42 0.56 0.29 0.20

MS, RI MS, RI MS, RI MS, RI MS, RI MS, RI MS, RI MS, RI MS, RI, Syn MS, RI MS, RI, Syn MS, RI MS, RI – – MS, RI, Syn MS, RI, Syn – MS, RI MS, RI – – MS, RI MS, RI, Syn MS, RI, Syn

a

Numbers match the structures in Figure 2. indeces (RI) on an apolar column. RI on mid-polar column. d Relative percentage determined by TIC on a mid-polar column. e Identification based on mass spectral comparison (MS) and RI on 2 column phases, along with synthesized reference standards (Syn) where noted. b Retention c

Figure 2–The proposed pathway for the formation of the thiol compounds identified in onion.

potential difficulties with any remaining plant solids in the onion juice during partitioning or centrifugation. The juice was salted to assist in driving analytes into the organic phase. The recovered extract was concentrated by Vigreux column distillation and passed through a mercuric agarose gel eluting with DCM/dithiothreitol to yield a thiol enriched fraction, which is a common method for this type of analysis (Steinhaus and others 2014; Pavez and others 2016). Chromatographic analysis of the extract by on-column GC-MS was attempted, however this resulted in a weak signal allowing the 4 Journal of Food Science r Vol. 00, Nr. 0, 2017

detection of only a few compounds. While acknowledging that cool on-column injection is the default technique for the analysis of reactive sulfur compounds, we considered a solvent elimination injection under mild thermal conditions since the thiols of interest are more thermally stable than thiosulfinates (Block and others 1993). Caution was taken when developing the methodology to maintain mild injector and GC-MS transfer line temperatures to suppress the generation of thermal artifacts (Block and others 1992). In order to increase signal, a multi-mode inlet operating in the solvent vent mode was used to further concentrate the


Food Chemistry

Figure 3–Mass spectra and chemical structures of (A) 1,1-propanedithiol 9; (B) 1-methylsulfanyl-1-propanethiol 11; (C) 1-allylsulfanyl-1-propanethiol 17; (D) 1-propylsulfanyl-1-propanethiol 18; (E) bis-(1-sulfanylpropyl)-sulfide isomers 25 and 26.

Vol. 00, Nr. 0, 2017 r Journal of Food Science 5


Figure 3–Continued.

Food Chemistry sample in the GC inlet liner at low temperature and then sweep only the volatile compounds to the column under mild thermal conditions. This sample introduction technique avoids contaminating the column by depositing non-volatile materials (which can occur with on-column techniques) and reduces artifact formation (as compared to hot split/splitless injections) while lowering detection limits due to the loading of a larger sample volume. After the solvent elimination period, the inlet was heated to 150 °C to transfer the analytes to the column. Following the analyte transfer, the inlet was raised to a traditional hot temperature as the inlet was set to purge, venting any remaining semi-volatile material out of the split vent instead of onto the analytical column. This allowed for the loading a large volume of sample without introducing non-volatiles or forming significant thermal artifacts. Large volume injections require matching of the sample introduction and solvent elimination rates. These parameters were determined using the vapor pressure of the solvent at the temperature of the injection port (calculated from Antoine’s equation) and applying the calculations to determine saturated vapor volumes and maximum injection speed detailed by Staniewski and Rijks (1992). For this work an injection rate of 0.3 μL/s was selected, and delivered using an Agilent Technologies GC Sampler 80 rail that was 6 Journal of Food Science r Vol. 00, Nr. 0, 2017

configured to signal the beginning of the chromatographic run after the syringe had reached the plunger down position. Because the chromatographic run does not start until all of the sample has been dispensed from the syringe, this setup allows for the varying of injection time through changes to injection volume and/or rate without the need to alter any inlet or oven conditions or causing changes to analyte retention times. Using this configuration, the GC remains at the ready state with the oven and inlet cooled to initial temperature settings, and the inlet venting while the sample is slowly introduced onto an inlet liner packed with glass wool. Glass wool packed inlet liners have been reported to increase the solvent elimination rate by increasing the gas-solvent contact area (Staniewski and Rijks 1992).Under the described conditions, chromatographic runs employing deactivated glass wool packed inlet liners produced increased signal as compared to empty baffled liners which we suspect may have allowed some of the liquid sample to be blown out of the vent during the injection process. Analysis of the GC-MS data acquired using this sample introduction technique resulted in 26 compounds identified in the thiol enriched onion extract. The full results are detailed in Table 1. Schutte and Koenders (1972) determined that hydrogen sulfide, methanethiol, and acetaldehyde were amino acid breakdown

Identification of thiols in yellow onion . . . Table 2–Odor and taste evaluations of the synthesized molecules identified in onion. Compound 1,1-Propanedithiol 9 1-Methylsulfanyl-1-propanethiol 11 1-Allylsulfanyl-1-propanethiol 17 1-Propylsulfanyl-1-propanethiol 18 Bis-(1-sulfanylpropyl)-sulfide 25 and 26 a

Odor qualitya

Taste qualityb

onion, durian, cooked alliaceous, sl eggy, sulfury sweet onion, vegetative, cooked green, onion, garlic, savory, sulfury pungent, onion

onion, vegetative, tropical, sulfury meaty onion, alliaceous, eggy, sauteed, buttery, creamy, dairy sweet onion, cooked, sl sulfurous, mild, sl sauteed onion, green, fresh, savory, cooked, garlic alliaceous green onion, tropical

Orthonasal evaluations of each sulfur compound were conducted at 0.01% in ethanol using paper smelling strips. Retronasal evaluations were evaluated at 25 ppb in water.

products present in beef broth and found to be precursors of 1-methylsulfanyl-1-ethanethiol, which was previously identified by Brinkman and others (1972). In model reactions, Boelens and others (1974) demonstrated that propanal and hydrogen sulfide can readily generate a multitude of high impact flavor compounds, including 1,1-propanedithiol 9 and bis-(1-sulfanylpropyl) sulfide isomers 25 and 26. The addition of thiol compounds to the reaction mixtures led to the formation of 1-alkylsulfanyl1-alkanethiols, including 1-propylsulfanyl-1-propanethiol 18 and 1-methylsulfanyl-1-propanethiol 11. Using proton transfer reaction mass spectrometry (PTR-MS) to perform real time analysis of freshly cut samples, Feilberg and others (2012) demonstrated that propanethiol, methanethiol, and hydrogen sulfide account for large proportions of the headspace of sliced onions. The authors also reported that propanethiol, the largest thiol constituent, was also the main compound responsible for fresh onion odor. Analysis by PTR-MS requires no GC separation prior to detection and can be carried out under mild thermal conditions. Our results are in agreement with this research as propanethiol 2 was the most abundant thiol observed. As onions are known to produce the starting materials of propanal and hydrogen sulfide as well as thiols, it follows that the presence of dithiols and alkylthio alkanethiols should be expected (Deng and others 2015; Poll and others 2003). 3-Sulfanyl-2-methylpentan-1-ol 21 was identified by Widder and others (2000) and then quantitated later by Granvogl and others (2004) as a potent odor active compound in onions. Both authors proposed 21 formed from the hydrogen sulfide reaction with 2-methyl-2-pentenal, which is the aldol selfcondensation product of propanal. In the presence of hydrogen sulfide, a Michael addition of a thiol group adds to the double bond of 2-methyl-2-pentenal. This compound was also identified in our extract. By generating the precursor molecules, this propanal / hydrogen sulfide reaction pathway may also be involved in the formation of other thiols observed in the onion extract, including 1-allylsulfanyl-1-propanethiol 17. Allylthiol was the least concentrated thiol identified in the extract that was also found as a 1alkysulfanyl-1-alkanethiol. This reaction pathway has also been suggested as the route to the formation of bis-(1-sulfanylpropyl)sulfide and ultimately 3,5-diethyl-1,2,4-trithiolane (Flaig and Granvogl 2015). It is suspected that other trace level thiols are present in onions and may react in this manner to produce the corresponding alkyl thiols. The proposed formation of the identified thiols are described in Figure 2. Compound 11 has previously been reported in durian fruit and described to have a roasted onion character (Steinhaus and others 2012), and compound 18 has previously been tentatively identified in onion oil (Farkas and others 1992). Compounds 9, 17, 18, 25, and 26 have been confirmed for the first time in nature and each spectrum is shown in Figure 3. The presence of 3,4-dimethyl-2,3dihydrothiophene-2-thiol isomers 20 and 24 is notable because

the isomers have been identified as thermal degradation products formed from bis-(2-propenyl)-disulfide (Block and Shu 1990). However small quantities are thought to be present in onions, especially when the activity of the LFS enzyme is suppressed (Block 2013). Both isomers were found at low concentrations despite the absence of any detectable bis-(2-propenyl)-disulfide precursor in the thiol enriched extract. Since bis-(2-propenyl)-disulfide does not contain a thiol group its presence in the thiol enriched extract was not to be expected. Therefore for this study, it is believed that both isomers are not artifacts of the chromatographic analysis, but recognized as potential artifacts of extraction.

Conclusions The odor and taste descriptors for each of the synthesized compounds are shown in Table 2. Each possessed some level of onionlike aroma or flavor. Based on the perceived intensity of these synthesized molecules in our sensory evaluations, it is likely that one or more contribute to the aroma of yellow onion juice. To the best of our knowledge, there has not been a quantitative gas chromatography-olfactometry study completed on yellow onion juice. In order to investigate the contribution of these new odorants to yellow onions, additional studies including aroma extract dilution analysis along with quantitation of accurate concentrations and determination of odor-activity values would be the next logical steps.

Acknowledgments We would like to thank the flavor creation and corporate analytical teams at IFF for their support of this research.

References Block E, Shu H-Z. 1990. Onion essential oil chemistry. Cis- and trans-2-mercaptopropyl-3,4dimethyl-2,3-tetrahydrothiophene from pyrolysis of bis(1-propenyl) disulfide. Tetrahedron Lett 31:4999–5002. Block E, Putman D, Zhao S-H. 1992. Allium chemistry: gc-ms analysis of thiosulfinates and related compounds from onion, leek, scallion, shallot, chive, and Chinese chive. J Agric Food Chem 40:2431–8. Block E, Naganathan S, Putnam D, Zhao S-H. 1993. Organosulfur chemistry of garlic and onion: recent results. Pure Appl Chem 65:625–32. Block E, Calvey EM. 1994. Facts and artifacts in allium chemistry. In Sulfur compounds in foods. Washington DC: American Chemical Society. p 63–79. Block E. 2013. Fifty years of smelling sulfur. J Sulfur Chem 34:158–207. Boelens M, van der Linde LM, de Valois PJ, van Dort HM, Takken HJ. 1974. Organic sulfur compounds from fatty aldehydes, hydrogen sulfide, thiols, and ammonia as flavor constituents. J Agric Food Chem 22:1071–6. Brinkman HW, Copier H, de Leuw JJM, Tjan SB. 1972. Components contributing to beef flavor. Analysis of the headspace volatiles of beef broth. J Agric Food Chem. 20:177–81. Cairns TL, Evans GL, Larchar AW, McKusick BC. 1952. Gem-dithiols. J Am Chem Soc 74:3982–5. Deng L-W, Dong L, Wang C, Tocmo R, Wu H, Huang D. 2015. Hydrogen sulfide (H2 S) releasing capacity of essential oils isolated from organosulphur rich fruits and vegetables. J Funct Foods 14:634–340. Farkas P, Hradsky P, Kovac M. 1992. Novel flavour components identified in the steam distillate of onion (Allium cepa L.). Z Lebensm Unters Forsch 195:459–62. Feilberg A, Lokke MM, Edelenbos M, Larsen E. 2012. Investigation of volatiles emitted from freshly cut onions (Allium cepa L.) by real time proton-transfer reaction-mass spectrometry (PTR-MS). Sensors 12:16060–76. Flaig M, Granvogl M. 2015. Quantitation of cis- and trans-3,5-diethyl-1,2,4-trithiolanes in cooked Allium varieties using a stable isotope dilution assay. In Importance of chirality to flavor compounds. Washington DC: American Chemical Society p 109–22.

Vol. 00, Nr. 0, 2017 r Journal of Food Science 7

Food Chemistry

b

Identification of thiols in yellow onion . . . Granvogl M, Christlbauer M, Schieberle P. 2004. Quantitation of the intense aroma compound 3-mercapto-2-methylpentan-1-ol in raw and processed onions (Allium cepa) of different origins and in other Allium varieties using a stable isotope dilution assay. J Agric Food Chem 52:2797– 802. Iranshahi M. 2012. A review of volatile sulfur-containing compounds from terrestrial plants: biosynthesis, distribution and analytical methods. J Essent Oil Res 24:393– 434. Pavez C, Agosin E, Steinhaus M. 2016. Odorant screening and quantitation of thiols in Carmenere red wine by gas chromatography-olfactometry and stable isotope dilution assays. J Food Chem. 64:3417–21. Poll L, Nielsen G, Larsen L. 2003. Formation of aroma compounds and lipoxygenase (EC 1.13.11.12) activity in unblanched leek (Allium ampeloprasum Var. Bulga) slices during longterm frozen storage. J Agric Food Chem 51:1970–76. Schreier P, Full G. 1994. Covalent chromatography. A valuable method for the aroma analysis of thiols at trace levels. Lebensmittelchemie 48:1–4 (in German). Schutte L, Koenders EB. 1972. Components contributing to beef flavor. Natural precursors of 1-methylthio-ethanethiol. J Agric Food Chem 20:181–4.

Food Chemistry 8 Journal of Food Science r Vol. 00, Nr. 0, 2017

Staniewski J, Rijks J-A. 1992. Solvent elimination rate in temperature-programmed injections of large sample volumes in capillary gas chromatography. J Chrom 623:105–13. Steinhaus M, Li J-X, Schieberle P. 2012. Characterization of the major odor-active compounds in Thai durian (Durio zibethinus L. ‘Monthong’) by aroma extract dilution analysis and headspace gas chromatography-olfactometry. J Agric Food Chem 60:11253–262. Steinhaus M, Munafo JP, Didzbalis J, Schnell RJ, Schieberle P. 2014. Characterization of the major aroma-active compounds in mango (Mangifera indica L.) cultivars Haden, White Alfonso, Praya Sowoy, Royal Special, and Malindi by application of a comparative aroma extract dilution analysis. J Agric Food Chem 62:4544–51. Storsberg J, Schulz H, Keusgen M, Tannous F, Dehmer K, Keller E-R. 2004. Chemical characterization of interspecific hybrids between Allium cepa L. and Allium kermesinum Rchb. J Agric Food Chem 52:5499–505. Timokhina L, Voronokov M, Shagun L, Ermolyuk LJ. 2004. Review: germinal dithiols. Sulfur Chem 25:131–53. Widder S, Luntzel CS, Dittner T, Pickenhagen W. 2000. 3-Mercapto-2-methylpentan-1-ol, a new powerful aroma compound. J Agric Food Chem. 48:418–23.