Taste-Active Maillard Reaction Products in Roasted Garlic (Allium

Subscriber access provided by The University of British Columbia Library

Article

Taste-Active Maillard Reaction Products in Roasted Garlic (Allium Sativum) Junichiro Wakamatsu, Timo D. Stark, and Thomas Hofmann J. Agric. Food Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jafc.6b02396 • Publication Date (Web): 05 Jul 2016 Downloaded from http://pubs.acs.org on July 9, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Journal of Agricultural and Food Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 42

Journal of Agricultural and Food Chemistry

1

Taste-Active Maillard Reaction Products in

2

Roasted Garlic (Allium Sativum)

3

Junichiro Wakamatsu, Timo D. Stark, Thomas Hofmann*

4 5

Chair of Food Chemistry and Molecular Sensory Science, Technische

6

Universität München, Lise-Meitner-Straße 34, D-85354 Freising, Germany

7 8 9 10 11 12 13 14 15 16

*

To whom correspondence should be addressed

17

PHONE

+49-8161/71-2902

18

FAX

+49-8161/71-2949

19

E-MAIL

[email protected]

20 21 22 23 1 ACS Paragon Plus Environment


Page 2 of 42

1

ABSTRACT

2

In order to gain first insight into candidate Maillard reaction products formed

3

upon thermal processing of garlic, mixtures of glucose and S-allyl-L-cysteine,

4

the major sulfur-containing amino acid in garlic, were low-moisture heated, and

5

nine major reaction products were isolated. LC-TOF-MS, 1D/2D-NMR, and CD

6

spectroscopy led to their identification as acortatarin A (1), pollenopyrroside A

7

(2), epi-acortatarin

8

hydroxymethyl-2-furanyl)methyl]-1H-pyrrole-2-carbalde-hyde (5), 3-(allylthio)-2-

9

(2-formyl-5-hydroxymethyl-1H-pyrrol-1-yl)propanoic

A

(3),

xylapyrroside

A

(4),

5-hydroxymethyl-1-[(5-

acid

(6),

(4S)-4-

10

(allylthiomethyl)-3,4-dihydro-3-oxo-1H-pyrrolo[2,1-c][1,4]oxazine-6-

11

carbaldehyde

12

3,4-dihydropyrrolo-[1,2-a]pyrazin-2(1H)-yl]propanoic

13

(allylthio)-2-((4S)-4-(allylthiomethyl)-6-formyl-3-oxo-3,4-dihydropyrrolo-[1,2-

14

a]pyrazin-2(1H)-yl)propanoic acid (9). Among the Maillard reaction products

15

identified, compounds 5-9 have not previously been published. The thermal

16

generation of the literature known spiroalkaloids 1-4 is reported for the first time.

17

Sensory analysis revealed bitter taste with thresholds between 0.5 and 785

18

µmol/kg for 1-5 and 7-9. Compound 6 did not show any intrinsic taste (water)

19

but exhibited a strong mouthfullness (kokumi) enhancing activity above 186

20

µmol/kg. LC-MS/MS analysis showed 1-9 to be generated upon pan-frying of

21

garlic with the highest concentration of 793.7 µmol/kg found for 6, thus

22

exceeding its kokumi threshold by a factor of 4 and giving evidence for its

23

potential taste modulation activity in processed garlic preparations.

(7),

(2R)-3-(allylthio)-2-[(4R)-4-(allylthiomethyl)-6-formyl-3-oxoacid

(8),

and

24 25

KEYWORDS: S-allyl-L-cysteine, Maillard reaction, garlic, taste, kokumi

26 2 ACS Paragon Plus Environment

(2R)-3-

Page 3 of 42


1

INTRODUCTION

2 3

Garlic (Allium Sativum) is used as a flavorful spice and a key ingredient in many

4

dishes of various regions, including Asia, the Middle East, northern Africa,

5

southern Europe, and parts of South and Central America, respectively. While a

6

vast majority of studies addressed the chemical structures, formation pathways,

7

and sensory attibutes of volatiles released from freshly crushed garlic1-4 and

8

thermally processed garlic,5-6 the knowledge on non-volatile components is

9

much more fragmentary.

10

Upon crushing of garlic, S-allyl-L-cysteine is enzymatically released from its γ-glutamyl-S-allyl-L-cysteine

by

means

of

the

γ-glutamyl-

11

precursor

12

transpeptidase, e.g. the levels of S-allyl-L-cysteine increase from ~60 µg/g (f.w.)

13

in intact garlic7 to 1000 µg/g after crushing.8 Multiple studies addressed the

14

physiological activities of S-allyl-L-cysteine like anti-cancer, anti-diabetic,

15

cholesterol and blood pressure lowering effects.9-12 Moreover, in-depth chemical

16

analysis revealed new insights in the formation of odor-active volatiles in model

17

reactions between S-allyl-L-cysteine and glucose.13-14 However, investigations

18

on non-volatile reaction products of S-allyl-L-cysteine formed upon thermal

19

processing of garlic are lacking.

20

The objective of the present study was, therefore, to thermally process

21

binary mixtures of S-allyl-L-cysteine and glucose, to identify the non-volatile

22

Maillard reaction products generated, to propose reaction pathways on their

23

formation, to validate their natural occurrence in pan-fried garlic preparations

24

and, finally, to investigate their taste attributes.

25

3 ACS Paragon Plus Environment


1 2

MATERIALS AND METHODS

3 4

Chemicals. The following reagents were obtained commercially: S-allyl-L-

5

cysteine (TCI Europe), S-allyl-D-cysteine (Akos GmbH, Germany), D-(+)-

6

glucose, maltodextrin, monosodium glutamate monohydrate, sodium chloride,

7

potassium phosphate dibasic, sodium hydroxide solution (Sigma Aldrich);

8

potassium dihydrogen phosphate (Merck, Darmstadt, Germany). The yeast

9

extract (Gistex X-II LS) was obtained from FID (Werne, Germany). Water for

10

chromatographic preparations was purified with a Milli-Q Gradient A10 system

11

(Millipore, Schwalbach, Germany), and solvents used were of HPLC-grade

12

(Merck, Darmstadt, Germany). Deuterated solvents were obtained from Euriso-

13

Top (Saarbrücken, Germany). Fresh raw garlic and garlic powder (non-fried and

14

without food additives) were purchased in a local supermarket in Germany.

15

General Experimental Procedures. 1D/2D-NMR spectroscopy 1H, 1H-1H 13

16

COSY, HSQC, HMBC, and

C NMR experiments were performed on an

17

Avance III 500 MHz spectrometer with a CTCI probe and an Avance III 400

18

MHz spectrometer with a BBO probe (Bruker, Rheinstetten, Germany),

19

respectively. Data processing was performed by using Topspin software

20

(version 2.1; Bruker) as well as MestReNova software (version 5.2.3; Mestrelab

21

Research, Santiago de Compostella, Spain). Mass spectra of the compounds

22

were measured on a Waters Synapt G2-S HDMS mass spectrometer (Waters,

23

Manchester, UK) coupled to an Acquity UPLC core system (Waters, Milford, MA,

24

USA). For circular dichroism (CD) spectroscopy, sample solutions of

25

compounds were analyzed by means of a Jasco J-810 spectropolarimeter


Page 4 of 42

Page 5 of 42


1

(Hachioji, Japan). Specific optical rotation was measured employing a P-2000

2

digital polarimeter (Jasco, Gross-Umstadt, Germany). HPLC separations were

3

performed with an analytical HPLC system (Jasco, Gross-Umstadt) consisting

4

of LG-2080-02S pumps, a DG-2080-53 degasser, an AS-2055 Plus auto

5

sampler and a MD-2010 Plus photodiode array detector for analysis of reaction

6

mixtures, and a preparative HPLC system (Jasco, Gross-Umstadt) consisting of

7

two PU-2087 Plus pumps, a DG-2080-53 degasser, a LG-2080-02 gradient unit

8

and a 2010 Plus multiwavelength detector for isolations.

9

Model Reactions. Binary solutions of D-(+)-glucose (125 mmol/L) and S-

10

allyl-L-cysteine (62.5 mmol/L) and individual solutions of D-(+)-glucose (125

11

mmol/L) and S-allyl-L-cysteine (62.5 mmol/L), respectively, in phosphate buffer

12

(100 mmol/L, pH 6.0) were freeze-dried to obtain amorphous powders that were

13

dry-heated for 60 min at 100 °C in closed glass vials using a laboratory oven. In

14

a comparative experiment, D-(+)-glucose (125 mmol/L) and S-allyl-D-cysteine

15

(62.5 mmol/L) were reacted as detailed above. After cooling, each reaction

16

mixture was taken up with water/acetonitrile (80/20, v/v; 1.5 mL) and separated

17

by HPLC on a 4.6 × 250 mm, 5 µm, Phenylhexyl column (Phenomenex,

18

Aschaffenburg, Germany) using the following gradient system with 0.1%

19

aqueous formic acid as solvent A and methanol as solvent B (flow rate: 0.9

20

mL/min): 0-5 min (0% B), 5-25 min (060% B), 25-35 min (6095% B), 35-40

21

min (95% B). Comparing the chromatograms recorded for the individuals and

22

binary reaction mixtures by monitoring the HPLC effluent by UV/Vis-detection

23

between 200 and 800 nm revealed specific peaks at 23.8, 24.9, 30.5, 34.0, 37.2

24

and 38.4 min when D-(+)-glucose was reacted with S-allyl-L-cysteine (Figure 1,

25

A).



1

Identification of D-(+)-Glucose/S-Allyl-L-cysteine Reaction Products.

2

The D-(+)-glucose/S-allyl-L-cysteine reaction mixture, prepared as described

3

above, was dissolved with water/acetonitrile (80/20, v/v; 150 mL) and extracted

4

with ethyl acetate (6 x 300 mL). The organic layer was separated from solvent

5

in vacuum, the residue taken up in water/acetonitrile (70/30, v/v), and the target

6

compounds showed in Figure 1-A were isolated by means of preparative HPLC

7

using a 21.2 × 250 mm, 5 µm, Phenylhexyl column (Phenomenex) using an

8

optimized linear gradient of 0.1% aqueous formic acid as solvent A and

9

methanol as solvent B: 0-3 min (50% B), 3-21 min (5095% B), 21-26 min

10

(95% B). Monitoring the effluent (15 mL/min) at 290 nm, the six target peaks,

11

eluting after 7.0 (GS-1), 8.1 (GS-2), 12.9 (GS-3), 17.5 (GS-4), 21.2 (GS-5), 22.6

12

min (GS-6), were collected, separated from solvent under reduced pressure and,

13

finally, freeze-dried. Thereafter, each fraction was further purified by means of

14

semi-preparative HPLC using a 10 × 250 mm, 5 µm, ODS-Hypersil column

15

(Thermo) with a linear gradient of 0.1% aqueous formic acid as solvent A and

16

acetonitrile as solvent B (Supporting Information, Table S1) to afford the

17

reaction products 1-9 (Figure 2). After removing solvents in vacuum, the

18

isolated compounds 1-5 and 7-9 were freeze-dried twice, while the pH of the

19

fraction containing compound 6 was adjusted to 7.0 with NaOH prior to

20

lyophilization. Structure determination of compounds 1 (25.7 mg), 2 (9.2 mg), 3

21

(23.9 mg), 4 (8.9 mg), 5 (10.8 mg), 6 (1206 mg, as Na salt), 7 (178.8 mg), 8

22

(13.0 mg) and 9 (29.9 mg), isolated in a high purity of at least 95% (HPLC-

23

ELSD), was conducted by means of 1D/2D-NMR and high resolution mass

24

spectrometry with electrospray ionization, and absolute configuration of the


Page 6 of 42

Page 7 of 42


1

compounds was determined by comparing specific optical rotation and CD

2

spectroscopic data to those of references and literature data, respectively.

3

Acortatarin A, 1 (Figure 2): [α]24D +295.9 (c 0.028, MeOH); UV/Vis

4

(MeCN/H2O, 15/85, v/v): λmax = 300 nm; TOF-MS(ESI+): m/z 254.1030 [M+H]+

5

(m/z 254.1031, calcd. for C12H16NO5); 1D/2D-NMR data are given in Table 1.

6

Pollenopyrroside A, 2 (Figure 2): [α]22D +122.7 (c 0.04, MeOH); UV/Vis

7


8

(m/z 254.1031, calcd. for C12H16NO5); 1D/2D NMR data are given in Table 2.

9

epi-Acortatarin A, 3 (Figure 2): [α]21D －138.9 (c 0.05, MeOH); UV/Vis

10


11


12

Xylapyrroside A, 4 (Figure 2): [α]23D － 156.0 (c 0.09, MeOH); UV/Vis

13


14


15

5-Hydroxymethyl-1-[(5-hydroxymethyl-2-furanyl)methyl]-1H-pyrrole-2-

16

carbaldehyde, 5 (Figure 2): UV/Vis (MeCN/H2O, 50/50, v/v): λmax = 228, 296

17

nm; TOF-MS(ESI+): m/z 258.0739 [M+Na]+ (m/z 258.0742, calcd. for

18

C12H13NO4Na); 1D/2D NMR data are given in Table 3.

19

3-(Allylthio)-2-(2-formyl-5-hydroxymethyl-1H-pyrrol-1-yl) propanoic acid (as

20

sodium salt), 6 (Figure 2): UV/Vis (MeOH/H2O, 75/25, v/v): λmax = 204, 296 nm;

21

TOF-MS (ESI－): m/z 268.0642 [M-H]－ (m/z 268.0644, calcd. for C12H14NO4S);

22

1D/2D NMR data are given in Table 3. CD spectrum is shown in Figure 3.

23

(4S)-4-(Allylthiomethyl)-3,4-dihydro-3-oxo-1H-pyrrolo[2,1-c][1,4]oxazine-6-

24

carbaldehyde, 7 (Figure 2): UV/Vis (MeOH/H2O, 80/20, v/v): λmax = 204, 292

25

nm; TOF-MS(ESI+): m/z 252.0695 [M+H]+

(m/z 252.0694, calcd.


for


1

C12H14NO3S); CD (50% MeOH, 0.398 mmol/L): λmax(∆ε) = 295 (+65.1), 247 (－

2

20.6); 1D/2D NMR data are given in Table 3. CD spectrum is shown in Figure

3

3.

4

(2R)-3-(Allylthio)-2-[(4R)-4-(allylthiomethyl)-6-formyl-3-oxo-3,4-

5

dihydropyrrolo-[1,2-a]pyrazin-2(1H)-yl] propanoic acid, 8, (Figure 2): UV/Vis

6

(MeOH/H2O, 30/70, v/v): λmax = 200, 296 nm; TOF-MS (ESI－): m/z 393.0949

7

[M-H] － (m/z 393.0943, calcd. for C18H21N2O4S2); CD (50% MeOH, 0.538

8

mmol/L): λmax(∆ε) = 299 (－108.7), 252 (+49.6), 210 (－53.8); 1D/2D NMR data

9

are given in Table 4.

10

(2R)-3-(Allylthio)-2-((4S)-4-(allylthiomethyl)-6-formyl-3-oxo-3,4-

11

dihydropyrrolo-[1,2-a]pyrazin-2(1H)-yl) propanoic acid, 9 (Figure 2): UV/Vis

12

(MeOH/H2O, 30/70, v/v): λmax = 200, 296 nm; TOF-MS(ESI－): m/z 393.0948 [M-

13

H]－ (m/z 393.0943, calcd. for C18H21N2O4S2); CD (50% MeOH, 0.538 mmol/L):

14

λmax(∆ε) = 299 (+139.3), 251 (－46.9), 240 (－38.5) 229 (－41.0), 207 (－11.2);

15

1D/2D NMR data are given in Table 4.

16

Synthesis of Decarboxy-Derivatives (10, 11) of Compounds 8 and 9.

17

To identify the absolute configuration of compounds 8 and 9, by means of CD

18

spectroscopy, decarboxylated forms of 8 and 9 were synthesized as reference

19

compounds exhibiting only a single chiral carbon atom. A solution of S-allyl-L-

20

cysteine (2137.2 µmol) in phosphate buffer (100 mmol/L, pH 6.0; 20 mL) was

21

added to a solution of compound 7 (712.4 µmol) in acetone (1 mL), the mixture

22

was freeze-dried and then, dry-heated for 50 min at 100 °C in a closed glass vial.

23

After cooling, the reaction mixture was taken up in water/acetonitrile (50/50, v/v;

24

20 mL) and separated by means of HPLC using a 21.2 × 250 mm, 5 µm,


Page 8 of 42

Page 9 of 42


1

Phenylhexyl column as the stationary phase. Monitoring the effluent at 290 nm,

2

chromatography was performed using the following linear gradient with 0.1%

3

aqueous formic acid as solvent A and methanol as solvent B (flow rate: 15

4

mL/min): 0-3 min (70% B), 3-23 min (7095% B). Next to compounds 8 (Rt =

5

17.9 min; yield: 10.2 mg) and 9 (Rt = 19.5 min; yield: 17.1 mg), their racemic

6

decarboxylated homologs 10 and 11 eluted after 20.8 min (yield: 15.5 mg). The

7

chemical structure of the racemic compounds 10 and 11 was determined by

8

means of LC-TOF-MS and 1D/2D-NMR spectroscopy. Chiral resolution of

9

racemic compounds 10 (Rt = 11.2 min; Figure 4) and 11 (Rt = 12.4 min; Figure

10

4) was performed by chiral chromatography using a 4.6 × 250 mm, 5 µm,

11

Cellulose-2 column (Phenomenex) and the following gradient system with 0.1%

12

aqueous acetic acid as solvent A and methanol as solvent B (flow rate: 0.8

13

mL/min): 0-3 min (95% B), 3-18 min (95100% B), then 10 and 11 were further

14

analyzed using LC-TOF-MS, NMR and CD spectroscopy.

15

(4S)- and (4R)-2-(Allylthioethyl)-4-(allylthiomethyl)-3-oxo-3,4-dihydropyrrolo

16

[1,2-a]pyrazin-2(1H)-6-carbaldehyde, 10 and 11 (Figure 4): UV/Vis (100%

17

MeOH): λmax = 204, 296 nm; TOF-MS(ESI+): m/z 351.1201 [M+H]+ (m/z

18

351.1201, calcd. for C17H23N2O2S2); CD of (4S)-configured 10 (100% MeOH,

19

0.571 mmol/L): λ max(∆ε) = 296 (+172.1), 240 (－66.2), 206 (－7.6); CD of (4R)-

20

configured 11 (100% MeOH, 0.571 mmol/L): λ

21

(+77.1), 208 (+10.7). 1D/2D NMR data are given in Table 4.

max(∆ε)

= 296 (－201.7), 237

22

Quantitation of Compounds 1-9 in Garlic Preparations. To mimic a

23

kitchen-type preparation of garlic, raw garlic (28 g) was cut in about 1 mm slices

24

and pan-fried at 250 °C for 8 min until its color reached light brown. Moreover,

25

garlic powder (15 g) was pan-fried at 250 °C for 3 min. After cooling, the garlic 9 ACS Paragon Plus Environment


were

frozen

in

liquid

nitrogen,

ground,

then

Page 10 of 42

1

samples

mixed

with

2

water/acetonitrile (60/40, v/v; 30 mL), and, finally, made up to 100 mL with the

3

same solvent. After ultrasonication of each garlic suspension for 10 min,

4

aliquots were taken, filtered (0.45 µm syringe filters), and analyzed by means of

5

UPLC-MS/MS (ESI) analysis using an Acquity UPLC i-class core system

6

equipped with a 2.0 × 150 mm, 1.7 µm, BEH phenyl column (Waters, Eschborn,

7

Germany) connected to a Xevo T-Q-S mass spectrometer (Waters).

8

Chromatography was performed at 50 °C with a flow rate of 0.4 mL/min using a

9

gradient of 0.1% aqueous formic acid as solvent A and 0.1% formic acid in

10

acetonitrile as solvent B (Supporting Information, Table S2). Using the positive

11

electrospray ionization mode, the ion source parameters were set as follows:

12

capillary voltage (2.5 kV), sampling cone voltage (24 V), source offset (50 V),

13

source temperature (150 °C), desolvation temperature (600 °C), cone gas (150

14

L/h), desolvation gas (850 L/h), collision gas (0.15 mL/min), and nebulizer gas

15

(7.0 bar). The MS/MS parameters of each compound were tuned using

16

IntelliStart of MassLnyx 4.1 software (Waters), and quantitative analysis was

17

conducted through the multiple reaction monitoring mode with fixed mass

18

transitions as follows: m/z 254.2206.1 for 1-4, m/z 258.3200.1 for 5, m/z

19

270.2228.1 for 6, m/z 252.2178.0 for 7, and m/z 395.299.1 for 8-9.

20

Standard solutions (1 mg/mL) of each compound were prepared in

21

water/acetonitrile (60/40, v/v) and proper dilutions of them were analyzed the

22

same way for standard curves of the compounds.

23

Analytical Sensory Experiments. Panel Training. Seven subjects (4

24

women and 3 men, aged 25-41), who gave consent to participate in the sensory

25

tests of the present investigation and have no history of known taste disorders,


Page 11 of 42


1

were trained to evaluate the taste of aqueous solutions of the following standard

2

taste compounds in bottled water (pH 6.8): sucrose (50 mmol/L) for sweet taste,

3

lactic acid (20 mmol/l) for sour taste, NaCl (20 mmol/L) for salty taste, caffeine

4

(1 mmol/L) for bitter taste, monosodium L-glutamate (3 mmol/L) for umami

5

taste, using the sip-and-spit method.15 Viscosity perception and kokumi activity

6

(mouthfulness enhancement) were trained as reported earlier with a gelatin

7

solution (0.5% in water) and a defined model broth spiked with glutathione (5.0

8

mmol/L).16-18 Prior to sensory analyses, purified compounds were analytically

9

confirmed to be essentially free of solvents and buffer compounds,15,18 and the

10

purity of each tested compound was checked by 1H NMR spectroscopy as well

11

as UPLC-MS. To prevent cross-modal interactions with olfactory inputs, the

12

panelists used nose clips and sensory sessions were performed at 22 °C in an

13

air-conditioned room.

14

Bitter Taste Thresholds. Detection thresholds of compounds 1-6 were

15

determined in bottled water (pH 6.8) and of compounds 7-9 in bottled water

16

containing 5% ethanol by means of a three-alternative forced choice test.19

17

Values between individuals and three separate sessions differed not more than

18

plus or minus one dilution step; that is, a threshold value of 20.0 µmol/L for

19

compound 8 represents a range of 10 – 40 µmol/L.

20

Kokumi activity. Taste-modulating activity of compound 6 was further

21

assessed in a model broth which consisted of monosodium glutamate

22

monohydrate (1.9 g/L), yeast extract (2.1 g/L), maltodextrin (6.375 g/L) and

23

sodium chloride (2.9 g/L) in bottled water.20 The pH value of the broth was

24

adjusted to 5.9 using trace amounts of formic acid (0.1 mmol/L). Several

25

concentrations of 6 in the broth were prepared. Sensory analysis was



1

conducted by means of the three-alternative forced choice test with the trained

2

panelists without nose clips in three independent sessions.

3 4

RESULTS AND DISCUSSIONS

5 6

Since S-allyl-L-cysteine is the major sulphur-containing amino acid in garlic

7

and non-enzymatic reactions between amino components and reducing

8

carbohydrates have been reported to generate tastants as well as taste

9

modulating compounds upon food processing,18,20-29 the potential of Maillard-

10

type reactions between S-allyl-L-cysteine and D-glucose should be investigated

11

in order to gain some first insight into process-induced formation of taste-active

12

compounds in garlic.

13

Identification of Maillard Reaction Products from S-Allyl-L-cysteine

14

and D-Glucose. Binary mixtures of S-allyl-L-cysteine and D-glucose were heat-

15

treated for 60 min at 100 °C under low-moisture conditions, the reaction

16

products analyzed by means of HPLC-UV/Vis (Figure 1, A) and compared to

17

the reaction products spectrum obtained after heating the precursors separately

18

(Figure 1, B, C). Six characteristic peaks were detected at 23.8, 24.9, 30.5,

19

34.0, 37.2 and 38.4 min only in the binary reaction mixture (Figure 1, A), thus

20

indicating S-allyl-L-cysteine/D-glucose reaction products. To identify these

21

compounds, the binary reaction mixture was extracted with ethyl acetate to

22

separate these reaction products from more polar components. Iterative

23

preparative and semipreparative HPLC revealed a total of nine purified reaction

24

products, namely 1 (Rt= 23.8 min), 2-5 (Rt= 24.9 min), 6 (Rt= 30.9 min), 7 (Rt=

25

34.0 min), 8 (Rt= 37.2 min), and 9 (Rt= 38.4 min).


Page 12 of 42

Page 13 of 42


1

High-resolution mass spectrometric analysis of compound 1 revealed m/z

2

254.1030 as the pseudomolecular ion ([M+H]+), suggesting an elemental

3

composition of C12H15NO5. The 1H NMR spectrum of 1 showed a proton signal

4

of an aldehyde group [H-C(14)] resonating at 9.42 ppm and two olefinic protons

5

observed at 7.03 [H-C(12)] and 6.08 ppm [H-C(11)]. In addition, six non-

6

equivalent heteroatom-bound methylene protons showed resonances at 4.44

7

[H-C(10b)], 4.17 [H-C(10a)], 3.54 [H-C(15b)], 3.44 [H-C(15a)], 2.30 [H-C(4b)],

8

and 1.97 ppm [H-C(4a)], two equivalent methylene protons were detected at

9

4.85 ppm [H-C(7)], the protons H-C(3) and H-C(2) of secondary alcohols/ethers

10

(4.26 and 3.88 ppm) were observed in the 1H NMR spectrum. Interpretation of

11

the

12

the olefinic quaternary carbons C(8) and C(13) at 135.0 and 130.6 ppm, and the

13

two olefinic carbons C(12) and C(11) at 123.8 and 104.8 ppm could be

14

assigned and indicated a pyrrole ring with an aldehyde function based on the

15

homonuclear COSY correlation between H-C(11) and H-C(12) and the

16

heteronuclear

17

C(12)C(8,11,13,14), and H-C(14)C(13). In addition, a morpholine moiety

18

with a quaternary carbon at 102.4 ppm (C(5)) and two heteroatom-bearing

19

methylene carbons at 57.2 (C(7)) and 50.8 ppm (C(10)) could be deduced from

20

the spectroscopic data. Moreover, a deoxysugar moiety was proposed by

21

assigning the carbons resonating at 102.4 (C(5)), 87.5 (C(2)), 69.9 (C(3)), 61.1

22

(C(15)), and 44.5 ppm (C(4)) and considering the COSY correlations H-C(2)/H-

23

C(3), H-C(2)/H-C(15), H-C(3)/H-C(4), as well as the heteronuclear (HMBC)

24

connectivities

25

C(4)C(2,3,5). Taking all spectroscopic data into account, the reaction product

13

C NMR data and HSQC correlations, the aldehyde C(15) at 179.0 ppm,

HMBC

correlations

H-C(7)C(5,8),

H-C(11)C(8,12,13),

H-C(10)C(5,8),


H-C(2)C(3),

H-

H-


Page 14 of 42

1

1 was identified to have a acortatarin-like structure (Figure 2), that was further

2

confirmed by comparing the NMR data (Table 1) measured in methanol-d4 with

3

those reported in the literature recently for acortatarin A, an antioxidative

4

sporoalkaroid isolated from

5

configuration of 1 was determined by measurements of specific optical rotation.

6

The value ([α]24D +296 (MeOH)) of 1 was very similar to that of acortatarin A

7

([α]22D +255 (MeOH)) and, consequently, the structure of 1 was unequivocally

8

identified as acortatarin A.31 This is the first time that the generation of such a

9

sugar-morpholine spiroketal pyrrole alkaloid is reported by means of a Maillard-

10

tatarinowii.30

Acorus

Finally, the

absolute

type reaction.

11

Using the same analytical strategy, the chemical structures of compounds

12

2 - 4 were elucidated. While compound 3 showed almost identical MS and NMR

13

data (Table 1) when compared to acortatarin A (1), the specific optical rotation

14

value of ([α]21D –138.9 (MeOH)) recorded for 3 was not in agreement with that

15

of acortatarin A ([α]22D +255 (MeOH)) but with that reported in the literature for

16

epi-acortatarin A ([α]19D –111 (MeOH)).32 The 1H NMR spectrum additionally

17

measured in CD3OD also matched that of epi-acortatain A (Table 1),32 thus

18

leading to the identification of the Maillard reaction product 3 as epi-acortatarin

19

A (Figure 2). Although MS, 1H, and

20

were very similar to those of 1 and 3, the chemical shift of 93.4 and 94.9 ppm

21

recorded for the quaternary carbon C(6) in 2 and 4, respectively, differed from

22

the data found for acortatarin A (1, 102.4 ppm) and epi-acortatarin A (3, 102.4

23

ppm). Whereas the spiro-carbon atom C(5) in compounds 1 and 3 showed a

24

HMBC correlation to two equivalent methylene protons (H-C(4)) in the

25

deoxysugar moiety, the quaternary carbon C(6) in 2 and 4 exhibited HMBC

13

C NMR spectra of compounds 2 and 4


Page 15 of 42


1

correlations to two pairs of methylene protons (H-C(5), H-C(2)). These

2

differences clearly indicated that the deoxysugar moieties of 2 and 4 consisted

3

of six heteroatoms, hence the structures of 2 and 4 were assigned as

4

pollenopyrroside A and xylapyrroside A.32,33 This was confirmed by comparing

5

the optical rotation values recorded for compound 2 ([α]22D +123 (MeOH)) and 4

6

([α]23D –156 (MeOH)) and the 1H NMR spectra (Table 2) with those reported

7

recently for pollenopyrroside A and xylapyrroside A isolated from bee-collected

8

Brassica campestris pollen and from Xylaria nigripes, respectively.32,33 To the

9

best of our knowledge, these sugar-morpholine spiroketal pyrrole alkaloids have

10

not been earlier reported as Maillard reaction products.

11

The high resolution mass spectrum of compound 5 showed m/z 258.0739

12

as the pseudomolecular ion ([M+Na]+), thus suggesting a molecular formula of

13

C12H13NO4. The 1H NMR spectrum and HSQC correlations of 5 exhibited an

14

aldehyde proton signal at 9.50 ppm (H-C(15)), four olefinic protons at 6.14 (H-

15

C(4), 6.17 (H-C(3)), 6.23 (H-C(10)) and 6.98 ppm (H-C(11)), and three pairs of

16

equivalent methylene protons at 4.30 (H-C(6)), 4.61 (H-C(13)) and 5.58 ppm (H-

17

C(7)). In addition, four carbon atoms resonating at 107.7 (CH(3)), 108.3 (CH(4)),

18

150.2 (C(5)) and 155.2 ppm (C(2)) as well as the coupling constant of 3.1 Hz

19

recorded for the olefinic protons H-C(3) and H-C(4) indicated the presence of a

20

furan structure.34 Then, four carbon signals detected at 109.6 (CH(10)), 124.1

21

(CH(11)), 131.5 (C(11)) and 143.7 ppm (C(9)) and the coupling constant of 4.0

22

Hz found for the olefinic protons H-C(10) and H-C(11) indicated the presence of

23

a pyrrole ring,35 which was supported by HMBC correlations H-C(3)C(2,4,5),

24

H-C(4)C(2,3,5), H-C(10)C(9,11,12) and H-C(11)C(9,10,12). Moreover,

25

the connection of both moieties and the positions of two hydroxymethyl and one 15 ACS Paragon Plus Environment


1

aldehyde functions were confirmed through HMBC correlations of H-

2

C(7)C(4,5,9,12), H-C(14)C(12), H-C(13)C(9,10) and H-C(6)C(2,3). In

3

summary, compound 5 was identified as 5-hydroxymethyl-1-[(5-hydroxymethyl-

4

2-furanyl)methyl]-1H-pyrrole-2-carbaldehyde (Figure 2) that, to the best of our

5

knowledge, has not yet been reported in literature.

6

The mass spectra of compound 6 showed a pseudomolecular ion of m/z

7

268.0646 [M-H]-, thus suggesting a molecular formula of C12H15NO4. 1H NMR

8

spectroscopy of compound 6 revealed four non-equivalent methylene protons at

9

2.76 (H-C(8)), 2.92 (H-C(8)), 2.92 (H-C(7)) and 3.37 ppm (H-C(7)), one olefinic

10

methine proton at 5.63 ppm (H-C(9)) and two non-equivalent olefinic methylene

11

protons at 4.38 (H-C(10)) and 4.43 ppm (H-C(10)). These chemical shifts and

12

the homo- (COSY) and hetereonuclear correlation experiments (HSQC, HMBC)

13

indicated the presence of an allylthiomethyl group as found in S-allyl-L-cysteine

14

(Figure 4). Moreover, the carbon resonances at 55.7 (C(11)), 110.2 (C(3)),

15

123.4 (C(4)), 132.8 (C(5)), 144.5 (C(2)), and 179.2 ppm (C(12)) were very

16

similar to those of the pyrrole motif in compound 5. Also the proton signals

17

observed at 4.99 (2H, H-C(11)), 6.19 (H-C(3)), 6.88 (H-C(4)) and 9.30 ppm (H-

18

C(12)) indicated the existence of a 5-hydroxymethyl-1H-pyrrole-2-carbaldehyde

19

moiety. HMBC correlations confirmed that S-allyl-L-cysteine is incorporated via

20

its nitrogen atom into the pyrrole motif, thus leading to the proposed structure of

21

3-(allylthio)-2-(2-formyl-5-hydroxymethyl-1H-pyrrol-1-yl)propanoic acid (Figure

22

2). Although the proton signal of the chiral carbon C(6) could not be assigned in

23

the NMR spectra by means of 2D correlations, the expected signal resonated at

24

5.76 ppm as a very broad singlet in the 1H NMR spectrum. As no specific cotton

25

effects could be observed in the CD spectrum (Figure 3), compound 6 seems


Page 16 of 42

Page 17 of 42


1

to be present as a racemate. This is well in line with the high acidity of C-H

2

bond and explained the fast protium/deuterium exchange observed during 1H

3

NMR measurements. To the best of our knowledge, this compound has not

4

been reported before in literature, but two analogs with glycine and alanine

5

were reported.36,37

6

1D/2D NMR data recorded for compound 7 were rather similar to those of

7

compound 6. However, compound 7 showed some differences in the

8

resonances of the methine proton H-C(5) at 5.59 ppm as well as in the two non-

9

equivalent methylene protons of H-C(2) resonating at 5.59 and 5.77 ppm.

10

Moreover, HMBC correlations showed heteronuclear connectivity between the

11

non-equivalent methylene protons H-C(2) and the quaternary carbon C(6),

12

which was not detected for compound 6. Taking all these date into

13

consideration, the chemical structure of 7 was determined as 4-(allylthiomethyl)-

14

3,4-dihydro-3-oxo-1H-pyrrolo[2,1-c][1,4]oxazine-6-carbaldehyde,

15

corresponding lactone of compound 6. In addition, all expected COSY and

16

HMBC correlations in 7 could be assigned (Figure 3) and the pseudomolecular

17

ions m/z 252.0697 [M+H]+ and 274.2748 [M+Na]+, detected by means of LC-

18

TOF-MS, confirmed an elemental composition of compound 7. The CD

19

spectrum of 7 was well in line with those of structurally related S-configured

20

derivatives.38 Therefore, the stereochemistry of the chiral center C(5) could be

21

deduced as S-configuration (Figure 3).

the

22

LC-TOF-MS and NMR spectroscopic analysis of compounds 8 and 9

23

revealed rather similar data and indicated diastereomers with an empirical

24

formula of C18H22N2O4S2 (Supporting Information, Figure S51-52, 58-59). The

25

NMR signals recorded for compound 8 for the aliphatic methylene protons at



1

2.59 (H-C(16)), 2.88 (H-C(16)) and 3.08 ppm (H-C(15)), the olefinic methylene

2

protons at 4.93 and 5.01 ppm (H-C(18)), the olefinic methine protons at 5.55 (H-

3

C(17)), 6.29 (H-C(7)) and 7.20 ppm (H-C(8)), the aliphatic methine proton at

4

5.54 ppm (H-C(5)), as well as the aldehyde proton at 9.47 (H-C(19)) were quite

5

similar to those found for compound 7. While also the carbon resonances at

6

34.1 (C(16)), 34.2 (C(15)), 58.3 (C(5)), 106.5 (C(7)), 117.8 (C(18)), 125.4 (C(8)),

7

129.9 (C(9)), 133.9 (C(17)), 134.7 (C(3)), 166.0 (C(6)) and 179.0 ppm (C(19))

8

were similar to those detected for compound 7, the methylene carbon C(2)

9

showed a lower chemical shift (41.8 ppm) than that of the corresponding carbon

10

in compound 7 (64.2 ppm). Two methylene protons of H-C(2) resonating at

11

4.69/4.93 ppm were also different from the corresponding methylene protons in

12

7 (5.59/5.77 ppm), thus suggesting a hetero-lactone structure of the target

13

compounds related to the lactone ring in compound 7. Further interpretation of

14

the NMR spectra revealed the presence of a S-allyl-L-cysteine moiety with

15

typical signals for the methine proton at 4.93 ppm (H-C(10)), four aliphatic

16

methylene protons at 2.96 (H-C(11)), 3.07 (H-C(11)) and 3.12 ppm (2H, H-

17

C(12)), the olefinic methine proton at 5.71 ppm (H-C(13)), and two olefinic

18

protons at 5.10 ppm [2H, H-C(14)]. The HMBC experiment showed the aliphatic

19

methine proton (H-C(10)) possessing heteronuclear connectivity to carbons

20

C(2) and C(6) and confirmed the S-allyl-L-cysteine moiety to be connected via

21

the nitrogen atom being a part of a 2-ketopiperazine ring system. Therefore, the

22

structure of the diastereomers 8 and 9 was identified as 3-(allylthio)-2-(4-

23

(allylthiomethyl)-6-formyl-3-oxo-3,4-dihydropyrrolo[1,2-a]pyrazin-2(1H)-yl)

24

propanoic acid (Figure 2).


Page 18 of 42

Page 19 of 42


1

In order to determine the absolute configuration of 8 and 9, the two

2

structurally related compounds 10 and 11 (Figure 4), differing from 8 and 9 by

3

lacking the carboxyl group, were prepared by reacting purified compound 7 with

4

S-allyl-L-cysteine. After purification, the racemic mixture of 10 and 11 was

5

separated by means of chiral HPLC separation and the individual stereoisomers

6

were analyzed by means of mass spectrometry and NMR spectroscopy (Table

7

4). The absolute configuration of 10 and 11 was then determined by comparing

8

the CD spectroscopic data obtained to those obtained for the S-configured

9

compound 7 showing a positive cotton effect at 295 nm (Supporting Information

10

S71). As the CD-spectrum of 10 was the mirror image of that recorded for 11,

11

the stereochemistry of 11 was deduced as the R-configuration. Comparison of

12

the CD-spectra of compounds 8-11 enabled the deduction of the steric

13

configuration at the chiral carbon C(5). As shown in Figure 4, the CD spectrum

14

of 8 was well in line with that of compound 11 and that of 9 matched the

15

spectrum obtained for 10, thus indicating the stereo chemistry at C(5) in

16

compound 8 to be R- and in compound 9 to be S-configured, respectively. As

17

the CD spectra of 8 and 9 were not mirror images, both compounds were

18

deduced to be diastereomers with the same configuration at the remaining

19

chiral carbon C(10). Absolute configuration of the chiral center C(10) was

20

assigned by means of further model reactions with S-allyl-L-cysteine and S-allyl-

21

D-cysteine,

22

and 11 were generated upon the reaction of compound 7 with S-allyl-L-cysteine,

23

thus indicating that formation of the ketopiperazine ring in 8 - 11 would start by

24

nucleophilic attack with the amino function of S-allyl-L-cysteine to the carboxyl

25

function of 7. As S-allyl-L-cysteine acts as a nucleophile in the reaction, the

respectively. As reported above, next to 8 and 9 also compounds 10



1

original R-configuration of S-allyl-L-cysteine will be maintained in the S-allyl-L-

2

cysteine moiety of 8 and 9. S-allyl-L-cysteine was more reactive than S-allyl-D-

3

cysteine for production of the sulfur compounds 6 - 9 (Figure 1, A, D).

4

Proposed Reaction Pathways Governing the Formation of Compoun

5

ds 1-9. Considering the joint 5-hydroxymethyl- or aminomethyl-pyrrol-2-

6

carbaldehyde moiety structure in compounds 1-9 and the relatively high

7

amounts of 5-hydroxymethylfurfural found in the model systems (Figure 1, A,

8

D), a reaction cascade was proposed starting from the 3-deoxyhexosone (1)

9

generated from glucose in course of the Maillard reaction (Figure 5). Reaction

10

of 1 with amino acids like S-allyl-L-cysteine reveals the 1-amino-1-deoxy-

11

fructose (2) that reacts with the 3,4-dideoxyosone (3), generated from 1 upon

12

water elimination and tautomerization, to give the N-(1-amino-1-deoxy-

13

fructosyl)-5-hydroxymethyl-pyrrol-2-carbaldehyde (4). This key intermediate can

14

be either cyclized via the hydroxyl group of the 5-hydroxymethyl-pyrrol-2-

15

carbaldehyde moiety to give the target furanosides (1, 3) and pyranosides (2, 4)

16

via the common intermediate 5 (Figure 5, flow: bc/d), or via the hydroxyl

17

group at position 5 of the N-1-amino-1-deoxy-fructosyl) moiety to give the 5-

18

hydroxymethyl-1-[(5-hydroxymethyl-2-furanyl)methyl]-1H-pyrrole-2-

19

carbaldehyde (5) via the intermediate 6 (Figure 5, flow: a). On the other hand,

20

reaction of the 3,4-dideoxyosone (3) with S-allyl-L-cysteine (SAC) generates the

21

3-(allylthio)-2-(2-formyl-5-hydroxymethyl-1H-pyrrol-1-yl)propanoic acid (6) and

22

its corresponding lactone (7) upon intramolecular esterification (Figure 5).

23

Reaction of the lactone (7) with another S-allyl-L-cysteine forms a dipeptide

24

intermediate 7 that, upon dehydration at the 5-hydroxymethyl pyrrol moiety,

25

followed by another ring closure via the peptide nitrogen atom to give rise to the


Page 20 of 42

Page 21 of 42


1

3-(allylthio)-2-(4-(allylthiomethyl)-6-formyl-3-oxo-3,4-dihydropyrrolo[1,2-

2

a]pyrazin-2(1H)-yl) propanoic acids 8 and 9.

3

Quantitation of Maillard Reaction Products in Garlic Preparations. As

4

compounds 1-9 were found to be generated upon the Maillard reaction of

5

glucose and S-allyl-L-cysteine, a major sulfur-containing amino acid in garlic,

6

their natural occurrence in garlic preparations should be shown and their

7

concentrations determined accordingly. To mimic a kitchen-type preparation of

8

garlic, freshly cut garlic slices were pan-fried at 250 °C for 8 min and

9

commercial garlic powder was pan-fried at 250 °C for 3 min and, then, extracted

10

with water/acetonitrile and analyzed by means of UPLC-MS/MS (ESI). As the

11

control, non-treated fresh garlic and garlic powder was analyzed. All

12

compounds (1-9) could be identified in pan-fried garlic powder, compounds 1-6,

13

but not 7-9, were detected in pan-fried garlic slices, while the non-thermally

14

processed control samples were essentially free of any of the Maillard reaction

15

products (data not shown). Quantitative analysis revealed rather low amounts of

16

6.8 - 91.6 and 2.1-12.5 nmol/kg for non-sulfur containing compounds 1-5 in

17

roasted garlic slices and roasted garlic powder (Table 5). In comparison, the

18

sulfur-containing compounds 6-9 were found in high concentrations in roasted

19

garlic

20

yl)propanoic acid (6) as the major component with a concentration of 793.7

21

µmol/kg, followed by its lactone 7 (51.4 µmol/kg), and the diastereomers 8 (5.9

22

µmol/kg) and 9 (2.8 µmol/kg), respectively.

powder

with

3-(allylthio)-2-(2-formyl-5-hydroxymethyl-1H-pyrrol-1-

23

Sensory Evaluation of Maillard Reaction Products. Sensory analysis of

24

the purified compounds 1-9 revealed an intrinsic bitter taste for all compounds

25

with the exception of 6 and 7, that did not show any taste activity on their own



1

until the highest tested concentration of 1000 and 500 µmol/kg (Table 5). The

2

bitter taste threshold concentrations of the Maillard reaction products 1-5, 8 and

3

9 ranged from 0.5 (9) to 785 µmol/kg (5). Interestingly, the diastereomers 1 (300

4

µmol/kg) and 3 (370 µmol/kg) as well as 2 (471 µmol/kg) and 4 (263 µmol/kg)

5

showed similar bitter thresholds, whereas the threshold of the (S,R)-configured

6

9 was 40 times below the threshold found for its (R,R)-configured diastereomer

7

8, thus indicating the absolute configuration of the ketopiperazine motif in 8 and

8

9 to play an important role in their taste activity. As compound 6 did not show

9

any intrinsic taste in water and was found as the quantitatively predominant

10

Maillard reaction product in thermally treated garlic (Table 5), this compound

11

was evaluated in a savory tasting model broth for taste modulatory activity.

12

Most interestingly, at a concentration of 500 µmol/kg the 3-(allylthio)-2-(2-

13

formyl-5-hydroxymethyl-1H-pyrrol-1-yl)propanoic acid (6) showed a strong

14

mouthfullness enhancing activity and increased complexity and long-lastingness

15

of the savory sensation, a phenomenon that is referred to kokumi activity.39 In

16

the model broth, a threshold concentration of 186 µmol/kg was determined for

17

this kokumi effect that is even somewhat below the threshold data found for the

18

kokumi effect of alliin (283 µmol/kg) and glutathione (326 µmol/kg).39

19

In summary, four spiroalkaloids (1-4), one new pyrrole derivative (5) and

20

four new sulfur containing Maillard reaction products (6-9) could be isolated

21

from a glucose/S-allyl-L-cysteine model system, their structures determined,

22

and their concentrations quantitated in roasted garlic preparations. The high

23

concentration

24

hydroxymethyl-1H-pyrrol-1-yl)propanoic acid (6) in pan-fried garlic powder

25

exceeded its kokumi threshold by a factor of 4, thus giving evidence for a taste

of

793.7

µmol/kg

found

for

3-(allylthio)-2-(2-formyl-5-


Page 22 of 42

Page 23 of 42


1

modulating activity of this compound in thermally processed garlic preparations.

2

Moreover, the concentration of (R,S)-3-(allylthio)-2-(4-(allylthiomethyl)-6-formyl-

3

3-oxo-3,4-dihydropyrrolo[1,2-a]pyrazin-2(1H)-yl)

4

µmol/kg) in pan-fried garlic powder exceeded its bitter taste threshold (0.5

5

µmol/kg) by a factor of 5, thus indicating this compound as a potential

6

contributor to the slight bitter taste of pan-fried garlic. To fully exploit the

7

potential of taste modulators, future programs will have to be focused on the

8

activity-guided fractionation of fresh and thermally treated garlic preparations.

propanoic

acid

(9)

(2.8

9 10 11

Author information

12

Corresponding author

13

Phone: +49-8161-71-2902. Fax: +49-8161-71-2949.

14

E-mail: [email protected]

15 16

Funding

17

We are grateful to Wakunaga Pharmaceutical Co. Ltd., for financial support.

18 19

Notes

20

The authors declare no competing financial interest.

21 22

Acknowledgments

23

We thank the NMR team, which is managed by Dr. Oliver Flank, of the Chair of

24

Food Chemistry and Molecular Sensory Science for performing the NMR

25

measurements on the isolated compounds. 23 ACS Paragon Plus Environment


1 2

Supporting information

3

Analytical conditions for HPLC separations and quantitative analysis on the

4

isolated compounds, 1D/2D NMR and mass spectra of all compounds, and CD

5

spectra of compounds 10 and 11 are shown in the supporting information. This

6

information is available free of charge via the Internet http://pubs.acs.org.

7 8


Page 24 of 42

Page 25 of 42


1

LITERATURE CITED

2

1.

Block, E.; Dane, J.J.; Thomas, S.; Cody, R.B. Applications of direct

3

analysis in real time-mass spectrometry (DART-MS) in Allium chemistry. 2-

4

Propene-sulfenic and –sulfinic acids, diallyl trisulfane S-oxide and other

5

reactive sulfur compounds from cut garlic (Allium sativum), Elephant Garlic

6

(A. ampeloprasum), Leek (A. porrum) and Chinese Chive (A. tuberosum). J.

7

Agric. Food Chem. 2010, 58, 4617-4625.

8

2.

Block, E.; Dane, A.J.; Cody, R.B. Crushing garlic and slicing onions:

9

detection of sulfenic acids and other reactive organosulfur intermediates

10

from garlic and other Alliums using direct analysis in real time-mass

11

spectrometry (DART-MS). Phosphorus Letters. 2011, 186, 1085-1093.

12

3.

Block, E. Fifty years of smelling sulfur. J. Sulfur Chem. 2013, 34, 158-207.

13

4.

Block, E.; Cody, R.B.; Dane, A.J.; Sheridan, R.; Vattekkatte, A.; Wank, K.

14

Allium chemistry: use of new instrumental techniques to “see” reactive

15

organosulfur species formed upon crushing garlic and onion. Pure Appl.

16

Chem. 2010, 82, 535-539.

17

5.

garlic. J. Agric. Food Chem. 1989, 37, 725-730.

18 19

Tsung-Hsi, Y.; Chung-May, W.; Yoh-Cherng, L. Volatile compounds from

6.

Tsung-Hsi, Y.; Chung-May, W.; Chi-Tang, Ho. Volatile compounds of deep-

20

oil fried, microwave-heated and oven-baked garlic slices. J. Agric. Food

21

Chem. 1993, 41, 800-805.

22 23

7.

Block, E. Garlic and other alliums The lore and the science. The Royal Society of Chemistry, UK, 2010. pp169.



1

8.

Lee, S.; Yoo, M.; Kim, S.; Shin, D. Identification and quantification of S-

2

allyl-L-cysteine in heated garlic juice by HPLC with ultraviolet and mass

3

spectrometry detection. LWT. 2014, 57, 516-521.

4 5

9.

Thomson, M.; Ali, Muslim. A review of its potential use as an anti-cancer agent. Curr. Cancer Drug Targets. 2003, 3(1), 67-81.

6

10. Saravanan, G.; Ponmurugan, P.; Begum, M, S. Effect of S-allylcysteine, a

7

sulphur containin amino acid on iron metabolism in streptozotocin induced

8

diabetic rats. J Trace Elem Med Biol. 2013, 27(2), 143-147.

9

11. Malekpour-Dehkordi, Z.; Javadi, E.; Doosti, M.; Paknejad, M.; Nourbakhsh,

10

M.; Yassa, N.; Gerayesh-Nejad, S.; Heshmat, R. S-Allylcycteine, a garlic

11

compound, increases ABCA1 expression in human THP-1 macrophages.

12

Phytother. Res. 2013, 27(3), 357-361.

13

12. Shouk, R.; Abdou, A.; Shetty, K.; Sarkar, D.; Eid, A, H. Mechanisms

14

underlying the antihypertensive effects of garlic bioactives. Nutr. Res. 2014,

15

34(2), 106-115.

16

13. Yu, T, H.; Wu, C, M.; Ho, C, T. Meat-like flavor generated from thermal

17

interactions of glucose and alliin or deoxyalliin. J. Agric. Food Chem. 1994,

18

42, 1005-1009.

19

14. Kimura, K.; Iwata, I.; Nishimura, H.; Kasuya, M.; Yamane, A.; Mizutani, J.

20

Thermal degradation products of S-alkyl-L-cysteine occurring in the Allium

21

species with D-glucose. Agric Biol. Chem. 1990, 54, 1893-1990.

22

15. Toelstede, S; Hofmann, T. Quantitative studies and taste re-engeneering

23

experiments toward the decoding of the nonvolatile sensometabolome of

24

Gouda cheese. J. Agric. Food Chem. 2008, 56, 5299-5307.


Page 26 of 42

Page 27 of 42


1

16. Dunkel, A; Hofmann, T. Sensory-directed identification of b-alanyl

2

dipeptides as contributors to the thick-sour and white-meaty orosensation

3

induced by chicken broth. J. Agric. Food Chem. 2009, 57, 9867-9877.

4

17. Toelstede, S.; Hofmann, T. Sensomics mapping and identification of the

5

key bitter metabolites in Gouda cheese. J. Agric. Food Chem. 2008, 56,

6

2795-2804.

7

18. Ottinger, H.; Hofmann, T. Identification of the taste enhancer alapyridaine

8

in beef broth and evaluation of its sensory impact by taste reconstitution

9

experiments. J. Agric. Food Chem. 2003, 51, 6791-6796.

10

19. Dunkel, A; Koester, J.; Hofmann, T. Molecular and sensory characterization

11

of γ-glutamyl peptides as key contributors to the kokumi taste of Edible

12

beans. J. Agric. Food Chem. 2007, 55, 6712-6719.

13

20. Sonntag, T.; Kunert, C.; Dunkel, A.; Hofmann, T. Sensory-guided

14

identification of N-(1-methyl-4-oxoimidazolidin-2-ylidene)-α-amino acids as

15

contributors to the thick-sour and mouth-drying orosensation of stewed

16

beef juice. J. Agric. Food Chem. 2010, 58, 6341-6350.

17

21. Beksan, E.; Schieberle, P.; Robert, F.; Blank, I.; Fay, L.B.; Schlichtherle-

18

Cerny, H.; Hofmann, T. Synthesis and sensory characterization of novel

19

umami-tasting glutamate glycoconjugates. J. Agric. Food Chem. 2003, 51,

20

5428-5436.

21

22. Rotzoll, N.; Dunkel, A; Hofmann, T. Activity-guided identification of (S)-

22

malic acid 1-O-D-glucopyranoside (morelid) and g-aminobutyric acid as

23

contributors to umami taste and mouth-drying oral sensation of morel

24

mushrooms (Morchella deliciosa Fr.). J. Agric. Food Chem. 2005, 53, 4149-

25

4156.



Page 28 of 42

1

23. Ottinger, H.; Soldo, T.; Hofmann, T. Discovery and structure determination

2

of a novel Maillard-derived sweetness enhancer by application of the

3

comparative taste dilution analysis (cTDA). J. Agric. Food Chem. 2003, 51,

4

1035-1041.

5 6

24. Soldo, T; Blank, I; Hofmann, T. (+)-(S)-Alapyridaine – a general taste enhancer ?. Chem. Senses. 2003, 28, 371-379.

7

25. Soldo, T.; Frank, O.; Ottinger, H.; Hofmann, T. Systematic studies of

8

structure and physiological activity of alapyridaine. A novel food-born taste

9

enhancer. Mol. Nutr. Food Res. 2004, 48, 270-281.

10

26. Soldo, T.; Hofmann, T. Application of HILIC-HPLC/cTDA for the

11

identification of a bitter inhibitor using a combinatorial approach based on

12

Maillard reaction chemistry. J. Agric. Food Chem. 2005, 63, 9165-9171.

13

27. Festring, D.; Hofmann, T. Discovery of N2-(1-carboxyethyl)guanosine 5’-

14

monophosphate as an umami-enhancing Maillard-modified nucleotide in

15

yeast extracts. J. Agric. Food Chem. 2010, 58, 10614-10622.

16

28. Festring, D.; Hofmann, T. Systematic studies on the chemical structure and

17

umami

enhancing

activity

of

Maillard

modified

18

monophosphates. J. Agric. Food Chem. 2011, 59, 665-676.

guanosine

5’-

19

29. Frank, O.; Ottinger, H.; Hofmann, T. Characterization of an intense bitter-

20

tasting 1H,4H-quinolizinium-7-olate by application of the taste dilution

21

analysis, a novel bioassay for the screening and identification of taste-

22

active compounds in foods. J. Agric. Food Chem. 2001, 49, 231-238.

23

30. Tong, X. G.; Zhou, L. L.; Wang, Y. H.; Xia, C.; Wang, Y.; Liang, M.; Hou, F.

24

F.; Cheng, Y. X. Acortatarin A and B, two novel antioxidative


Page 29 of 42


1

sporoalkaroids with a naturally unusual morphoine morif from Acorus

2

tatarinowii. Org. Lett. 2010, 12, 1844-1847.

3

31. Li, M.; Xiong, J.; Huang, Y.; Wang, L. J.; Tang, Y.; Yang, G. X.; Liu, X. H.;

4

Wei, B. G.; Fan, H.; Zhao, Y.; Zhai, W. Z.; Hu, J. F. Xylapyrrosides A and B,

5

two rare sugar-morpholine spiroketal pyrrole-derived alkaloids from Xylaria

6

nigripes: isolation, complete structure elucidation, and total syntheses.

7

Tetrahedron. 2015, 71, 5285-5295.

8

32. Borrero, N. V.; Aponick, A. Total synthesis of acortatarin A using a Pd(II)catalyzed spiroketalization strategy. J. Org. Chem. 2012, 77, 8410-8416.

9 10

33. Guo, J. L.; Feng, Z, M.; Yang, Y. J.; Zhang, Z. W.; Zhang, P. C.

11

Pollenopyrroside A and B, novel pyrrle ketohexoside derivatives from bee-

12

collected Brassica campestris pollen. Chem. Pharm. Bull. 2010, 58, 983-

13

985.

14

34. Murai,

N.;

Yonaga, of

M.; aryl

Tanaka, halides

K. and

Palladium-catalyzed

15

hydroxymethylation

triflates

with

16

acetoxymethyltrifluoroborate. Org. Lett. 2012, 14, 1278-1281.

direct

potassium

17

35. Sudhakar, G.; Kadam, V. D.; Bayya, S.; Pranitha, G.; Jagadeesh, B. Total

18

synthesis and stereochemical revision of acortatarins A and B. Org. Lett.

19

2011, 13, 5452-5455.

20

36. Olsson, K.; Pernemalm, P.; Theander, O. Formation of aromatic

21

compounds from carbohydrates. VII. Reaction of D-glucose and glycine in

22

slightly acidic, aqueous solution. Acta Chem. Scand. B32, 1978, 249-256.

23

37.

Kim, S.B.; Chang, B.Y.; Hwang, B.Y.; Kim, S.Y.; Lee, M.K. Pyrrole

24

alkaloids from the fruits of Morus alba. Bioorg. Med. Chem. Lett. 2014,

25

24(24), 5656-5659.



1

38. Wang, P.; Kong, F.; Wei, J.; Wang, Y.; Wang, W.; Hong, K.; Zhu, W.

2

Alkaloids from the mangrove-derived actinomycete Jishengella endophytica

3

161111. Mar. Drugs. 2014, 12, 477-490.

4

39. Ueda, Y.; Sakaguchi, M.; Hirayama, K.; Miyajima, R.; Kimizuka, A.

5

Characteristic flavor constituents in water extract of garlic. Agric Biol. Chem.

6

1990, 54, 163-169.

7


Page 30 of 42

Page 31 of 42


Table 1. NMR Data of Compounds 1 and 3 Compound 1

No. 2 3 4a 4b 5 7a 7b 8 10a 10b 11 12 13 14 15a 15b

a

δH (J values in Hz) 3.88 (td, 5.0, 5.0, 3.1) 4.12 (m) 1.97 (dd, 13.9, 3.4) 2.30 (dd, 13.9, 8.5) 4.85 (s, 2H)

4.17 (d, 13.9) 4.44 (d, 13.9) 6.08 (d, 4.0) 7.03 (d, 4.0) 9.42 (s) 3.44 (m) 3.54 (m)

δC 87.5 69.9 44.5

a

102.4 57.2 135.0 50.8 104.8 123.8 130.6 179.0 61.1

a

In DMSO-d6 (1H: 400 MHz, 13C: 100 MHz)

b

In CD3OD (1H: 500 MHz, 13C: 125 MHz)

Compound 3 b

δH (J values in Hz) 4.07 (td, 4.7, 4.7, 3.3) 4.26 (ddd, 8.3, 4.6, 2.7) 2.15 (dd, 14.0, 2.6) 2.35 (dd, 14.0, 8.3) 4.86 (d, overlapped) 5.02 (d, 15.8)

δH (J values in Hz) 3.86 (m) 4.23 (m) 2.00 (dd, 13.4, 6.2) 2.34 (dd, 13.4, 6.7) 4.82 (d, 15.8) 4.95 (d, 15.8)

4.23 (d, 14.0) 4.59 (d, 14.0) 6.08 (d, 4.2) 7.02 (d, 4.1)

4.17 (d, 13.8) 4.54 (d, 13.8) 6.06 (d, 4.1) 7.03 (d, 4.1)

9.37 (s) 3.62 (dd, 12.1, 4.9) 3.71 (dd,12.1, 3.3)

9.42 (s) 3.40 (m) 3.49 (m)

31

ACS Paragon Plus Environment

a

δCa 88.4 70.2 44.3 102.4 57.8 135.1 51.4 104.7 123.8 130.6 178.6 62.3

δH (J values in Hz)b 4.01 (td, 6.7, 4.7, 4.7) 4.38 (td, 6.9, 6.9, 5.1) 2.09 (dd, 13.3, 6.9) 2.35 (dd, 13.3, 6.9) 4.81 (d, 15.9) 5.09 (d, 15.8) 4.22 (d, 14.0) 4.67 (d, 13.9) 6.06 (d, 4.2) 7.02 (d, 4.1) 9.37 (s) 3.61 (dd, 11.7, 6.8) 3.70 (dd, 11.7, 4.4)


Page 32 of 42

Table 2. NMR Data of Compounds 2 and 4 Compound 2

No. δH (J values in Hz) 2a 2b 3 4 5a 5b 6 8a 8b 9 11a 11b 12 13 14 15 a

3.37 (m) 3.67 (t, 10.3, 10.3) 3.56 (m) 3.88 (m) 1.88 (dd, 14.5, 3.9) 2.03 (dd, 14.5, 3.8)

60.1 65.9 65.5 37.5

134.7 51.0

3.96 (d, 13.9) 4.36 (d, 14.0) 6.06 (d, 4.1) 7.01 (d, 4.1)

104.8 123.5 130.6 178.5

9.41 (s)

In DMSO-d6 ( H: 500 MHz,

δC

a

93.4 56.9

4.69 (d, 15.9) 4.85 (d, 15.9)

1

a

13

Compound 4 δH (J values in Hz)

3.54 (m) 3.76 (m) 3.70 (m) 4.05 (m) 2.07 (m) 2.22 (dd, 14.8, 3.4) 4.84 (d, 15.8) 4.90 (d, 15.7)

b

δH (J values in Hz) 3.62 (m, 2H)

65.3

3.62 (m) 3.88 (m) 1.75 (dd, 12.8, 5.1) 1.85 (m)

66.8 64.2 34.9

4.62 (d, 15.8) 4.86 (d, 16.0)

4.03 (d, 14.1) 4.46 (d, 14.2) 6.06 (d, 3.9) 6.97 (d, 3.9)

3.95 (d, 14.0) 4.44 (d, 13.9) 6.07 (d, 4.1) 7.02 (d, 4.1)

9.46 (s)

9.42 (s)

C: 125 MHz)

b

In CD3COCD3 ( H: 500 MHz, 13C: 125 MHz)

c

In CDCl3 (1H: 500 MHz, 13C: 125 MHz)

δC

1

32


94.9 57.1 134.7 51.8 104.8 123.3 130.6 178.6

δH (J values in Hz)c 3.81 (dd, 12.8, 1.7) 3.88 (dd, 12.8, 2.2) 3.89 (m, overlapped) 4.17 (ddd, 11.4, 5.3, 3.2) 1.91 (dd, 13.0, 11.4) 2.04 (dd, 13.0, 5.4) 4.74 (d, 15.6) 4.82 (d, 15.4) 4.02 (d, 14.2) 4.70 (d, 14.2) 6.01 (d, 4.1) 6.91 (d, 4.1) 9.45 (s)

Page 33 of 42


Table 3. NMR Data of Compounds 5-7 No.

Compound 5 δH (J values [Hz])

2 3 4 5 6 7 9 10 11 12 13 15

6.17 (d, 3.1) 6.14 (d, 3.2) 4.30 (d, 4.9, 2H) 5.58 (s, 2H) 6.23 (d, 4.0) 6.98 (d, 4.0) 4.61 (d, 4.5, 2H) 9.50 (s)

a

a

δC 155.2 107.7 108.3 150.2 55.6 41.1 143.7 109.6 124.1 131.5 55.0 179.5

No.

Compound 6 δH (J values [Hz])

a

a

2 3 6.19 (d, 4.0) 4 6.88 (d, 4.0) 5 6 7a 2.92 (m) 3.37 (m) 8a 2.76 (dd, 13.6, 7.0) 8b 2.92 (m) 9 5.63 (ddt, 17.1, 10.3, 7.2, 7.2) 10a 4.38 (d, 13.9) 10b 4.43 (d, 13.6) 11 4.99 (m, 2H) 12 9.30 (s) 13

a


b


δC 144.5 110.2 123.4 132.8 60.8 34.1 33.4 134.6 117.0 55.7 179.2 171

33


No. 2a 2b 3 5 6 7 8 9 10a 10b 11a 11b 12 13a 13b 14

Compound 7 δH (J values [Hz])b 5.59 (d, 15.2) 5.77 (d, 14.8) 5.59 (m) 6.35 (d, 4.0) 7.23 (d, 4.0) 3.16 (dd, 14.8, 4.8) 3.22 (dd, 14.8, 4.5) 2.62 (dd, 13.6, 7.8) 2.96 (dd, 13.6, 6.6) 5.78 (m) 4.92 (dq, 16.9, 1.4, 1.4, 1.4) 5.04 (m) 9.51 (s)

δCb 64.2 132.6 57.6 166.5 106.5 125.0 130.5 33.3 34.0 133.6 118.0 179.5


Page 34 of 42

Table 4. NMR Data of Compounds 8, 9 and racemate of 10 and 11 Compound 8

No. 2a 2b 3 5 6 7 8 9 10a 10b 11a 11b 12a 12b 13 14a 14b 15a 15b 16a 16b 17 18a 18b 19 20

δH (J values [Hz]) 4.69 (d, 16.9) 4.93 (d, 16.9)

a

5.54 (m) 6.29 (d, 4.0) 7.20 (d, 4.0) 4.93 (m)

Compound 9 δC 41.8

a

134.7 58.3 166.0 106.5 125.4 129.9 56.6

δH (J values [Hz]) 4.55 (d, 16.4) 4.81 (d, 16.4)

a

Racemate of 10 and 11 δC 42.4

a

134.3 58.3 165.8 106.6 125.2 130.0 56.1

5.58 (m) 6.31 (d, 4.0) 7.21 (d, 4.0)

5.51 (m) 6.29 (d, 4.0) 7.22 (d, 4.0)

δCb 44.6 134.2 58.1 165.1 106.4 125.6 129.7 45.8

28.7

33.5

2.92 (dd, 14.2, 10.7) 3.11 (m) 3.17 (m, 2H)

3.58 (ddd, 13.8, 8.2, 6.4) 3.66 (ddd, 13.6, 8.2, 6.5) 2.68 (m, 2H)

33.5

3.21 (m, 2H)

33.4

5.71 (ddt, 17.1, 10.0, 7.2, 7.2) 5.10 (m, 2H)

134.3 117.4

5.74 (ddt, 17.0, 10.0, 7.2, 7.2) 5.11 (m, 2H)

134.1 117.6

134.3 117.4

3.08 (m, 2H)

34.2

3.09 (m, 2H)

34.2

2.59 (dd, 13.6, 7.6) 2.88 (dd, 13.6, 6.8) 5.55 (m) 4.93 (m) 5.01 (m) 9.47 (s)

34.1

2.52 (m) 2.83 (dd, 13.6, 6.6) 5.54 (m) 4.90 (m) 5.01 (m) 9.48 (s)

34.2

5.76 (ddt, 17.1, 9.9. 7.2, 7.2) 5.09 (dt, 10.0, 0.9, 0.9) 5.18 (dq, 17.0, 1.7, 1.7, 1.7) 3.05 (m) 3.18 (m) 2.41 (dd, 13.6, 7.9) 2.80 (dd, 13.5, 6.5) 5.53 (m) 4.87 (m) 5.00 (dt, 10.0, 0.9, 0.9) 9.45 (s)

2.96 (dd, 14.2, 10.5) 3.07 (m) 3.12 (m, 2H)

28.4

133.9 117.8 179.0 170.1

a


b


4.98 (m)

δH (J values [Hz])b 4.73 (d, 17.0) 4.90 (d, 17.1)

133.9 117.7 179.1 170.4

34


26.3

34.7 34.0 133.8 117.8 178.8

Page 35 of 42


Table 5. Sensory Quality, Threshold Concentrations, and Concentration of Maillard compounds 1-9 in garlic preparations concentration in garlic preparations [µmol/kg] threshold conc.

garlic powder

natural garlic

compd no.

sensory quality

[µmol/kg]

intact

roasted

intact

roasted

1

bitter

300

n.d*1

0.013

n.d

0.092

2

bitter

370

n.d

0.003

n.d

0.009

3

bitter

471

n.d

0.008

n.d

0.025

4

bitter

263

n.d

0.005

n.d

0.007

5

bitter

785

n.d

0.002

n.d

0.040

6

tasteless

> 1000

n.d

793.7

n.d

0.2

kokumi

186

7

tasteless

> 500

n.d

51.4

n.d

n.d

8

bitter

20

n.d

5.9

n.d

n.d

9

bitter

0.5

n.d

2.8

n.d

n.d

*1 n.d: not detected

35



Figure Legends

Figure 1.

Comparison of HPLC-UV (λ = 290 nm) chromatograms of (A) a mixture of glucose and S-allyl-L-cysteine, (B) glucose, (C) S-allylL-cysteine,

and (D) a mixture of glucose and S-allyl-D-cysteine,

dry-heated for 60 min at 100 °C. Figure 2.

Chemical structures of Maillard reaction products 1-9 identified in the thermally processed glucose/S-allyl-L-cysteine mixture.

Figure 3.

CD spectra, chemical structure and selected HMBC correlations of target compounds 6 (blue) and 7 (green).

Figure 4.

CD spectra of Maillard reaction products 8 (red) and 9 (blue), and synthesized references 10 (green) and 12 (purple).

Figure 5.

Reaction pathways proposed for the formation of compounds 1-9. via 3-deoxyhexosone (1) as the joint intermediate.


Page 36 of 42

Page 37 of 42


Figure 1 (Wakamatsu et al.)

6

A 1

absorption at 290 nm

HMF

2-5

7

8 9

B

C

D

5

10

15

20

25

30

retention time (min)


35

40

45




Page 38 of 42

Page 39 of 42



O

90

OH

OH

S N

O O S

6

⊿ε

60

O

N

30

O

7

0 200

225

250

275

300

325

350

-30 wavelength [nm]


375

400


Page 40 of 42


18

250

17

16

10

O S

5

15

N

O

O

12

S 2

4

200 150

N1

6

(S)

11

13

HO O

14

3

9

S

7

19

8

10

100

O

⊿ε

50

(R)

N

(S)

S

N 9

0 200

225

250

275

300

325

350

375

400

-50

O

-100

HO O

-150 -200

O S

S

N

S

(R)

-250

O

O

N 11

wavelength [nm]


(R)

N

(R)

N 8

S

Page 41 of 42

1 2



3 4



TOC graphic 254x190mm (96 x 96 DPI)


Page 42 of 42

Taste-Active Maillard Reaction Products in Roasted Garlic (Allium

Recommend Documents