Quantitation and sensory properties of three newly identified pyroglutamyl oligopeptides in sake

ABSTRACT

Three new peptides: (pGlu)L-ethyl, (pGlu)LFGP-ethyl and (pGlu)LFNP-ethyl, were identified in the search for pyroglutamyl oligopeptide ethyl esters in sake. The ethyl esterified peptides in sake were quantitated using stable isotope dilution analysis and additional quantitation of (pGlu)L was performed using an external standard method. The concentrations of (pGlu)L-ethyl and (pGlu)L in 33 commercial sake samples ranged from 0.16 to 1.57 mg/L and 1.49 to 7.55 mg/L, respectively. The sensory properties of the pyroglutamyl oligopeptide ethyl esters and corresponding non-esterified peptides were examined: the estimated difference threshold of (pGlu)L (2.0 mg/L) and (pGlu)L-ethyl (0.267 mg/L) was exceeded in 32 and 26 samples, respectively. Estimated thresholds of (pGlu)LFGP-ethyl and (pGlu)LFNP-ethyl were often lower than the levels in quantitated sake samples. The sensory effects of these pyroglutamyl dipeptides on a model sake quality may be negative because of their unpleasant taste, however, (pGlu)LFNP-ethyl may be positive because of its mild taste.

Graphical Abstract

Three new pyroglutamyl oligopeptides ethyl esters were identified in sake. Their quantitated levels often exceeded the estimated threshold values.

KEYWORDS:

• Pyroglutamyl oligopeptide ethyl ester
• sake
• taste
• threshold value

During the food manufacturing process, proteins are sometimes hydrolyzed to small peptides which can become taste-active, for example bitter, umami, sweet and/or astringent tasting; the taste-active peptides affect the sensory quality of food. [1–4] The taste and functional properties [5] of food peptides have been vigorously investigated and many taste-active [6–10] and functional [11–15] peptides have been found in Japanese fermented foodstuffs. Peptides found in sake and sake lee have pyroglutamyl N-termini which may form from heating glutamine residues in an aqueous solution. [13] Bitter-tasting pyroglutamyl peptides, consisting of 6–13 amino acids, in sake may be directly derived from the N-termini of rice glutelin acidic subunits, an endosperm storage protein, [6] because they have exactly the same amino acid sequences and glutelin has a pyroglutamyl N-terminal residue. [16,17] The bitter-tasting peptides are described as strong bitter or astringent tasting, producing a negative impact on the sensory quality of sake. They are formed in sake mash but their levels weakly correlate with the glutelin content of rice grains. [18–20] The taste-active ethyl esterified pyroglutamyl decapeptides (PGDPE) have amino acid sequences exactly the same as bitter-tasting peptides and are also found in sake. [8] The threshold levels of PGDPE in sake are lower than their corresponding non-esterified bitter-tasting peptides and their levels in sake often exceed the threshold values in charcoal untreated commercial sake, affecting the sensory quality. [10] Bitter-tasting peptides and PGDPE have some hydrophobic properties and their molecular masses are between 680–1510 Da, which makes them easily removed by the charcoal treatment that is commonly applied in the sake refining process. [6,8] In recent years, the number of charcoal untreated sake products has increased because the treatment damages the harmony of the sake flavor.

Kiyono et al (2003) reported that many di- and tri- peptides were present in sake, and (pGlu)L and (pGlu)Q were the major constituents. It was reported that (pGlu)L has hepatoprotective and anti-colitic activities [13,14], and (pGlu)Y and (pGlu)NI have anti-colitic activity; [15] the sensory properties of these peptides were not reported. It is suggested that (pGlu)L may have a bitter taste because of its similarity to the amino acid sequences of the bitter-tasting peptides in sake. The bitter-tasting peptides of sake are further hydrolyzed to smaller peptides, [19] and the formation of ethyl esterified oligopeptides from these bitter-tasting sake peptides may be similar to PGDPE.

We investigated the pyroglutamyl ethyl esterified oligopeptides in sake and examined their sensory properties using a model sake solution; we then determined the level of these oligopeptides in commercial sake samples in order to evaluate their sensory significance in sake. The results of this study will increase the fundamental knowledge of taste-active peptides in sake.

Materials and methods

Chemicals and materials

Pyroglutamyl oligopeptides, (pGlu)L, (pGlu)LF, (pGlu)LFGP and (pGlu)LFNP, were obtained from BEX Co., Ltd. (Tokyo, Japan). (pGlu)LFNPSTNPW was obtained from Medical & Biological Laboratories Co., Ltd. (Nagoya, Japan). Super dehydrated ethanol (99.5 %), anhydrous ethanol (D, 99 %), acetyl chloride (98 %), ethanol (95 %), and charcoal were purchased from Wako Pure Chemical Industries Co., Ltd. (Osaka, Japan). Acetonitrile (HPLC grade) was obtained from Kanto Chemical Co., Inc. (Tokyo, Japan). Sake samples were purchased from local markets in Tokyo or Akita city.

Preparation of pyroglutamyl oligopeptide ethyl esters

Ethyl esterification of pyroglutamyl oligoptides was performed as previously described [21]. Acetyl chloride (75 µL) was dropped into dehydrated ethanol (425 µL) to produce alcoholic HCl before the solution was mixed with 1 mg of pyroglutamyl oligopeptides in a glass vessel. The reaction proceeded at 37 °C for 3 h before the volatiles were removed by rotary evaporation. The residue was dissolved in 0.5 mL of 47.5 % (v/v) aqueous ethanol solution before it was purified by HPLC. Preparation was scaled up if required. Target fractions were collected and analyzed by mass spectrometry (MS) and the solvent was removed under vacuum. MS analyzed data of synthesized peptides are as follows: (pGlu)L-ethyl, MS (ESI⁺) m/z: 271.17 [M + H]⁺, 293.15 [M+ Na]⁺. (pGlu)LF-ethyl, MS (ESI⁺) m/z: 418.23 [M + H]⁺, 440.22 [M+ Na]⁺. (pGlu)LFGP-ethyl, MS (ESI⁺) m/z: 572.31 [M + H]⁺, 594.29 [M+ Na]⁺. (pGlu)LFNP-ethyl, MS (ESI⁺) m/z: 629.33 [M + H]⁺, 651.31 [M+ Na]⁺. Synthesized compounds were stored at −20 °C until use and were dissolved in 47.5 % (v/v) aqueous ethanol solution when used. Deuterated ethyl esterified peptides were synthesized using anhydrous ethanol-D₆ (D, 99 %). Obtained residues were dissolved in 100 µL of 50 % (v/v) acetonitrile solution and purified by HPLC. MS analyzed data of synthesized peptides are as follows: (pGlu)LCOOC₂D₅, MS (ESI⁺) m/z: 276.20 [M + H]⁺, 298.18 [M+ Na]⁺. (pGlu)LFCOOC₂D₅, MS (ESI⁺) m/z: 443.27 [M + H]⁺, 445.25 [M+ Na]⁺. (pGlu)LFGPCOOC₂D₅, MS (ESI⁺) m/z: 577.341 [M + H]⁺, 599.32 [M+ Na]⁺. (pGlu)LFNPCOOC₂D₅, MS (ESI⁺) m/z: 634.36 [M + H]⁺, 656.34 [M+ Na]⁺.

Rice koji enzymatic digestion of pyroglutamyl decapeptide

Enzyme solutions were prepared from rice koji according to a previous report [22] using 40 % polished Akitasakekomachi rice and Aspergillus oryzae RIB128. The reaction mixture (80 µL) consisted of enzyme solution (10 µL), 26.0 % (v/v) ethanol, 12.5 mM sodium lactate buffer (pH 3.0), and 12 µg (pGlu)LFNPSTNPW. After incubation at 30 °C for 2 d, the reaction mixture was analyzed by MS. MS analysis was performed as described above using heated enzyme solution (100 °C, 5 min) and non-ethanol solutions as controls.

Detection and identification of pyroglutamy oligopeptide ethyl esters in sake. Sake samples of 25 mL were applied to pre-conditioned (7 mL of methanol and then 7 mL of water) Bond Elute C18 LRC 500MG columns (Agilent Technologies, Santa Clara, CA, USA). After the column was washed with 7 mL of water, trapped constituents were eluted with 1 mL of 95 % (v/v) aqueous ethanol solution. The eluents of four batches were combined and the solvent was removed under vacuum. The residue was dissolved in 200 µL of 47.5 % (v/v) aqueous ethanol solution and 100 µL was applied to HPLC fractionation. The HPLC fraction was analyzed using tandem mass spectrometry (MS/MS) and compared with synthesized compounds, using a TSQ Quantum Ultra instrument (Thermo Fisher Scientific, Bremen, Germany) under the following conditions: 15–30 eV collision energy and 1.5 mTorr CID gas of Ar. The spray voltage of electrospray ionization (ESI) was 2.0 kV, and the capillary temperature was maintained at 250 °C.

Quantitation of pyroglutamyl oligopeptides in sake

Ethyl esterified peptides were quantitated using stable isotope dilution analysis (SIDA) and non-esterified peptides were quantitated using an external standard method. Deuterated ethyl esterified peptides of 0.32 – 0.98 µg in 40 µL of 50 % (v/v) aqueous acetonitrile solution were added to 2.0 mL of sake sample. The samples were diluted in 4 mL of 50 mM sodium succinate buffer (pH 4.3) before they were applied to a pre-conditioned (7 mL of methanol and then 7 mL of water) Bond Elute C18 LRC 500MG column. The column was successively washed with 1 mL of water and then 1 mL of 10 % (v/v) acetonitrile solution. Trapped constituents were eluted into a glass vial using 1.0 mL of 30 % acetonitrile solution and the non-esterified peptides were quantitated using MS. The resulting residue was then eluted into a glass vial using 0.5 mL of 100 % acetonitrile which was then removed under vacuum. The obtained residue was dissolved in 40 µL of 50 % (v/v) acetonitrile solution and 5 µL was analyzed using MS for quantitation of ethyl esterified peptides. The ESI-MS analysis was conducted using an Exactive mass spectrometer (Thermo Fisher Scientific, Bremen, Germany). The voltage for ESI was +4.0 kV. Non-esterified peptides were analyzed in negative mode and the following MS signal intensities: (pGlu)L MS (ESI⁻) m/z: 241.12 [M-H]⁻, (pGlu)LF MS (ESI⁻) m/z: 388.19 [M-H]⁻, (pGlu)LFGP MS (ESI⁻) m/z: 542.26 [M-H]⁻, (pGlu)LFNP MS (ESI⁻) m/z: 599.28 [M-H]⁻. Ethyl esterified peptides were analyzed in positive mode and the following MS signal intensities from + Na additives of natural and deuterated peptides: (pGlu)L-ethyl m/z: 293.15, (pGlu)LF-ethyl m/z:440.22, (pGlu)LFGP-ethyl m/z:594.29, (pGlu)LFNP-ethyl m/z:651.31, (pGlu)L-ethyl-d₅ m/z: 298.18, (pGlu)LF-ethyl-d₅ m/z: 445.25, (pGlu)LFGP-ethyl-d₅ m/z: 599.32, (pGlu)LFNP-ethyl-d₅ m/z: 656.34. Calibration curves were constructed using a model sake solution consisting of 17 % (v/v) ethanol, 2 % (w/v) glucose, 300 mg/L lactic acid, 300 mg/L succinic acid, and 300 mg/L malic acid; the pH was adjusted to 4.3 using NaOH. Natural and deuterated peptides were added to the calibration solution; the natural/deuterated ratio of ethyl esters ranged from 0.25–2.1. The concentration of peptides used in this experiment was adjusted according to the UV (200 nm) -monitored HPLC data. Accuracy and precision assessments were conducted using a commercial sake. Natural peptides were added to the sake at 1 mg/L for (pGlu)L and at 200 µg/L for the other peptides, and 6 sample preparations and 3 MS analyses were conducted for each sample to calculate the relative standard deviation (RSD) and recovery ratio. The detection limit was calculated from standard deviation values of quantitated data in which peptides were added at 1 mg/L for (pGlu)L or at 100 µg/L for the other peptides. Data were obtained from 6 sample preparations and 3 MS analyses for each sample.

HPLC

HPLC for purification and analysis of synthesized peptides was conducted using a Capcell Pak C18 Type MG column (4.5 mm x 150 mm) (Shiseido, Tokyo, Japan). Solvent A was acetonitrile, and solvent B was 0.1 % phosphoric acid/water. A linear gradient was used from A: B = 30: 70 to A: B = 45: 55 over 30 min at a flow rate of 1.0 mL/min. The absorbance was monitored at 200 nm. Injection volume was 20–200 µL for purification and 5–20 µL for analysis. Analysis of sake peptides and purification for the sensory test were conducted using the columns described and the following conditions: solvent A was 90 % (v/v) ethanol, and solvent B was 0.01 % aqueous HCl solution. A linear gradient was used from A: B = 30: 70 to A: B = 50: 50 over 30 min at a flow rate of 1.0 mL/min. The absorbance was monitored at 200 nm. Injection volume was 20–200 µL. The volume of one fraction was 1 mL. The solvent ratio was arranged according to the hydrophobicity of the target peptide.

Charcoal treatment of sake

Sake samples were spiked with synthesized peptides at 0.1–1.0 mg/L and treated with 100, 200, 400, 800, 1600, and 3200 mg/L of charcoal for 30 min. Treated samples were filtered with a mixed cellulose ester membrane of 0.45 μm before peptide analysis.

Threshold assessment

The threshold value of peptides in a model sake solution was examined according to the BCOJ sensory analysis method (Brewery Convention of Japan, 2002) [23] which is founded on the forced-choice ascending concentration series method of limits of ASTM E-679. The model sake solution was the same as that used for the calibration curves. Assessors were made up of 12 undergraduate students, of 20–22 years old, from Akita prefectural university who were trained using synthesized peptides. Assessors dropped 0.2 mL of sample on their tongues using a pipette and assessed the taste. [10] We presented 4 sets of the triangular test, although the official method demanded 6 sets of the triangular test: this was to prevent sensory fatigue of the assessor. If the threshold was not found in the presented test set, another set of triangular tests which included the lower or the higher level was presented to the assessor. We obtained threshold values using this method, in addition, we also estimated recognition thresholds using the concentration at which the assessor recognized the taste.

Results and discussion

Rice koji enzymatic digestion of pyroglutamyl decapeptide

Before detection of ethyl esterified pyroglutamyl oligopeptides, enzymatic digestion was conducted under sake mash simulated conditions using (pGlu)LFNPSTNPW as a substrate. After 2 days, almost all of the substrate had disappeared and the following m/z signals were clearly detected, in the positive mode, using MS analysis: m/z 293.15, m/z 651.31 and m/z 1236.57 which may correspond to the signals for (pGlu)L-ethyl [M+Na]⁺, (pGlu)LFNP-ethyl [M+Na]⁺ and (pGlu)LFNPSTNPW-ethyl [M+Na]⁺, respectively. Other signals of expected ethyl esterified candidates were not detected. MS (ESI⁻) m/z 241.12 was detected in negative mode analysis and may correspond to (pGlu)L [M-H]⁻. The substrate was completely digested under non-ethanol conditions, and signals for other ethyl esterified peptide candidates was not detected. In the non-ethanol mixture, MS (ESI⁻) m/z 241.12 was detected, corresponding to (pGlu)L [M-H]⁻. There was no digestion of the substrate by the heat-treated enzyme. These results suggest enzymatic formation of the target peptides occurred under sake mash conditions.

Preparation of pyroglutamyl oligopeptide ethyl esters

The expected ethyl esterified pyroglutamyl peptides, consisting of 2–5 amino acids, were synthesized. Both synthesized natural and deuterated peptides showed the same retention times during HPLC analysis. In the deuterated peptide samples, the d₅ isotope was exclusively formed and H-type peptide was not detected. Synthesized d₅ isotopes were stable in the 50 % (v/v) aqueous acetonitrile solution.

Detection and identification of pyroglutamyl oligopeptide ethyl esters in sake

Hydrophobic condensate prepared from two sake samples, daiginjo-shu and ordinary-type sake made from rice and neutral spirits, was fractionated using HPLC. All fractions were analyzed using MS in positive mode and expected m/z signals were searched. Based on HPLC retention times, the analysis suggested the presence of (pGlu)L-ethyl, (pGlu)LFNP-ethyl, (pGlu)LFGP-ethyl, and (pGlu)LF-ethyl in the daiginjo-shu sample (Figure 1). The characteristic m/z signals were analyzed using tandem mass spectro-metry (MS/MS) (Table 1): (pGlu)L-ethyl was clearly identified and the major signal peaks of (pGlu)LFNP-ethyl and (pGlu)LFGP-ethyl in the HPLC fractions were well matched with the synthesized samples. (pGlu)LF-ethyl was not clearly identified because of its low signal level and impure noise signals, however some common major signals were observed.

Table 1. MS/MS analysis of precursor ion signals in HPLC fraction and synthesized pyroglutamyl oligopeptide ethyl esters.

Peptide	Precursor ions m/z	Sample	Observed ions m/z (%)
(pGlu)L-ethyl	271	Fraction No.10	197 (100), 86 (24), 160 (22), 225 (14)
		Synthesized	197 (100), 86 (26), 160 (24), 225 (22)
(pGlu)LFNP-ethyl	651	Fraction No.16	651 (100), 480 (26), 463 (27), 508 (17), 263 (14)
		Synthesized	651 (100), 480 (28), 463 (26), 508 (18), 263 (25)
(pGlu)LFGP-ethyl	572	Fraction No.18	144 (100), 348 (58), 233 (58), 201 (56), 397 (46), 572 (38), 461 (36), 120(18)
		Synthesized	144 (100), 348 (55), 201 (65), 572 (42), 461 (27), 120 (17)
(pGlu)LF-ethyl	418	Fraction No.23	418 (100), 194 (88), 290 (82), 245 (65), 322 (60), 120 (26)
		Synthesized	194 (100), 120 (40), 354 (24), 400 (16), 418 (15)

Figure 1. HPLC fractionation and MS analysis of pyroglutamyl oligopeptide ethyl esters. (A): daiginjo-shu. (B): ordinary-type sake. Analyzed m/z signals were those observed in the preliminary analysis of the synthesized peptides.

Glutelin is one of the most abundant protein in the highly polished rice grains for sake brewing and it is easily digested by rice koji enzymes, [20] therefore identified peptides may be derived from the glutelin that has the exact common amino acid sequences. However, rice glutelin might not be a sole source protein because the amino acid sequences of QLFNP and QLFGP are also contained in other proteins according to the results of BlastP search which was conducted on the nonredundant protein database of Oryza sativa (japonica cultivar-group) (taxid:39,947). These hit proteins are predicted proteins from the genome data.

Quantitation of pyroglutamyl oligopeptides in sake

Due to the high selectivity and intensity of the m/z signals, the mode of MS analysis was positive for ethyl esterified peptides and negative for non-esterified peptides. Representative MS analysis charts are shown in Figure 2. Verification data of developed quantitation are shown in Table 2. Calibration curves demonstrated good linearity. The RSDs were under 7.5 % and the recovery ratios of spiked peptides ranged from 93.5 to 101.1 %. The estimated detection limits were lower than the estimated thresholds (Table 4). These results indicated that the developed methods were accurate and reliable in the quantitation of pyroglutamyl oligopeptides in sake. During the quantitation procedure, the column washing step is crucial as this affects the amount of residual coexistent constituents, especially during quantitation of the external standard: coexistent constituents significantly affect ionization of the target compound in the ESI-MS analysis. [24] (pGlu)LF, (pGlu)LFGP and (pGlu)LFNP were not detected in the sake samples tested in the preliminary examination; therefore, these peptides were not quantitated in this study.

Table 2. Verification data of quantitation.

	External standard	Stable isotope dilution analysis (SIDA)
Method	(pGlu)LCOOH	(pGlu)L-ethyl	(pGlu)LF-ethyl	(pGlu)LFGP-ethyl	(pGlu)LFNP-ethyl
Calibration curve
R ^2 of calibration curve	0.997	0.999	0.999	0.999	0.997
Response factor ^a		1.09	0.948	1.01	0.958
RSD ^b for 6 analyses (%)	3.2	3.3	4.4	5.0	7.2
Recovery ratio (%)	97.5±8.1	98.2±10.5	100.0±10.0	100.1±4.3	93.5±9.5
Detection limit (mg/L)	0.33	0.038	0.012	0.017	0.019

^a Response factor means (ratio of concentration)/(ratio of peak signals)

^b Relative standard deviation.

Table 3. Groups of analyzed sake samples.

Group	Number	Class or type	Polishing rate of rice (%)^a
a)	7	Daiginjo-shu+Junmai-daiginjo-shu	46.4 ± 3.8^b
b)	10	Junmai-ginjo-shu	54.4 ± 3.0
c)	8	Tokubetsu-junmai-shu+Junmai-shu	62.8 ± 8.0
d)	7	Ordinary type	-

^a The polishing rate of rice in the sake samples was taken from the label on the bottle. The polishing rate of sake samples in group d) are unknown.

^b means ±SD

Table 4. Best estimated difference threshold (BET) values and their range values in the model sake solution.

	BET values
Peptides	mg/L	µM	Panelist BET range (mg/L)
(pGlu)L	2.000	8.2	0.2 ～11.3
(pGlu)LF	0.101	0.26	0.01 ～ 0.91
(pGlu)LFGP	0.604	1.1	0.11 ～ 7.24
(pGlu)LFNP	0.456	0.76	0.11 ～ 3.62
(pGlu)L-ethyl	0.267	0.99	0.04 ～ 1.13
(pGlu)LF-ethyl	0.014	0.035	0.0007～ 0.091
(pGlu)LFGP-ethyl	0.027	0.047	0.003 ～ 0.181
(pGlu)LFNP-ethyl	0.068	0.11	0.01 ～ 0.36

Figure 2. MS charts in the quantitation of sake samples. (A): negative mode analysis for non-esterified pyroglutamyl peptides. (B): positive mode analysis for pyroglutamyl oligopeptide ethyl esters.

The concentrations of pyroglutamyl oligopeptides in 33 commercial sake samples, summarized in Table 3, were quantitated. Sake samples were divided into groups based on the manufacturing classes or types, as described on the label. Group a) contained 7 samples and the average polishing rate was 46%; group b) contained 10 samples and the average polishing rate was 54%; group c) contained 8 samples and the average polishing rate was 63% and group d) contained 7 samples but the average polishing rate was unknown. The concentration of (pGlu)L-ethyl ranged from 0.16 to 1.57 mg/L. The average concentration of (pGlu)L-ethyl in the groups was a) 0.39, b) 0.63, c) 0.84, and d) 0.31 mg/L (Figure 3(a)). The concentration of (pGlu)L ranged from 1.49 to 7.55 mg/L which was lower than the concentration of 10–15 mg/L previously reported. [13] This might be due to the use of different sake samples; the former report’s samples are made from 60% polished rice, [13] while a half of samples in this study are made from more highly polished rice (Table 3). It is also possible that the lower quantitated data in this study was due to ion suppression or matrix effects of coexistent constituents in the ESI-MS analysis using an external standard method. The average level for each sample group was a) 2.92, b) 3.61, c) 4.28, and d) 3.54 mg/L. The average ratio of (pGlu)L-ethy to (pGlu)L was, group a) 13, b) 17, c) 20, and d) 9 %. (pGlu)LFGP-ethyl was detected in 12 samples, and the concentration ranged from 0.0 to 0.14 mg/L. (pGlu)LFNP-ethyl was detected in 12 samples, and the concentration ranged from 0.0 to 0.30 mg/L. Neither pentapeptide was detected in group d) samples (Figure 3(b)). The quantitation data suggested that the peptide levels are affected by charcoal treatment; group d) samples are commonly refined by charcoal treatment, therefore, the effects of charcoal treatment were analyzed (Figure 4). Pentapeptide ethyl esters were removed by small amounts of charcoal, whereas dipeptides were hardly removed, even using charcoal at 3200 mg/L. Removal of the dipeptides during charcoal treatment correlates with the abundance of dipeptides present in commercial sake: [25] ethyl esterified dipeptides, present at low concentrations in sake, were removed more efficiently than non-esterified dipeptides which are present at higher concentrations. The low ratio of (pGlu)L-ethyl to (pGlu)L in group d) may be due to the greater removal (pGlu)L-ethyl during charcoal treatment. Levels of both (pGlu)L and (pGlu)L-ethyl in sample groups a), b) and c) showed weak positive relationships between the polishing ratio of rice (R [2]= 0.178 and 0.169, respectively). The level of pentapeptides in groups a), b) and c) increased according to the average polishing rate of the rice. These results suggested that protein, especially glutelin, content in the polished rice grains may affect the levels of these peptides in sake because glutelin decreases during the polishing of rice grains.

Figure 3. Quantitated levels of pyroglutamyl oligopeptides in sake samples from each group.

(A): dipeptides ■; (pGlu)L, □; (pGlu)L-ethyl(B): pentapeptides ■; (pGlu)LFGP-ethyl, □; (pGlu)LFNP-ethylData are means and S.D. for each sample group. “nd”: peptide was not detected.

Figure 4. Effects of charcoal treatment on the concentrations of pyroglutamyl oligopeptides in the sake sample. The percentage value of the residual to original is shown. Data are means and S.D. of three MS determinations.

Sensory analysis of pyroglutamyl oligopeptides in sake

The estimated difference thresholds are shown in Table 4. The panelists’ best estimated thresholds were widely distributed. The panelists answered the taste character, in almost all cases, at the lowest concentration, so the recognition threshold may be close to the estimated difference threshold. Ethyl esters had lower threshold values than the non-esterified peptides. Similar findings have been reported elsewhere for other small peptides. [26,27] The threshold values of the tested peptides were higher than the threshold values of PGDPE, [10] which may be due to their short peptide chain lengths. [28,29] (pGlu)LF-ethyl had the lowest threshold which may be due to its hydrophobic side chains and ethyl ester group. A cyclized bitter dipeptide (anhydrous L-prolyl-L-leucine) has been found in sake, [30] and its threshold (259 mg/L) was far higher than that of (pGlu)L-ethyl. The difference may be due to the presence of a pyroglutamyl N-terminus and ethyl esterified C-terminus in (pGlu)L-ethyl. (pGlu)L-ethyl was described as having an unpleasant after taste, bitter and/or astringent (Table 5). (pGlu)L was described as rich or sour tasting. The response to (pGlu)LFGP-ethyl was varied, with both negative and positive taste impressions. (pGlu)LFNP-ethyl received positive taste impressions from all panelists.

Table 5. Taste characteristics of tested pyroglutamyl oligopeptides in the model sake solution.

		From 12 panelists^b
Peptides	Taste characteristics^a	Like	Dislike
(pGlu)L	Rich (4), Sour (2)	1	3
(pGlu)LF	Bitter (3), Dry (2), Mild (2)	1	2
(pGlu)LFGP	Rich (2), Mild (2)	2	2
(pGlu)LFNP	Bitter (4), Mild (2)	0	2
(pGlu)L-ethyl	Unpleasant aftertaste (6), Bitter (5), Astringent (3), Dry (2)	1	5
(pGlu)LF-ethyl	Bitter (3), Dry (2)	4	0
(pGlu)LFGP-ethyl	Pleasant aftertaste (2), Mild (2), bitter (2)	3	2
(pGlu)LFNP-ethyl	Mild (4), Rich (3), Sweet (2), Sweet aftertaste (2)	4	0

^a Characteristics obtained from more than two panelists were shown. Number in () indicates the number of panelists’ that commented.

^b Number indicates the number of panelists’ that declared likes and dislikes.

The sensory test was conducted using a model sake solution; the results may also be useful in assessing the sensory effects of real sake because the model sake solution contained the major taste-active constituents. The levels of (pGlu)L and (pGlu)L-ethyl exceeded the threshold values in almost all of the samples, 32 and 26 samples, respectively, suggesting they may be effective taste-active constituents. The highest levels of (pGlu)L and (pGlu)L-ethyl were 3.8-times and 5.9-times higher than the threshold values, respectively. As these dipeptides are hardly removed by charcoal refining, they may be effective in a wide range of sake products. The pentapeptide ethyl esters were not always detected in the sake samples, but the maximum concentration detected was 5.1 times higher than the threshold value. The pentapeptides may play a significant role in the sensory quality of charcoal untreated sake. The preferable taste characteristic of (pGlu)LFNP-ethyl is noteworthy. Further research using real sake samples is needed to elucidate the role of the pentapeptides in the quality of sake. The young student panelists in this study were trained and may have a high sense of taste but they have less experience: evaluations by more experienced panelists may elucidate more precise sensory characteristics of the peptides. Two of the taste-active pyroglutamyl pentapeptide ethyl esters have a C-terminal proline in common; it was reported that C-terminal proline is hardly hydrolyzed by the acid carboxypeptidase-O from Aspergillus oryzae. [31,32] The high ethanol conditions in sake mash may affect hydrolysis and the production of ethyl esters, such as ethyl ferulate. [22] Further research on the formation mechanism of these ethyl esterified peptides is required.

Conclusion

We found three pyroglutamyl oligopeptide ethyl esters: (pGlu)L-ethyl, (pGlu)LFGP-ethyl and (pGlu)LFNP-ethyl, in sake. SIDA, using deuterated isotopes, and high resolution MS analysis accurately quantitated these peptides in sake. Using a model sake solution, we estimated the threshold values of four pyroglutamyl oligopeptide ethyl esters and their corresponding non-esterified peptides. The threshold values were often under the quantitated levels in sake, suggesting these taste-active peptides may be involved in the sensory quality of sake.

Author Contribution

K.H. designed the study. All authors performed the experiments. K.H. wrote the manuscript.

Acknowledgments

We are grateful to the students of the Food and Brewing Group, Department of Biological Resource Sciences of Akita Prefectural University for their assistance with the sensory analysis.

Disclosure statement

No potential conflict of interest was reported by the authors.