Formation of taste-active pyroglutamyl peptide ethyl esters in sake by rice koji peptidases

ABSTRACT

Formation of taste-active pyroglutamyl (pGlu) peptide ethyl esters in sake was investigated: 2 enzymes (A and B) responsible for the esterification were purified from a rice koji extract. MADLI-TOF/TOF analysis after deglycosylation identified enzyme (A) as peptidase S28 (GenBank accession number OOO13707.1) and enzyme (B) as serine-type carboxypeptidase (accession number AO090010000534). Both enzymes hydrolyzed pGlu peptides and formed ethyl esters under sake mash conditions: acidic pH (3-4) and in ethanol (5%-20% v/v) aqueous solutions. Enzyme (A) formed pGlu penta-peptide ethyl esters from pGlu undeca-peptides by a prolyl endo-type reaction. Enzyme (B) formed (pGlu) deca-peptide and its ethyl esters from pGlu undeca-peptides in an exo-type reaction. We are the first to report the enzymatic ethyl esterification reaction in the formation of pGlu peptides by rice koji peptidases.

Graphical Abstract

Peptidases of Aspergillus oryzae in rice koji form taste-active ethyl esters of pyroglutamyl peptides under sake mash conditions.

pyroglutamyl peptide ethyl ester, koji, peptidase, Aspergillus oryzae, sake

During the sake brewing process, proteins in the steamed rice grains are digested to peptides and free amino acids by rice koji enzymes (Okuda 2019). The N-termini of the glutelin acidic subunit, the most abundant endosperm storage protein in rice grains, provides a group of pyroglutamyl (pGlu) peptides in sake which are called bitter-tasting peptides because of their strong bitter and/or astringent taste (Hashizume et al. 2007a, 2007b; Maeda et al. 2011). After the discovery of bitter-tasting peptides, 2 ethyl esterified pGlu deca-peptides (PGDPE1: (pGlu)LFGPNVNPWCOOC₂H₅ and PGDPE2: (pGlu)LFNPSTNPWCOOC₂H₅) were reported in sake (Hashizume et al. 2012). PGDPE1 and PGDPE2 are unique taste-active components because of their bilateral tastes; both positive (rich, full mild, and/or good sake likeness) and negative (bitter and/or astringent) taste sensations were reported (Hashizume et al. 2012; Hashizume, Ito and Igarashi 2017). The maximum levels of PGDPE1 and PGDPE2 in sake exceeds their recognition threshold values (3.8 and 8.1 µg/L, respectively) (Hashizume, Ito and Igarashi 2017). Taste-active ethyl esterified pGlu penta-peptides have also been recently found in sake (Hashizume et al. 2019). They have favorable taste characteristics, such as sweetness or richness, with very low difference threshold values (27-68 µg/L) which may have sensory impacts on sake quality; however, the mechanisms of their formation are unexplained.

Aspergillus oryzae, a koji mold, produces many peptidases in rice koji. Before genome analysis of A. oryzae, 4 acid carboxypeptidases (ACPs) had been found in rice koji for sake brewing (Mikami et al. 1987). Genome analysis revealed that A. oryzae RIB 40 strain has 134 genes encoding proteinases, including 24 genes for carboxypeptidases (Machida et al. 2005; Yamagata 2016). ACP activities were assayed using Cbz-Glu-Tyr as a substrate (Nakadai, Nasuno and Iguchi 1973; Takeuchi, Ushijima and Ichishima 1982); however, Cbz-Tyr-Ala has become a new common substrate (Suzuki, Imai and Suzuki 1999). The ACP activity assayed using Cbz-Glu-Tyr did not show any relationships with the arginine forming activity of rice koji (Iwano et al. 2001), or with the protein constituent (prolamin, glutelin, and globulin) contents of koji rice (Takahashi et al. 2008). We recently reported that ACP activity for Cbz-Glu-Tyr did not show any relationships with the crude protein content in koji rice, or with bitter-tasting peptide levels in the formed sake; however, ACP activity for (pGlu)LFGPNVNPWHCOOH ((pGlu)LFGPNVNPWH) showed significant relationships with both (Hashizume et al. 2020).

It was reported that proteinase and carboxypeptidase are able to catalyze both hydrolysis and esterification of the C-termini of peptides under suitable solvent conditions (Breddam, Widmer and Johansen 1981; Kise and Shirato 1985; Rolland-Fulcrand, Lazaro and Viallefont 1994). These findings suggested that proteolytic enzyme(s) in rice koji might be able to ethyl esterify the C-termini of peptides in sake mash when ethanol is present at high levels.

The aims of this study were to confirm the contribution of rice koji enzyme(s) to the ethyl esterification of the pGlu peptides, then identify the enzyme protein(s) after purification from the rice koji extract, and elucidate the properties of the purified enzyme(s). The results obtained in this study will provide useful knowledge of the taste-active pGlu peptide ethyl esters in sake.

Materials and methods

Chemicals and materials

Pyroglutamyl peptides of 10-11 amino acid residues were obtained from Medical & Biological Laboratories Co., Ltd. (Nagoya, Japan). Pyroglutamyl peptides of 2-6 amino acids residues were purchased from BEX Co. Ltd. (Tokyo, Japan). N-[(Benzyloxy) carbonyl]-l-α-glutamyl-l-tyrosine (Cbz-Glu-Tyr) was purchased from Sigma-Aldrich Japan (Tokyo, Japan). ACP assay kit for rice koji was obtained from Kikkoman Biochemifa Company (Tokyo, Japan). Ethyl ester of pGlu peptides and their deuterium (D) labeled peptides were synthesized as previously reported (Hashizume, Ito and Igarashi 2017; Hashizume et al. 2019). Silver staining reagents kit was obtained from Cosmo Bio Co., Ltd. (Tokyo, Japan). PNGase F was purchased from New England Bio labs Japan Inc. (Tokyo, Japan). Sequi-Blot TM PVDF membrane and Coomassie G-250 stain (CBB) were obtained from Bio-Rad Laboratories, Inc. (CA, USA). All HPLC columns for enzyme purification were purchased from Tosoh Co. (Tokyo, Japan). Rice koji was prepared from 40% or 50% polished Akitasakekomachi steamed rice grains using A. oryzae RIB128 strain according to the standard sake koji making process (Ito et al. 2014).

Enzymatic formation of pGul deca-peptide ethyl esters by rice koji extract

Rice koji extract was prepared as reported previously (Ito et al. 2014), then treated with a PD-10 column (GE Healthcare Life Sciences, Buckinghamshire, England). Formation of ethyl esterified peptide was tested using the following reaction mixture: 5 µL of 0.1 m sodium lactate (pH 3.0), 5 µL of 0.4 mg/mL pGlu peptide in 47.5% (v/v) ethanol aqueous solution, 25 µL of water, 5 µL of 47.5% (v/v) ethanol aqueous solution, and 10 µL of rice koji extract. The reaction was incubated at 30°C for 30 min before 0.25 µg of D labeled internal standard and 20 µL of 95% (v/v) ethanol was added. The ethyl esterified peptides were immediately analyzed by high resolution MS, as reported previously (Hashizume, Ito and Igarashi 2017; Hashizume et al. 2019). Other peptides were analyzed by HPLC (Hashizume et al. 2012).

Enzyme assays

Using (pGlu)LFGPNVNPWH as a common substrate, 2 types of ester forming and/or hydrolyzing activities were assayed: type (A) activity was (pGlu)LFGPCOOC₂H₅ ((pGlu)LFGP-ethyl) and (pGlu)LFGPCOOH ((pGlu)LFGP) forming activity and type (B) activity was PGDPE1 and (pGlu)LFGPNVNPWCOOH ((pGlu)LFGPNVNPW) forming activity. The reaction mixture was composed of 5 µL of 150 µm (pGlu)LFGPNVNPWH in 47.5% (v/v) ethanol aqueous solution, 5 µL of 0.1 m sodium lactate (pH 3.0), 2-10 µL of enzyme solution, and water to a total volume of 20 µL. The reaction was conducted at 30°C for 15 min to 6 h, then 20 µL of acetonitrile was added to stop the reaction. After the addition of D labeled internal standard (5-50 ng) in 20 µL of 50% (v/v) acetonitrile aqueous solution, the ethyl esterified peptides were analyzed by high resolution MS (Hashizume, Ito and Igarashi 2017; Hashizume et al. 2019) and other peptides were analyzed by HPLC (Hashizume et al. 2012). ACP activity for Cbz-Glu-Tyr was measured according to the method of the National Research Institute of Brewing (https://www.nrib.go.jp/bun/nribanalysis.htm). ACP activity for Cbz-Tyr-Ala was measured according to the manufacturer's instructions. One katal of enzyme activity was defined as the amount of enzyme required to form product peptide at the rate of 1 mol s^–1 at 30°C.

Analysis of pGlu peptide ethyl ester forming activities in rice koji extract using AX-300 weak anion exchange column chromatography

Lyophilized koji extract from 2 g of rice koji was dissolved in 100 µL of water, and applied to a SynChropac AX-300 column (250 × 4.6 mm) (Eichrom Industries Inc., IL, USA). The HPLC conditions were, solvent A: 10 m m sodium phosphate (pH 6.8) and solvent B: 1 m NaCl and 10 m m sodium phosphate (pH 6.8). Samples were held for 10 min using A:B at 100:0, and then eluted using a linear gradient of A:B at 100:0 to A:B at 0:100 for 40 min using a flow rate of 1.0 mL/min. The volume of 1 fraction was 1 mL. Fractions with a high NaCl content (>0.35 m) were desalted using PD-10 columns before enzyme assays. The pGlu peptide ethyl ester forming activities in the obtained fractions were analyzed.

Purification of pGlu peptide ethyl ester forming enzymes

Purification was conducted by monitoring 2 ethyl ester forming activities: type (A) and (B). Rice koji (1 kg) was extracted with 6 L of 0.5% (w/v) NaCl in 10 m m sodium acetate buffer (pH 5.0), giving a resulting volume of 4.75 L. After concentrating to 80 mL, using Vivaflow 50 R, 10 000 MWCO Hydrosart (Sarutorius Japan, Tokyo, Japan), ammonium sulfate was added to 90% saturation. The supernatant was dialyzed against 10 m m sodium phosphate buffer (pH 6.8), then concentrated to 32 mL using the Vivaflow 50 R. The condensate was applied to TSKgel SQ-5PW (7.5 × 75 mm) column chromatography (3.5 mL per batch). The column was equilibrated with 10 m m sodium phosphate buffer (pH 6.8) and eluted using a NaCl gradient (0-1.0 m). Flow rate was 1 mL/min, and fraction volume was 1 mL. Active fractions were pooled and desalted. The solution (51 mL), which contained both type (A) and (B) activities, was subjected to a second round of TSKgel SQ-5PW column chromatography (18 mL per batch); the conditions were as described above except the flow rate was 0.5 mL/min, and fraction volume was 0.5 mL. Obtained active fraction was desalted and the buffer was changed to 10 m m sodium phosphate (pH 8.2). The solution was then applied to TSKgel DEAE-5PW (7.5 × 75 mm) column chromatography (8 mL per batch). The column was equilibrated with 10 m m sodium phosphate buffer (pH 6.8) and eluted using a NaCl gradient (0-1.0 m). Flow rate was 0.5 mL/min, and fraction volume was 0.5 mL. The active fraction was divided into type (A) abundant pool and type (B) abundant pool. Each pool was purified separately by a second round of TSKgel DEAE-5PW column chromatography under the conditions described above. During the purification procedure, protein concentration was measured using the Bradford method and BSA as the standard (Bradford 1976). The active fraction from the second DEAE-5PW column chromatography was applied to a TSKgel SuperSW3000 column (4.6 × 300 mm) equilibrated with 0.2 m NaCl and 10 m m phosphate buffer (pH 6.8). The column was eluted with the same buffer. Flow rate was 0.15 mL/min, and fraction volume was 0.15 mL. The active fractions were collected and condensed by an ultrafiltration membrane unit (Advantec Toyo Roshi Kaisya, Ltd., Tokyo, Japan) and applied to a second round of TSKgel SuperSW3000 chromatography using 10 m m phosphate buffer (pH 6.8) without NaCl. Flow rate was 0.15 mL/min, and fraction volume was 0.15 mL.

SDS-PAGE and N-terminal amino acid sequence analysis

SDS-PAGE was performed according to the method of Laemmli (1970). The PNGase F untreated sample was stained with silver, and the PNGase F treated sample was stained with CBB. Samples for N-terminal amino acid sequence analysis were blotted onto PVDF membrane and stained using CBB. Analysis was performed using a Shimadzu PPSQ-31A protein sequencer.

Identification of proteins by MALDI-TOF/TOF analysis

The CBB stained gel bands were enzymatically digested in-gel, similar to that previously described by Schevchenko et al. (1996) using a modified porcine trypsin (Promega Co., Madison, AL, USA). Samples were analyzed using the Bruker Autoflex Speed with LIFTTM ion optics. Both MS and MS/MS data were acquired with a Smartbeam Laser with 2 kHz repetition rate, and up to 400 shots were accumulated for each spectrum. MS/MS mode was operated with 2 keV collision energy; air was used as the collision gas such that nominally single collision conditions were achieved. Both MS and MS/MS data were acquired using the instrument default calibration. MS/MS ions searches were performed with the license Mascot for in-house use.

Results and discussion

Formation of pyroglutamyl deca-peptide ethyl esters by rice koji extract

PGDPE1 and PGDPE2 were formed in rice koji extract from their corresponding substrates (Table 1); however, when koji extract was heated to 100°C (5 min) PGDPEs were not formed. The molar ratios of ethyl esterification to hydrolysis were 2.1%-2.8%. Ethyl esterification of (pGlu)LFGPNVNPW was very low (0.02%-0.19%). The results indicated that PGDPE forming enzymes exist in the rice koji extract.

Table 1.

Pyroglutamyl deca-peptide ethyl esters (PGDPEs) formation activity in koji extract

Reaction	PGDPE formation activity (nmol/min/mL koji extract)	Ratio to hydrolysisa (%)
(pGlu)LFGPNVNPWH → PGDPE1b	2.37 ± 0.09	2.8
(pGlu)LFGPNVNPW → PGDPE1	0.02 ± 0.02	0.02
(pGlu)LFNPSTNPWH → PGDPE2b	1.45 ± 0.01	2.1
(pGlu)LFNPSTNPW → PGDPE2	0.13 ± 0.04	0.19

a

Data indicate % ratio of (esterified mol)/(hydrolyzed mol).

b

PGDPE1 and PGDPE2 mean (pGlu)LFGPNVNPW-ehtyl and (pGlu)LFNPSTNPW-ethyl, respectively.

Analysis of pGlu peptide ethyl ester forming activities in rice koji extract by AX-300 weak anion exchange column chromatography

After confirmation of enzymatic formation of PGDPEs (Table 1), enzymatic formation of (pGlu) penta-peptide ethyl esters from (pGlu)LFGPNVNPWH was investigated. The (pGlu) peptide ethyl ester forming activities in rice koji extract were analyzed by AX-300 weak anion exchange column chromatography. More than 7 peaks were identified during the elution of ACP activity for Cbz-Glu-Tyr; however, the elution of ethyl esterification activity was limited (Figure 1). PGDPE forming activity (type (B) activity) was eluted along with (pGlu)LFGPNVNPWH hydrolyzing activity, which suggested that 1 enzyme might perform both hydroxylation and ethyl esterification. (pGlu)LFGP-ethyl forming activity (type (A) activity) was eluted within the PGDPE1 forming fractions. The elution patterns of type (A) and (B) activities differed, which suggested that different enzymes performed these reactions.

Figure 1.

AX-300 weak anion exchange column chromatography of rice koji extract. Black filled circles and white open circles indicate hydrolysis activity for Cbz-Glu-Ty and for (pGlu)LFGPNVNPWH, respectively. Black filled triangle and white open triangle indicate formation activity of (pGlu)LFGP-ethyl (type (A) activity) and PGDP1 from (pGlu)LFGPNVNPWH (type (B) activity), respectively.

Purification of pGlu peptide ethyl ester forming enzymes

The purification procedure is summarized in Table 2. Type (A) and (B) activities were separated using DEAE-5PW chromatography (Figure S1), but were overlapped using SQ-5PW chromatography. The highest activity fractions in the second SuperSW3000 chromatography were applied to SDS-PAGE analysis. Purified preparations with type (A) and (B) activity were named as enzyme (A) and (B), respectively.

Table 2.

Purification of pyroglutamyl peptide ethyl ester forming enzymes

		(A) typea				(B) typeb
	Total protein	Total units	Specific activity	Recovery	Purification	Total units	Specific activity	Recovery	Purification
Step	(mg)	(nkat)	(nkat/mg)	(%)	(fold)	(nkat)	(nkat/mg)	(%)	(fold)
Crude extract	500	0.470	0.0009	100	1	85.7	0.17	100	1
Ultramembrane condensate	495	0.542	0.0011	115	1	71.0	0.17	83	1
90% (NH4)2SO4 supernatant	263	0.328	0.0012	70	1	81.3	0.31	95	2
1st SQ-5PW chromatography	11.3	0.242	0.0214	51	23	82.0	7.2	96	42
2nd SQ-5PW chromatography	4.36	0.240	0.0550	51	59	62.2	13.1	73	76
1st DEAE-5PW chromatography	3.36	0.165	0.0491	35	52	37.5	18.6	44	109
2nd DEAE-5PW chromatography	0.85	0.112	0.1314	24	140	18.0	21.2	21	123
1st Super SW3000 chromatographyc			0.1717		183		25.8		150
2nd Super SW3000 chromatography			0.3767		401		41.2		240

a

(pGlu)LFGP-ethyl forming activity was measured.

b

(pGlu)LFGPNVNPWCOOH forming activity was measured.

c

Part of the sample from 2nd DEAE-5PW chromatography fraction was applied.

SDS-PAGE and N-terminal amino acid sequence analysis

Purified native samples showed smearing bands with estimated mass weights of ca. 105 kDa for enzyme (A) and 120 kDa for (B) (Figure 2). After deglycosylation by PNGase F, each sample showed a major band at about 60 kDa. Some minor bands were observed in the enzyme (B) lane, but they were not analyzed in this study. Previously reported apparent mass weights of native A. oryzae ACP were over 100 kDa (Nakadai, Nasuno and Iguchi 1972; Nakadai, Nasuno and Iguchi 1973; Yamagata 2016), suggesting glycosylation. Overexpressed ACP proteins of A. oryzae ACP genes using other fungi were also heavily glycosylated, while the mass weights of their peptide chains ranged from ca. 55-70 kDa (Morita et al. 2009, 2010, 2011). The relationship of mass weights between native and deglycosylated samples reported in this study was similar to previous reports. N-terminal amino acid sequence analysis of both deglycosylated enzyme samples was unsuccessful; this might be due to the modification of N-terminal amino acids or the presence of impure proteins.

Figure 2.

SDS-PAGE analysis of the active fractions of the second SuperSW3000 size-exclusion chromatography. Proteins before and after deglycosylation were stained with silver and CBB, respectively. Deglycosylation of the purified enzyme samples was carried out using PNGase F following the manufacturer's instructions. Lane M, molecular mass marker; lane (A), enzyme (A); lane (B), enzyme (B). Identification analysis was performed on arrow indicated bands.

Identification of enzyme proteins by MALDI-TOF/TOF analysis

Samples were analyzed using Mascot and the BLAST database. Although some impure proteins, for example glucoamylase (accession number AO090003000321), existed in both purified samples, only 1 peptidase was identified in each preparation. Results of the MALDI-TOF/TOF analysis are summarized in Table 3. Amino acid sequences of 2 peptides from sample (A) were matched with a putative serine-type peptidase S28 of A. oryzae BCC7051 (Genbank accession number OOO13707.1). The amino acid sequence closely matched to the nucleotide sequence of AoS28A prolyl endopeptidase of A. oryzae NRRL 2220 (Eugster et al. 2015). The AoS28A prolyl endopeptidase gene is not common among A. oryzae strains; it is not present in A. oryzae RIB40 (Eugster et al. 2015). Amino acid sequences of 4 peptides in sample (B) were matched with a putative serine carboxypeptidase of A. oryzae RIB40 (accession number AO090010000534). Several serine type ACP genes and proteins of A. oryzae have been identified (Morita et al. 2009, 2010, 2011; Blinkovsky et al. 1999); however, this is the first report of a gene whose function was suggested to be a serine type ACP.

Table 3.

Results of MALDI TOF/TOF analysis

Band	m/z	Sequences identified	Mascot score	Proteins identified	Range	Description
(A)	1180.587	R.TINAEYWQR.Q	37	OOO13707.1	402-409	Peptidase S28
	1546.787	R.LLWVNGQYDPWR.D	23		451-462
(B)	1022.617	R.HLQWLLGR.V	36	AO090010000534	532-538	Serine-type carboxypeptidase
	1095.606	K.NFQELFGIK.N	20		193-201
	1215.632	K.IYVTGESYAGR.Y	41		205-305
	2440.069	R.GLMWAETYQSGHMQPQYQPR.V + 2 Oxidation (M)	27		497-517

Enzymatic properties of purified enzymes

Effects of pH

The optimum pH for (pGlu) LFGP formation from (pGlu)LFGPNVNPWH by enzyme (A) was 3.5, and that for (pGlu)LFGP-ethyl ((pGlu)LFGP ethyl ester) formation was 4.0 (Figure S2A). The optimum pH for (pGlu)LFGPNVNPWH exo-type hydrolysis was 4.0 for enzyme (B), and that of PGDPE formation was 3.0 (Figure S2B). The observed differences between the 2 enzyme reactions for the same substrate might be due to the differing conditions required for the different types of enzymatic reaction. The observed pH profile suggested that ethyl esterification may proceed well under sake mash pH (4.3) conditions.

Effects of temperature

Temperature profiles for both enzymes were clearly different (Figure S3). The optimum temperatures for both hydrolysis and ethyl esterification were 40°C for enzyme (A), and 45°C for enzyme (B). The lower optimum temperature of enzyme (A) was similar to the prolyl endopeptidase from A. oryzae isolated in China whose optimum temperature was 30°C (Kang, Yu and Xu 2014). It was reported that an ACP obtained by the overexpression of A. oryzae OcpC gene was unstable over 45°C (Morita et al. 2011), which was similar to enzyme (B). Our results suggest that hydrolysis and ethyl esterification might proceed slowly in sake mash where the temperature ranges from 5 to 15°C.

Enzyme activity for various peptide substrates

Enzymatic activity on various peptides was investigated (Table 4). Enzyme (A) formed pGlu hexa-peptides and their ethyl esters from the (pGlu) undeca-peptides; this might be due to prolyl endo-type reactions. It was reported that an overexpressed protein of the gene LN866855 of A. oryzae NRRL 2220, a homologous protein of OOO13707.1 of A. oryzae BCC7051, was a prolyl endopeptidase (Eugster et al. 2015). Our data agrees with the previous report on the reaction type of enzyme (A). Trace exo-type hydrolysis activity for the undeca-peptides was also observed in enzyme (A); however, further analysis is required to elucidate the reaction type. Enzyme (B) showed high activity with (pGlu)LFGPNVNPWH but not (pGlu)LFGPNVNPW (data not shown), indicating that this enzyme is an ACP and not a proline Xaa carboxypeptidase (Salamin et al. 2017). It has been reported that A. oryzae ACP hardly hydrolyzes proline and proline-Xaa residues (Takeuchi, Ushijima and Ichishima 1982; Takeuchi and Ichishima 1986). The slow hydrolysis step might favor the use of ethanol instead of water in the reaction.

Table 4.

Relative activity of purified enzymes for various peptides

	Enzyme (A)		Enzyme (B)
Reaction	Activity	E/H ratioa	Activity
(Substrate → detected product)	(nkat/mg protein)	(%)	(nkat/mg protein)	E/H ratio
(pGlu)LFGPNVNPWH
→(pGlu)LFGPNVNPW	0.1		100b
→(pGlu)LFGPNVNPW-ethyl	ndc		1.5	1.5
→(pGlu)LFGP	1.74		nd
→(pGlu)LFGP-ethyl	0.186	10.9	nd
(pGlu)LFNPSTNPWH
→(pGlu)LFNPSTNPW	0.1		41
→(pGlu)LFNPSTNPW-ethyl	nd		0.48	1.2
→(pGlu)LFNP	1.92		nd
→(pGlu)LFNP-ethyl	0.053	2.6	nd
(pGlu)LFGPN
→(pGlu)LFGP	nd		nd
→(pGlu)LFGP-ethyl	nd		nd
(pGlu)LFNPS
→(pGlu)LFNP	nd		nd
→(pGlu)LFNP-ethyl	nd		nd
(pGlu)LF
→(pGlu)L	nd		27
→(pGlu)L-ethyl	nd		nd	0.0
Hydrolysis of Cbz-Glu-Tyr	nd	–	617	–
Hydrolysis of Cbz-Tyr-Ala	nd	–	790	–

Concentration of all pyroglutamyl peptides and ethanol was 37.5 µm and 11.9% v/v, respectively.

Concentration of Cbz-Glu-Tyr and Cbz-Tyr-Ala was 0.5 and 1.8 m m, respectively.

a

“E/H ratio” indicates esterification activity/hydrolysis activity.

b

All activities were standardized by this data.

c

“nd” indicates that products were not detected.

The effects of substrate concentration were analyzed using (pGlu)LFGPNVNPWH as a common substrate (Table 5 and Figure S4). Km values for endo-type hydrolysis and (pGlu)LFGP-ethyl formation by enzyme (A) was 0.40 and 0.28 m m, respectively, while those for exo-type hydrolysis and PGDPE1 formation by enzyme (B) were both 0.27 m m. The agreements between Km values for hydrolysis and esterification for both enzymes suggested that the hydrolysis and esterification reactions might be due to the same reaction intermediates. The values for both enzymes were comparable with previously reported values of A. oryzae ACPs (Takeuchi, Ushijima and Ichishima 1982; Morita et al. 2011). The kcat values of enzyme (A) were very low, while those of enzyme (B) were high and comparable with former ACP data. The Km values of enzymes (A) and (B) suggested that both enzymes might form (pGlu) penta-peptide ethyl esters and PGDPEs in sake mash.

Table 5.

Kinetic parameters for (pGlu)LFGPNVNPWH as a substrate

	Enzyme (A)			Enzyme (B)
	Km	kcat	kcat/Km	Km	kcat	kcat/Km
Formed peptides	(m m)	(s−1)	(s−1 m m−1)	(m m)	(s−1)	(s−1 m m−1)
(pGlu)LFGP	0.40	0.28	0.70
(pGlu)LFGP-ethyl	0.28	0.019	0.068
(pGlu)LFGPNVNPW				0.27	6.5	24
(pGlu)LFGPNVNPW-ethyl				0.27	0.058	0.21

Effects of ethanol concentration

The effects of ethanol concentration on the enzyme reactions were examined (Figure 3). Hydrolysis and ethyl esterification of enzyme (A) decreased with increasing ethanol concentration (Figure 3a). The ratio of ethyl esterification to hydrolysis increased with increasing ethanol concentration (from 5% to 15%) but decreased over 25%. The hydrolysis activity of enzyme (B) decreased with increasing ethanol concentration, while the rate of ethyl esterification increased with increasing ethanol concentration (5%-10%) and then decreased (Figure 3b). The ratio of ethyl esterification to hydrolysis increased up to 15% ethanol concentration and reached a constant ratio thereafter, up to 30% ethanol. These results indicated that both enzymes are able to form ethyl esters in sake mash where ethanol concentration ranges from 5% to 20%.

Figure 3.

Effects of ethanol concentration on enzyme activities. Substrate was (pGlu) LFGPNVNPWH. Enzyme (a); black and white circles indicate formation of (pGlu)LFGP and (pGlu) LFGP-ethyl (10×), respectively. Enzyme (b); black and white circles indicate C-terminal hydrolysis and ethyl esterification (50×), respectively. Black triangles indicate percent molar ratio of esterification/hydrolysis.

Effects of metals and reagents

The effects of metals and reagents on the ester forming and hydrolysis reactions were examined using (pGlu)LFGPNVNPWH as a substrate (Table 6). PMSF, a serine type ACP inhibitor, significantly inhibited enzyme (B), but did not inhibit enzyme (A). Previous studies revealed that the inhibition of A. oryzae peptidases by PMSF differed among the enzymes (Morita et al. 2009, 2010, 2011), and that the S28 family peptidase of A. oryzae was not inhibited by PMSF (Kang, Yu and Xu 2014). EDTA treatment weakly inhibited both enzymes, and addition of Fe³⁺ weakly inhibited enzyme (B). Inhibition by Cu⁺¹ was not observed for both enzymes although it was reported in a former study (Mikami et al. 1987).

Table 6.

Effects of metal ions or reagents on the hydrolysis and ester forming activities of enzymes^a

	Concentration	Enzyme (A)		Enzyme (B)
	(m m)	Hydrolysis	Esterification	Hydrolysis	Esterification
None		100.0	100.0	100.0	100.0
Metal ionsb
CuCl	1	109 ± 8	103 ± 6	106 ± 3	98 ± 2
CuSO4	1	106 ± 4	86 ± 3	98 ± 4	106 ± 8
FeSO4	1	97 ± 11	90 ± 4	100 ± 2	87 ± 4
FeCl3	1	111 ± 7	97 ± 6	75 ± 3	79 ± 5
Reagentsc
EDTA	10	79 ± 2	78 ± 5	80 ± 3	84 ± 10
PMSF	2	114 ± 6	106 ± 10	14 ± 2	29 ± 4

a

All enzyme activities were assayed using (pGlu)LFGPNVNPWH as a common substrate. Data were shown as relative % values standardized by none addition control.

b

Metal ions were added in the reaction mixture.

c

Enzymes were treated by the reagents at 30°C for 30 min before the reaction.

General consideration

Enzymatic formation of taste-active pGlu peptide ethyl esters in sake was confirmed. Two enzymes (A and B) responsible for ester formation were purified from a rice koji extract. Enzyme (A) was identified as a putative peptidase S28 (Genbank: OOO13707.1), and enzyme (B) as a putative serine-type carboxypeptidase (AO090010000534), by MALDI-TOF/TOF analysis. Both enzymes formed ethyl esterified peptides under sake mash conditions. Enzyme (A) may contribute to the formation of pGlu penta-peptide ethyl esters in sake. Enzyme (B) may contribute to the formation of PGDPEs and decrease the bitter-tasting peptides in sake. The abundance of (pGlu)LFGP-ethyl, (pGlu)LFNP-ethyl ester, and PGDPEs in sake might be explained by the reaction properties of the rice koji peptidases. The low enzymatic reactivity for proline and proline-Xaa C-terminal might give ethanol the chance to participate in the reaction to form ethyl esters. The 2 enzymes reported in this study showed both unique and interesting enzymatic properties. Further research focusing on the enzymatic reaction using pure enzyme preparations are needed. Formation activity of (pGlu)L ethyl ester; an abundant dipeptide in sake, was not detected in the 2 enzymes of this study, suggesting that some other peptidase(s) are required for esterification. The level of ethyl esterified pGlu peptides in sake are very low, but their threshold values are low enough to have sensory impacts on sake sensory quality; therefore, the esterification steps may have a significant effect on quality in sake brewing.

Data availability

The data underlying this article will be shared on reasonable request to the corresponding author.

Author contribution

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

Funding

None declared.

Disclosure statement

No potential conflict of interest was reported by the authors.