Isolation and characterization of a β-primeverosidase-like enzyme from Penicillium multicolor

Article abstract of DOI:10.1271/bbb.70.691

p-Nitrophenyl and eugenyl β-primeveroside (6-O-β-D-xylopyranosyl- β-D-glucopyranoside) hydrolytic activity was found in culture filtrate from Penicillium multicolor IAM7153, and the enzyme was isolated. The enzyme was purified as a β-primeverosidase-like enzyme by precipitation with ammonium sulfate followed by successive chromatographies on Phenyl Sepharose, Mono Q, and β-galactosylamidine affinity columns. The molecular mass was estimated to be 50 kDa by SDS-PAGE and gel filtration. The purified enzyme was highly specific toward the substrate p-nitrophenyl β-primeveroside, which was cleaved in an endo-manner into primeverose and p-nitrophenol, but a series of β-primeveroside as aroma precursors were hydrolyzed only slightly as substrates for the enzyme. In analyses of its hydrolytic action and kinetics, the enzyme showed narrow substrate specificity with respect to the aglycon and glycon moieties of the diglycoside. We conclude that the present enzyme is a kind of β-diglycosidase rather than β-primeverosidase.

Full text of DOI:10.1271/bbb.70.691

Biosci. Biotechnol. Biochem., 70 (3), 691–698, 2006
Isolation and Characterization of a ꢀ-Primeverosidase-Like
Enzyme from Penicillium multicolor
Kazutaka TSURUHAMI,1 Shigeharu MORI,2 Satoshi AMARUME,4 Shigetaka SARUWATARI,4
y
1;4;
Takeomi MURATA,4 Jun HIRAKAKE,3 Kanzo SAKATA,3 and Taichi USUI
1Science of Biological Resources, The United Graduate School of Agricultural Science,
Gifu University, Gifu 501-1193, Japan
2Amano Enzyme, Inc., Gifu R&D Center, 4-179-35 Sue-cho Kakamigahara, Gifu 509-0108, Japan
3Institute for Chemical Research, Kyoto University, Kyoto 611-0011, Japan
4Department of Biological Chemistry, Faculty of Agriculture, Shizuoka University, Shizuoka 422-8529, Japan
Received September 22, 2005; Accepted November 11, 2005
p-Nitrophenyl and eugenyl ꢀ-primeveroside (6-O-ꢀ-D-
xylopyranosyl-ꢀ-D-glucopyranoside) hydrolytic activity
was found in culture filtrate from Penicillium multicolor
IAM7153, and the enzyme was isolated. The enzyme was
purified as a ꢀ-primeverosidase-like enzyme by precip-
itation with ammonium sulfate followed by successive
chromatographies on Phenyl Sepharose, Mono Q, and
ꢀ-galactosylamidine affinity columns. The molecular
mass was estimated to be 50 kDa by SDS–PAGE and gel
filtration. The purified enzyme was highly specific
toward the substrate p-nitrophenyl ꢀ-primeveroside,
which was cleaved in an endo-manner into primeverose
and p-nitrophenol, but a series of ꢀ-primeveroside as
aroma precursors were hydrolyzed only slightly as
substrates for the enzyme. In analyses of its hydrolytic
action and kinetics, the enzyme showed narrow sub-
strate specificity with respect to the aglycon and glycon
moieties of the diglycoside. We conclude that the present
enzyme is a kind of ꢀ-diglycosidase rather than ꢀ-
primeverosidase.
used for Japanese green tea,3,4) Shuixian (C. sinensis
var. sinensis), used for oolong tea,5) and a cultivar
(C. sinensis var. assamica), used for black tea.6) The
purified enzyme showed high substrate specificity
toward ꢀ-Psds to produce primeverose and aglycones.7)
These results demonstrate that ꢀ-primeverosidase plays
an important role in the development of tea aroma
during black tea, oolong tea, and green tea produc-
tion.8–10) The amount of aroma precursors and ꢀ-
primeverosidase activity both decrease as tea leaves
mature.11) In tea leaves, alcoholic aroma precursors are
predominantly conjugated to disaccharides, such as
primeverose, rather than monosaccharides.8)
Recently we reported that a ꢀ-Psd hydrolyzing
enzyme from Aspergillus fumigatus AP-20 cleaved in
an endo-manner pNP (p-nitrophenyl) ꢀ-Psd into prime-
verose and p-nitrophenol,12) but enzymes derived from
pathogenic microbes are strictly regulated in their use as
food processing aids and it is difficult to overcome the
regulatory obstacles. Demonstrating the safety of en-
zymes from such microbes is expensive and time-
consuming, and the enzyme might not be readily
accepted by the consumer. Hence we searched for
another source showing ꢀ-Psd hydrolyzing activity with
pNP ꢀ-Psd among microbes already commonly used in
traditional food processing. The availability of such an
enzyme from these sources would expand its potential
applications in biotechnology, e.g., in the control of
aroma release, the high recovery of aromas in extracts,
and the enzymatic synthesis of ꢀ-Psds as aroma
precursors. This paper describes screening for a micro-
organism producing ꢀ-Psd hydrolyzing enzyme, the
purification and characterization of a unique ꢀ-diglyco-
sidase from Penicillium multicolor IAM7153.
Key words: ꢀ-primeverosidase-like enzyme; ꢀ-diglyco-
sidase; plant aroma precursors; ꢀ-prime-
veroside; Penicillium multicolor IAM7153
Various tea-aroma constituents, such as geraniol,1)
benzyl alcohol, 2-phenylethanol,2) and (Z)-3-hexenol
have been shown to be present mainly as diglycoside
precursors such as ꢀ-primeverosides (6-O-ꢀ-D-xylopyr-
anosyl-ꢀ-D-glucopyranoside, Psd) in tea plant leaves
used to produce oolong tea. Such ꢀ-Psds are known to
be cleaved in an endo-manner into primeverose and
aglycones by ꢀ-primeverosidase. A ꢀ-primeverosidase
was purified from crude enzyme extract from fresh
leaves of cv. Yabukita (Camellia sinensis var. sinensis),
y
To whom correspondence should be addressed. Fax: +81-54-238-4873; E-mail: actusui@agr.shizuoka.ac.jp
Abbreviations: pNP, p-nitrophenyl; Psd, primeveroside; Glcd, glucopyranoside; Gald, galactopyranoside; Xyld, xylopyranoside; Lamd,
laminaribioside; Gend, gentiobioside; Celd, cellobioside; Xyl, xylopyranose; Glc, glucopyranose

692
K. TSURUHAMI et al.
Materials and Methods
JMC5560, Aspergillus niger IAM2107, Aspergillus
awamori IFO4033, Rhizopus oryzae JCM5560, Talar-
omyces emersonii IFO9747, Trichoderma reesei
QM9414B, Penicillium camembertii FERMP-10452,
and Penicillium multicolor IAM7153). P. multicolor
IAM7153, Institute of Molecular and Cellular Bioscien-
ces, The University of Tokyo, had the highest enzyme
production of all the microorganisms tested. The fungus
was streaked on a Potato Dextrose Agar slant and
incubated at 27 ꢀC for 14 d. An agar block (5 ꢁ 5 mm)
taken from the Potato Dextrose Agar slant was inocu-
lated into the production media in 500 ml-shakers and
incubated at 27 ꢀC for 8 d on a reciprocal shaker (140
strokes/min). The culture broth from multiple cultures
was combined (5-liter total) and filtered through a filter
paper to remove the mycelia. The culture filtrate (4-liter)
was concentrated to 400 ml with an ultrafiltration module
SIP-1010 (Asahi Kasei, Osaka, Japan). The enzyme
solution was centrifuged to remove insoluble material.
The supernatant was lyophilized to yield crude enzyme
powder, which was used for enzyme purification.
General methods. Furcatin was isolated from leaves
of Viburnum furcatum Blume by a previously published
method with slight modifications.13) pNP ꢀ-Psd, prime-
verose, pNP ꢀ-gentiobioside (Gend), and 2-phenylethyl
ꢀ-Gend were prepared by our method.7,14) Sophorose,
cellobiose, gentiobiose, pNP ꢀ-glucopyranoside (Glcd),
and pNP ꢀ-galactopyranoside (Gald) were purchased
from Sigma Chemical (St. Louis, MO). Cellooligosac-
charides (DP 2–5), Laminarioligosaccharides (DP 2–5),
pNP ꢀ-cellobioside (Celd), and pNP ꢀ-laminaribioside
(Lamd) were purchased from Yaizu Suisankagaku
Industry (Shizuoka, Japan). Gentiotriose and gentiote-
traose were prepared by previous method.15) The affinity
adsorbents with ꢀ-galactosylamidine as the ligand were
prepared by our method.16,17) All other chemicals were
obtained from commercial sources.
Media. The formula for the medium used for seed
cultures for screening was as follows: Defatted soy meal
[Honen, Tokyo, Japan (2.0%)], glucose (3.0%), KH2PO4
(0.5%), (NH4)2SO4 (0.4%), and dry yeast powder
(0.3%). The medium for the main cultures used for
screening consisted of Gentose 80# [Nihon Shokuhin
Kako, Tokyo, Japan (3.0%)], KH2PO4 (2.0%),
(NH4)2SO4 (1.0%), and Meast P1G [Asahi Food and
Healthcare, Tokyo, Japan (3.1%)]. The medium for
enzyme production consisted of Solulys A-ST [Roquette
Japan, Tokyo, Japan (3.0%)], Pinedex No. 2 [Matsutani
Chemical Industry, Hyogo, Japan (1.0%)], and KH2PO4
(0.5%). The pH of the medium was adjusted to 5.5.
Enzyme assay. During purification, the enzyme
fractions were assayed for hydrolyzing activities toward
pNP ꢀ-Psd, pNP ꢀ-Glcd, and pNP ꢀ-Gald (assays A, B,
and C respectively): a mixture containing 80 ml of each
substrate (2 mM) in 20 mM sodium acetate buffer,
pH 5.5, and 20 ml of enzyme solution was incubated in
a 96-well microplate for 30 min at 40 ꢀC. The reaction
was stopped by adding 100 ml of 1 M sodium carbonate.
The p-nitrophenol released was measured photometri-
cally at 405 nm in a microplate reader (Biolumin 960,
Amersham Pharmacia Biotech, Sweden). One unit of the
enzyme was defined as the amount releasing 1 mmol of
p-nitrophenol per min. Assays to determine the optimum
temperature, the thermal stability, the optimum pH, and
the pH stability of the enzyme essentially followed
previous methods.12)
Isolation of microorganisms and fermentation.
Screening for microorganisms producing the ꢀ-Psd
hydrolyzing enzyme was carried out with 30 type
cultures of strains traditionally used for food processing.
Each type culture was streaked on a potato dextrose
agar slant and incubated at 27–32 ꢀC for 3–14 d. Agar
blocks (5 ꢁ 5 mm) were taken from the potato dextrose
agar slant, inoculated into the screening medium in a
glass test tube, and incubated at 27 ꢀC for 8 d on a
reciprocal shaker (140 strokes/min). The culture was
then centrifuged at 10,000 rpm for 10 min and the
supernatant was used to determine ꢀ-Psd hydrolyzing
activity. The supernatant (100 ml) was mixed with 100 ml
of eugenyl ꢀ-Psd (10 mg/ml in 20 mM acetate buffer) or
pNP ꢀ-Psd (5 mg/ml in 20 mM acetate buffer) and
incubated at 37 ꢀC for 96 h. The reaction was stopped by
incubation in boiling water for 10 min. Each assay
Analytical methods. TLC analysis was carried out on
a silica gel 60 F254 plate using a solvent system of ethyl
acetate/acetic acid/water (3:1:1, v/v). Sugars on the
plate were visualized by heating at 120 ꢀC for 10 min
after spraying with 20% sulfuric acid-methanol solution.
HPLC analysis was performed using a Mightysil RP-
18(H) column (ꢁ150 ꢁ 4:6 mm) on a JASCO LCS-905
HPLC System station with an ultraviolet detector
(absorbance at 210 nm). Elution was performed with
H2O at a flow rate of 0.6 ml/min at 65 ꢀC. Protein was
measured by the method of Lowry et al.19) Sodium
dodecyl sulfate–polyacrylamide gel electrophoresis
(SDS–PAGE) was done with a PhastSystem (Amersham
Pharmacia Biotech, Sweden). ‘‘Perfect’’ protein markers
15–150 kDa (Novagen, Darmstadt, Germany) were used
for mass calibration. Proteins were detected with
PhastGel Blue R (Coomassie R 350 stain). Gel filtration
was done with a HiPrep Sephacryl S-200 HR column
(ꢁ2:6 ꢁ 60 cm) equilibrated with 20 mM sodium acetate
mixture (20 ml) was analyzed by TLC [Silica gel 60 F254
,
Merck, Frankfurt, Germany; ethyl acetate/acetic acid/
H2O = 3:1:1 (v/v); detected by the orcinol-sulfuric acid
method].18) Spots of reaction products were compared
with those of authentic samples [primeverose, pNP ꢀ-
Psd, eugenyl ꢀ-Psd, glucose (Glc), and xylose (Xyl)]. A
ꢀ-Psd hydrolyzing enzyme was detected in the super-
natant from several microorganisms (Aspergillus oryzae

Characterization of ꢀ-Primeverosidase-Like Enzyme
693
buffer (pH 4.7). The protein was eluted with a same
buffer (flow rate, 1.0 ml/min).
lyzed by the purified enzyme of various kinds of ꢀ-
mono- and diglycosides and ꢀ-linked disaccharides were
investigated by incubating a mixture (0.5 ml) containing
20 mM substrate in 20 mM acetate buffer pH 5.5 with
8 mU of enzyme for pNP ꢀ-mono and diglycosides,
or 80 mU for aroma precursors, with assay A at 40 ꢀC.
Samples (50 ml) were taken at 5 min intervals (0, 5, 10,
15, 20, and 25 min) and inactivated by adding 50 ml of
1 M sodium carbonate or by boiling for 5 min. The
amount of hydrolysates formed at an early stage of
incubation (up to 15% hydrolysis) was analyzed by
spectrophotometry (405 nm) for p-nitrophenol, or by
HPLC for aroma components, and a determination kit.
From these results, the relative rate of hydrolysis for
each substrate was calculated, with the rate of pNP
ꢀ-Psd hydrolysis taken as 100%.
Purification of ꢀ-Psd hydrolyzing enzyme. The crude
enzyme powder (1.5 g) mentioned above was dissolved
in 15 ml of 20 mM sodium acetate buffer (pH 5.0).
Proteins were precipitated in 50% ammonium sulfate.
The precipitate was dissolved in 15 ml of 20 mM sodium
acetate buffer (pH 5.0) containing 30% ammonium
sulfate. After centrifugation, the supernatant was de-
salted and concentrated with Amicon PM10 membrane.
The lyophilized powder (178 mg) was dissolved in
20 mM sodium acetate containing 30% ammonium
sulfate (pH 5.0) and applied to a HiLoad 16/10 Phenyl
Sepharose H. P. column (ꢁ1:6 ꢁ 10 cm) equilibrated
with 20 mM sodium acetate buffer (pH 5.5) containing
30% ammonium sulfate. The column was washed with
350 ml of the equilibration buffer (flow rate, 2 ml/min)
and eluted with a linear gradient of ammonium sulfate
from 30 to 0% in 480 ml of the same buffer, followed by
150 ml of 20 mM sodium acetate buffer (pH 5.0). The
elution pattern showed the enzyme mixture to be
fractionated into two peaks (F-1 and F-2) showing
enzyme activity with assay A. F-2 (tubes 36–42), which
showed the same elution profile of the activity as assays
A, B, and C, was pooled, concentrated with Amicon
PM10 membrane, and lyophilized. This fraction was
next applied to a Mono Q R5/5 column (ꢁ 0:5 ꢁ 5 cm)
equilibrated with 20 mM sodium acetate buffer (pH 5.0).
The column was washed with 50 ml of the same buffer
(flow rate, 0.5 ml/min), and eluted with a linear gradient
from 0 M (40 ml) to 0.2 M NaCl (40 ml) in the same
buffer. The elution pattern showed two fractions (F-1a
and F-1b) with assay A. F-1a (tubes 18–21), which
showed higher activity than F-1b with assay A, was
pooled and concentrated to a small volume. The solution
was then applied to a ꢀ-galactosylamidine affinity
column (2 ml bed volume) equilibrated with 20 mM
Tris–HCl buffer (pH 7.0). After the column was washed
with the same buffer, proteins were eluted stepwise by
lowering the pH of the 20 mM sodium acetate buffer to
5.0, 4.0, and 3.0 (flow rate, 0.5 ml/min). A high level of
activity with assay A, associated with the protein peak,
was recovered from the column at pH 5.0. More tightly
bound proteins, mainly showing activity with assays B
and C, were eluted with successive washes with 50 ml of
20 mM sodium acetate buffer (pH 4.0) and 50 ml of
20 mM sodium acetate buffer (pH 3.0), followed by
20 ml of the same buffer containing 1 M NaCl. The
fractions eluted at pH 5.0 (tubes 53–61) were pooled,
concentrated with Amicon PM10 membrane, and
lyophilized to give a purified enzyme. The purity of
the enzyme was checked by SDS–PAGE (PhastSystem,
Sweden Pharmacia Biotech) and gel filtration. This
preparation was used as purified enzyme for this study.
Action of the purified enzyme on ꢀ-diglycosides.
Incubations with substrates (20 mM) were carried out
at 40 ꢀC in 0.5 ml of 20 mM acetate buffer (pH 5.5).
Reactions were initiated by the addition of the enzyme
(8 or 80 mU with assay A). Samples (50 ml) were taken
at 5-min intervals (0, 5, 10, 15, and 20 min) during
incubation, and inactivated by adding 50 ml of 1 M
sodium carbonate. The amount of each hydrolysate
formed from the initial substrates by an early stage of
incubation with the enzyme (up to 15–20% hydrolysis)
was determined by HPLC. The amount of each product
increased linearly with time in the initial stage of the
reaction. Based on these data, the frequency of the
purified enzyme-catalyzed cleavage of glycosidic link-
ages was determined.
Kinetic parameters of the purified enzyme. The
amount of p-nitrophenol or eugenol liberated from the
initial substrates (pNP ꢀ-Glcd, pNP ꢀ-Gend, pNP ꢀ-
Psd, and eugenyl ꢀ-Psd) was determined by spectropho-
tometry (405 nm) for p-nitrophenol and by HPLC for
eugenol. The initial rates of the enzymatic reaction were
derived from kinetic curves of product accumulation,
as described above. The Michaelis–Menten parameters
were determined by 1/v–1/[S] plots and the least-
squares method. Five different substrate concentrations
(0.1–40 mM) were used for each experiment.
Results
A total of 30 type cultures were grown in a liquid
medium and screened for secretion of a ꢀ-Psd hydro-
lyzing activity with pNP ꢀ-Psd, namely the hydrolysis
of pNP and eugenyl ꢀ-Psds to primeverose and the
corresponding aglycones. Each culture supernatant was
reacted with pNP and eugenyl ꢀ-Psds, and the reaction
mixture was analyzed by TLC. Hydrolysis of the
substrates was detected by the release of primeverose.
ꢀ-Psd hydrolyzing activity was detected in the super-
natant from eight different type cultures; P. multicolor
IAM7153 had the highest enzyme activity. The super-
Relative rate of hydrolysis of various substrates by the
purified enzyme. The relative rates of hydrolysis cata-

694
K. TSURUHAMI et al.
pH 3.0
with 1M NaCl
pH 7.0
pH 5.0
pH 4.0
pH 3.0
0.5
0.4
0.3
0.2
0.1
0.0
0.12
0.10
0.08
0.06
0.04
0.02
0.00
0
50
100
Fraction no.
150
200
Fig. 1. Separation of the ꢀ-Psd Hydrolyzing Enzyme on a Column of ꢀ-Galactosylamidine Affinity Resin.
After Mono Q column chromatography, the sample was chromatographed on a ꢀ-galactosylamidine affinity column (see ‘‘Materials and
Methods’’). The pH of the buffer was changed stepwise: tubes 0–50, elution with 20 mM Tris–HCl buffer (pH 7.0); tubes 51–100, elution with the
same buffer at pH 5.0; tubes 101–150, elution with the same buffer at pH 4.0; tubes 151–200, elution with the same buffer at pH 3.0. Tubes
containing ꢀ-primeverosidase activity (tubes 51–61) were collected and lyophilized. , ꢀ-primeverosidase activity with assay A;
ꢀ-glucosidase activity with assay B; , ꢀ-galactosidase activity with assay C; ꢁ, A280
,
.
natant of the culture filtrate was lyophilized to give a
crude enzyme powder.
Table 1. Purification of the Enzyme
Specific Total Purification
activity activity
Purification
step
Protein
(mg)
Yield
(%)
factor
(fold)
Purification and characteristics of ꢀ-Psd hydrolyzing
enzyme
(U/mg)
(U)
111
30
Crude enzymea
30–50% ammonium
sulfate
1500
179
0.07
1.0
2.3
100
27
The crude enzyme preparation from P. multicolor
IAM7153 was precipitated with ammonium sulfate at
50% saturation and subjected to Phenyl Sepharose
column chromatography [HiLoad 16/10 Phenyl Sephar-
ose H. P., sodium acetate buffer (pH 5.5) containing 30%
ammonium sulfate]. Two fractions of this column (F-1
and F-2) showed enzyme activity with assays A and B.
The major fraction (F-1) was then applied to a Mono Q
HR 5/5 column [sodium acetate buffer (pH 5.0)] and
fractionated into two peaks (F-1a and F-1b), with assay A
activity overlapping protein absorbance. The major
fraction (F-1a) was finally loaded onto a ꢀ-galactosyl-
amidine affinity column, as shown in Fig. 1. After the
column was washed with 20 mM Tris–HCl at pH 7.0, the
adsorbed pNP ꢀ-Psd hydrolyzing activity was eluted
stepwise with 20 mM sodium acetate buffer at pHs 5.0,
4.0, and 3.0. Most of the enzyme activity adsorbed to the
ꢀ-galactosylamidine affinity resin was eluted at pH 5.0
with a very small amount of ꢀ-Glcd activity. ꢀ-
Galactosidase, which was more tightly bound to the ꢀ-
galactosylamidine affinity resin, was eluted below
pH 3.0. The fractions eluted at pH 5.0 were combined
and lyophilized to give a purified enzyme powder. The
purification is summarized in Table 1. The enzyme was
purified 20-fold based on assay A, with a specific activity
of 1.5 U/mg and 1.0% yield. The ꢀ-Psd hydrolyzing
enzyme eluted at pH 5.0 gave a single protein band
corresponding to 50 kDa on SDS–PAGE (Fig. 2) and gel
filtration using a column of Hi Prep Sephacryl S-200 HR
0.17
Phenyl Sepharose
Mono Q
ꢀ-Galactosylamidine
20
0.54
0.91
1.5
11
7.3
12
9.9
2.5
1.4
3.0
1.1
2.7
1.6
20
aLyophilized powder of a culture broth was used as starting material.
(data not shown). This indicates that the enzyme does not
possess a subunit structure. The enzyme had a broad pH
optimum in the range of pH 4.5–5.5 and was stable at all
pHs tested. The optimal temperature was 55 ꢀC, with
near-optimum activities observed from 50 to 60 ꢀC.
Hydrolytic activities on various substrates
The purified enzyme was found to hydrolyze pNP ꢀ-
Psd into primeverose and p-nitrophenol. Based on this
result, a series of naturally occurring ꢀ-diglycosides,
pNP ꢀ-diglycosides, and related compounds were tested
as substrates for the enzyme (Table 2). In this experi-
ment, the amounts of p-nitrophenol or aglycones formed
from initial substrates were analyzed at an early stage of
incubation with the enzyme (up to 15–20% hydrolysis).
The best substrate for the enzyme was the artificial
substrate pNP ꢀ-Psd; five other ꢀ-Psds and furcatin
(p-allylphenyl ꢀ-acuminoside) were not readily hydro-
lyzed. The relative rates of attack on these ꢀ-Psds were
compared with that on pNP ꢀ-Psd. The relative rate of
hydrolysis of eugenyl ꢀ-Psd (2.5%) was greater than that

Characterization of ꢀ-Primeverosidase-Like Enzyme
695
hydrolysis of primeverose, other ꢀ-linked glucobioses,
or glucooligosaccharides (DP 2–5) was observed. The
rate of hydrolysis of pNP ꢀ-Gend, in which Xyl is
replaced with Glc, was only 1/50 that of pNP ꢀ-Psd.
pNP ꢀ-Glcd still acted as a substrate despite the absence
of another sugar residue at the terminal position. 2-
Phenylethyl ꢀ-Gend, pNP ꢀ-Lamd, and pNP ꢀ-Celd did
not act as substrates. The substrate specificity of the
P. multicolor enzyme was also compared to that of ꢀ-
primeverosidase from tea leaves.7) The best substrate for
the tea enzyme was 2-phenylethyl ꢀ-Psd, and pNP ꢀ-Psd
was also hydrolyzed well. pNP ꢀ-Glcd was only a poor
substrate for the tea enzyme. Therefore, the substrate
specificity of P. multicolor enzyme appears to differ
appreciably from that of tea ꢀ-primeverosidase.
E
M
97.0 (kDa)
66.0
-
-
45.0
30.0
-
-
20.1
14.4
-
-
Fig. 2. SDS–PAGE of the Purified Enzyme from P. multicolor
IAM7153.
Mode of hydrolysis of chromogenic ꢀ-diglycosides
The mode of action of the purified enzyme on three
different ꢀ-diglycosides was further examined. From
HPLC chromatograms of the digests of each substrate,
the amounts of the products were calculated from their
peak areas. The amount of each product increased
linearly during the initial stage of incubation as the
enzyme reaction proceeded (data not shown). Based on
the product analysis, the frequency of enzyme-catalyzed
cleavage of each of the glycosidic linkages in the three
substrates was determined (Fig. 3). pNP ꢀ-Psd was
hydrolyzed exclusively, and pNP ꢀ-Gend predominant-
ly, at the glucosidic bond nearest the pNP group (bond
1). Eugenyl ꢀ-Psd was hydrolyzed to form aglycon and
ꢀ-Glcd in a ratio of roughly 0.8:0.2. The stereochemistry
of the hydrolysis catalyzed by the purified enzyme was
SDS–PAGE was performed with a 12.5% polyacrylamide gel
(PhastGel Homogeneous 12.5) and the protein visualized by
Coomassie Brilliant Blue staining. The migration of markers (15–
150 kDa) is shown on the right. E, the purified enzyme; M, size
marker.
Table 2. Relative Hydrolytic Rates of the Purified Enzyme on
Various Kinds of ꢀ-Glycosides and Reducing Saccharide Substrates
Hydrolytic rate on pNP ꢀ-Psd was set at 100%.
Rate of hydrolysis (%)
Substrate
Purified enzyme Tea ꢀ-primeverosidasea
pNP ꢀ-Psdb
100
2.5
NDd
ND
15
100
—
pNP ꢀ-Gendb
pNP ꢀ-Lamdb
pNP ꢀ-Celdb
pNP ꢀ-Glcdb
Eugenyl ꢀ-Psde
Benzyl ꢀ-Psde
2-Phenylethyl ꢀ-Psde
(Z)-3-Hexenyl ꢀ-Psde
Geranyl ꢀ-Psde
Furcatine
c
—
—
1
analyzed by H-NMR (Fig. 4), essentially as described
by Wong et al.20) 1H-NMR spectra of a reaction mixture
0.4
2.5
0.1
0.1
0.1
0.1
1.0
ND
ND
ND
ND
ND
—
—
185
—
—
—
Substrate
S1 S2
S3
S4
Gentiooligo.
—
—
Laminarioligo.
Cellooligo.
Sophorose
[100]
G
X
P
X
—
—
pNP β-Psd
G
P
P
Primeverose
—
G
G
P
G
[96]
[4]
aData are cited from the report by Ma et al.7)
bRelease of p-nitrophenol was detected photometrically at 405 nm.
cNot tested in the cited report.
pNP β-Gend
G
dNo hydrolytic products were detected.
eRelease of each aroma compound was detected by HPLC as the hydrolytic
product.
[81]
[19]
G
X
E
X
Eugenyl β-Psd
G
E
of furcatin (1.0%), despite the structural similarity of
their respective aglycones. The relative rate of hydrol-
ysis of ꢀ-Psds having alcoholic aglycones (benzyl, 2-
phenylethyl, (Z)-3-hexenyl, and geranyl) was 0.1%
(1,000-fold slower than pNP ꢀ-Psd). Of the naturally
occurring aroma precursors, eugenyl ꢀ-Psd was the most
rapidly hydrolyzed, although with 1/40 activity relative
to pNP ꢀ-Psd. It is notable that the alcoholic aroma
precursors were very poor substrates. No detectable
S1 S2
S3
S4
Fig. 3. Proposed Structure of the Cleavage Site of the Purified
Enzyme.
The arrow shows the cleavage site of the glycosidic linkage of
each substrate. Values in parenthesis indicate the frequency of
enzyme-catalyzed cleavage of each glycosidic linkage. pNP ꢀ-Psd
was hydrolyzed exclusively to the primeverose unit by the enzyme.
X, xylose; G, glucose; P, p-nitrophenol; E, eugenol.

696
K. TSURUHAMI et al.
A
B
H-1’
100
80
60
40
20
0
H-1
0 min
Ha
J = 8 Hz
20 min
40 min
0
30 60 840 870
Time (min)
He
J = 4 Hz
850 min
5.0
ppm
5.25
4.75
4.5
Fig. 4. Time Course of the Hydrolysis of pNP ꢀ-Psd by the Purified Enzyme from P. multicolor IAM7153.
Reaction samples were analyzed by 1H-NMR to study the stereochemistry of the enzymatic hydrolysis by the purified enzyme. A, 1H-NMR
spectra recorded at 0, 20, 40, and 850 min after the enzyme reaction. B, Generation of Ha and He after hydrolysis of pNP ꢀ-Psd by the purified
enzyme. The relative amounts of Ha and He were calculated from the intensity of each signal. H-1 and H-10, resonances of pNP ꢀ-Psd; Ha and
He, resonances of primeverose. , H-1 resonance of pNP ꢀ-Psd; , Ha resonance of primeverose; , He resonance of primeverose.
Table 3. Kinetic Paramerters of the Purified Enzyme from P. multicolor IAM7153
The parameters of Michaelis–Menten-type kinetics were evaluated by 1/v–1/[S] plots and the least square method.
Km
Vmax
(mmolꢃminꢂ1ꢃmg proteinꢂ1
kcat
(sꢂ1
kcat=Km
Substrate
Vmax=Km
ꢂ1
(mM)
)
)
(sꢂ1ꢃmM
)
pNP ꢀ-Glcda
pNP ꢀ-Genda
Eugenyl ꢀ-Psdb
pNP ꢀ-Psda
47
21
3.8
12
1.8
0.31
0.26
12
0.04
0.02
0.07
0.99
1.5
0.03
0.01
0.06
0.82
0.26
0.21
9.9
aRelease of p-nitrophenol was detected photometrically at 405 nm.
bRelease of eugenol was detected by HPLC as the hydrolytic product.
containing pNP ꢀ-Psd and the enzyme were measured at
the time intervals as shown in Fig. 4A. The ꢀ-anomer
(Ha, ꢂ 4.63, J ¼ 8:1 Hz) of primeverose was formed
first and increased with time during the early phase of
the reaction. The ꢃ-anomer (He, ꢂ 5.21, J ¼ 3:6 Hz) of
primeverose appeared later as a consequence of muta-
rotation. Therefore, the P. multicolor enzyme is a
retaining glycosidase, as has been observed for other
members of the ꢀ-glycosidase family 1.21,22)
The specificity constant of pNP ꢀ-Psd (kcat=Km
¼
0:82 sꢂ1 mMꢂ1) was 13.7-fold higher than that of
eugenyl ꢀ-Psd (kcat=Km ¼ 0:06 sꢂ1 mMꢂ1), 27.3-fold
higher than that of pNP ꢀ-Glcd (kcat=Km ¼ 0:03 sꢂ1
mMꢂ1), and 82-fold higher than that of pNP ꢀ-Gend
(kcat=Km ¼ 0:01 sꢂ1 mMꢂ1). These results indicate that
the synthetic diglycoside pNP ꢀ-Psd is the best substrate
for hydrolysis by the enzyme. Eugenyl ꢀ-Psd and pNP
ꢀ-Gend are not always good substrates for the enzyme.
The enzyme also hydrolyzes pNP ꢀ-Glcd, a monosac-
charide-glycoside, despite the absence of a Xyl residue
at the terminal position such as pNP ꢀ-Psd. Because
ꢀ-D-glucopyranosides of aroma compounds are present
in tea leaves and fruits, this characteristic of the enzyme
might be useful for improving the aroma quality of tea
drinks and fruit juices.
Kinetic parameters for chromogenic diglycoside sub-
strates
The kinetic parameters of the hydrolysis of three ꢀ-
diglycosides and ꢀ-Glcd were determined by colorimet-
ric analysis of p-nitrophenol or by HPLC analysis of
eugenol liberated from the substrate by the action of the
P. multicolor enzyme (Table 3). The affinity of the
enzyme for pNP ꢀ-Psd (Km ¼ 12 mM) was 3.2-fold
higher than that for eugenyl ꢀ-Psd (Km ¼ 3:8 mM), 3.9-
fold lower than that for pNP ꢀ-Glcd (Km ¼ 47 mM), and
1.8-fold lower than that for pNP ꢀ-Gend (Km ¼ 21 mM).
Discussion
The supernatant of culture filtrate from P. multicolor
IAM7153 showed the highest ꢀ-Psd hydrolyzing activity

Characterization of ꢀ-Primeverosidase-Like Enzyme
697
of the other culture filtrates of the 30 different type
cultures tested. Hence the supernatant was followed
by a purification procedure, in order to clarify the entity
of the ꢀ-Psd hydrolyzing enzyme. The crude enzyme
preparation was precipitated with ammonium sulfate
precipitation, followed by successive column chroma-
tographies (Table 1). In the final step, ꢀ-galactosylami-
dine affinity chromatography was highly effective for
separating the ꢀ-Psd hydrolyzing enzyme from contam-
inants. After the column was washed with 20 mM Tris–
HCl buffer at pH 7.0, the enzyme remained on the
adsorbent, suggesting that the active site of the enzyme
was bound to the positively-charged amidine base of the
affinity ligand. Most of the ꢀ-Psd hydrolyzing enzyme
with a small amount of ꢀ-glucosidase activity were
easily eluted by lowering the pH to 5.0 (Fig. 1). Under
these conditions, contaminating ꢀ-galactosidase re-
mained on the column, and then eluted below pH 3.0
(Fig. 1). The purity of the resulting enzyme was judged
by SDS–PAGE and gel filtration.
The hydrolytic action of the purified enzyme on
various ꢀ-Psds, including aroma precursors and related
compounds, was investigated to clarify the substrate
specificity of the enzyme with respect to the glycon and
aglycon moieties. pNP ꢀ-Psd, which is not a natural
substrate, was overwhelmingly the best substrate among
the ꢀ-glycosides listed in Table 2. Six naturally occur-
ring aroma precursor ꢀ-Psds and furcatin were hydro-
lyzed only to a limited extent. Among these, the four
alcoholic aroma precursor ꢀ-Psds synthesized as above
were hydrolyzed 10 to 25-fold slower than the other two
phenolic aroma eugenyl ꢀ-Psd and furcatin. Further-
more, three positional isomers of pNP ꢀ-glucobiosides
were compared with pNP ꢀ-Psd to define the substrate
specificity of the enzyme with respect to the glycon
moiety. The hydrolysis rate of pNP ꢀ-Gend was 1/50
that of pNP ꢀ-Psd. pNP ꢀ-Lamd and pNP ꢀ-Celd did
not act as substrates. No detectable hydrolysis of
primeverose, other ꢀ-linked glucobioses, or glucooligo-
saccharides (DP 2–5) was observed. On the other hand,
we recently reported that a ꢀ-Psd hydrolyzing enzyme
from A. fumigatus has broad substrate specificity with
respect to ꢀ-diglycosides and ꢀ-linked disaccharides.
pNP ꢀ-Gend acts as a fairly good substrate for the
enzyme (a rate of 41/100 relative to pNP ꢀ-Psd),
whereas primeverose and ꢀ-linked glucobioses act only
slightly as substrates. This suggests that the substrate
specificity of ꢀ-Psd hydrolyzing enzyme from A. fumi-
gatus differs somewhat from that of the present
P. multicolor enzyme. Ma et al.7) has reported that tea
ꢀ-primeverosidase shows a very narrow substrate speci-
ficity with respect to the glycon moiety, and a strong
preference for the ꢀ-primeverosyl moiety (Table 2). 2-
Phenylethyl ꢀ-Psd is a fairly good substrate for the tea
enzyme (185/100 relative to pNP ꢀ-Psd), whereas pNP
ꢀ-Glcd is a poor substrate (0.4/100 relative to pNP ꢀ-
Psd). Tea ꢀ-primeverosidase recognizes aroma precur-
sors more effectively than the P. multicolor enzyme.
From these results, the present enzyme was shown to be
highly specific for only one substrate, pNP ꢀ-Psd.
From the kinetic parameters and bond cleavage
frequency for the hydrolysis of three ꢀ-diglycosides,
we propose a subsite structure for the enzyme, that
matches the shape of pNP ꢀ-Psd and can accommodate
a chain of three residues to fit into the active site
(Fig. 3). As a matter of convenience, it is assumed that
the active site of the enzyme consists of four subsites
(S1, S2, S3, and S4), and that the glycosidic bonds of the
substrates are split between S2 and S3. The present
enzyme did not act on primeverose or other ꢀ-linked
glucobioses as substrates. These results show that the
disaccharide must have an aglycon moiety to accom-
modate at S3 and allow for binding to the active site.
The enzyme is also strict for binding the aglycon group,
because the catalytic efficiencies of eugenyl and 2-
phenylethyl ꢀ-Psds were dramatically lower than that of
pNP ꢀ-Psd (Table 3). The enzyme is also strict for
binding at S1, since the catalytic efficiency of pNP ꢀ-
Gend, in which a Xyl residue is replaced with Glc,
decreased remarkably as compared to that of pNP ꢀ-Psd
(Table 3). pNP ꢀ-Glcd, in which a Xyl residue is absent
at the terminal position, still acted as a substrate, but
primeverose and ꢀ-linked glucobioses did not. This
suggests that the p-nitophenyl group of pNP ꢀ-Glcd is
predominantly bound to S3 and gives some affinity to
the subsite. Thus a structural change to any part of pNP
ꢀ-Psd can affect interaction with the active site of the
enzyme. We concluded that pNP ꢀ-Psd binds best,
among the substrates tested, to the subsites geometri-
cally complementary to three residues, S1, S2, and S3.
Systematic trends in kinetic data obtained after changes
in the substrate structure were helpful in revealing the
structural requirements for binding and catalytic speci-
ficity. We found that the enzyme was able to hydrolyze
pNP ꢀ-Psd to liberate a primeverose unit and an
aglycon, and that the hydrolysis was specific to the
primeverose-aglycon ꢀ-glycosidic linkage and not the
inter-glycosidic bond between Xyl and Glc. The enzyme
was shown to catalyze hydrolysis of the glycosyl bond
with retention of the anomeric configuration, a trait
consistent with the ꢀ-primeverosidase in the glycosyl
hydrolase family 1.21,22) In analyses of the hydrolytic
action and kinetics, the present enzyme showed narrow
substrate specificity with respect to the aglycon and
glycon moieties of the diglycoside. The fungal enzyme
was highly specific for pNP ꢀ-Psd rather than for
eugenyl ꢀ-Psd. On the other hand, we also confirmed
that the purified enzyme induced a ꢀ-primeverosyl
transfer reaction from pNP ꢀ-Psd as donor to aromas as
acceptors to produce a series of ꢀ-Psds, which were only
slightly hydrolyzed by the present enzyme, mentioned
above. In this case, it is notable that aromas such as
benzyl alcohol, 2-phenylethanol, and (Z)-3-hexenol
serve as suitable acceptors for transglycosylation. The
efficient synthesis of ꢀ-Psds and its synthetic mechanism
has been reported recently.23)

698
K. TSURUHAMI et al.
Ingredients J. (in Japanese), 208, 991–1003 (2003).
In conclusion, by screening for microorganisms
9) Sakata, K., ꢀ-Primeverosidase relationship with floral tea
aroma formation during processing of oolong tea and
black tea. In ‘‘Caffeinated Beverages,’’ American Chem-
ical Society, Washington, DC, pp. 327–336 (2000).
10) Sakata, K., Watanabe, N., and Usui, T., Molecular basis
of alcoholic aroma formation during tea processing. 3rd
International Conference on Food Science and Technol-
ogy, 93–105 (1999).
11) Ogawa, K., Moon, J. H., Guo, W., Yagi, A., Watanabe,
N., and Sakata, K., A study on tea aroma formation
mechanism: alcoholic aroma precursor amounts and
glycosidase activity in parts of the tea plant. Z.
Naturforsch., 50c, 493–498 (1995).
producing any ꢀ-Psd hydrolyzing enzymes, we found a
P. multicolor IAM7153 strain that produces a unique ꢀ-
primeverosidase-like enzyme that hydrolyzes ꢀ-Psd in
an endo-manner into primeverose and aglycon. The
enzyme showed much narrower substrate specificity
with respect to the aglycon moiety than tea ꢀ-primever-
osidase. We conclude that the present P. multicolor
enzyme is a kind of ꢀ-diglycosidase rather than ꢀ-
primeverosidase from tea leaves, because naturally
occurring aroma precursors such as benzyl, 2-phenyl-
ethyl, and geranyl ꢀ-Psds are hydrolyzed only slightly as
substrates for the enzyme.
12) Yamamoto, S., Okada, M., Usui, T., and Sakata, K.,
Isolation and characterization of a ꢀ-primeverosidase-
like endo-manner ꢀ-glycosidase from Aspergillus fumi-
gatus AP-20. Biosci. Biotechnol. Biochem., 66, 801–807
(2002).
13) Imaseki, H., and Yamamoto, T., A furcatin hydrolyzing
glycosidase of Viburnum furcatum blume. Arch. Bio-
chem. Biophys., 92, 467–474 (1961).
Acknowledgment
This work was supported by a research grant (Agri-
business) from the Ministry of Agriculture, Forestry, and
Fisheries of Japan.
14) Murata, T., Shimada, M., Watanabe, N., Sakata, K., and
Usui, T., Practical enzymatic synthesis of primeverose
and its glycoside. J. Appl. Glycosci., 46, 431–437 (1999).
15) Nanjo, F., Usui, T., and Suzuki, T., Mode of action of an
exo-ꢀ-(1!3)-D-glucanase on the laminaran from Eise-
nia bicyclis. Agric. Biol. Chem., 48, 1523–1532 (1984).
16) Inoue, K., Hiratake, J., Mizutani, M., Takada, M.,
Yamamoto, M., and Sakata, K., ꢀ-Glycosylamidine as
a ligand for affinity chromatography tailored to the
glycon substrate specificity of ꢀ-glycosidases. Carbohyd.
Res., 338, 1477–1490 (2003).
17) Hiratake, J., and Sakata, K., Glycosylamidines as potent
selective and easily accessible glycosidase inhibitors and
their application to affinity chromatography. Methods
Enzymol., 363, 421–443 (2003).
18) Svennerholm, L. J., The quantitative estimation of
cerebrosides in nervous tissue. J. Neurochem., 1, 42–53
(1956).
19) Lowry, O. H., Rosebrough, N. J., Farr, A. L., and
Randall, R. J., Protein measurement with the folin
phenol reagent. J. Biol. Chem., 193, 265–275 (1951).
20) Wong, A. W., He, S., Grubb, J. H., Sly, W. S., and
Withers, S. G., Identification of Glu-540 as the catalytic
nucleophile of human ꢀ-glucuronidase using electro-
spray mass spectrometry. J. Biol. Chem., 273, 34057–
34062 (1998).
21) Mizutani, M., Nakanishi, H., Ema, J. I., Ma, S. J.,
Noguchi, E., Inohara-Ochiai, M., Fukuchi-Mizutani, M.,
Nakao, M., and Sakata, K., Cloning of ꢀ-primeverosi-
dase from tea leaves, a key enzyme in tea aroma
formation. Plant Physiol., 130, 2164–2176 (2002).
22) Ahn, O. Y., Mizutani, M., Saino, H., and Sakata, K.,
Furcatin hydrolase from Viburnum furcatum blume is a
novel disaccharide-specific acuminosidase in glycosyl
hydrolase family 1. J. Biol. Chem., 279, 23405–23414
(2004).
References
1) Guo, W., Sakata, K., Watanabe, N., Nakajima, R., Yagi,
A., Ina, K., and Luo, S., Geranyl 6-O-ꢀ-xylopyranosyl-
ꢀ-D-glucopyranoside isolated as aroma precursor from
tea leaves for oolong tea. Phytochemistry, 33, 1373–
1375 (1993).
2) Guo, W., Hosoi, R., Sakata, K., Watanabe, N., Yagi, A.,
Ina, K., and Luo, S., (S)-Linalyl, 2-phenylethyl, and
benzyl disaccharide glycosides isolated as aroma pre-
cursors from oolong tea leaves. Biosci. Biotechnol.
Biochem., 58, 1532–1534 (1994).
3) Guo, W., Yamauchi, K., Watanabe, N., Usui, T., Luo, S.,
and Sakata, K., A primeverosidase as a main glycosidase
concerned with the alcoholic aroma formation in tea
leaves. Biosci. Biotechnol. Biochem., 59, 962–964 (1995).
4) Guo, W., Ogawa, K., Yamauchi, K., Watanabe, N., Usui,
T., Luo, S., and Sakata, K., Isolation and characterization
of a ꢀ-primeverosidase concerned with alcoholic aroma
formation in tea leaves. Biosci. Biotechnol. Biochem.,
60, 1810–1814 (1996).
5) Ogawa, K., Ijima, Y., Guo, W., Watanabe, N., Usui, T.,
Dong, S., Tong, Q., and Sakata, K., Purification of a ꢀ-
primeverosidase concerned with alcoholic aroma for-
mation in tea leaves (cv. Shuixian) to be precessed to
oolong tea. J. Agric. Food Chem., 45, 877–882 (1997).
6) Ijima, Y., Ogawa, K., Watanabe, N., Usui, T., Ohnishi-
Kameyama, M., Nagata, T., and Sakata, K., Character-
ization of ꢀ-primeverosidase, being concerned with
alcoholic aroma formation in tea leaves to be processed
into black tea, and preliminary observation on its
substrate specificity. J. Agric. Food Chem., 46, 1712–
1718 (1998).
7) Ma, S. J., Mizutani, M., Hiratake, J., Hayashi, K., Yagi,
K., Watanabe, N., and Sakata, K., Substrate specificity of
ꢀ-primeverosidase, a key enzyme in aroma formation
during oolong tea and black tea manufacturing. Biosci.
Biotechnol. Biochem., 65, 2719–2729 (2001).
8) Mizutani, M., and Sakata, K., ꢀ-Primeverosidase and its
contribution to formation of floral tea aromas during
oolong and black tea manufacturing. Foods Food
23) Tsuruhami, K., Mori, S., Sakata, K., Amarume, S.,
Saruwatari, S., Murata, T., and Usui, T., Efficient
synthesis of ꢀ-primeverosides as aroma precursors by
transglycosylation of ꢀ-diglycosidase from Penicillium
multicolor. J. Carbohydr. Chem., 24, 849–863 (2005).