Causes of the Production of Multiple Forms of -Galactosidase

by Bacillus circulans
Jingyuan SONG,
1
Kiriko ABE,
1
Hiroyuki IMANAKA,
1
Koreyoshi IMAMURA,
1
Masashi MINODA,
2
Shotaro YAMAGUCHI,
3
and Kazuhiro NAKANISHI
1;f
1
Graduate School of Natural Science and Technology, Okayama University,
3-1-1 Tsushima-naka, Kita-ku, Okayama 700-8530, Japan
2
Daiwa Kasei Co., Ltd., 4-19 Hie-cho, Konan, Shiga 520-3203, Japan
3
Gifu R&D Center, Amano Enzyme Inc., 1-6 Technoplaza, Kakamigahara, Gifu 509-0109, Japan
Received August 9, 2010; Accepted November 2, 2010; Online Publication, February 7, 2011
[doi:10.1271/bbb.100574]
The presence of multiple types of -galactosidases
in a commercial enzyme preparation from Bacillus
circulans ATCC 31382 and dierences in their trans-
galactosylation activity were investigated. Four -gal-
actosidases, -Gal-A, -Gal-B, -Gal-C, and -Gal-D,
which were immunologically homologous, were isolated
and characterized. The N-terminal amino acid se-
quences of all of the enzymes were identical and
biochemical characteristics were similar, except for
galactooligosaccharide production. -Gal-B, -Gal-C,
and -Gal-D produced mainly tri- and tetra saccharides
at maximum yields of 2030 and 912%, while -Gal-A
produced trisaccharide with 7% with 5% lactose as
substrate. The Lineweaver-Burk plots for all of the
enzymes, except for -Gal-A, showed biphasic behavior.
-Gal-A was truncated to yield multiple -galactosidases
by treatment with protease isolated from the culture
broth of B. circulans. Treatment of -Gal-A with
trypsin yielded an active 91-kDa protein composed of
21-kDa and 70-kDa proteins with characteristics similar
to those for -Gal-D.
Key words: -galactosidase; Bacillus circulans; trans-
galactosylation; galactooligosaccharide
-Galactosidase (-D-galactoside galactohydrolase,
EC 3.2.1.23) is a widely distributed enzyme that is
found in microorganisms, plants, and animal tissues
15)
and plays important physiological roles, since it pro-
vides energy and a carbon source via the hydrolysis
of lactose to glucose and galactose. It is also utilized
in various areas of the food industry.
48)
It is used to
hydrolyze lactose to glucose and galactose to produce
lactose-free milk that can be used by lactose-intolerant
individuals, in addition to enhancing the sweetness of
the milk. Because it also has transglycosylation activity,
it is utilized to produce galactooligosaccharides (GOS)
and functional galactosylated compounds. GOS are
important prebiotics that enhance the proliferation
of intestinal bidobacteria and lactobacilli in the colon,
which in turn promote human health by protecting
individuals against infection, reducing the accumulation
of harmful bacteria/toxic compounds, and facilitating
the normal function of the gut.
811)
Because of this, a
number of studies have reported attempts to isolate a
-galactosidase with high transgalactosylation activity
from various microorganisms,
48)
including Bacillus
circulans,
12,13)
Bidobacterium adolescentis,
14)
Lacto-
bacillus reuteri,
15,16)
and others.
1720)
Furthermore, the
genes that encode -galactosidases from bacteria, in-
cluding B. circulans,
21)
Bidobacterium bidum
2225)
/
Bidobacterium infantis,
22,2628)
L. reuteri,
29)
L. planta-
rum,
30)
and others,
20,31)
have been cloned and sequenced,
and the enzymes have been overexpressed in heterolo-
gous cells.
Previously, Nakanishi (one of the authors of this study),
Mozaar, and co-workers
12,13)
isolated and puried two
-galactosidases, -galactosidase-1, with an estimated
molecular mass of approximately 240 kDa, and
-galactosidase-2, with a molecular mass of 160 kDa
by a commercial enzyme preparation produced by
B. circulans (Biolacta

, Daiwa Kasei, Shiga, Japan),

originally from B. circulans ATCC 31382. They re-
ported that -galactosidase-1 showed very low trans-
galactosylation activity, while -galactosidase-2 had the
highest level of transgalactosylation activity among a
number of enzymes isolated.
4,5,12,13)
In particular, -
galactosidase-2 does not require high lactose concen-
trations to produce high yields of oligosaccharides, in
contrast to most -galactosidases reported to date.
4)
-Galactosidase-2 produces GOS at a yield of about
41% from a 4.56% solution of lactose at 40

C. Hence,
during the past 20 years, the commercial enzyme
Biolacta

has frequently been utilized to synthesize

various functional galactosylated compounds
3242)
as
well as GOS.
5,9,12,13,43,44)
In a later study, Vetere et al.
45)
isolated a third enzyme, with a molecular mass of
86 kDa, from Biolacta

, but they did not study oligosac-

charide production. Ajisaka et al.
4648)
expressed a
-galactosidase gene in E. coli cells cloned from
f
To whom correspondence should be addressed. Tel/Fax: +81-86-251-8200; E-mail: kazuhiro@cc.okayama-u.ac.jp
Abbreviations: BSA, bovine serum albumin; EDTA, ethylenediaminetetraacetic acid; GOS, galactooligosaccharides; HPLC, high performance
liquid chromatography; MALDI/TOF MS, matrix-assisted laser desorption/ionization, time-of-ight mass spectrometry; MU, 4-methylumbelliferone;
MUGA, 4-methylumbelliferyl -galactoside; o-NP, o-nitrophenol; o-NPF, o-nitrophenyl -D-fucopyranoside; p-NPF, p-nitrophenyl -D-fucopyrano-
side; o-NPG, o-nitrophenyl -D-galactopyranoside; p-NPG, p-nitrophenyl -D-galactopyranoside; PCMB, p-chloromercuribenzoic acid; TFA,
triuoroacetic acid
Biosci. Biotechnol. Biochem., 75 (2), 268278, 2011
B. circulans ATCC 31382, and used it to synthesize
functional galactosylated compounds, but its calculated
molecular mass was only 66.9 kDa,
21)
much smaller than
the -galactosidase-1 and -galactosidase-2 isolated
from Biolacta

. Thus ambiguities concerning the -

galactosidases produced by B. circulans ATCC 31382
clearly exist.
The objective of this study was to clarify the causes
of the presence of multiple forms of -galactosidases
in Biolacta

and to examine the dierences in trans-

galactosylation activity among enzymes of dierent
molecular sizes, through biochemical characterization.
Materials and Methods
Materials. Biolacta

, a commercial enzyme preparation of -

galactosidase from B. circulans ATCC 31382, a product of Daiwa
Kasei was used as the enzyme source. CHTCeramic Hydroxyapatite
Type I, Sephadex G-100 Superne, and p-aminobenzyl-1-thio--
galactopyranosideagarose were purchased from Bio-Rad Laboratories
(Hercules, CA), Amersham Biosciences (Uppsala, Sweden), and
Sigma-Aldrich (St. Louis, MO) respectively. o-NPG, p-NPG, o-
nitrophenyl -D-glucopyranoside, o-NPF, p-NPF, and p-nitrophenyl
-L-arabinopyranoside were from Sigma-Aldrich. m-Nitrophenyl
-D-galactopyranoside, p-nitrophenyl -D-glucopyranoside, and m-
nitrophenyl -D-glucopyranoside were from Wako Pure Chemical
Industries (Osaka, Japan). MUGA and casein (from milk, acc.
Hammarsten) were from Wako. All other reagents were of analytical
grade and were purchased from Wako or Sigma-Aldrich, unless
otherwise noted.
Buers. The following buers were used: buer A, 100 mM sodium
phosphate buer, pH 6.0; buer B, 50 mM sodium acetate buer,
pH 5.8; buer C, 50 mM sodium acetate buer, pH 3.5; buer D,
20 mM sodium phosphate buer containing 150 mM NaCl, pH 6.0;
buer E, 10 mM TrisHCl buer containing 10 mM CaCl
2
, pH 8.0.
-Galactosidase assay. -Galactosidase activity was assayed using
o-NPG or lactose as substrate. When o-NPG was used as substrate, the
reaction was initiated by adding 10 mL of an enzyme solution dissolved
in buer A to 990 mL of 0.245% o-NPG in the same buer to give a
nal o-NPG concentration of 0.243% (8 mM), followed by incubation
for 10 min at 40

C. A 1-mL aliquot of 10% Na

2
CO
3
was then added to
the reaction mixture to terminate the reaction, and the o-NP released
was determined by measuring the absorbance at 420 nm using a
Benchmark Plus Microplate Spectrophotometer (Bio-Rad). One
enzyme unit with o-NPG as the substrate (U
o-NPG
) was dened as the
amount of enzyme required to produce 1 mmol of o-NP in 1 min at
40

C at pH 6. When lactose was used as substrate, 10 mL of the

enzyme solution dissolved in buer A was added to 400 mL of a 5%
lactose solution dissolved in the same buer to give a nal
concentration of 4.88% (143 mM) and this was incubated for 15 min
at 40

C. The reaction was terminated by boiling the reaction mixture

for 3 min, and the glucose produced was measured by the mutarotase-
glucose oxidase/peroxidase method using Glucose CII-test (Wako).
One enzyme unit with lactose as substrate (U
Lactose
) refers to the
amount of enzyme required to produce 1 mmol of D-glucose in 1 min at
40

C at pH 6.0.
Purication of -galactosidases. All purication steps were
performed at 4

C, except for hydroxyapatite chromatography. During

the purication steps, enzyme activity was measured using o-NPG as
substrate by the method described above, and the protein concentration
was monitored by measuring the absorbance at 280 nm with a UV-VIS
recording spectrophotometer (UV-2500 PC, Shimadzu, Kyoto, Japan).
Two and a half g of an enzyme powder of Biolacta

was rst dissolved

in 50 mL of 10 mM sodium phosphate buer, pH 6.0, and dialyzed
against the same buer. The dialyzed solution was then applied to a
hydroxyapatite column (CHT Ceramic Hydroxyapatite Type I, bed
volume; 2.5 cm i.d. 25 cm) at room temperature and eluted stepwise,
using 10, 100, 150, 200, 300, and 500 mM sodium phosphate buers,
pH 6.0, at a ow rate of 0.67 mL/min by a previously reported
method.
43)
Three active fractions that showed -galactosidase activity
eluted at 100 mM, 150 mM, and 300 mM sodium phosphate buers,
pH 6.0. These three active fractions were collected separately and
brought up to 75% saturation with (NH
4
)
2
SO
4
to precipitate the
enzyme. Each solution was then centrifuged at 9;800 g at 4

C for
20 min to recover the precipitated enzyme. The active fraction that
eluted at 300 mM sodium phosphate buer, pH 6.0, was found to be
homogeneous by SDSPAGE as explained in Results below and
designated -Gal-A. The active fractions that eluted with 100 mM and
150 mM sodium phosphate buers, pH 6.0 (designated as active
fractions I and II respectively), were separately dialyzed against buer
B at 4

C. Dialyzed active fractions I and II were then loaded onto an

anity chromatographic column (p-aminobenzyl-1-thio--galactopyr-
anoside-agarose, gel bed volume; 1.6 cm i.d. 13 cm) that had been
equilibrated with the same buer that was used for dialysis, and the
column was eluted using a pH gradient from 5.8 to 3.5 using the same
amounts of buer B and buer C after a wash-through elution using
buer B at a ow rate of 0.5 mL/min. In the anity chromatography
of active fraction I, -galactosidase activity appeared in the fractions
that eluted at about pH 4.4, and the enzyme was found to be
homogeneous (designated -Gal-C) as shown in Results. On the
other hand, anity chromatography of the active fraction II was
conducted in a manner similar to that for fraction I, in which -
galactosidase activity was detected in both the wash-through fractions
and the fractions that eluted at about pH 4.4. Both enzyme fractions
were found to be homogeneous by SDSPAGE, as explained below.
The former was designated -Gal-D, and the latter as -Gal-B. All four
of the puried -galactosidases were dialyzed against buer A and
stored at 4

C until used.
SDSPAGE and Western blotting. Protein concentrations were
determined using a BCA Protein Assay Reagent Kit (Pierce Chemical,
Rockford, IL) with BSA as standard. SDSPAGE was performed using
a 6% gel by the method of Laemmli,
49)
and the gel was stained with a
Quick-CBB plus kit (Wako). Myosin (220 kDa),
2
-macroglobulin
(170 kDa), E. coli -galactosidase (116 kDa), transferrin (76 kDa), and
glutamic dehydrogenase (53 kDa) (HMW-SDS Calibration Kit, GE
Healthcare Bio-Sciences, Piscataway, NJ) were used as standard
proteins. For Western blotting, the enzyme solution was rst subjected
to 6% SDSPAGE gel and electro-blotted onto a polyvinylidene
diuoride (PVDF) membrane (Invitrogen Japan, Tokyo) using an
iBlot gel transfer device (Invitrogen) and then detected using anti--
Gal-A antibody (rabbit) (GeneDesign, Osaka, Japan) as primary
antibody and goat anti-rabbit IgG-AP antibody (Sigma-Aldrich) as
second antibody at a dilution of 1/3,000.
Active staining. Active staining was carried out using 4-methyl-
umbelliferyl -galactoside (MUGA) by the method reported by
Manchenko.
50)
First, 10 mL of the sample solution was mixed with
10 mL of SDS sample buer (2% SDS, 5% 2-mercaptoethanol, 62.5 mM
TrisHCl pH 6.8, 10% glycerol, and 0.01% bromophenol blue) at 20,
40, and 60

C for 5 min and subjected to SDSPAGE on a 10% gel. A

staining paper that had been soaked in 0.35% MUGA dissolved in
50 mM sodium phosphate buer, pH 6.5, was placed on top of the gel.
This was followed by incubation at 37

C for 30 min. Finally, the paper

was removed and the gel was placed under a 365-nm UV light to detect
the active band that contains 4-methylumbelliferone (MU). Precision
Plus Protein Standards from Bio-Rad were used as marker proteins.
MALDI/TOF MS molecular mass measurements. The molecular
masses of the puried proteins were determined by MALDI/TOF MS
using a Voyager-DETM PRO (Applied Biosystems, Foster City, CA)
at an accelerating voltage of 25 kV. A spectrum of the puried enzyme
dissolved in a matrix solution (10 mg/mL sinapinic acid in 30%
acetonitrile containing 0.1% TFA) was obtained with calibration
standards and analyzed using Grams 32 Processing Software (Pre-
Septive Biosystems, Framinghan, MA).
The molecular masses of the oligosaccharides were determined by
MALDI/TOF MS. Each oligosaccharide produced by hydrolysis of
lactose using the puried -galactosidases was isolated with two
YMC-PACK NH
2
columns (6.0 mm i.d. 150 mm) connected in
series followed by concentration using a rotary evaporator (N-1001S-
-Galactosidase from Bacillus circulans 269
W, Tokyo Rikakikai, Tokyo) at 50

C. The concentrated solution of

each oligosaccharide was mixed with a matrix solution (10 mg/mL
2,5-dihydroxybenzonic acid in 10% ethanol) to measure its spectrum
with calibration standards.
HPLC gel chromatography. The molecular masses of -Gal-A,
-Gal-B, -Gal-C, and -Gal-D were estimated by HPLC gel
chromatography. In a typical experiment, a 0.1-mL portion of the
enzyme solution (1.0 mg/mL) was injected onto a Superose 6HR
column (1 30 cm) and a Superose 12HR column (1 30 cm)
connected in series, and eluted at a ow rate of 0.4 mL/min with
buer D, with monitoring at 280 nm. Pyranose oxidase (Sigma-
Aldrich), alcohol dehydrogenase (Sigma-Aldrich), Aspergillus niger
-glucosidase,
51)
and the High Molecular Weight (HMW) and
Low Molecular Weight (LMW) gel ltration calibration kits (GE
Healthcare) were used for calibration.
Analytical centrifugations. Estimation of the molecular mass of a
large protein such as -Gal-A by gel chromatography sometimes
includes experimental errors. Hence, we determined the molecular
state and molecular masses of -Gal-A and -Gal-C by carrying out
sedimentation velocity experiments using an analytical centrifuge. The
sedimentation velocity experiments were performed in buer D using
a Proteomelab XL-1 Analytical Ultracentrifuge (Beckman-Coulter,
Fullerton, CA) by a previously reported method.
52,53)
A 390-mL portion
of -Gal-A at a concentration of 0.56 mM, 2.8 mM, or 5.6 mM was
sedimented at 35,000 rpm at 20

C using 12-mm charcoal-lled double

sector centerpieces with quartz windows, while -Gal-C (0.9 mM,
1.9 mM, and 7.4 mM) was sedimented at 40,000 rpm. Sedimentation
behavior was monitored with absorbance detection optics, and all
the sedimentation velocity data were analyzed by the continuous
sedimentation coecient distribution c(s) model included in the
SEDFIT11.71 software program.
54)
The parameters used in the analysis
was estimated as follows: The partial specic volumes of -Gal-A and
-Gal-B were calculated to be 0.7329 cm
3
/g and 0.7309 cm
3
/g
respectively from the amino acid compositions. The amino acid
compositions for -Gal-A and -Gal-C will be reported elsewhere. The
density and viscosity of the buer were estimated to be 1.00727 g/cm
3
and 1.0261 cP respectively by the SEDNTERP ver.1.09 program. The
axial dimensions (length diameter) of the protein were determined
on the basis of a prolate ellipsoid model, assuming a hydration level of
0.3 g of water/g of protein using the value of the sedimentation
coecient converted to standard conditions (in water at 20

C).
Sequencing of the N-terminal peptide. The N-terminal amino acid
sequences of the puried proteins were analyzed using a protein
sequencer (Applied Biosystems, Model 491) by a previously reported
method.
55)
pH Dependencies of the activity and stability of -galactosidases.
The pH dependency of the activity of each puried enzyme was
measured by carrying out reactions at 40

C, using o-NPG and lactose

as substrates dissolved in 100 mM acetate buers, pH 3.05.0, 100 mM
sodium phosphate buers, pH 5.58.5, and 100 mM sodium carbonate
buers, pH 9.010.5, in which the nal enzyme concentration was in
the range of 24 mg/mL.
For pH stability measurements, each enzyme was rst pre-incubated
for 1 h at 40

C in a pH range of 3.0 to 10.0, using the same buers as

for the measurements of pH dependency of the activity, at nal enzyme
concentrations of 0.20.4 mg/mL. Then the remaining activity was
then assayed using o-NPG dissolved in buer A as substrate by the
method as explained above.
Eects of temperature and reagents on the stability of -
galactosidases. For thermal stability measurements, each enzyme
dissolved in buer A at nal enzyme concentrations of 0.20.4 mg/mL
was pre-incubated for 1 h in a temperature range of 2070

C and the
remaining activity was then assayed using o-NPG as substrate, as
described above. For measurements of the eects of reagents on
enzyme stability, a 20-mL portion of each enzyme solution (0.2
0.4 mg/mL) was added to 40 mL of 1.5 mM PCMB, DL-dithioerythritol,
AgNO
3
, MgSO
4
, 1, 10-phenanthroline, 1.5 or 15 mM EDTA2Na,
15 mM KCl, and 15 mM NaCl dissolved in buer A and pre-incubated
for 10 min at 20

C. As a control, 40 mL of buer A was added in place

of the reagent solution. The remaining activity was assayed using
o-NPG as substrate.
Substrate specicity. The hydrolytic activities of each -galactosi-
dase were measured with respect to various lactose analogs, viz.,
o-NPG, m-nitrophenyl -D-galactopyranoside, o-nitrophenyl -D-
glucopyranoside, p-nitrophenyl -D-glucopyranoside, m-nitrophenyl
-D-glucopyranoside, o-NPF, p-NPF, and p-nitrophenyl -L-arabino-
pyranoside by the same method used for o-NPG activity measurement.
Kinetic measurements. Initial reaction rates were determined for the
hydrolysis of lactose in a range of 2.5300 mM and those for the
hydrolysis of o-NPG over a concentration range of 130 mM in buer
A at 40

C, in which the nal enzyme concentrations were about 6.0

mg/mL and 0.62.6 mg/mL, respectively. Typically, a 10-mL aliquot of
the puried enzyme solution was added to 400 mL (lactose) or 990 mL
(o-NPG) of substrate solution at 40

C to start the reaction. At

appropriate time intervals during the reaction, a 100-mL aliquot of the
reaction mixture was withdrawn, and the concentration of glucose or
o-NP was determined by a previously described method. The initial
reaction rates were determined from the linear portion of the time
courses for various substrate concentrations. The kinetic parameters
were determined from the Lineweaver-Burk plot (LB plot) using the
initial kinetic data obtained for dierent substrate concentrations, and
were expressed as the mean SD of values obtained by duplicate or
triplicate measurements.
Courses of hydrolysis of lactose. Courses of the enzymatic reaction
for a nal 5% lactose solution dissolved in buer A were obtained for
all of the puried -galactosidases at 40

C. A 10-mL aliquot of each

enzyme solution was added to 1,490 mL of a lactose solution at a nal
enzyme concentration of about 2 U
Lactose
/mL to start the reaction. At
appropriate time intervals during the reaction, a 100-mL aliquot of the
reaction mixture was withdrawn and boiled for 3 min to terminate the
reaction. The amount of glucose produced was determined as described
previously, and galactose was measured using a Lactose Assay Kit
(BioVision, Mountain View, CA).
The amounts of GOS produced by the transgalactosylation reactions
were quantied by HPLC (LS-20AT, Shimadzu) using a refractive
index detector (RID-10A, Shimadzu), on a column (6.0 mm i.d. 150
mm YMC-PACK NH
2
, YMC, Kyoto) with a guard column (6.0 mm
i.d. 5 mm YMC-PACK NH
2
) that was placed in a constant-temper-
ature oven (C0-8010, Tosoh, Tokyo). For analysis of lactose and other
disaccharide, a 70% CH
3
CN/H
2
O solution was used as elution buer at
a ow rate of 0.8 mL/min at a column temperature of 25

C. For
analysis of GOS, elution was performed using 60% CH
3
CN/H
2
O at a
ow rate of 0.8 mL/min at 25

C. Maltotriose, maltotetraose, malto-

pentaose (Hayasibara Shoji, Okayama, Japan) and 4
/
-galactosyllactose
(Wako) were used as standard oligosaccharides for quantication.
Treatment of -Gal-A with endogenous protease isolated. The
culture broth of B. circulans ATCC 31382 was ultraltered to obtain
the supernatant, followed by drying. Then 100 mg of dried enzyme
powder dissolved in 5 mL of buer A was allowed to come into contact
with 2.5 g (wet weight) of p-aminobenzyl-1-thio--galactopyranoside-
agarose gels, and then incubated at room temperature for 2 h with
gentle shaking to remove -galactosidases, and nally the supernatant
was recovered by centrifugation. This operation was repeated 2 times
to prepare a partially puried endogenous protease. Protease activity
was assayed using nal 0.5% w/v casein dissolved in 100 mM sodium
phosphate buer, pH 7.0 as substrate by the method of Yano et al.,
56)
with a slight modication. One unit of protease activity was dened as
the amount of enzyme which catalyzes the solubilization of 1 mg
protein at 37

C at pH 7.0 in 1 min.
For proteolytic digestion of -Gal-A, 0.2 mL of the protease
solution (2.8 mg/mL) was added to 2 mL of -Gal-A (1 mg/mL)
dissolved in buer A and this was incubated at 37

C for 1 week. At
appropriate times, a 100-mL aliquot of the sample solution was
withdrawn and analyzed by SDSPAGE using a 6% gel.
Tryptic digestion of -Gal-A and purication of the active fraction.
Crystalline trypsin from bovine pancreas (7.5 mg) (Wako) was added
to 7.5 mL of puried -Gal-A (1.0 mg/mL) dissolved in buer E, and
270 J. SONG et al.
the resulting solution was incubated at 37

C for 10 h. At appropriate
time intervals during incubation, a 60-mL aliquot of the reaction
mixture was withdrawn and placed in buer E containing a nal
0.5 mg/mL soybean trypsin inhibitor (Wako) to stop the proteolytic
reaction by the method of Benjakul et al.
57)
The remaining -
galactosidase activity was then determined using o-NPG and lactose as
substrate by the method described above. SDSPAGE of trypsin-
digested solutions taken at appropriate times was carried out on a 10%
gel. In addition, after a 12-h incubation, we added an additional 35 mg
of trypsin to the reaction mixture to promote the reaction. Using the
trypsin-digested solution of -Gal-A thus obtained, a major protein
component that showed -galactosidase activity was puried. To
accomplish this, the digested solution was rst dialyzed against 10 mM
phosphate buer pH 6.0, and then applied to a hydroxyapatite column
(1.6 cm i.d. 8 cm), followed by stepwise elution, changing the
sodium phosphate buer (pH 6.0) concentration at 10, 50, 100, 200,
and 300 mM at a ow rate of 0.5 mL/min. The -galactosidase fraction
that eluted at 200 mM sodium phosphate buer pH 6.0 was collected
and dialyzed against buer A. The dialyzed enzyme solution was
nally applied to a column of Sephadex G-100 Superne (1.6 cm
i.d. 40 cm) and eluted using buer A at a ow rate of 0.3 mL/min to
remove all traces of trypsin.
Results
Isolation and purication of the -galactosidases
from a commercial product
Figure 1A shows the results of SDSPAGE/Western
blotting for a commercial -galactosidase preparation
(Biolacta

) using an anti--Gal-A antibody (rabbit). The

SDSPAGE and Western blotting results indicated that
Biolacta

contains multiple forms of -galactosidases

of dierent molecular sizes, but that the forms are
immunologically homologous. Figure 1B shows the
SDSPAGE results for the puried -Gal-A, -Gal-B,
-Gal-C, and -Gal-D. The yields of puried -Gal-A,
-Gal-B, -Gal-C, and -Gal-D were approximately 15,
8, 3, and 5% as assayed using o-NPG as substrate, in
which the total activity included in the crude enzyme
preparation was set at 100%. Furthermore, the N-
terminal amino acid sequences of all of the puried
-galactosidases were determined to be GNSVSYDG-
ERRVNFNEN.
Biochemical characteristics of puried -galactosi-
dases
The molecular masses, as determined by MALDI/
TOF MS, SDSPAGE, and HPLC gel chromatography,
were similar for the various -galactosidases in
Biolacta

, as shown in Table 1, in which -Gal-91 is

a complex composed of 21-kDa and 70-kDa proteins
obtained by tryptic treatment that shows -galactosidase
activity, as described below. Furthermore, analysis using
sedimentation velocity data indicated that 91% of -Gal-
A is composed of a molecule with a sedimentation
-Gal-A
-Gal-B
-Gal-C
-Gal-D
A B
Fig. 1. Western Blotting and SDSPAGE of -Galactosidases from Bacillus circulans.
A, Results of SDSPAGE and Western blotting of the crude enzyme preparation. Lane 1, SDSPAGE results; lane 2, Western blotting results;
lane 3, marker proteins. B, SDSPAGE results of puried -Gal-A. Lane 1, crude enzyme; lane 2, -Gal-A; lane 3, -Gal-C; lane 4, -Gal-D;
lane 5, -Gal-B; lane 6, marker proteins.
Table 1. Properties of the Various -Galactosidases Present in Biolacta

-Gal-A -Gal-B -Gal-C -Gal-D -Gal-91

a
MALDI/TOF MS 189.283 153.932 134.788 91.627 91.307
b
Molecular SDSPAGE 195 160 135 86 20/70
c
mass (kDa) Gel chromatography 210 190 145 105
Ultracentrifugation 187 132
Optimum pH 5.56.5 5.56.5 5.56.5 5.56.5 5.56.5
pH Stability
d
5.78.5 4.58.5 5.08.5 4.58.5 4.58.5
Thermal stability
e
<50

C <50

C <50

C <50

C <50

C
Maximum yield of disaccharide (%)
f
3.0 5.4 6.8 14.9 10.2
Maximum yield of trisaccharide (%)
f
7.2 25.5 26.4 23.0 26.4
Maximum yield of tetrasaccharide (%)
f
0 9.2 10.7 10.6 9.0
Maximum yield of pentasaccharide (%)
f
0 2.0 1.8 2.5 1.6
a
An enzyme complex composed of the 21-kDa and 70-kDa proteins obtained by tryptic treatment.
b
Sum of 21-kDa proteins (21.034 kDa) and 70-kDa (70.273 kDa) with a total mass of 91.307 kDa.
c
Composed of two proteins of approximately 20 kDa and 70 kDa.
d
Incubated at 40

C for 1 h.
e
Incubated at pH 6.0 for 1 h.
f
Values with respect to the initial amount of lactose at 40

C at pH 6.0.
-Galactosidase from Bacillus circulans 271
coecient of 7.5 s, and 98% of -Gal-C, of a molecule
with a coecient of 6.8 s. The results also indicate that
the concentration dependencies of the distribution
patterns and s values for the two proteins were similar.
The molecular mass of -Gal-A with a coecient of
7.5 s, was determined to be 187 kDa, consistent with
the MALDI/TOF MS results (189.283 kDa), indicating
that the majority of -Gal-A molecules are present as
monodisperse monomers in solution, which is in con-
sistent with the HPLC gel chromatography data. The
axial dimension of -Gal-A was determined to be 4:7
28:6 nm using a determined frictional ratio of 1.48. By a
similar analysis, -Gal-C was found to be a monodis-
perse monomer of 132 kDa, close to the value deter-
mined by MALDI/TOF MS (134.788 kDa). The axial
dimension of -Gal-C was determined to be 4:8 19:1
nm using a frictional ratio of 1.35. These ndings
indicate that -Gal-A, -Gal-B, -Gal-C, and -Gal-D
are undoubtedly monomeric proteins, although most
bacterial -galactosidases reported to date are multi-
meric, except for that from Streptococcus mitis.
31)
-
Galactosidases from L. reuteri,
15)
Enterobacter agglom-
erans B1,
20)
and L. plantarum WCFS
30)
are dimeric,
those from B. adolescentis DSM20083,
14)
B. infantis
HL96,
28)
and E. coli,
58)
tetrameric, and that from
B. bidum NCIMB41171
22,25)
is hexameric.
Some of the biochemical characteristics determined
for puried -Gal-A, -Gal-B, -Gal-C, and -Gal-D
are summarized in Table 1. All four of the -galacto-
sidases showed an optimum reaction pH of 5.56.5
when activity was assayed using both o-NPG and lactose
as substrates and a similar pH stability at pH 4.5 or
5.08.5 at 40

C, except for -Gal-A, which showed a

slightly narrower pH stability of pH 5.78.5, slightly
dierent from those for -Gal-B, -Gal-C, and -Gal-D.
All four enzymes were stable up to a temperature of
approximately 50

C at pH 6.0, while the stability of -

Gal-B, -Gal-C, and -Gal-D was higher than -Gal-A
at 60

C, as shown in Fig. 2. Among the reagents tested,

only HgCl
2
was found to inactivate the enzymes. -Gal-
A was inactivated to some extent by EDTA2Na
(Table 2), while -Gal-B, -Gal-C, and -Gal-D re-
tained activity at 93% or higher. On the other hand, all
the enzymes were generally stable in the presence of
1,10-phenanthroline (Table 2).
Table 3 shows specic activities towards various
substrates in addition to lactose and o-NPG at 40

C at
pH 6.0. All of the enzymes showed appreciable activity
towards p-NPG, 34 times higher than that towards
o-NPG, while they showed very low or negligible
activity towards m-nitrophenyl -D-galactopyranoside,
o-nitrophenyl -D-glucopyranoside, p-nitrophenyl -D-
glucopyranoside, and p-nitrophenyl -L-arabinopyrano-
side (data not shown). The ratio of specic activity
towards o-NPG to that towards lactose for -Gal-A was
45 times higher than the corresponding values for
-Gal-B, -Gal-C, and -Gal-D.
Kinetic analyses
The kinetic data for all of the enzymes using lactose
and o-NPG showed biphasic LB plots, with the exception
for -Gal-A. Figure 3 shows biphasic LB plots for the
-Gal-C-catalyzed reactions using lactose and o-NPG as
substrates, which could be expressed approximately by
two straight lines (dotted lines), a line with a higher slope
in the higher substrate concentration range (smaller 1/[S]
values) and one with a lower slope in the lower substrate
concentration range (larger 1/[S] values). Hence, we
determined two sets of kinetic parameters for the
Michaelis-Menten equation, the maximum velocity
(V
max
) and Michaelis constant (K
m
) from the straight
line for the lower substrate concentration region, V
max,L
and K
m,L
, and those for the higher substrate concentration
region, V
max,H
and K
m,H
.
59,60)
On the other hand, the LB
0
20
40
60
80
100
10 20 30 40
50 60 70
Temperature (C)
R
e
l
a
t
i
v
e

r
e
m
a
i
n
i
n
g

a
c
t
i
v
i
t
y

(
%
)
Fig. 2. Thermal Stability of -Gal-A ( ), -Gal-B ( ), -Gal-C ( ),
-Gal-D ( ), and -Gal-91 ( ) Incubated at Various Temperatures
at pH 6.0 for 1 h.
-Gal-91 is an enzyme complex composed of the 21-kDa and 70-
kDa proteins obtained by tryptic treatment.
Table 2. Eects of Metal Ions and Other Reagents on the Stability of the Various -Galactosidases in Biolacta
;a
Metal ions and Concn.
b Relative remaining activity (%)
c
reagents (mM)
-Gal-A -Gal-B -Gal-C -Gal-D -Gal-91
d
HgCl
2
1 1 5 2 1 3
KCl 10 104 100 99 100 103
NaCl 10 102 101 129 104 105
MgSO
4
1 115 93 113 96 102
DL-Dithioerythiritol 1 115 90 113 98 105
EDTA2Na 1 84 95 104 98 105
10 57 98 101 93 98
PCMB 1 88 96 111 95 101
1,10-Phenanthroline 1 97 108 108 104 101
a
All values were determined at 40

C at pH 6.0.
b
Final concentration during pre-incubation.
c
The activity of enzyme pretreated in buer A was taken as 100%.
d
An enzyme complex composed of 21-kDa and 70-kDa proteins obtained by tryptic treatment.
272 J. SONG et al.
plot for the -Gal-A-catalyzed reaction using lactose as
substrate shows very weak biphasic behavior: A line with
a higher slope appears at very high substrate concen-
trations (very low 1/[S] values), as shown in Fig. 3C,
while the LB plot for the -Gal-A-catalyzed reaction
using o-NPG as the substrate is linear in the concen-
tration range tested, as shown in Fig. 3D. The kinetic
parameters determined for -Gal-B, -Gal-C, and -
Gal-D are similar for the substrate, as shown in Table 4,
indicating that these enzymes show similar kinetic
behavior although the parameters determined are appa-
rent values, as discussed below.
Relationship between the conversion of lactose and
saccharides produced
Figure 4 shows the relationship between the conver-
sion of lactose and saccharides produced in the course of
reactions using 5% lactose in buer A at 40

C with
puried -Gal-A, -Gal-B, -Gal-C, and -Gal-D,
which showed production of GOS in addition to glucose
and galactose. The molecular masses of the tri-, tetra-,
and pentasaccharides produced were conrmed by
MALDI/TOF MS. The molecular masses of the isolated
trisaccharide (theoretical molecular mass, 504.4 Da),
tetrasaccharide (666.6 Da), and pentasaccharide (828.7
Da) were determined to be 503.6 Da, 666.7 Da, and
828.7 Da by MALDI/TOF MS.
In the reaction catalyzed by -Gal-A, a trisaccharide
was predominantly formed at a maximum yield of
approximately 7% with respect to the initial amount of
lactose. On the other hand, a large amount of GOS,
composed mainly of tri- and tetrasaccharides was
produced when -Gal-B, -Gal-C, and -Gal-D were
used. The formation of di- and pentasaccharides was
also observed. The yield of pentasaccharide was
approximately 2% or less. In Table 1, the maximum
yields of tri-, tetra-, and pentasaccharides, and disac-
charide other than lactose produced are summarized. As
shown in Table 1 and Fig. 4, some dierences in the
yield and composition of GOS were observed among
-Gal-B, -Gal-C, and -Gal-D. -Gal-C appeared to
produce the highest yield of trisaccharides among the
three enzymes, while the use of -Gal-D resulted in the
highest yield of disaccharide other than lactose.
Table 3. Substrate Specicity of the Various -Galactosidases from Biolacta
;a
Substrate
Specic activity (U
b
/mg) and relative activity (%)
-Gal-A -Gal-B -Gal-C -Gal-D -Gal-91
c
Lactose 44.3 (100)
d
45.3 (100)
d
61.0 (100)
d
71.9 (100)
d
75.0 (100)
d
o-Nitrophenyl -D-galactopyranoside (o-NPG) 47.2 (107) 11.8 (26.0) 12.5 (20.5) 16.8 (23.4) 15.0 (20.0)
p-Nitrophenyl -D-galactopyranoside (p-NPG) 169.8 (383.3) 46.4 (102.4) 48.7 (79.8) 65.5 (91.0) 59.4 (79.2)
o-Nitrophenyl -D-fucopyranoside (o-NPF) 10.3 (23.3) 3.2 (7.1) 3.1 (5.1) 4.5 (6.3) 3.4 (4.5)
p-Nitrophenyl -D-fucopyranoside (p-NPF) 21.0 (47.4) 12.8 (28.3) 13.4 (22.0) 17.1 (23.8) 15.1 (20.0)
U
o-NPG
/U
e
lactose
1.07 0.26 0.20 0.23 0.20
a
All data were taken at 40

C at pH 6.0.
b
One enzyme unit for each substrate (lactose analogs) was dened as the amount of enzyme required to produce 1 mmol of o-NP or 1 mmol of glucose in 1 min at 40

C
at pH 6.0, in a manner similar to the enzyme unit using o-NPG as substrate (U
o-NPG
).
c
An enzyme complex composed of 21-kDa and 70-kDa proteins obtained by tryptic treatment.
d
Value in parenthesis indicates relative % activity taking the activity towards lactose as 100%.
e
The ratio of the activity towards o-NPG to that towards lactose.
0
400
800
1,200
-0.2 0.0 0.2 0.4 0.6 0.8
1
/
v

(
m
i
n
/
m
M
)
D
1/[S] (1/mM)
o-NPG
1
/
v

(
m
i
n
/
m
M
)C
0
20
40
60
80
100
-0.1 0 0.1 0.2 0.3 0.4 0.5
1/[S] (1/mM)
Lactose
0.2 0.4 0.6 0.8 1 0
1/[S] (1/mM)
0
100
200
300
B
o-NPG
1
/
v

(
m
i
n
/
m
M
)
1/[S] (1/mM)
0
10
20
30
40
50
60
1
/
v

(
m
i
n
/
m
M
)
A
Lactose
-0.1 0 0.1 0.2 0.3 0.4 0.5
Fig. 3. LB Plots for the Hydrolysis of Lactose and o-NPG Catalyzed by -Gal-C (A, B) and for the Hydrolysis of Lactose and o-NPG Catalyzed by
-Gal-A (C, D) at 40

C at pH 6.0.
The enzyme concentrations in Fig. A, B, C, and D were 27.0, 7.4, 6.0, and 0.2 nM respectively. The solid lines in Fig. 4A and B show the
results calculated by Eq. (2) using the values of the parameters determined.
-Galactosidase from Bacillus circulans 273
Treatment of -Gal-A with endogenous protease
isolated
The specic activity of the endogenous protease
prepared was approximately 140 U/mg. Figure 5 shows
the results of SDSPAGE for -Gal-A treated with the
protease solution at 37

C at pH 7.0. The lane 1 shows

the results for the sample obtained at start of the
incubation of the mixture of -Gal-A and protease. With
increasing incubation periods, -Gal-A was degraded
into plural smaller proteins, including -Gal-B and -
Gal-C, judging from their molecular sizes via a protein
with a molecular mass of approximately 170 kDa,
slightly higher than -Gal-B.
Tryptic treatment and purication of truncated pro-
teins
When treated with trypsin, -Gal-A underwent a
gradual degradation, as evidenced by SDSPAGE, and
its hydrolytic activities towards lactose and o-NPG
gradually decreased (data not shown). After a 10-h
incubation at 40

C, the activities towards lactose and o-

NPG decreased to 70% and 20% of that before
incubation respectively. SDSPAGE analysis showed
that two major bands, corresponding to approximately
70 kDa and 20 kDa, were generated (data not shown).
Even after a 12-h incubation at increased trypsin
concentrations, the two bands remained, and no appre-
ciable degradation was observed. In addition, the two
bands were also detected when -Gal-B, -Gal-C, and
-Gal-D were treated with trypsin under the same
conditions as for -Gal-A. Hence we puried the
proteins corresponding to the two bands by the method
described in Materials and Methods. The active
fraction showing -galactosidase activity that eluted
from the gel chromatographic column was collected and
subjected to SDSPAGE after pre-treatment of the
fraction at 20, 40, and 60

C in an SDS sample buer.

Figure 6A shows the results of SDSPAGE analysis of
the active fractions treated at 20

C (lane 1), 40

C
(lane 2), and 60

C (lane 3) for 5 min. The samples pre-

treated at 40

C and 60

C showed two bands, in

agreement with the SDSPAGE results for a sample
treated at 100

C for 5 min. On the other hand, the

sample incubated at 20

C showed a single major band

with a molecular mass of approximately 90 kDa
(lane 1). Furthermore, on active staining with MUAG,
this 90-kDa band emitted light under UV as a result
of the formation of MU that had been hydrolyzed by
the 70-kDa protein (lane 1 in Fig. 6B), while the bands
in lanes 2 and 3 did not. MALDI/TOF MS analysis
indicated the presence of three peaks, two major peaks
0
1
2
0 20 40 60 80 100
C
Conversion of lactose (%)
C
o
n
c
e
n
t
r
a
t
i
o
n

(
%
)
0
1
2
Conversion of lactose (%)
D
C
o
n
c
e
n
t
r
a
t
i
o
n

(
%
)
0 20 40 60 80 100
A
0
1
2
0 20 40 60 80
C
o
n
c
e
n
t
r
a
t
i
o
n

(
%
)
Conversion of lactose (%)
B
Conversion of lactose (%)
0
1
2
C
o
n
c
e
n
t
r
a
t
i
o
n

(
%
)
0 20 40 60 80 100
Fig. 4. Relationships between the Conversion of Lactose and Saccharides Produced with -Gal-A (A), -Gal-B (B), -Gal-C (C), and -Gal-D (D)
at 40

C at pH 6.0.
Glucose ( ) and galactose ( ) were assayed enzymatically. Disaccharide other than lactose ( ), trisaccharide ( ), tetrasaccharide ( ), and
pentasaccharide ( ) were analyzed by HPLC.
-Gal-A
-Gal-B
-Gal-C
-Gal-D
(kDa)
250
150
100
75
50
1 2 3 4 5
Fig. 5. Results of SDSPAGE for -Gal-A Treated with Endogenous
Protease for Various Incubation Periods at 37

C at pH 7.0.
Lane 1, marker proteins; lane 2, the sample (-Gal-A and
protease) at the start of incubation; lane 3, -Gal-A treated for 1 d;
lane 4, -Gal-A treated for 2 d; lane 5, -Gal-A treated for 7 d.
274 J. SONG et al.
of 21.034 kDa (21-kDa protein) and 70.273 kDa
(70-kDa protein), along with a small peak of
91.211 kDa. These ndings indicate that the 21-kDa
and 70-kDa proteins form an enzyme complex with a
molecular mass of approximately 91 kDa (-Gal-91) that
shows -galactosidase activity. The N-terminal amino
acid sequences of the 21-kDa and 70-kDa proteins were
determined to be GNSVSYDGERRVNFNEN and
EDRADVNIKTKISND respectively. The N-terminal
amino acid sequence of the 21-kDa protein was identical
to those of -Gal-A, -Gal-B, -Gal-C, and -Gal-D.
Puried -Gal-91, consisting of 21-kDa and 70-kDa
proteins showing -galactosidase activity, was charac-
terized by the same methods as for -Gal-A, -Gal-B,
-Gal-C, and -Gal-D. All of the biochemical character-
istics determined for -Gal-91, such as optimum
reaction pH, pH stability, substrate specicity, and
thermal stability, were similar to those for -Gal-B, -
Gal-C, and -Gal-D, as shown in Tables 13 and Fig. 2.
The LB plots for the hydrolysis of lactose and o-NPG
showed biphasic LB behavior. The apparent values of
the kinetic parameters determined are shown in Table 4.
The amount of GOS produced by -Gal-91 was much
higher than that by -Gal-A, and was similar to those by
-Gal-B, -Gal-C, and -Gal-D, as shown in Fig. 7. In
particular, disaccharide other than lactose was produced
in large amounts by -Gal-91 in the latter stage of
reaction in a way similar to -Gal-D, which has a
molecular mass similar to -Gal-91.
Discussion
Western blot analysis using an anti--Gal-A antibody
indicated that a commercial product derived from
B. circulans ATCC 3182 contains multiple forms of
-galactosidases that have dierent molecular sizes, but
are immunologically homologous. Among the immu-
nologically homologous enzymes, we puried four
enzymes, with molecular weights determined by
MALDI/TOF MS as follows: -Gal-A (189.283 kDa),
-Gal-B (153.932 kDa), -Gal-C (134.788 kDa), and -
Gal-D (91.627 kDa). Using the puried enzyme prep-
arations, we investigated the biochemical characteristics
of these enzymes in order to obtain information on the
causes of the presence of multiple forms in Biolacta

and dierences in the GOS productivity among the

enzymes. The N-terminal amino acid sequence of the
21-kDa peptide region of -Gal-91 obtained by tryptic
treatment as well as those of -Gal-A, -Gal-B, -Gal-
C, and -Gal-D were identical. -Gal-91 showed -
galactosidase activity, while a 70-kDa protein that was
generated by deletion of the N-terminal 21-kDa peptide
region of -Gal-A was inactive. Thus the N-terminal
peptide region might be important either in the
maintenance of the conformation of the enzyme or in
the formation of the substrate binding site and/or active
site, although the details are not known at present.
Furthermore, we found that -Gal-A was degraded into
smaller proteins, including ones having molecular
0
1
2
0 20 40 60 80 90
C
o
n
c
e
n
t
r
a
t
i
o
n

(
%
)
Conversion of lactose (%)
Fig. 7. Relationships between the Conversion of Lactose and
Saccharides Produced with -Gal-91 at 40

C at pH 6.0.
The keys are the same as those shown in the legend to Fig. 4.
Table 4. Kinetic Parameters for the Various -Galactosidases from Biolacta
;a
Substrate Kinetic constant
b
-Gal-A -Gal-B -Gal-C -Gal-D -Gal-91
c
Lactose K
m,H
(mM) 16:6 1:0 69:3 16:8 66:3 15:2 65:7 20:6 61:3 16:8
K
m,L
(mM) 7:0 0:3 7:6 0:8 8:0 0:8 10:8 3:3 10:7 3:8
V
max,H
10
3
=[E]
d
t
(1/min) 10:2 3:9 10:5 0:9 9:6 1:3 7:6 0:8 9:0 0:8
V
max,L
10
3
=[E]
d
t
(1/min) 6:6 0:4 3:6 0:5 3:1 0:2 3:6 0:5 3:2 0:4
o-NPG K
m,H
(mM) 10:8 1:8 12:6 1:4 12:5 0:7 11:6 2:0
K
m,L
(mM) 5:2 0:2 1:0 0:1 1:3 0:2 1:7 0:1 1:2 0:3
V
max,H
10
3
=[E]
d
t
(1/min) 3:7 0:42 4:6 1:1 4:4 1:6 3:1 0:1
V
max,L
10
3
=[E]
d
t
(1/min) 20:0 0:5 1:2 0:1 1:3 0:12 1:2 0:2 1:1 0:1
a
All values were determined at 40

C at pH 6.0.
b
Subscripts H and L in the kinetic constants indicate the values determined from the lines drawn through the LB plots of the high and low substrate regions
respectively.
c
An enzyme complex composed of 21-kDa and 70-kDa proteins obtained by tryptic treatment.
d
The molecular weight of the enzyme as determined by MALDI/TOF MS was used.
B
1 2 3
A
250
150
100
75
50
37
25
20
1 2 3
Marker
(kDa)
70-kDa
protein
21-kDa
protein
-Gal-91
Fig. 6. Results of SDSPAGE (A) and Active Staining (B) for
Puried -Gal-91.
A, Results of SDSPAGE. Lane 1, -Gal-91 treated at room
temperature; lane 2, -Gal-91 treated at 40

C; lane 3, -Gal-91
treated at 60

C; lane 4, protein markers. B, Results of active

staining of puried -Gal-91. Lane 1, -Gal-91 treated at room
temperature; lane 2, -Gal-91 treated at 40

C; lane 3, -Gal-91
treated at 60

C.
-Galactosidase from Bacillus circulans 275
masses similar to -Gal-B and -Gal-C, by incubation
with protease isolated from the culture broth of B.
circulans ATCC 31382 in a way similar to the tryptic
treatment of -Gal-A. However, the degradation process
of -Gal-A by the endogenous protease appear to be
more ecient than that by trypsin, judging from the
protein concentrations tested. The ndings obtained in
this study indicate that multiple forms of -galactosi-
dases of dierent molecular sizes are formed by the
action of the endogenous protease that is produced
during cultivation, which might truncate the C-terminal
peptide region of -Gal-A.
Although the biochemical characteristics for all of the
puried -galactosidases, including -Gal-91, were
similar, with only minor dierences, their activities
with respect to the biosynthesis of GOS were very
dierent. The productivity of GOS for -Gal-B, -Gal-
C, -Gal-D, and -Gal-91 was considerably higher than
that for -Gal-A, although some dierences in the yield
and composition of the GOS were found among the four
enzymes (Table 1). Jorgensen et al.
23)
found that the
transgalactosylation activity of recombinant -galacto-
sidase from B. bidum BIF3 expressed in E. coli cells
was greatly enhanced by deletion of 580 amino acid
residues from its C-terminus, including a putative
galactose-binding motif. In addition, the deletion mutant
of B. bidum BIF3 -galactosidase was found not to
require high lactose concentrations to produce higher
amounts of GOS, in a manner similar to -Gal-B, -Gal-
C, and -Gal-D. On the other hand, the specic activity
of the deletion mutant towards o-NPG decreased to 6%,
with a slightly lower thermal stability, as compared to
the native enzyme. Taking into consideration the
decreased thermal stability of the deletion mutant, they
suggested tentatively that one possible reason for the
increased transgalactosylation is the more open fragile
structure of the deletion mutant. However, they also
pointed out that some molecular mechanism that is not
known may be involved in the increased transgalacto-
sylation. On the other hand, the thermal stability of -
Gal-B, -Gal-C, and -Gal-D studied here was higher
than -Gal-A, judging from the remaining activity,
as determined at 60

C (Fig. 2), in contrast to the -

galactosidase from B. bidum BIF3. Ohdan et al.
61)
have
reported that -amylase (designated Ba-L) was degraded
by 28% of its primary structure to form a truncated
enzyme (Ba-S) by the action of endogenous protease
during the cultivation of B. subtilis X-23 cells. Although
the two forms of -amylase showed the same enzymatic
characteristics, the thermal stability of Ba-S was higher
than BA-L as with our results.
The kinetic properties of -Gal-B, -Gal-C, and -
Gal-D for the hydrolysis of lactose and o-NPG were
dierent from those of -Gal-A. The LB plots for -Gal-
B, -Gal-C, and -Gal-D with lactose and o-NPG as
substrate, showed biphasic behavior, while those for -
Gal-A showed a weak or non-biphasic plot. The biphasic
LB plot has been frequently observed for hydrolytic
reactions of disaccharides and glycosides catalyzed
by retaining glycosidases, such as -glucosidase and
-galactosidase,
19,59,60,62)
and is explained by the occur-
rence of the transglycosylation reaction at higher
substrate concentrations. The LB plot obtained for the
-Gal-A-catalyzed hydrolysis of o-NPG might be
ascribed to its lower transgalactosylation activity and
in addition to a low solubility of o-NPAG in the solution
(about 32 mM), which made it impossible to conduct the
reaction at higher substrate concentrations.
As reported in Results above, we determined
tentatively two sets of maximum velocities and Michaelis
constants from biphasic LB plots (Table 4). Here, we
analyzed the biphasic LB plot according to the scheme
shown in Fig. 8,
60,62)
which involves both the hydrolysis
and transgalactosylation reactions. Based on the scheme
shown in Fig. 8, the initial reaction rate for hydrolyses
of lactose and o-NPG with transgalactosylation reaction
is derived as follows:
v =
k
2
[E]
t
[S](k
w
k
t
[S])=(k
2
k
w
k
t
[S])
K
d
(k
w
k
t
[S])=(k
2
k
w
k
t
[S]) [S]
(1)
where [E]
t
is the total enzyme concentration, [S] the
substrate concentration, K
d
the dissociation constant
of the substrate, and k
2
, k
w
, and k
t
are the rate constants
(Fig. 8). Here, it should be noted that k
w
value contains
the contribution of water.
Equation (1) is rearranged as Eq. (2) as follows:
1
v
=
1
k
2
[E]
t
K
d
[S]
1

[S]
1
(1=[S] )

(2)
where = k
t
=k
w
and = k
2
=k
w
.
The k
2
[E]
t
value in Eq. (2) was set to the 1=v value
obtained by extrapolating the 1=[S] value to 0. The K
d
, ,
and values were determined by nonlinear least-squares
t of the experimental data of 1=v versus 1=[S] for the
-Gal-C-catalyzed hydrolyses of lactose (Fig. 3A), and
o-NPG (Fig. 3B) to Eq. (2) using KaleidaGraph (ver.
3.6.3, Synergy Software Systems, Dubai, UAE). The
determined values of k
2
[E]
t
, K
d
, , and are listed in
Table 5. The calculated LB plots using the values
determined for -Gal-A-catalyzed hydrolyses of lactose
and o-NPG are shown in solid lines in Fig. 3A and B
respectively, and were in good agreement with the
E, Free enzyme; Gal-OR, substrate (Lactose or o-NPG); HOR, glucose or nitrophenol
E Gal, galactosyl enzyme; Gal, galactose; Gal-Gal-OR, tri-saccharide or galactosyl o-NPG
K
d
, dissociation constant; k
2
, k
w
,k
t
, rate constant
E + Gal-OR E Gal-OR E Gal
HOR
Gal
Gal-Gal-OR
k
w
k
2
K
d
k
t
Water
Substrate
Fig. 8. Reaction Scheme for -Galactosidase-Catalyzed Hydrolysis of Lactose and o-NPG with Transgalactosylation.
276 J. SONG et al.
experimental data, which indicates the validity of the
scheme shown in Fig. 8. The fact that the determined K
d
value for o-NPG was lower than that for lactose indicates
the higher binding anity of o-NPG to the enzyme than
that of lactose. Furthermore, the (=k
t
=k
w
) value for
o-NPG was slightly higher than that for lactose. Hence, it
is possible that o-NPG is a better galactosyl acceptor
than lactose for B. circulans -galactosidase, although
structural analysis of the enzyme is needed to clarify the
details of the molecular mechanism.
The activities of -Gal-A, -Gal-B, -Gal-C, and -
Gal-D were not increased by treatment of MgCl
2
, an
essential metal ion for E. coli -galactosidase.
58)
On the
other hand, EDTA inactivated -Gal-A to some extent,
although -Gal-B, -Gal-C, and -Gal-D were not
aected. This remains as a question for the future.
Ito et al.
21)
cloned a -galactosidase gene from
B. circulans ATCC 31382, the same strain as the
enzyme source used in this study, and sequenced it.
The cloned gene of the -galactosidase contains an ORF
of 1,758 bp, encoding 586 amino acids with an N-
terminal amino acid sequence of MSQLTYDDSF.
Furthermore, the -galactosidase isolated by Ito et al.
21)
cleaves the -1,3 galactoside linkage selectively. Thus it
is obvious that B. circulans ATCC 31382 encodes at
least two dierent -galactosidases. Furthermore, the
genes of two -galactosidases from B. circulans G1,
bgaA and bgaB, are available at GenBank under
accession nos. L03424 and L03425, which encode
enzymes that are also dierent from our enzymes.
In a future paper, we intend to report on the cloning of
the B. circulans -galactosidase gene and expression in
recombinant E. coli cells.
Acknowledgments
We thank Professor Hidenori Yamada and Dr. Mizuki
Kitamatsu of the Graduate School of Natural Science
and Technology, Okayama University, for help in
determining the N-terminal peptide sequences and
molecular masses of proteins.
References
1) Pomeranz Y, Food Technol., 18, 682687 (1964).
2) Wallenfels K and Wei R, The Enzyme Vol. VII, ed. Boyer
PD, Academic Press, New York, pp. 616663 (1972).
3) Michael LS, Chem. Rev., 90, 11711202 (1990).
4) Park A-R and Oh D-K, Appl. Microbiol. Biotechnol., 85, 1279
1286 (2010).
5) Gosling A, Stevens GW, Barber AR, Kentish SE, and Gras SL,
Food Chem., 121, 307318 (2010).
6) Richmond ML, Gray JI, and Stine CM, J. Dairy Sci., 64, 1739
1771 (1981).
7) Mlichova Z and Rosenberg M, J. Food Nutr. Res., 45, 4754
(2006).
8) Mahoney R, Food Chem., 63, 147154 (1998).
9) Yamahira S, Kobayashi T, Suguri T, Nakakoshi M, Miura S,
Ishikawa H, and Nakajima I, Biosci. Biotechnol. Biochem., 59,
10211026 (1995).
10) Sako T, Matsumoto K, and Tanaka R, Int. Dairy J., 9, 6980
(1999).
11) Macfarlane GT, Steed H, and Macfarlane S, J. Appl. Microbiol.,
104, 305344 (2008).
12) Nakanishi K, Matsuno R, Torii K, Yamamoto K, and Kamikubo
T, Enzyme Microb. Technol., 5, 115120 (1983).
13) Mozaar Z, Nakanishi K, Matsuno R, and Kamikubo T, Agric.
Biol. Chem., 48, 30533061 (1984).
14) van Laere KMJ, Abee T, Schols HA, Beldman G, and Voragen
AGJ, Appl. Environ. Microbiol., 66, 13791384 (2000).
15) Nguyen T-H, Splechtna B, Steinbck M, Kneifel W, Lettner HP,
Kulbe KD, and Haltrich D, J. Agric. Food Chem., 54, 4989
4998 (2006).
16) Splechtna B, Nguyen T-H, Steinbck M, Kulbe KD, Lorenz W,
and Haltrich D, J. Agric. Food Chem., 54, 49995006 (2006).
17) Onishi N, Yamashiro A, and Yokozeki K, Appl. Environ.
Microbiol., 61, 40224025 (1995).
18) Reuter S, Nygaard AR, and Zimmermann W, Enzyme Microb.
Technol., 25, 509516 (1999).
19) Petzelbauer I, Reiter A, Splechtna B, Kosma P, and Nidetzky B,
Eur. J. Biochem., 267, 50555066 (2000).
20) Lu L, Xiano N, Xu X, Li Z, and Li Y, Biochem. Biophys. Res.
Commun., 356, 7884 (2007).
21) Ito Y and Sasaki T, Biosci. Biotechnol. Biochem., 61, 1270
1276 (1997).
22) Moller PL, Jorgensen F, Hansen OC, Madsen SM, and
Stoutgaard P, Appl. Environ. Microbiol., 67, 22762283 (2001).
23) Jorgensen F, Hansen OC, and Stoutgaard P, Appl. Microbiol.
Biotechnol., 57, 647652 (2001).
24) Goulas TK, Goulas AK, Tzortzis G, and Gibson GR, Appl.
Microbiol. Biotechnol., 76, 13651372 (2007).
25) Goulas T, Goulas A, Tzortzis G, and Gibson GR, Appl.
Microbiol. Biotechnol., 82, 10791088 (2009).
26) Hung MN and Lee BH, Biotechnol. Lett., 20, 659662 (1998).
27) Hung M-N, Xia Z, Hu N-T, and Lee BH, Appl. Environ.
Microbiol., 67, 42564263 (2001).
28) Hung M-N and Lee BH, Appl. Microbiol. Biotechnol., 58, 439
445 (2002).
29) Halbmayr E, Mathiesen G, Nguyen T-H, Maischberger T,
Peterbauer CH, Eijsink VGH, and Haltricht D, J. Agric. Food
Chem., 56, 47104719 (2008).
30) Iqbal S, Nguyen T-H, Nguyen TT, Maischberger T, and Haltrich
D, Carbohydr. Res., 345, 14081416 (2010).
31) Campuzano S, Serra B, Llull D, Garcia JL, and Garcia P, Appl.
Environ. Microbiol., 75, 59725980 (2009).
32) Kitahata S, Fujita K, Takagi Y, Hara K, Hashimoto H, Tanimoto
T, and Koizumi K, Biosci. Biotechnol. Biochem., 56, 242245
(1992).
33) Usui T, Morimoto S, Hayakawa Y, Kawaguchi M, Murata T,
Matahira Y, and Nishida Y, Carbohydr. Res., 285, 2939 (1996).
34) Vetere A and Paoletti S, Biochem. Biophys. Res. Commun., 219,
613 (1996).
35) Murata T, Akimoto S, Horimoto M, and Usui T, Biosci.
Biotechnol. Biochem., 61, 11181120 (1997).
36) Yanahira S, Yabe Y, Nakakoshi M, Miura S, Matsubara N, and
Ishikawa G, Biosci. Biotechnol. Biochem., 62, 17911794
(1998).
37) Murata T, Kosugi M, Nakamura T, Urashima T, and Usui T,
Biosci. Biotechnol. Biochem., 65, 24562464 (2001).
38) Higashiyama T, Watanabe H, Aga H, Nishimoto T, Kubota M,
Fukuda S, Kurimoto M, and Tsujisaka Y, Carbohydr. Res., 339,
16031608 (2004).
39) Miyasato M and Ajisaka K, Biosci. Biotechnol. Biochem., 68,
20862090 (2004).
40) Shimizu R, Shimabayashi H, and Moriwaki M, Biosci.
Biotechnol. Biochem., 70, 940948 (2006).
41) Zheng X and Uzawa H, Carbohydr. Res., 340, 24692475 (2004).
42) Zheng X, Sun Y, Ye H, Liu J, and Uzawa H, Biotechnol. Lett.,
Table 5. Values of Kinetic Parameters Determined for -Gal-C-
Catalyzed Hydrolyses
Substrate
k
2
[E]
a
t
K
d
=k
t
/k
w
=k
2
/k
w
(mM/min) (mM) (mM
1
) ()
Lactose 0.25 20.0 0.13 5.3
o-NPG 0.05 5.6 0.25 8.7
a
The enzyme concentrations for hydrolysis of lactose and o-NPG were 27.0
and 7.4 nM, respectively.
-Galactosidase from Bacillus circulans 277
29, 11051110 (2007).
43) Mozaar Z, Nakanishi K, and Matsuno R, Appl. Microbiol.
Biotechnol., 25, 224228 (1986).
44) Boon MA, Jansen AEM, and vant Riet K, Enzyme Microb.
Technol., 26, 271281 (2000).
45) Vetere A and Paoletti S, Biochim. Biophys. Acta, 1380, 223231
(1998).
46) Fujimoto H, Miyasato M, Ito Y, Sasaki T, and Ajisaka K,
Glycoconj. J., 15, 155160 (1998).
47) Naundorf A, Caussette M, and Ajisaka K, Biosci. Biotechnol.
Biochem., 62, 13131317 (1998).
48) Ajisaka K and Miyasato M, Biosci. Biotechnol. Biochem., 64,
17431746 (2000).
49) Laemmli UK, Nature, 277, 680685 (1970).
50) Manchenko GP, Handbook of Detection of Enzymes on
Electrophoretic Gels 2nd ed., CRC LLC Press, Boca Raton
(2003).
51) Song J, Imanaka H, Imamura K, Kajitani K, and Nakanishi K,
J. Biosci. Bioeng., 110, 281287 (2010).
52) Shiga D, Nakane D, Inomata T, Masuda H, Oda M, Noda M,
Uchiyama S, Fukui K, Takano Y, Nakamura H, Mizuno T, and
Tanaka T, Biopolymers, 91, 907916 (2009).
53) Nishio M, Kamiya Y, Mizushima T, Wakatsuki S, Sasakawa H,
Yamamoto K, Uchiyama S, Noda M, McKay AR, Fukui K,
Hauri H-P, and Kato K, Proc. Natl. Acad. Sci. USA, 107, 4034
4039 (2010).
54) Schuck P, Biophys. J., 78, 16061619 (2000).
55) Zhang D, Koreishi M, Imanaka H, Imamura K, and Nakanishi
K, J. Biotechnol., 128, 788800 (2007).
56) Yano T, Ashida S, Tachiki T, Kumagai H, and Tochikura T,
Agric. Biol. Chem., 55, 379385 (1991).
57) Benjakul S, Visessanguan W, and Thummaratwasik P, J. Food
Biochem., 24, 107127 (2000).
58) Matthews BW, C. R. Biologies, 328, 549556 (2005).
59) Lind DL, Daniel RM, Cowan DA, and Morgan HW, Enzyme
Microb. Technol., 11, 180186 (1989).
60) Fourage L, Dion M, and Colas B, Glycoconj. J., 17, 377383
(2000).
61) Ohdan K, Kuriki T, Kaneko H, Shimada J, Takada T, Fujimoto
Z, Mizuno H, and Okada S, Appl. Environ. Microbiol., 65,
46524658 (1999).
62) Namchuk MN and Withers SG, Biochemistry, 34, 1619416202
(1995).
278 J. SONG et al.