Báo cáo khóa học: UDP-galactose 4-epimerase from Kluyveromyces fragilis Evidence for independent mutarotation site pdf

UDP-galactose 4-epimerase from
Kluyveromyces fragilis
Evidence for independent mutarotation site
Amrita Brahma and Debasish Bhattacharyya
Division of Drug Development, Design and Molecular Modeling, Indian Institute of Chemical Biology, Calcutta, India
UDP-galactose 4-epimerases from the yeast Kluyvero-
myces fragilis and Escherichia coli are both homodimers
but the molecular mass of the former (75 kDa/subunit) is
nearly double that of the latter (39 kDa/subunit). Protein
databank sequence homology revealed the possibility of
mutarotase activity in the excess mass of the yeast
enzyme. This was confirmed by three independent assay
protocols. With the help of specific inhibitors and chem-
ical modification reagents, the catalytic sites of epimerase
and mutarotase were shown to be distinct and inde-
pendent. Partial proteolysis with trypsin in the presence of
specific inhibitors, 5¢-UMP for epimerase and galactose
for mutarotase, protected the respective activities. Similar
digestion with double inhibitors cleaved the molecule into
two fragments of 45 and 30 kDa. After separation by
size-exclusion HPLC, they manifested exclusively epi-
merase and mutarotase activities, respectively. Epimerases
from Kluyveromyces lactis var lactis, Pachysolen tanno-
philus and Schizosaccharomyces pombi also showed asso-
ciated mutarotase activity distinct from the constitutively
formed mutarotase activity. Thus, the bifunctionality of
homodimeric yeast epimerases of 65–75 kDa/subunit
appears to be universal. In addition to the inducible
bifunctional epimerase/mutarotase, K. fragilis contained a
smaller constitutive monomeric mutarotase of % 35 kDa.
Keywords: bifunctional enzyme; domain separation; muta-

rotase; UDP-galactose 4-epimerase; yeast.
UDP-galactose 4-epimerase (hereafter called epimerase),
which reversibly converts UDP-galactose into UDP-glu-
cose, is the first enzyme of the ÔLeloir pathwayÕ of galactose
metabolism [1,2]. Clinically, this is related to the disease
ÔgalactosemiaÕ [3]. It belongs to the rare class of enzymes that
utilizes noncovalently bound NAD as cofactor through the
transient formation of enzyme-bound NADH (class II
oxidoreductase). This is unlike classical dehydrogenases
where NAD acts as cosubstrate (class I oxidoreductase).
Although epimerase is ubiquitously present from bacteria to
mammals, its quaternary structure, size and NAD require-
ment vary. The bacterial and yeast enzymes are homo-
dimers with bound NAD, whereas, mammalian enzymes
are monomeric and require extraneous NAD. The X-ray
crystallographic structures of human and Escherichia coli
epimerases have been determined at high resolution [4–6].
An intriguing fact of epimerase biochemistry is the
significant difference in the size of protein isolated from
different sources even though their mechanisms of action
are the same. Whereas the molecular mass of the yeast
enzyme varies between 65 and 75 kDa/subunit (homo-
dimeric), those of E. coli and mammalian systems are
39 kDa/subunit (homodimeric) and 40 kDa (monomeric),
respectively. With the availability of gene sequencing data
and the development of the amino-acid sequence homology
search facility of the Data Bank, it was possible to compare
these enzymes with nonrelated proteins [7,8]. This revealed
that all epimerases contain a conserved ÔRossman foldÕ
sequence identified as the NAD-binding site at the extreme

N-terminus; the E. coli enzyme has strong sequence homo-
logy with the yeast enzyme constituting the N-terminal
half, and the C-terminal part of the yeast enzymes bears
homology with mutarotase (Scheme 1, where N and C
represent the N-terminus and C-terminus) [9]. Mere
sequence homology, however, does not predict manifesta-
tion of enzyme activity.
Mutarotase is another ubiquitous enzyme found in
organisms from microbes to mammals and of molecular
mass 34–38 kDa, with 10-fold variation in specific activity
[10,11]. This enzyme is also known in yeast and is of
comparable size [12,13]. Thus Ôepimerase associated muta-
rotaseÕ, if that exists, should be a second one. A preliminary
report on epimerase from Saccharomyces cerevisiae supports
this hypothesis [14]. In continuation of our studies on the
epimerase of Kluyveromyces fragilis [15–18], we examined its
bifunctionality. The results are presented here. Further, we
show that epimerases of comparable size from other yeast
strains are also associated with mutarotase activity.
Materials and methods
Reagents
1,2-Cyclohexanedione, Gly-Gly, hydroxylamine-HCl,
p-chloromercuribenzoate, UDP-Gal, UDP-Glu, 5¢-UMP,
Correspondence to D. Bhattacharyya, Division of Drug Development,
Design and Molecular Modeling, Indian Institute of Chemical
Biology, 4, Raja S.C. Mallick Road, Jadavpur, Calcutta 700032,
India. Fax: 91 33 2473 5197/0284,
E-mail:
(Received 13 September 2003, revised 16 October 2003,
accepted 28 October 2003)

Eur. J. Biochem. 271, 58–68 (2004) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03902.x
protein molecular mass markers (for SDS/PAGE and
HPLC), all sugars and their derivatives were from Sigma.
5,5¢-dithiobis-(2-nitrobenzoic acid) was from Pierce, and
diethylpyrocarbonate was from Aldrich. Glucose dehydro-
genase, glucose oxidase, horseradish peroxidase and trypsin
were from SRL, Mumbai, India. Hydroxyapatite was
synthesized as described [19]. Yeast nitrogen base, peptone,
yeast extract, malt yeast extract and agar were from
HiMedia, Mumbai, India. Urea was recrystallized from
hot ethanol to remove cyanate contamination, if any.
Yeast strains
Kluyveromyces fragilis (renamed Kluyveromyces marxianus
var marxianus, ATCC strain no. 10022), Kluyveromyces
lactis var lactis and Pachysolen tannophilus were purchased
from Microbial Type Collection Center and Gene Bank,
IMTECH, Chandigarh, India. Yeast Gal 10 mutant was
from ATCC (Manassas, VA, USA; strain no. 204836).
Schizosaccharomyces pombi was a gift from D. J. Chattyo-
padhyay, University of Calcutta.
Growth of cells
Liquid media containing 0.3% yeast extract, 0.3% malt
extract, 0.5% peptone and 3% galactose were used. Cells
were grown for 16 h at 30 °C under aerobic conditions with
shaking until the turbidity reached 0.4 at 650 nm after
10-fold dilution with water. To induce epimerase, cells were
grown in the presence of galactose as the only carbon
source. Glucose was used instead of galactose for the Gal 10
mutant, which was devoid of the epimerase gene, and also
for K. fragilis where induction of epimerase was not sought.

Purification of enzymes
Epimerase from K. fragilis. Yeastcells(10gwetweight),
suspended in 20 m
M
potassium phosphate, pH 7.0, were
subjected to toluenization under shaking in two steps: with
15% (v/v) toluene at 30 °C for 90 min followed by dilution
to 4% toluene at 4 °C for 16 h. All subsequent steps were
carried out at 4 °C. Partially lysed cells were centrifuged at
16 500 g for 30 min to remove debris. Ammonium sulfate
was added to the supernatant up to 55% saturation, and the
precipitate after dialysis was slowly stirred with hydroxy-
apatite (5 g wet weight) where epimerase remained unab-
sorbed. After recovery of the enzyme, it was passed through
a DEAE-cellulose column (6.0 · 0.5 cm) equilibrated with
20 m
M
potassium phosphate, pH 8.0. After a complete
wash, the epimerase was eluted with the same buffer
containing 0.5
M
NaCl. The eluted enzyme in 500 lL
aliquots was passed through Centricon Nylon Membrane
filters (Millipore; cut off limit 100 kDa) under 200 g for
15 min to remove trace amounts of low molecular mass
contaminants. It was stored in aliquots in 20 m
M
potassium
phosphate, pH 8.0, containing 0.5
M

NaCl at )20 °C.
Partial purification of epimerase from other yeast strains.
Yeast cells (5 g wet wt) were subjected to toluenization and
centrifugation as described above. Ammonium sulfate was
added up to 80% saturation to the supernatant. The
precipitate, after dialysis against 20 m
M
potassium phos-
phate, pH 7.4, was fractionated using a precalibrated
Sephadex G-200 column (118.0 · 0.5 cm), equilibrated
with the same buffer, at a flow rate of 15 mLÆh
)1
.Fraction
size was 1.5 mL. Elution was followed in terms of epimerase
and mutarotase activities.
Purification of mutarotase from capsicum (bell or green
pepper, Capsicum frutescens). The enzyme was purified
from fresh capsicums up to the dialysis step as described
in [11].
Purification of UDP-glucose dehydrogenase. This was
partially purified from bovine liver up to the heat denatur-
ation step as described in [20] when the preparation was free
of contaminating epimerase activity.
Assay of enzymes
Epimerase. Continuous conversion of UDP-Gal into UDP-
Glu by epimerase was monitored spectrophotometrically at
340 nm in the presence of the coupling enzyme, UDP-
glucose dehydrogenase, and NAD at 25 °C [21]. In brief,
22.5 m
M

UDP-Gal, 12 m
M
NAD and 20 m
M
UDP-glucose
dehydrogenase together with 500 lL0.2
M
Gly-Gly,
pH 8.8, were made up to 1 mL with water. Epimerase
was then added. One unit of enzyme was defined as the
amount that converts 1 lmol UDP-Gal into UDP-Glu per
min under standard assay conditions.
Mutarotase. Conversion of a-
D
(+) to b-
D
(+) Glu by
mutarotase was followed by three independent assay
protocols [10,22]. In all presentations, the spontaneous rate
of conversion served as control. One unit of mutarotase
activity corresponds to the conversion of 1 lmol of a-Glu to
b-Glu per min at 25 °C, pH 7.4.
Polarographically. The enhanced rate of change of specific
rotation for the said conversion by mutarotase was followed
spectropolarimetrically. The sugar (28 m
M
)wasdissolvedin
ice-cold 1 m
M
Tris/HCl, pH 7.4, immediately before addi-

tion of the enzyme, and the reaction rate for 10 min was
recorded. The first-order rate constant was obtained from
theslopeofthestraight-lineplot:
ln½ða
0
À a
e
Þ=ða
t
À a
e
Þ ¼ Kt ð1Þ
where K is the rate constant, a
0
, a
t
and a
e
are the observed
angular rotations at time zero, t, and equilibrium, respect-
ively [10].
Scheme 1. Alignment of peptides indicating amino acid sequence
homology.
Ó FEBS 2003 Mutarotase activity of epimerase (Eur. J. Biochem. 271)59
With glucose oxidase. The same conversion was followed
using glucose oxidase in 0.1
M
potassium phosphate,
pH 6.0, which specifically converted the b-anomer into
glucuronic acid with quantitative liberation of hydrogen

peroxide. The latter was spectrophotometrically estimated
at 460 nm with horseradish peroxidase in the presence of
1% o-dianisidine [10].
With glucose dehydrogenase. The anomer conversion was
also followed using the NAD-dependent glucose dehydro-
genase reaction, which specifically converts the b-form into
glucuronic acid [22]. The reaction was carried out in 5 m
M
Tris/HCl, pH 7.4, containing 20 m
M
NAD and 2.8 m
M
substrate. Formation of NADH was followed spectro-
photometrically taking e
M
340nm
¼ 6.3 · 10
)3
M
)1
Æcm
)1
.For
convenience and reliability, mutarotase was assayed using
glucose dehydrogenase, unless otherwise stated.
Modification reactions
Epimerase (0.5 mgÆmL
)1
) was incubated with 1.25 m
M

5¢-UMP and 100 m
ML
-arabinose at 25 °C in the presence
of 50 m
M
sodium phosphate, pH 7.5, to reduce NAD to
NADH (ÔReductive inhibitionÕ) [23]. For dissociation of the
dimeric structure, epimerase (0.05 mgÆmL
)1
) was dialyzed
against low-salt buffer (1 m
M
potassium phosphate,
pH 7.5) at 4 °C for 16 h [24]. Cysteine residues were
modified by allowing epimerase (0.5 mgÆmL
)1
) to react with
20 m
M
p-chloromercuribenzoate in the presence of 100 m
M
Tris/HCl, pH 7.4, at 25 °C for 2 h [17,25]. Alternatively,
the enzyme (0.25 mgÆmL
)1
)wastreatedwith0.2m
M
5,5¢-dithiobis-(2-nitrobenzoic acid) in the presence of
20 m
M
sodium phosphate, pH 7.0 [26]. Histidine residues

were modified by treating the enzyme (0.5 mgÆmL
)1
)with
0.25 m
M
diethylpyrocarbonate in 20 m
M
sodium phos-
phate, pH 7.5, at 20 °C. Modification of histidine was
reversed by treatment with hydroxylamine/HCl [27]. Argi-
nine residues were modified after reaction of epimerase
(0.5 mgÆmL
)1
) with 2.0 m
M
1,2-cyclohexanedione in 0.2
M
sodium borate, pH 9.0, for 3 h at 37 °C [28]. Except for the
p-chloromercuribenzoate reaction, all modifications were
monitored spectrophotometrically.
Reversible folding
Epimerase (1.0 mgÆmL
)1
) was denatured with 8
M
urea in
20 m
M
sodium phosphate, pH 7.5, containing 2 m
M

2-mercaptoethanol at 25 °C for 10 min. Under these
conditions, the molecule is known to be denatured and
inactivated, with dissociation of the constituent molecules.
Refolding/re-activation was initiated by 20-fold dilution of
the denaturant with the same buffer in the presence or
absence of 1 m
M
extraneous NAD at 25 °C [15,16].
Partial proteolysis
Epimerase (1.0 mgÆmL
)1
)wastreatedwithtrypsin(50:1;
w/w) in 20 m
M
potassium phosphate, pH 8.0, in the
presence of 2.5 m
M
5¢-UMP or 6.7 m
MD
(+)-galactose or
both at 4 °C for 4 h. The reaction was followed in terms of
epimerase and mutarotase activities, and the digest was
analyzed by SDS/PAGE (15% gel) and HPLC.
HPLC
The purity of epimerase was determined by using a Waters
Protein Pak300 size-exclusion HPLC column (fractionation
range 20–300 kDa). The column was equilibrated with
20 m
M
sodium phosphate, pH 7.0, containing 0.2

M
NaCl
or the same buffer containing 8
M
urea at a flow rate of
0.5 mLÆmin
)1
. It was then calibrated with the marker
proteins: alcohol dehydrogenase (150 kDa), BSA (66 kDa),
ovalbumin (43 kDa), lysozyme (14 kDa) and cytochrome c
(14.3 kDa).
To separate fragments of epimerase after partial trypsi-
nization, a Waters Protein Pak125 size-exclusion HPLC
column (fractionation range 5–80 kDa) was used. The
column was equilibrated with 20 m
M
sodium phosphate,
pH 7.0, and elution was followed at 280 nm at a flow rate of
0.5 mLÆmin
)1
. It was calibrated with the marker proteins:
BSA (66 kDa), ovalbumin (43 kDa), trypsin (21 kDa),
myoglobin (19 kDa) and cytochrome c (14.3 kDa). In both
cases of HPLC, linear dependence of log (molecular mass)
vs. elution volume was observed.
Other methods
PAGE was performed by following standard procedures,
and gels were stained with Coomassie Brilliant Blue RC-250
(Sigma) or silver nitrate. The following markers were used
in SDS/PAGE; phosphorylase B (92 kDa), BSA (66 kDa),

ovalbumin (43 kDa) and RNase A (10.5 kDa). Optical
absorbance was recorded with a SICO 200 UV-VIS (India)
or Analytik Jena Specord 200 (Germany) spectrophoto-
meter. A Jasco P 1020 spectropolarimeter was used to
measure specific rotation of sugars. Protein concentration
was determined with Bio-Rad Protein Assay Reagent
(catalog no. 10044) with BSA as reference.
Results
Purification of epimerase
Epimerase from K. fragilis has so far been purified using salt
fractionations [21]. The degree of purity achieved was
variable, and occasional contamination of proteases was
suspected. To remove these, a protocol has been developed
using ammonium sulfate fractionation, hydroxyapatite
treatment, DEAE-cellulose chromatography, and ÔfiltrationÕ
by Centricon. A 40-fold purification gave a homogeneous
preparation (Table 1). Epimerase after hydroxyapatite
treatment was found to be free from proteases, as the
SDS/PAGE profile of the fraction remained unchanged
after incubation at 37 °C for 6 h or at 4 °Cfor96hto
account for thermolabile proteases [29]. Hydroxyapatite is
known to effectively remove proteases from yeast cell
extracts [30]. The specific activity recovered was
72–75 U per mg protein compared with 70 UÆper mg
protein reported previously [15,17].
Demonstration of purity
This was verified by production of a single band on SDS/
PAGE and PAGE (10% gels) at pH 8.8, even after
overloading of the samples, after staining with Coomassie
60 A. Brahma and D. Bhattacharyya (Eur. J. Biochem. 271) Ó FEBS 2003

Blue or silver nitrate (Fig. 1, upper panel). On SDS/PAGE,
the molecular mass was found to be % 75 kDa with respect
to the markers, consistent with previous observations
[15,16].
The elution profile of the purified enzyme from Protein
Pak300 size-exclusion HPLC showed a single sharp
and symmetrical peak with retention time,
R
t
¼ 12.5 ± 0.1 min under conditions stated above. The
profiles were indistinguishable when followed at 280 or
220 nm, indicating the absence of contaminating proteins
or peptides of different sizes and probably any bio-organic
compounds. To ensure that no protein was coeluted,
epimerase was pretreated with 8
M
urea to dissociate
adhering proteins. The denatured protein emerged as a
single peak of R
t
¼ 11.25 ± 0.1 min (Fig. 1, lower panel,
A, B and C). The lower retention time of the unfolded
protein was due to expansion of the molecule in spite of
dissociation of the subunits [31].
The Waters 745B data module system used in HPLC, can
identify peaks of area abundance 0.01%. No peaks of such
low intensity appeared in the chromatograms even after
overloading of the samples. As the possibility of the
coexistence and copurification of two proteins of identical
molecular mass, charge and subunit composition at such an

advanced stage of purity was insignificant, homogeneity of
the preparation was confirmed.
Assay for mutarotase
Strong mutarotase activity associated with the purified
epimerase was demonstrated by three independent proce-
dures. In all cases, the mutarotase from capsicum served as
control.
Polarographically. The enhanced rate of change in the
specific rotation of a-
D
(+)-Glu over the spontaneous rate of
conversion was found to be dependent on epimerase
concentration (Fig. 2A). A linear dependence of log (specific
rotation) with time (up to 10 min) indicated first-order
dependence of the reaction (Fig. 2A, inset). Further, the
slopes of the straight lines were found to be linearly
dependent on epimerase concentration. When the same data
were plotted according to Eqn (1), linear dependence was
again observed for the initial 10 min. The derived specific
activity was 600–700 U per mg protein depending on the
batch. Whereas the specific activity of mammalian
mutarotase was 425–1500 U per mg protein [10], that of
Fig. 1. Demonstration of purity of epimerase. (A) SDS/PAGE (10% gel)
of 20 lg phosphorylase B (97 kDa) (lane 1) and 7 lg epimerase (lane 2).
(B)PAGE(10%gel)atpH8.8of10lg BSA (lane 1) and 10 lg
epimerase (lane 2). (C) Elution profiles of epimerase from Protein
Pak300 size-exclusion HPLC column: (a) 5 lg, monitored at 280 nm,
(b) 0.25 lg, monitored at 220 nm and (c) 5 lg after equilibration of the
column and the sample with 8
M

urea in the presence of the buffer.
Table 1. Purification of epimerase from K. fragilis.
Fraction
Total
activity
(units)
Total
protein
(mg)
Specific
activity
(units/mg)
Fold
purity
Crude 2110 1125 1.87 –
55% (NH
4
)
2
SO
4
precipitation
627 163 3.86 2.1
Hydroxyapatite
treatment
550 12.5 44.0 23.5
DEAE-cellulose
chromatography
500 8.3 60.2 32.2
Centricon ÔfiltrationÕ 387 5.3 73.0 39.0

Ó FEBS 2003 Mutarotase activity of epimerase (Eur. J. Biochem. 271)61
capsicum was 150 U per mg protein [11]. Thus the low
activity of yeast mutarotase was comparable to its phylogenic
position.
With glucose oxidase. The enhanced rate of conversion of
a-
D
(+)-Glu by epimerase was observed using the coupling
enzyme glucose oxidase. The reactions followed linear
kinetics up to 2 min (Fig. 2B). The initial rates were found
to be related to epimerase concentration in a linear manner,
thereafter reaching a plateau (Fig. 2B, inset). The derived
specific activity was 450–550 U per mg protein.
With glucose dehydrogenase. Mutarotation was demon-
strated most conveniently with the coupling enzyme glucose
dehydrogenase. Under the assay conditions, the reaction
followed linear kinetics for at least 5 min with epimerase.
Also the rates were linearly dependent on enzyme concen-
tration (Fig. 3). The specific activity observed was 550–
650 U per mg protein. When the rates of conversion
(turnover) were compared with that of capsicum mutarotase
on a weight by weight basis, the dependencies were very
similar (Fig. 3, inset A). As the molecular mass of yeast
mutarotase is about fourfold higher than that of the
capsicum enzyme, its catalytic efficiency appears to be
higher by the same factor. A similar result was obtained
with the polarographic assay. Yeast mutarotase showed a
typical Michaelis–Menten relation with the substrate
a-
D

(+)-Glu, yielding a linear Lineweaver–Burk plot
(Fig.3,insetB,C).TheK
m
derived was 22.2 m
M
,compared
with 19 m
M
for the capsicum mutarotase.
The specific activity and purity data for the yeast enzyme
indicated that it was bifunctional.
Stability of mutarotase activity
The pH optima of mutarotase from both kidney cortex and
capsicum are broad, with maxima at % 7.4, and activity at
pH 4.0 and 8.0 was 70–72%. For yeast mutarotase, the pH
optimum was 7.5, with 60–70% of activity under those
conditions. When incubated at 30, 40, 50, and 60 °Cfor3 h,
residual activities of capsicum mutarotase were 100%,
100%, 90%, and 11%, respectively; for epimerase residual
activities were 85%, 30%, 10%, and 0.5% and for yeast
mutarotase they were 89%, 45%, 30%, and 0.8%, respect-
ively. Thus the yeast mutarotase was less stable than its
capsicum counterpart.
Substrate specificity
Mutarotases from different sources show significant sub-
strate specificity, with 60-fold variation of turnover [10,11].
D
(+)-Glu and
D
(+)-Gal are effective substrates for yeast

mutarotase, as are also standard substrates for most of the
mutarotases. Catalytic activity was lower with
D
(+)-fucose
and
D
(–)-fructose, which are poor substrates, if at all, for
capsicum or pig kidney mutarotase. The substitution or
removal of the equatorial OH on C2 is known to abolish all
substrate interactions. 2-Deoxy-
D
(+)-glucose and 2-deoxy-
D
(+)-galactose, which act as inhibitors for the kidney
enzyme, appeared to be poor substrates for yeast mutarotase.
Thus the substrate specificity of yeast mutarotase appeared
to be comparable to that of other mutarotases (Table 2).
Substrate specificity was determined polarographically in all
cases because of restricted use of coupling enzymes.
Interactions with inhibitors
The addition of a large number of sugars to the mutarotase
assay using
D
(+)-Glu as substrate often markedly reduces
Fig. 2. Demonstration of mutarotase activity by epimerase observed
using polarography and the coupling enzyme galactose oxidase. (A) Time
kinetics of specific rotation of a-
D
(+)-glucose (28 m
M

) in the presence
of epimerase: (d) nil (spontaneous hydrolysis); (s)33n
M
;(n)66n
M
;
(h)99n
M
. As the initial rotation of 111 ° could not be maintained
exactly because of manual mixing of the reagents, the results were
normalized. Inset: First-order rate dependency of these reactions for
the initial 10 min. Symbols were the same as stated. Derived rates were
0.043, 0.095, and 0.152 min
)1
. (B) Time kinetics of the same mutaro-
tase assay using the coupling enzyme galactose oxidase. The initial rate
of conversion for 2 min are presented with 33 n
M
(h), 57 n
M
(n), and
75 n
M
(s) epimerase. Inset: Dependence of reaction rate on enzyme
concentration. The plateau was apparently due to limitation of the
coupling enzyme in the assay. An enzyme concentration of 1 lgÆmL
)1
corresponds to 6.66 n
M
.

62 A. Brahma and D. Bhattacharyya (Eur. J. Biochem. 271) Ó FEBS 2003
enzyme-catalyzed mutarotation in a competitive fashion,
although some of these sugars themselves are substrates
[10,11]. Of the five inhibitors tested,
D
(+)-Gal,
L
-arabinose
and
D
-fructose were found to be effective in inhibiting yeast
mutarotase to varying degrees, while
L
-fucose was ineffective.
L
-Deoxyglucose could not be tested because of its interfer-
ence with the coupling enzyme glucose dehydrogenase. This
inhibition pattern was similar to that of capsicum muta-
rotase. None of these sugars was inhibitory to epimerase.
Inhibitory effects of different nucleotides on epimerase but
not on mutarotase was also observed (Table 3).
Modification reactions
The architecture of the functional site of epimerase has been
mapped in detail, and involvement of several amino acids,
e.g. cysteines [17,26,32], histidine [33] and arginine [34], has
been suggested. Apart from these, Ôreductive inhibitionÕ was
performed to reduce the surface NAD to NADH with
inactivation [23]. Also dialysis of epimerase against low-salt
buffer led to spontaneous dissociation to monomer with
irreversible inactivation [24]. In the case of mutarotase, only

histidine residues are known to be involved in catalysis [35].
These experiments were repeated to verify the effects on
yeast mutarotase.
During Ôreductive inhibitionÕ, epimerase was inactivated
by 4 h (residual activity 1 ± 1%; k ¼ 0.0114 min
)1
). Low-
salt incubation also led to elimination of residual epimerase
activity. In all chemical modification reactions, epimerase
was inactivated to 0–5% of residual activity by 30–60 min in
a time-dependent manner. The k derived for the p-chloro-
mercuribenzoate, 5,5¢-dithiobis-(2-nitrobenzoic acid) and
diethylpyrocarbonate reactions were 0.035, 0.025 and
0.022 min
)1
, respectively. Upon hydroxylamine/HCl treat-
ment after diethylpyrocarbonate modification, 92% of
activity was recovered. Under identical conditions, yeast
and capsicum mutarotase resisted inactivation to the extent
of 85–100%, except in the case of diethylpyrocarbonate
where complete inactivation occurred. The extent of
reversible re-activation by hydroxylamine/HCl treatment
Fig. 3. Demonstration of mutarotase activity
by epimerase observed using the coupling
enzyme glucose dehydrogenase. Enhanced rates
of mutarotation of a-
D
(+)-glucose in the
presence of 33 n
M

(n), 66 n
M
(h), and 99 n
M
(s) epimerase for the initial 4.5 min are
shown. The relative rates were 0.9, 1.7 and
2.74, respectively. Inset A: Rate of the same
reaction expressed with an equal amount (wt/
wt) of yeast epimerase (s) and capsicum
mutarotase (d). Inset B: Dependence of the
reaction rate on substrate concentration at an
epimerase concentration of 200 n
M
.InsetC:
Lineweaver–Burk plot of the same reaction.
Table 2. Substrate specificity of mutarotase. The concentrations of the
sugars and deoxy-sugars were 28 and 30 m
M
, respectively. The con-
centration of capsicum mutarotase was 100 n
M
. Reactions were fol-
lowed polarimetrically, and the initial rates, where first-order kinetics
were observed, were used.
Substrate
Activity (U per mg enzyme)
Mutarotase
(associated with
epimerase)
Capsicum

mutarotase
D
(+)-Glucose 600 150
D
(+)-Galactose 300 260
D
(+)-Fucose 50 Not a substrate
D
())-Fucose 30 Not a substrate
2-Deoxy-
D
(+)-galactose 6 Not a substrate
2-Deoxy-
D
(+)-glucose 7 Not a substrate
Table 3. Inhibitory effects of different sugars and nucleotides on muta-
rotase and epimerase activities. Concentrations of the substrate
D
(+)-glucose, sugar inhibitors and nucleotide inhibitors were 5.6, 5.5,
and 5.0 m
M
, respectively. ND, not determined because of interference
with coupling enzyme. Capsicum mutarotase assay was performed
with glucose dehydrogenase as coupling enzyme.
Inhibitor
% Inhibition
Mutarotase
(associated with
epimerase)
Capsicum

mutarotase
a
Epimerase
D
-Galactose 33 40 n.d.
D
-Fructose 22 25 0
L
-Fucose 1.5 2 1
L
-Arabinose 90 61 0
L
-Deoxyglucose n.d. 0 0
UMP 0 0 44
UDP 0 0 25
UTP 0 0 13
a
From ref [11].
Ó FEBS 2003 Mutarotase activity of epimerase (Eur. J. Biochem. 271)63
was 90%. Similar results were observed for other mutaro-
tases [35]. Thus cysteine and arginine residues could not be
involved in the catalytic site of yeast mutarotase. As
treatments with low-salt buffer and p-chloromercuribenzo-
ate led to dissociation of the molecule without affecting
mutarotase activity, it was certain that mutarotase func-
tionality did not require a dimeric structure. These results
have been summarized in Table 4.
Reversible refolding
The refolding and re-activation pattern of epimerase after
denaturation by 8

M
urea is known in detail [15,16,18],
although similar studies on mutarotase are still scarce. To
optimize the re-activation yield, refolding was performed at
200 l
M
epimerase with capsicum mutarotase serving as
control. An initial linear dependence of re-activation rate
was observed in all cases. Relative re-activation rates and
maximum recoveries were as follows: 1.1 and 82% for yeast
mutarotase, 3.2 and 88% for epimerase, and 6.8 and 91%
for capsicum mutarotase. In another set, refolding was
initiated in the absence of extraneous NAD. In this case, no
recovery for epimerase was observed without affecting the
recovery of mutarotase activity. This indicated that matur-
ation of the mutarotase site has no relation to formation of
the epimerase site (results not shown).
Partial proteolysis with trypsin
Epimerase was sensitive to trypsin as it was degraded to
small peptides without accumulation of a stable intermedi-
ate [36]. However, when trypsinized in the presence of
2.5 m
M
5¢-UMP, its mutarotase activity was lost in a time-
dependent manner without affecting the epimerase activity
(Fig. 4A). SDS/PAGE (15% gel) of the digest showed the
disappearance of the original 75-kDa band with appearance
of a single 45-kDa fragment (result not shown). Proteolysis
under identical conditions in the presence of 6.7 m
M

galactose or fructose led to reversal of this protection;
epimerase activity was completely lost whereas mutarotase
activity was 90% protected (Fig. 4B). In this case, the
molecular mass of the reduced fragment was 30 kDa. Thus
it was apparent that the inhibitors were capable of
protecting the respective functional domains but not the
rest of the molecule. Finally, when epimerase was trypsi-
nized in the presence of the two inhibitors together,
Table 4. Inhibition of mutarotase and epimerase activities of yeast
enzyme after chemical modifications. Reaction conditions have been
described in the text. 1,2-CHD, 1,2-Cyclohexanedione; p-CMB,
p-chloromercuribenzoate; DTNB, 5,5¢-dithiobis-(2-nitrobenzoic acid);
DEPC, diethylpyrocarbonate.
Reactions
Residual activities (%)
Epimerase
Mutarotase
(associated
with epimerase)
Capsicum
mutarotase
ÔReductive inhibitionÕ 582 93
Incubation with
low salt buffer
0 90 100
p-CMB 0 95 95
DTNB 10 96 90
DEPC (re-activation by
hydroxylamine/HCl)
10 (92) 0 (91) 0 (90)

1,2-CHD 0 97 90
Fig. 4. Retention of epimerase and mutarotase activities after partial proteolysis with trypsin in the presence of inhibitors. Time-dependence of residual
activities of yeast enzyme: epimerase (d) and mutarotase (s) activities after partial proteolysis with trypsin in presence of (A) 2.5 m
M
5¢-UMP, (B)
6.7 m
M
galactose, or (C) the inhibitorstogether.(D)SDS/PAGE (15% gel)profileofthedouble inhibitor digestfor4(lane1), 8 (lane 2) and12h(lane3).
64 A. Brahma and D. Bhattacharyya (Eur. J. Biochem. 271) Ó FEBS 2003
protection of either of the activities was observed (Fig. 4C).
SDS/PAGE of the digest demonstrated the disappearance
of the parent molecule, with generation of the two fragments
of 45 and 30 kDa (Fig. 4D). As the catalytic efficiencies of
the 45-kDa and 30-kDa fragments (in terms of efficiency/
mol of catalytic site) remained within ± 5% with respect to
epimerase and mutarotase, respectively, an allosteric rela-
tion between the two sites did not appear to exist.
Separation of catalytic domains
The result shown in Fig. 4C suggest the existence of two
functionally independent domains in epimerase. Based on
the different molecular masses, their elution was followed
using a precalibrated Protein Pak125 size-exclusion HPLC
column. Whereas the native epimerase was eluted at the
void volume (R
t
¼ 5.88 ± 0.05 min), the 5¢-UMP-protec-
ted and galactose-protected fragments were eluted at
R
t
¼ 9.40 ± 0.05 and 10.42 ± 0.07 min, which corres-

pond to molecular masses of 43 and 28 kDa, respectively,
with respect to the molecular mass markers. The absence of
any detectable fraction at the void volume indicated
complete digestion of the parent molecule. HPLC of the
digest generated in presence of the two inhibitors led to
separation of two peaks of identical retention times as stated
above. The fractions eluted first and second expressed
exclusively epimerase and mutarotase activities, respectively
(Fig. 5A–D). Recovery of the epimerase and mutarotase
domains was 80 ± 5 and 90 ± 5%, respectively, in terms
of activity.
Existence of two yeast mutarotases
During induction of epimerase in K. fragilis, mutarotase
activity corresponding to 150 kDa but not 38 kDa was
expected to be increased. Thus K. fragilis was grown in
galactose (to induce epimerase) or glucose (to retain
epimerase at basal level) medium. Further, three other
strains, K. lactis var. lactis, P. tannophilus and S. pombi,
were grown in galactose medium. A Sephadex G-200
column was used to separate the two mutarotase activities
from the crude cell lysates. To confirm the absence of
mutarotase activity corresponding to epimerase but the
presence of its constitutively formed mutarotase, the Gal 10
mutant strain served as a control (Fig. 6).
In the case of the four wild-type yeast strains, the
chromatograms clearly separated two mutarotase activities
corresponding to 150 and 35 kDa. It further showed that
with induction of epimerase, parallel induction of mutaro-
tase corresponding to 150 kDa occurred (Fig. 6A,D,E). As
expected, growth of K. fragilis and Gal 10 mutant in glucose

medium did not show induction or manifestation of
mutarotase activities corresponding to a molecular mass
of 150 kDa (Fig. 6B,C). To reduce the possibility that the
lower molecular mass mutarotase was not derived from its
higher counterpart, a cocktail of protease inhibitors was
added at the time of lysis of the cells. The chromatographic
patterns thereby remained unaltered. Also, to rule out salt-
induced multimerization of the smaller mutarotase [24], one
set of extraction and gel filtration chromatography was
performed in 5 m
M
sodium phosphate, pH 7.5, with
K. fragilis strain. No difference was observed.
Discussion
Among epimerases, yeast and E. coli enzymes are well
studied. However, it was an enigma that, although an
identical mechanistic pathway and many enzymatic prop-
erties are shared, the size of the yeast enzyme is almost
double that of the E. coli. enzyme. Being extracellular and
devoid of cysteine bridges, synthesis of bigger molecules is
not warranted for stability unless specific requirements are
attributed.
Mehta [36] reported that, when trypsinized in the
presence of 5¢-UMP under specified conditions, the epi-
merase from K. fragilis was reduced to half its size
(homodimer, % 38 kDa/subunit) with retention of activity.
This indicated that nearly half of the molecule had no role in
epimerization. A plausible explanation subsequently came
from the sequence homology profiles where the C-terminal
part was suspected to be associated with mutarotase activity

(Scheme 1). This hypothesis has been verified by applying
three independent mutarotase assays to this enzyme (Figs 2
and 3) after assessing its physical homogeneity (Fig. 1). The
Fig. 5. Size-exclusion HPLC of epimerase after partial proteolysis with
trypsin. The digest was fractionated isocratically on a Waters Protein-
Pak125 column and was followed at 280 nm. (A) Elution profile of
catalase (240 kDa) served to determine the void volume (V
o
). (B–D)
Elution profiles of epimerase treated with trypsin in the presence of
5¢-UMP, galactose, or both, respectively. Proteolytic conditions are
described in the text. Inset: Calibration line of log (molecular mass) vs.
retention time for standard proteins as described in the text. Down-
ward and upward arrows indicate the void volume and the positions of
fragmented epimerase, respectively. ÔRÕ stands for regression coeffi-
cient.
Ó FEBS 2003 Mutarotase activity of epimerase (Eur. J. Biochem. 271)65
substrate and inhibitor specificity of this newly reported
function is in general agreement with those of other
mutarotases (Tables 2 and 3).
It is well known that large proteins consist of multiple
domains [37,38], and multifunctional proteins are Ôgenerated
by folding of contiguous stretches of chains to yield
autonomous domainsÕ [39]. Also the mutarotases from
most of the sources are monomers of 37–38 kDa [10–12]. As
about half of the K. fragilis enzyme was not involved in
epimerization, it was pertinent to ask whether the two
functions of K. fragilis epimerase operated from different
sites. Interaction with specific inhibitors and chemical
modification reactions suggested two independent catalytic

sites (Tables 2 and 4). Differences in the kinetics of refolding
of this bifunctional enzyme with reappearance of activity
also indicated this.
As 5¢-UMP protects epimerase against trypsinization
both in yeast [36] and E. coli [40], we investigated
whether galactose (or fructose), a competitive inhibitor
for mutarotase, could induce similar stabilization. The
results were indeed positive (Fig. 4). When partial
proteolysis was carried out in the presence of the
inhibitors together, both activities were retained. SDS/
PAGE of the digest showed complete disappearance of
Fig. 6. Sephadex G-200 gel filtration profiles
of different yeast extracts with respect to
epimerase and mutarotase activities. In all sets,
open and filled circles represent epimerase and
mutarotase activities, respectively. (A) K. fra-
gilis grown in galactose medium. V
o
was
measured using blue dextran and (1) catalase
(240 kDa), (2) alcohol dehydrogenase
(150 kDa), (3) BSA (66 kDa), (4) haemoglo-
bin (64 kDa), (5) ovalbumin (43 kDa), and (6)
myoglobin (19 kDa). Inset: log (molecular
mass) vs. V
e
calibration curve for standard
proteins. (B) K. fragilis growninglucose
medium. (C) Gal 10 mutant grown in glucose
medium. (D) K. lactis var lactis grown in

galactose medium. (E) P. tannophilus also
grown in galactose medium. A result identical
with those in (D) and (E) was obtained with
S. pombi grown in galactose medium. The
amount of sample loaded in these sets was
similar but not identical.
66 A. Brahma and D. Bhattacharyya (Eur. J. Biochem. 271) Ó FEBS 2003
the parent molecule, with generation of two fragments
of molecular mass 45 and 30 kDa (Fig. 4). This was
consistent with the notion that multidomain proteins are
connected by a proteolytically sensitive hinge region [41].
Finally, the functional domains were separated by size-
exclusion HPLC (Fig. 5). This proved conclusively the
bifunctional character of yeast epimerase operating from
two independent regions. As enhancement or inhibition
of any one of the activities was observed in none of the
partial proteolysis experiments, allosteric regulation
between the sites was unlikely.
We also wanted to know whether 130–150-kDa epi-
merases from other yeast strains were also associated with
mutarotase activities and that too in addition to the
constitutively formed mutarotase of 38 kDa. For this, cell
lysates of K. fragilis (with or without induction of epi-
merase), K. lactis var lactis [42], P. tannophilus [43] and
S. pombi [12] were fractionated using a Sephadex G-200
column. In all cases, mutarotase activity was found
associated with the epimerase which was distinct from the
second one appearing in the 35–39-kDa region (Fig. 6).
Thus the bifunctional character of yeast epimerase appeared
to be universal.

The biological significance of epimerase and mutarotase
gene fusion in yeast may be questioned in the light of its
constitutively expressed mutarotase gene. It is evident that
the mutarotase gene is co-induced when cells are grown in
galactose medium. In the metabolic pathway, galactose is
first phosphorylated by galactokinase before entering the
Leloir pathway. Galactokinase has an absolute specificity
for the a-anomer and thus rapid conversion of b fi a
anomer is essential to utilize the b-form [44]. When cells
grow in the exponential phase, constitutively formed
mutarotase may not be adequate, leading to the necessity
of induced mutarotase. Moderate to low catalytic efficiency
of yeast mutarotase compared with mammalian sources
may be a supportive reason; for example, the specific
activity of galactose mutarotase from Lactococcus lactis,
which is very similar to glucose mutarotase in terms of
substrate specificity, is only 134 UÆper mg protein for the
b-anomer [45]. This speculation may be confirmed with
mutant strains with epimerase devoid of associated muta-
rotase activity. Other unknown factors may also play
important roles.
Acknowledgements
This research was funded by the Department of Science and
Technology (DST) grant No. SP/SO/D-107/98 awarded to D.B. A.B.
was supported as a Junior/Senior Research Fellow by the DST and the
Council of Scientific and Industrial Research in different phases. We are
grateful to Dr Basudeb Achariya of this institute for critical reading of
the manuscript.
References
1. Frey, P.A. (1987) Pyridine nucleotide coenzymes. Chemical, Bio-

chemical and Medical Aspects (Dolphin, D., Poulson, R. &
Avarmovie, O., eds), Vol. 2B, pp. 462–447. Wiley, New York.
2. Frey, P.A. (1996) The Leloir pathway: a mechanistic imperative
for three enzymes to change the stereochemical configuration of a
single carbon in galactose. FASEB J. 10, 461–470.
3. Segel, S. (1995) The Metabolic Basis of Inherited Diseases (Scriver,
C.R.,Beaudt,A.L.,Sly,W.S.&Valle,D.,eds),pp.967–1000.
McGraw-Hill, New York.
4. Thoden, J.B., Wohlers, T.M., Fridovich-Keil, J.L. & Holden,
H.M. (2000) Crystallographic evidence for Tyr157 functioning as
theactivesitebaseinhumanUDP-galactose4-epimerase.
Biochemistry 39, 5691–5701.
5. Frey, P.A. & Vanhooke, J.L. (1994) Characterization and acti-
vation of naturally occurring abortive complexes of UDP-ga-
lactose 4-epimerase from E. coli. J. Biol. Chem. 269, 31496–31504.
6. Thoden, J.B., Frey, P.A. & Holden, H.M. (1996) High-resolution
X-ray structure of UDP-galactose 4-epimerase complexed with
UDP-phenol. Protein Sci. 5, 2149–2161.
7. Smith, T.F. & Waterman, M.S. (1981) Overlapping genes and
information theory. Adv. Appl. Math. 2, 482–489.
8. Devereux, J., Haeberli, P. & Smithies, O. (1984) A comprehensive
set of sequence analysis programs for the VAX. Nucleic Acids Res.
12, 387–395.
9. Thoden, J.B., Frey, P.A. & Holden, H.M. (1996) Molecular
structure of the NADH/UDP-glucose abortive complex of UDP-
galactose 4 epimerase from E. coli: implication for the catalytic
mechanism. Biochemistry 35, 5137–5144.
10. Bailey, J.M., Fishman, P.H., Kusiak, J.W., Mulhern, S. &
Pentchev, P.G. (1975) Mutarotase (Aldose 1-epimerase) from
kidney cortex. Methods Enzymol. 41, 471–484.

11. Fishman, P.H., Pentchev, P.G. & Bailey, J.M. (1975) Mutarotase
from higher plants. Methods Enzymol. 41, 484–487.
12. Smits, P.H.M., deHaan, M., Maat, G. & Grivell, L.A. (1994) The
complete sequence of a 33 kb fragment on the right arm of
chromosome II from Saccharomyces cerevisiae reveals 16 open
reading frames, including ten new open reading frames, five pre-
viously identified genes and a homologue of the SCO1 gene. Yeast
10, S75–S80.
13. Poolman, D., Royer, P.K., Mainzer, S.E. & Schmidt, B.F. (1990)
Carbohydrate utilization in Streptococcus thermophilus:Char-
acterization of the genes for aldose 1-epimerase (mutarotase) and
UDP-glucose 4-epimerase. J. Bacteriol. 172, 4037–4047.
14. Majumdar, S. (2000) Studies on assembly pathway and active site
of UDP-glucose 4-epimerase. PhD Thesis, Jadavpur University,
Calcutta, India.
15. Bhattacharyya, D. (1993) Reversible folding of UDP-galactose
4-epimerase from yeast Kluyveromyces fragilis. Biochemistry 32,
9726–9734.
16. Dutta, S., Maity, N.R. & Bhattacharyya, D. (1997) Multiple
unfolded states of UDP galactose 4-epimerase from yeast Kluy-
veromyces fragilis: involvement of proline cis-trans isomerisation
in reactivation. Biochim. Biophys. Acta 1343, 251–262.
17. Majumder, S., Bhattacharjee, H., Bhattacharyya, D. & Bhaduri,
A. (1998) UDP-galactose-4-epimerase from Kluyveromyces fragi-
lis: reconstitution of holoenzyme structure after dissociation with
parachloromercuribenzoate. Eur. J. Biochem. 257, 427–433.
18. Maity, N.R., Barart, B. & Bhattacharyya, D. (1999) UDP-
galactose 4-epimerase from Kluyveromyces fragilis: equilibrium
unfolding studies. Indian J. Biochem. Biophys. 36, 433–441.
19. Bernardi, G. (1971) Chromatography of proteins on hydroxy-

apatite. Methods Enzymol. 22, 325–339.
20. Zalitis, J., Uram, M., Bowser, A.M. & Feingold, D.S. (1972)
UDP-glucose dehydrogenase from beef liver. Methods Enzymol.
28, 430–435.
21. Darrow, R.A. & Rodstrom, R. (1968) Purification and properties
of UDP-galactose 4-epimerase from yeast. Biochemistry 7, 1645–
1654.
22.Gatz,C.,Altschmied,J.&Hillen,W.(1986)Cloningand
expression of the Acinetobacter calcoaceticus mutarotase gene in
Escherichia coli. J. Bacteriol. 168, 31–39.
Ó FEBS 2003 Mutarotase activity of epimerase (Eur. J. Biochem. 271)67
23. Kalckar, H.M., Bertland, A.U. & Bagge, B. (1970) The reductive
inactivation of UDP-galactose 4-epimerase from yeast and E. coli.
Proc.NatlAcad.Sci.USA65, 1113–1119.
24. Darrow, R.A. & Rodstorm, R. (1970) Uridine diphosphate
galactose-4-epimerase from yeast: studies on the relationship
between quaternary structure and catalytic activity. J. Biol. Chem.
245, 2036–2042.
25. Darrow, R.A. & Rodstorm, R. (1966) Subunit association and
subunit assembly of uridine diphosphate galactose-4-epimerase
form yeast. Proc.NatlAcad.Sci.USA55, 205–212.
26. Bhattacharjee, H. & Bhaduri, A. (1992) Distinct functional roles
of two active site thiols in UDP-galactose 4-epimerase from
Kluyveromyces fragilis. J. Biol. Chem. 267, 11714–11720.
27. Miles, E.W. (1977) Modification of histidyl residues in proteins by
diethylpyrocarbonate. Methods Enzymol. 47, 431–442.
28. Smith, E.L. (1977) Reversible blocking of arginine by cyclo-
hexanedione. Methods Enzymol. 47, 156–161.
29. Nayar, S. (2000) Studies on the subunit structure and folding
behaviour of a complex multimeric enzyme: UDP-galactose

4-epimerase. PhD Thesis, Jadavpur University, Calcutta.
30. Gorbunoff, M.J. (1985) Protein chromatography of hydroxyapa-
tite columns. Methods Enzymol. 182, 329–339.
31. Dutta, S. & Bhattacharyya, D. (2001) Size of unfolded and dis-
sociated subunits versus that of native multimeric proteins. J. Biol.
Phys. 27, 59–71.
32. Nath, S., Brahma, A. & Bhattacharyya, D. (2003) Extended
application of gel-permeation chromatography by spin column.
Anal. Biochem. 320, 199–206.
33. Mukherjee, S. & Bhaduri, A. (1992) An essential histidine residue
for the activity of UDP-galactose 4-epimerase from Kluyver-
omyces fragilis. J. Biol. Chem. 267, 11709–11713.
34. Mukherjee, S. & Bhaduri, A. (1986) UDP-glucose 4-epimerase
from Saccharomyces fragilis: presence of an essential arginine
residue at the substrate-binding site of the enzyme. J. Biol. Chem.
261, 4519–4524.
35. Beebe, J.A. & Frey, P.A. (1998) Galactose mutarotase: purifica-
tion, characterization, and investigations of two important histi-
dine residues. Biochemistry 37, 14989–14997.
36. Mehta, S. (1994) Studies on the molecular architecture and envi-
ronment around the active site of UDP-glucose 4-epimerase from
Kluyveromyces fragilis. PhD Thesis, Jadavpur University, Cal-
cutta.
37. Garel, J R. (1992) Folding of large proteins: multidomain and
multisubunit proteins. In Protein Folding (Creighton, T.E., ed.),
pp. 405–454. W.H. Freeman, New York.
38. Creighton, T.E. (1991) Stability of folded proteins. Curr. Opin.
Struct. Biol. 1, 5–16.
39. Kirschner, K. & Bisswanger, H. (1976) Multifunctional proteins.
Annu. Rev. Biochem. 45, 143–166.

40. Barat, B. & Bhattacharyya, D. (2001) UDP-galactose 4-epimerase
from Escherichia coli: formation of catalytic site during reversible
folding. Arch. Biochem. Biophys. 391, 188–196.
41. Morrice, N.A. & Carrey, E.A. (1997) Peptide mapping. In Protein
Structure: a Practical Approach (Creighton, T.E., ed.), pp. 117–
149. IRL Press, Oxford University, Oxford.
42. Webster, T.D. & Dickson, R.C. (1988) Nucleotide sequence of the
galactosegeneclusterofKlueyveromyces lactis. Nucleic Acids Res.
16, 8192–8195.
43. Skrzypek, M. & Maleszka, R. (1994) A gene homologous to that
encoding UDP galactose-4-epimerase is inducible by xylose in the
yeast Pachysolen tannophilus. Gene 140, 127–129.
44. Heinrich, M.R. & Howard, S.M. (1966) Galactokinase. Methods
Enzymol. 11, 407–412.
45. Thoden, J.B. & Holden, H.M. (2002) High resolution X-ray
structure of galactose mutarotase from Lactococcus lactis. J. Biol.
Chem. 277, 20854–20861.
68 A. Brahma and D. Bhattacharyya (Eur. J. Biochem. 271) Ó FEBS 2003