Studies of UDP-galactopyranose mutase from Escherichia coli: An unusual role of reduced FAD in its catalysis

Article abstract of DOI:10.1021/ja001333z

The galactofuranose moiety found in many surface constituents of microorganisms is derived from UDP-D-galactopyranose (UDP-Galp) via a unique ring contraction reaction catalyzed by UDP-Galp mutase. This enzyme, which has been isolated from several bacteri

Full text of DOI:10.1021/ja001333z

J. Am. Chem. Soc. 2000, 122, 9065-9070
9065
Studies of UDP-Galactopyranose Mutase from Escherichia coli: An
Unusual Role of Reduced FAD in Its Catalysis
Qibo Zhang and Hung-wen Liu*
Contribution from the Department of Chemistry, UniVersity of Minnesota, Minneapolis, Minnesota 55455
ReceiVed April 17, 2000
Abstract: The galactofuranose moiety found in many surface constituents of microorganisms is derived from
UDP-D-galactopyranose (UDP-Galp) via a unique ring contraction reaction catalyzed by UDP-Galp mutase.
This enzyme, which has been isolated from several bacterial sources, is a flavoprotein where the FAD coenzyme
is noncovalently bound. Since its catalysis does not appear to involve a redox mechanism, whether the enzyme-
bound FAD plays an active role in the reaction mechanism has been obscure. To study this transformation, the
corresponding E. coli mutase was purified, and the ring contraction product, UDP-Galf, was chemically
synthesized. Using UDP-Galf as the substrate, a Km of 194 µM and a kcat of 1.5 s-1 for the catalysis in the
reverse direction were obtained. The preference of the reaction toward the pyranose product was confirmed by
an equilibrium constant of 0.057 in the forward direction. Interestingly, when the enzyme was reduced by
sodium dithionite, its catalytic efficiency was increased by more than 2 orders of magnitude. A comparable
rate enhancement was also noted when the flavin coenzyme was selectively reduced by photoreduction in the
presence of 5-deazariboflavin under anaerobic conditions. Since mutase with either oxidized or reduced FAD
is active, the change of the redox state in FAD appears to affect only the activity, but not the catalytic mechanism.
It is conceivable that reduction of FAD may induce a favorable conformational change of the enzyme that
may be more conducive to catalysis. It is also possible that the reduced flavin bears a higher electron density
at N-1, which may then be used to stabilize the transiently formed oxocarbenium ion intermediates to facilitate
catalysis. Whether structural effects, electronic effects, or a combination of both dictates the ability of FAD
to enhance the rate of the mutase reaction is an interesting, albeit challenging question. Nevertheless, the
present work has provided, for the first time, evidence indicating the active involvement of FAD in regulating
the catalytic efficiency of this enzyme.
UDP-D-galactofuranose (2, UDP-Galf) is the precursor of the
galactofuranose moiety1 found in a variety of surface constitu-
ents of microorganisms.2 UDP-Galf is biosynthetically derived
from UDP-D-galactopyranose (1, UDP-Galp) via a unique ring
contraction reaction catalyzed by UDP-galactopyranose mutase
(Scheme 1).3 The fact that galactofuranose has been found in
many pathogens but not in human tissues makes UDP-Galp
mutase an attractive target for therapeutic agents. This enzyme,
which has been isolated from several bacterial sources,3,4 is a
flavoprotein, with the flavin adenine dinucleotide (FAD)
coenzyme bound noncovalently. Few mechanistic studies of this
mutase had been described until a recent report,5 in which the
UDP-Galf and UDP-Galp interconversion was elegantly dem-
onstrated to involve cleavage of the anomeric C-O bond. As
illustrated in Scheme 1, the catalysis of the forward reaction is
likely to be initiated by distortion of the ring which allows attack
of O4 on C-1 to release UDP, or by elimination of UDP first to
form an oxocarbenium ion 3 followed by O4 attack on C-1.
Subsequent ring opening between C-1 and O5 of the bicyclo-
acetal intermediate 4, via either 5 or 6, followed by the rebound
of UDP at C-1 leads to the formation of 2. It must be noted
that the proposed mechanism does not involve a redox reaction,
nor is there a change in the redox states of the substrate or
product. Thus, these two considerations do not clearly ascribe
an active role to the enzyme-bound FAD in the reaction
mechanism. Hence, the role of FAD in UDP-Galp mutase
remains obscure. The present work describes our study on the
mutase from Escherichia coli, and the initial evidence indicating
the active involvement of FAD in regulating the catalytic
efficiency of this enzyme.
* To whom correspondence should be addressed. Current address:
Division of Medicinal Chemistry, College of Pharmacy, University of Texas,
Austin, TX 78712. Telephone: 512-232-7811. Fax: 512-471-2746. E-
mail: h.w.liu@mail.utexas.edu.
(1) Trejo, A. G.; Chittenden, G. J. F.; Buchanan, J. G.; Baddiley, J.
Biochem. J. 1970, 117, 637-639.
(2) (a) de Lederkremer, R. M.; Colli, W. Glycobiology 1995, 5, 547-
552. (b) Brennan, P. J.; Nikaido, H. Annu. ReV. Biochem. 1995, 64, 29-
63. (c) Whitfield, C. Trends Microbiol. 1995, 3, 178-185. (d). Previato, J.
O.; Mendonca-Previato, L.; Jones, C.; Wait, R. Glycoconjugate J. 1993,
10, 340.
(3) Nassau, P. M.; Martin, S. L.; Brown, R. E.; Weston, A.; Monsey,
D.; McNeil, M. R.; Duncan, K. J. Bacteriol. 1996, 178, 1047-1052.
(4) (a) Koplin, R.; Brisson, J.-R.; Whitfield, C. J. Biol. Chem. 1997,
272, 4121-4128. (b) Weston, A.; Stern, R. J.; Lee, R. E.; Nassau, P. M.;
Monsey, D.; Martin, S. L.; Scherman, M. S.; Besra, G. S.; Duncan, K.;
McNeil, M. R. Tuberc. Lung Dis. 1998, 78, 123-131. (c) Chen, P.; Bishai,
W. Infect. Immun. 1998, 66, 5099-5106. (d) McMahon, S. A.; Leonard,
G. A.; Buchanan, L. V.; Giraud, M.-F.; Naismith, J. H. Acta Crystallogr.
1999, D55, 399-402.
Results and Discussion
Expression of glf Gene and Purification of Glf Protein.
The UDP-Galp mutase encoding gene from E. coli K-12, glf,3
was amplified by the polymerase chain reaction (PCR) and
cloned into the expression vector pET24b(+). The recombinant
construct was used to transform E. coli BL21 (DE3), and the
resulting cells were incubated at 37 °C in LB medium. This
C-terminal His6-tagged enzyme was purified to near homogene-
(5) Barlow, J. N.; Girvin, M. E.; Blanchard, J. S. J. Am. Chem. Soc.
1999, 121, 6968-6969.
10.1021/ja001333z CCC: $19.00 © 2000 American Chemical Society
Published on Web 09/08/2000

9066 J. Am. Chem. Soc., Vol. 122, No. 38, 2000
Zhang and Liu
Scheme 1
synthesis of the phosphonate analogue of UDP-Galf has recently
been reported,8 UDP-Galf itself has only been obtained enzy-
matically in small quantities.4a,7 To prepare sufficient amounts
of UDP-Galf (2) for the intended biochemical studies, a synthetic
sequence based on the general strategy of de Lederkremer et
al. to make R-D-Galf-1-phosphate (12)9 was developed. As
delineated in Scheme 2, the key intermediate, per-O-benzoy-
lated-D-galactofuranose (10), was prepared from D-galactono-
γ-lactone (7) in three steps. A reported procedure for direct
benzoylation of D-galactose at elevated temperature failed in
our hand to generate the desired furanose derivatives 10.10-12
Selective derivatization at C-1 of 10 with bromotrimethylsilane
(TMSBr) followed by phosphorylation gave 11. The final
conversions involved hydrogenation, hydrolysis, and coupling
with uridine phosphomorpholidate in anhydrous pyridine in the
presence of 1H-tetrazole to afford 2 in 4% overall yield (7 f
2).13 The final product was purified by size exclusion chroma-
tography and reversed phase HPLC. The structure of 2 was
confirmed spectroscopically by NMR (1H, 13C, 31P) and high-
resolution mass spectra.
Figure 1. Lane A: Molecular weight marker, bovine serum albumin
(66 kDa), ovalbumin (45 kDa), glyceraldehyde-3-phosphate dehydro-
genase (36 kDa), carbonic anhydrase (29 kDa), trypsinogen (24 kDa).
Lane B: Purified UDP-galactopyranose mutase.
ity (see Figure 1) using Ni-NTA resin (Qiagen) and found to
be stable in phosphate buffers containing 15% glycerol.6
N-Terminal amino acid sequencing confirmed that the first 10
residues (MYDYIIVGSG) of this protein are identical to the
translated glf sequence. The purified UDP-Galp mutase (Glf)
was determined to be a homodimer by judging from a Mr of 72
kDa estimated by gel filtration chromatography (Superdex 200)
and a calculated mass of 44 kDa based on the translated peptide
sequence. The enzyme exhibits a typical unresolved flavin
spectrum with absorbance maxima at 450 and 376 nm. A
stoichiometry of 0.52 of bound FAD per protein monomer was
estimated by quantitation of FAD released from an enzyme
sample of known concentration.
Synthesis of UDP-Galf (2). Since the equilibrium of the
interconversion between UDP-Galp (1) and UDP-Galf (2)
catalyzed by UDP-Galp mutase greatly favors the formation of
UDP-Galp (1),7 it has been difficult to assay the enzyme activity
using commercially available UDP-Galp. This problem, how-
ever, can be circumvented by studying the catalysis in the
reverse direction using UDP-Galf (2) as the substrate. Although
Characterization of Glf Protein. The activity of the purified
enzyme was assayed by incubating a mixture of UDP-Galf (2)
and Glf at 37 °C, and analyzing the extent of interconversion
by reversed phase HPLC. The equilibrium constant (Keq
)
[2]/[1]) was determined to be 0.057, providing the first quantita-
tive measurement of the preference of the reaction toward the
pyranose product. A Km of 194 µM for UDP-Galf and a kcat of
1.5 s-1 for the catalysis in the reverse direction (2 f 1) were
also obtained. None of the common redox cofactors, such as
NAD(P)+, NAD(P)H, FAD, or the combination of NAD(P)H
and FAD (2 mM each) showed a significant effect on the
reaction rate.14 Likewise, assays performed under anaerobic
conditions gave results comparable to those obtained aerobically.
(8) Ovensky, J.; McNeil, M.; Sinay, P. J. Org. Chem. 1999, 64, 6202-
6205.
(9) de Lederkremer, R. M.; Nahmad, V. B.; Varela, O. J. Org. Chem.
1994, 59, 690-692.
(10) de Lederkremer, R. M.; Litter, M. I. Carbohydr. Res. 1971, 20, 442-
444.
(11) Kohn, P.; Samaritano, R. H.; Lerner, L. M. J. Am. Chem. Soc. 1965,
87, 5475-5480.
(6) The wild-type enzyme was reported to be unstable even in the
presence of glycerol.3
(12) D’Accorso, N. B.; Thiel, I. M. E. Carbohydr. Res. 1983, 124, 177-
184.
(7) Lee. R.; Monsey, D.; Weston, A.; Duncan, K.; Rithner, C.; McNeil,
M. Anal. Biochem. 1996, 242, 1-7.
(13) Chen. H.; Guo, Z.; Liu, H.-w. J. Am. Chem. Soc. 1998, 120, 11796-
11797.

FAD and UDP-Galactopyranose Mutase Catalysis
J. Am. Chem. Soc., Vol. 122, No. 38, 2000 9067
Scheme 2
In addition, no redox-active metal ion was detected by induc-
tively coupled plasma (ICP) analysis of Glf, and dialysis of the
enzyme against 10 mM EDTA did not result in any loss of
activity. All of these findings suggest that the interconversion
between 1 and 2 catalyzed by the mutase is unlikely to be a
redox process.
with Ellman’s reagent (DTNB)15 revealed that half ( ∼52%) of
the thiols in the native enzyme are in the reduced form with
the remaining half involved in disulfide linkage. Thus, the
increase of catalytic efficiency upon dithionite treatment may
be a combined effect of the reduction of FAD and the disulfide
linkage. Indeed, after treatment with dithionite, the chromophore
of the flavin coenzyme was bleached, and almost all thiols (91%)
were in the reduced form. Interestingly, the bleached FAD in
the reduced mutase could be readily oxidized by air, and the
activity of the resulting enzyme lowered to the same level as
the control without dithionite treatment. Because most of the
cysteines (85%) remained reduced upon brief exposure to air,
the activity enhancement after treatment with dithionite must
be largely attributed to the reduction of FAD. This conclusion
is consistent with the facts that these cysteines are not conserved
among mutases of different origins,16 and thiol-modifying
agents, such as iodoacetamide and iodoacetate, have no inhibi-
tory effect on the enzyme at concentration as high as 20 mM.
Photoacceleration. To further study the role of reduced FAD
in enhancing the activity of mutase, the flavin coenzyme was
selectively reduced by photoreduction of mutase in the presence
of 5-deazariboflavin under anaerobic conditions at room tem-
perature.17 It was found that the specific activity of the bleached
Activity Enhancement Upon Dithionite Reduction. Al-
though none of the above data are indicative of a redox process
for the mutase reaction, a mechanism in which substrate
oxidation followed by rearrangement and product reduction are
coupled to the recycling of FAD between its oxidized and
reduced forms remains a viable scenario. If such a redox
recycling is indeed the mechanism for Glf, reduction of the
enzyme bound FAD should inactivate the mutase since the flavin
coenzyme assumes an active role in this postulated oxidation-
reduction sequence. To gain further insight into whether the
enzyme bound FAD plays a role in the catalysis, the mutase
was treated with dithionite to reduce the flavin coenzyme. To
our great surprise, when sodium dithionite (20 mM) was
included in the assay mixture, the catalytic efficiency (kcat/Km)
of the mutase was increased by more than 2 orders of magnitude
(kcat of 27 s-1 and Km of 22 µM). This result clearly indicated
that the reduced enzyme is more active than the oxidized
enzyme.
(15) Riddles, P. W.; Blakeley, R. L.; Zerner, B. Methods Enzymol. 1983,
91, 49-60.
(16) The mutase from Mycobacterium tuberculosis contains no cysteine
residues at all,4b and that from Klebsiella pneumoniae serotype O1 contains
six cysteines.4a Since these cysteine residues are not conserved, they are
unlikely to play active roles in the proposed mechanism.
It should be noted that in addition to FAD, UDP-Galp mutase
also contains two cysteine residues (C16 and C45).3 Titration
(14) NAD(P)H was reported to be required for the activity of mutase
isolated from Klebsiella pneumoniae.4a

9068 J. Am. Chem. Soc., Vol. 122, No. 38, 2000
Zhang and Liu
enzyme was 29-fold higher than that of the control without
illumination. The magnitude of enhancement is comparable to
that observed in the case of dithionite reduction (36-fold
enhancement at room temperature). Most importantly, the free
thiol content of the photoreduced Glf was determined to be 51%,
which is essentially the same as that of the native enzyme. These
data firmly established that reduction of the enzyme-bound FAD
is the most significant factor in the large enhancement of UDP-
Galp mutase activity. Since mutase with either oxidized or
reduced FAD is active, this fact strongly suggests that the change
of the redox state of FAD affects only the activity, but not the
catalytic mechanism.
Several enzymes that contain oxidized flavin and yet catalyze
reactions involving no net redox chemistry are known. These
include chorismate synthase,18 acetolactate synthase,19 mande-
lonitrile lyase,20 tartronate-semialdehyde synthase,21 and YerE.13
While the catalytic roles for the flavin in these enzymes remain
uncertain, it has been shown that chorismate synthase, which
catalyzes the conversion of 5-enolpyruvylshikimate-3-phosphate
to chorismate, requires a reduced FMN for activity. Recently,
the reduced flavin coenzyme in chorismate synthase has been
demonstrated to be involved in the catalytic cycle mediating a
radical mechanism.22 A similar pathway for UDP-Galp mutase
invoking radicals appears unlikely since both the oxidized and
reduced enzymes are active.
Proposed Mechanism of Acceleration. Overall, the observed
significant effect of the reduced enzyme-bound FAD on the rate
of UDP-Galp mutase illustrates an interesting phenomenon. One
can propose that reduction of FAD, which involves transforma-
tion of the coenzyme from a highly conjugated planar frame to
a bent butterfly structure, may induce a conformational change
within the enzyme that may be more conducive to catalysis.
Likewise, it is also possible that the reduced flavin imparts a
more negative character at N-1 (14), which may then be used
Figure 2. Circular dichroism spectra of Glf (11 µM) in 50 mM
potassium phosphate buffer, pH 7.6 (bold line), and Glf (11 µM) in
the same buffer containing 20 mM sodium dithionite (plain line).
In summary, the coenzyme FAD in UDP-galactopyranose
mutase plays an unusual role in regulating the catalytic
efficiency of this enzyme. Further studies should reveal whether
this is a unique characteristic of UDP-Galp mutase or if this is
just the first example of an enzyme utilizing flavin in a
nonclassical way to achieve optimal activity.
Experimental Section
General. Unless otherwise specified, all chemicals were purchased
from Aldrich or Sigma, and used without further purification. NMR
spectra were recorded on Varian 200, 300, or 500 MHz spectrometers.
13C NMR spectra were recorded with proton broad-band decoupling
and the reported spin couplings are from phosphorus. 31P NMR spectra
were recorded with proton decoupling and externally referenced with
85% phosphoric acid. The J-values are given in Hz. For NMR spectra
recorded in D2O, t-BuOH was added as an internal reference; the
1
chemical shifts are 1.27 and 31.2 ppm for H NMR and 13C NMR,
respectively.23 All UV-visible spectra were taken on a Beckman
DU650 spectrophotometer. Fast-atom bombardment (FAB) and elec-
trospray ionization (ESI) mass spectra were recorded by the MS facility
at the Department of Chemistry of the University of Minnesota. High-
performance liquid chromatography (HPLC) analysis and purification
was conducted with a Hewlett-Packard 1090A system equipped with
a photodiode array detector. To detect nucleotide sugar, the detector
wavelength was set at 262 nm. Circular dichroism spectra were recorded
on a Jasco 710 instrument. Analytical thin-layer chromatography was
carried out on Merck silica gel 60 G-254 plates, and the spots were
visualized either under UV light or by heating plates previously stained
with solutions of vanillin/ethanol/H2SO4 (1:98:1) or phosphomolybdic
acid (7% in EtOH). ICP analysis was performed by the Research
Analytical Laboratory at the Department of Soil, Water, and Climate,
of the University of Minnesota. N-Terminal amino acid sequencing
was performed by the Microchemical Facility at the Institute of Human
Genetics of the University of Minnesota.
Construction of the glf Expression Plasmid. The sequence of the
glf gene in Escherichia coli has been reported,3 and this information
allowed the design of oligonucleotide primers complementary to the
sequence at each end of the gene. The start primer, 5′-TGTTTTGCT-
GAGGATCATATGTACGATTATATCATT-3′, contained a NdeI re-
striction site (in bold) and the codons for the first six amino acid residues
of Glf. The halt primer, 5′-TTACAAGATAGACTCGAGATCCG-
TACTCATTATATT-3′, introduced a XhoI restriction site (in bold)
which replaced the stop codon of the glf gene. These primers were
used to amplify the glf gene from the genomic DNA of E. coli C600
(ATCC 23724) by PCR. The PCR-amplified fragment was purified,
digested with NdeI and XhoI and ligated into NdeI/XhoI sites of the
transcription vector, pET-24b(+) (Novagen) to give pQZ-1. This
recombinant plasmid was isolated and was used to transform E. coli
DH5R. Positive clones were identified by digestion of the plasmid DNA
to stabilize the transiently formed oxocarbenium ion intermedi-
ates (3 and/or 5) to facilitate catalysis. A comparison of the
circular dichroism (CD) spectra of the native and the reduced
enzyme failed to discern a major change of the enzyme
conformation associated with the bleaching of FAD (Figure 2).
Thus, this makes the charge stabilization theory more appealing.
Still, one cannot fully discount the possibility that the invoked
favorable conformational change may be highly localized in the
active site and hence may not be detected by CD. Therefore,
understanding whether it is structural or electronic effects (or a
combination of both) that dictate the novel capacity of reduced
FAD for mutase rate enhancement is clearly a challenging task
for future exploration.
(17) Massey, V.; Hemmerich, P. Biochemistry 1978, 17, 9-17.
(18) Kerwin, J. F. Jr.; Lancaster, J. R., Jr.; Feldman, P. L. J. Med. Chem.
1995, 38, 4343-4362.
(19) Chipman, D.; Barak, Z.; Schloss, J. V. Biochim. Biophys. Acta 1998,
1385, 401-419.
(20) Xu, L.-L.; Singh, B. K.; Conn, E. E. Arch. Biochem. Biophys. 1986,
250, 322-328.
(21) Cromartie, T. H.; Walsh, C. T. J. Biol. Chem. 1976, 251, 329-
333.
(22) Macheroux, P.; Schmid, J.; Amrhein, N.; Schaller, A. Planta 1999,
207, 325-334 and references therein.
(23) Zhang, Q.; Cone, M. C.; Gould, S. J.; Zabriskie, T. M. Tetrahedron
2000, 56, 693-701.

FAD and UDP-Galactopyranose Mutase Catalysis
J. Am. Chem. Soc., Vol. 122, No. 38, 2000 9069
with NdeI and XhoI and visualization of the excised insert by staining
an agarose gel of the DNA with ethidium bromide after electrophoresis.
The plasmid DNA from positive clones was used to transform E. coli
BL21 (DE3). The general methods and protocols for recombinant DNA
manipulations were as described by Sambrook et al.24
temperature and was then poured into ice water to quench the reaction.
The aqueous layer was decanted, and the remaining syrup was dissolved
in ethyl acetate (300 mL). The organic solution was washed with dilute
sodium bicarbonate and brine and dried over magnesium sulfate. The
solvent was removed under reduced pressure, and the residue was
chromatographed (ethyl acetate/hexanes ) 1:5) on silica gel to give 8
Expression and Purification of UDP-galactopyranose Mutase. E.
coli BL21 (DE3)/pQZ-1 was grown in Luria-Bertani (LB) broth
supplemented with kanamycin (50 µg/mL). Six 1 L batches were
inoculated with an overnight culture (1 mL each), and the cells were
grown at 37 °C for 18 h with vigorous agitation. The cells were
harvested by centrifugation (5000g, 10 min), washed with 50 mM
potassium phosphate buffer (pH 7.0), collected again by centrifugation
(5000g, 10 min), and stored at -80 °C. A typical yield was
approximately 60 g (wet weight) of cells/6 L of culture. Isopropyl â-D-
thiogalactoside (IPTG) was found to have little effect on the level of
overexpression, and thus was not added. The C-terminal His6-tagged
protein was purified to near homogeneity using Ni-NTA resin (Qiagen),
according to the protocols recommended by the manufacturer. The
purified mutase (Glf) was found to be stable in phosphate buffers
containing 15% glycerol. Neither reducing agent (such as dithiothreitol)
nor protease inhibitor was included in any purification steps. The yield
was about 100 mg of the pure protein from 6 L of culture.
Protein Assay. Protein concentrations were routinely determined
by the Bradford method25 using bovine serum albumin as the standard.
This assay was calibrated by comparing the results from quantitative
amino acid analysis performed on aliquots of the same sample by the
Microchemical Facility at the Institute of Human Genetics of the
University of Minnesota. The purity of the enzyme was assessed by
SDS-polyacrylamide gel electrophoresis. Electrophoresis was carried
out in the discontinuous buffer system of Laemmli,26 and the separating
gel and stacking gel were 12 and 4% polyacrylamide, respectively.
Prior to electrophoresis, protein samples were heated in 62.5 mM Tris-
HCl buffer (pH 6.8) containing 10% glycerol, 2% SDS, 5% â-mer-
captoethanol and 0.0025% bromophenol blue. Electrophoresis of the
heated samples was run in 25 mM Tris-HCl, 192 mM glycine, and
0.1% SDS (pH 8.3) at 25 mA. Gels were stained with Coomassie blue
and destained with acetic acid-ethanol-water (15:20:165 by volume).
Molecular Weight Determination. The molecular weight of the
native enzyme was determined by gel filtration chromatography
performed on a Pharmacia FPLC Superdex S-200 HR 10/30 column
eluted with 50 mM potassium phosphate buffer (pH 7.6) containing
0.15 M NaCl. The column was calibrated by separate chromatographic
runs with the following protein standards: cytochrome c (12 kDa),
carbonic anhydrase (29 kDa), ovalbumin (45 kDa), bovine serum
albumin (66 kDa), alcohol dehydrogenase (160 kDa), â-amylase (200
kDa), and blue dextran (2000 kDa). Each standard was dissolved in
50 mM potassium phosphate buffer (pH 7.6) containing 0.15 M NaCl
and injected separately to determine the retention time. The data were
analyzed by the method of Andrews.27 The subunit molecular mass
was estimated by SDS-PAGE as described by Laemmli.26 Protein
standards included trypsinogen (24 kDa), carbonic anhydrase (29 kDa),
glyceraldehyde-3-phosphate dehydrogenase (36 kDa), ovalbumin (45
kDa), and bovine albumin (66 kDa).
1
in 72% yield (24 g). H NMR of 8 is identical to the reported data.10
2,3,5,6-Tetra-O-benzoyl-D-galactofuranose (9). Disiamylborane
was prepared by adding 2-methyl-2-butene (2.0 M in THF, 40 mL) to
a solution of borane-dimethyl sulfide complex (10 M borane, 4 mL) at
0 °C under nitrogen atmosphere. After stirring for 3 h, 2,3,5,6-tetra-
O-benzoyl-D-galactono-1,4-lactone (8, 5.94 g, 10 mmol) in THF (15
mL) was added to this disiamylborane solution, and the resulting
mixture was allowed to react for an additional 18 h at room temperature.
Water (15 mL) was slowly added, and the mixture was stirred for 1 h
to quench the reaction. After removal of solvent under reduced pressure,
the organic residue was purified by silica gel column chromatography
with ethyl acetate/hexanes (1:3) as the eluent. The product 9 was isolated
as a mixture of R and â isomers (4.7 g, 79%). 1H NMR of 9 is identical
to the reported data.10
Penta-O-benzoyl-D-galactofuranose (10). 2,3,5,6-Tetra-O-benzoyl-
D-galactofuranose (9, 10.1 g, 17 mmol) was dissolved in pyridine (80
mL) and cooled with an ice-water bath. Excess benzoyl chloride (30
mL) was slowly added to this solution and the resulting mixture was
stirred for an additional 3 h at 0 °C. The mixture was then poured into
ice water, and the aqueous layer was decanted. The remaining syrup
was dissolved in ethyl acetate (300 mL) and washed with dilute sodium
bicarbonate and brine and dried over sodium sulfate. After the solvent
was removed under reduced pressure, the crude product was chro-
matographed on silica gel to afford the desired product 10 in 84% yield
1
(9.98 g). H and 13C NMR of 10 are identical to the reported data.12
Dibenzyl 2,3,5,6-Tetra-O-benzoyl-r-D-galactofuranosyl-1-phos-
phate (11). To a solution of 10 (3 g, 4.3 mmol) in anhydrous CH2Cl2
(25 mL) was slowly added excess bromotrimethylsilane (6 mL) at 0
°C. The cooling bath was removed 1 h later, and the reaction was stirred
for 24 h at room temperature. The mixture was evaporated to dryness
under reduced pressure, and the residue was mixed with dibenzyl
phosphate (1.8 g, 6.5 mmol) and Et3N (895 µL) in anhydrous toluene
(10 mL). After stirring overnight at room temperature, the mixture was
evaporated to dryness under reduced pressure. The crude product was
chromatographed on silica gel (ethyl acetate/toluene ) 1:9) to give
1
the product 11 in 50% yield (1.79 g). H NMR (CDCl3) δ 4.63 (1H,
dd, J ) 12.0, 6.3), 4.74 (1H, dd, J ) 6.9, 4.5), 4.77 (1H, dd, J ) 12.0,
4.5), 4.85 (1H, dd, J ) 11.7, 7.8), 4.96 (1H, dd, J ) 11.7, 6.9), 4.99
(1H, dd, J ) 12.0, 8.7), 5.06 (1H, dd, J ) 11.7, 7.5), 5.76 (1H, ddd,
J ) 7.5, 4.5, 1.8), 5.86 (1H, dt, J ) 6.3, 4.5), 6.19 (1H, t, J ) 6.9),
6.38 (1H, dd, J ) 6.0, 4.5), 7.0-8.2 (m, aromatic-Hs). 13C NMR
(CDCl3) δ 62.8, 69.4 (d, J ) 5), 69.5 (d, J ) 5), 70.8, 73.4, 76.6 (d,
J ) 6), 80.0, 97.7 (d, J ) 4), and the aromatic signals 127-134, 165.5,
165.6, 165.7, 165.9. 31P NMR (CDCl3) δ -1.8.
Triethylammonium 2,3,5,6-Tetra-O-benzoyl-r-D-galactofurano-
syl-1-phosphate (12). Compound 11 (1.65 g, 2 mmol) in a mixture of
ethyl acetate (20 mL) and Et3N (1.6 mL) was hydrogenated overnight
in the presence of catalytic amount of palladium (10%) on charcoal.
After being filtered through a Celite pad, the solution was concentrated
to dryness to yield 12 as a monotriethylammonium salt (1.41 g, 98%
Quantitation of Bound FAD. The stoichiometry of bound FAD
per subunit of the mutase was estimated by measuring the quantity of
released FAD from a denatured mutase sample of known concentration.
In this experiment, a sample of mutase (94 µM) in 100 mM potassium
phosphate buffer, pH 7.6, was denatured by boiling for 10 min in a
foil-covered centrifuge tube. The precipitate was removed by centrifu-
gation, and portions of the supernatant were used for the measurement
of absorption at 450 nm against a blank that consisted of the final
dialysis buffer. An extinction coefficient of 11 300 M-1 cm-1 was used
for the calculation.
1
yield). H NMR (CDCl3) δ 1.15 (9H, t, J ) 7.2), 2.89 (6H, q, J )
7.2), 4.59 (1H, dd, J ) 12.0, 4.8), 4.75 (1H, dd, J ) 12.0, 4.2), 4.88
(1H, dd, J ) 12.0, 4.2), 5.67 (1H, ddd, J ) 7.2, 4.5, 0.9), 5.85 (1H, dt,
J ) 6.6, 4.2), 6.17 (2H, m), 7.26-7.35 (8H, m), 7.37-7.49 (4H, m),
7.85-8.18 (8H, m). 13C NMR (CDCl3) δ 8.5, 45.3, 63.1, 71.5, 74.3,
76.6 (d, J ) 6), 78.7, 96.0 (d, J ) 4), and the aromatic signals 127-
134, 165.5, 165.8, 165.9, 166.0.
Synthesis of UDP-Galf (2). 2,3,5,6-Tetra-O-benzoyl-D-galactono-
1,4-lactone (8). To a solution of D-galactono-1,4-lactone (7, 10 g, 56
mmol) in pyridine (150 mL) was slowly added excess benzoyl chloride
(40 mL) at 0 °C. The resulting mixture was stirred for 3 h at room
Bis(triethylammonium) r-D-Galactofuranosyl-1-phosphate (13).
The protected galactofuranosyl-1-phosphate salt (12, 0.71 g, 0.84 mmol)
was kept in a solution of MeOH/H2O/Et3N (5:2:1, 32 mL) at 32 °C for
3 days. The resulting mixture was concentrated to dryness under reduced
pressure, and the residue was chromatographed on Sephadex LH-20
(1.5 × 45 cm) using MeOH as the eluent. Fractions containing the
product were pooled and evaporated to dryness to yield 13 as a bis-
triethylammonium salt (341 mg, 88% yield). H NMR (D2O) δ 1.30
(18H, t, J ) 7.2), 3.23 (12H, q, J ) 7.2), 3.68 (1H, dd, J ) 11.7, 7.2),
3.76 (1H, dd, J ) 11.7, 4.2), 3.78-3.91 (2H, m), 4.18 (1H, ddd, J )
(24) Sambrook, J.; Fritsch, E. F.; Maniatis, T. Molecular Cloning: A
Laboratory Manual, 2nd ed.; Cold Spring Harbor Laboratory Press: Spring
Harbor, NY, 1989.
(25) Bradford, M. M. Anal. Biochem. 1976, 72, 248-254.
(26) Laemmli, U. K. Nature 1970, 227, 680-685.
(27) Andrews, P. Biochem. J. 1964, 91, 222-223.
1

9070 J. Am. Chem. Soc., Vol. 122, No. 38, 2000
Zhang and Liu
8.4, 4.2, 2.4), 4.29 (1H, dd, J ) 8.4, 7.2), 5.57 (1H, t, J ) 4.5). 13C
NMR (D2O) δ 9.8, 48.2, 63.9, 73.4, 75.1, 78.1 (d, J ) 8), 83.1, 98.4
(d, J ) 6). 31P (D2O) NMR δ 0.8.
determined by the constant product/substrate ratio by HPLC. The same
procedure was repeated with a 200 µL incubation of 1 mM UDP-Galp
and 8 µM enzyme in 100 mM potassium phosphate buffer (pH 7.6).
UDP-Galf (2). The triethylammonium salt of R-D-galactofuranosyl
phosphate (13, 200 mg, 0.43 mmol) was dried by coevaporating with
anhydrous pyridine (3 mL) a few times. To this residue was added
UMP-morpholidate (760 mg, 1.1 mmol) in anhydrous pyridine (3 mL),
and the solution was evaporated again to dryness under vacuum. This
was followed by the addition of 1H-tetrazole (100 mg, 1.4 mmol) in
anhydrous pyridine (5 mL), and the resulting solution was stirred at
room temperature for 40 h. Removal of the solvent under vacuum gave
a solid residue which was dissolved in methanol and chromatographed
on a Sephadex LH-20 column (2.5 × 120 cm) using methanol as the
eluent. Fractions containing the desired product, judging by TLC
analysis (EtOH:NH4OH:H2O ) 5:3:1), were pooled and evaporated to
dryness under reduced pressure. Compound 2 was further purified by
HPLC using a C18 column (Alltech, 10 × 250 mm). The eluent was
1.5% acetonitrile in 50 mM triethylammonium acetate buffer, pH 6.8,
and the flow rate was 5.0 mL/min. The retention time of 2 under these
conditions was 4.7 min. 1H NMR (D2O) δ 1.30 (18H, t, J ) 7.2, Et3N-
Me), 3.23 (12H, q, J ) 7.5, Et3N-CH2), 3.65 (1H, dd, J ) 11.4, 6.9,
6-H), 3.74 (1H, dd, J ) 11.7, 4.2, 6-H), 3.76-3.87 (2H, m, 4-H and
5-H), 4.17 (1H, ddd, J ) 8.4, 4.2, 2.7, 2-H), 4.21-4.33 (4H, m, 3-H,
4′-H, and 5′-H2), 4.37-4.43 (2H, m, 2′-H and 3′-H), 5.66 (1H, dd, J )
5.9, 4.2, 1-H), 6.00 (1H, d, J ) 8.4, 5′′-H), 6.01 (1H, d, J ) 3.3, 1′-H),
8.00 (1H, J ) 8.4, 6′′-H). 13C NMR (D2O) δ 9.8 (Et3N-Me), 48.2
(Et3N-CH2), 63.7 (C-6), 66.5 (d, J ) 6, C-5′), 71.3 (C-2′), 73.7 (C-5),
75.1 (C-3), 75.3 (C-3′), 78.2 (d, J ) 7, C-2), 83.3 (C-4), 84.9 (d, J )
9, C-4′), 89.9 (C-1′), 99.2 (d, J ) 6, C-1), 104.2 (C-5′′), 143.2 (C-6′′),
153.4 (C-2′′), 167.8 (C-4′′). 31P NMR (D2O) δ -12.2 (d, J ) 20.8),
-11.0 (d, J ) 20.8). Negative ion HRFABMS calcd for C15H23N2O17P2
[M - H]- 565.0472, found m/z 565.0499.
Enzyme Assay. The activity of the purified enzyme was assayed
by incubating a mixture of UDP-Galf (2, 1 mM) and an appropriate
amount of the mutase in 30 µL of 100 mM potassium phosphate buffer
(pH 7.5) for 2 min at 37 °C. Dithiothreitol was initially included in the
assay mixture but was later found unnecessary and was omitted. The
reaction was analyzed by HPLC using a C18 column (Microsorb-MV,
Varian, 4.6 × 250 mm) which was eluted with 1.5% acetonitrile in 50
mM triethylammonium acetate buffer, pH 6.8. The detector was set at
262 nm, and the flow rate was 1.0 mL/min. Baseline resolution was
achieved, and the retention times for UDP-Galf (2) and UDP-Galp (1)
were 7.8 and 6.3 min, respectively. The extent of conversion was
determined by comparing the integration of the substrate and product
peaks. The conversion was routinely controlled to be within 30% by
properly adjusting the enzyme concentration.
Preparation of 5-Deazariboflavin. 5-Deazariboflavin required for
the photoreduction experiment was synthesized according to a reported
1
procedure,28 and characterized by NMR. H NMR (DMSO-d6) δ 2.32
(3H, s), 2.44 (3H, s), 3.44 (1H, br), 3.60 (3H, br), 4.21 (1H, br), 4.46
(1H, s, exchangeable), 4.63 (1H, d, J ) 13.8), 4.78 (1H, d, J ) 5.7,
exch.), 4.87 (1H, s, exch.), 4.99 (1H, m), 5.11 (1H, d, J ) 3.9, exch.),
7.86 (1H, s), 7.94 (1H, s), 8.86 (1H, s), 11.0 (1H, s, exch.). 13C NMR
(DMSO-d6) δ 19.1, 21.4, 47.7, 63.9, 70.0, 73.4, 74.1, 114.2, 118.4,
120.1, 131.0, 134.1, 140.4, 141.5, 146.4, 156.8, 158.0, 162.7.
Photoreduction. To selectively reduce the enzyme bound flavin
coenzyme, a mixture of 180 nM mutase, 23 µM 5-deazariboflavin, and
12 mM EDTA in a total volume of 120 µL of 100 mM potassium
phosphate buffer (pH 7.6) was prepared. This reaction mixture was
made anaerobic through a Schlenk line with repeated cycles between
vacuum and argon on ice. To this mixture was added 30 µL of anaerobic
5 mM UDP-Galf solution (in 100 mM potassium phosphate buffer,
pH 7.6), and photoreduction was carried out promptly by irradiating
this sample with a slide projector light bulb (400 W, 82 V) controlled
by a variable autotransformer. The light source was approximately 10
cm away from the sample which was kept in a water bath at room
temperature. The total incubation time was 2 min, and the total
illumination time was 1 min, with 10 s cooling intervals after every 10
s illumination. Excess 5-deazariboflavin was used to ensure complete
reduction of the bound FAD. A control experiment was performed in
parallel without illumination.
Thiol Titration. All experiments were run in buffers saturated with
nitrogen gas. A typical experiment involved the addition of the
denaturing solution consisting of 1 mM EDTA and 6.4 M guanidine
hydrochloride in 80 µL of 100 mM potassium phosphate buffer (pH
7.3) to a solution of the mutase in 10 µL of 50 mM phosphate buffer
(pH 7.6). The exposed thiols in the denatured enzyme were titrated by
adding 10 µL of 10 mM 5,5′-dithiobis(2-nitrobenzoic acid) (DTNB)
to the solution and measuring the change in absorbance at 412 nm
(ꢀ412 ) 13 700 M-1 cm-1) against a control without DTNB. The
absorption of DTNB alone at 412 nm measured from a separate
experiment was subtracted. To study the effect of photoreduction of
Glf, a mixture of 110 µM enzyme, 20 µM 5-deazariboflavin, and 5
mM EDTA in 200 µL of 100 mM potassium phosphate buffer (pH
7.6) was illuminated for 15 s under anaerobic conditions, which resulted
in complete bleaching of the flavin chromophore. To this solution was
added 1.6 mL of the above denaturing solution and the exposed thiols
were similarly titrated. To determine the effect of dithionite reduction
on Glf, the purified enzyme (210 µM) was treated with 5 mM sodium
dithionite in 100 µL of potassium phosphate buffer (50 mM, pH 7.6).
To this solution was added 800 µL of the above denaturing solution
and the resulting mixture was loaded onto a 10DG desalting column
(Bio-Rad) preequilibrated with EDTA (1 mM) and guanidine hydro-
chloride (6.0 M) in potassium phosphate buffer (50 mM, pH 7.3). The
protein was eluted using the same solution. The protein fractions were
pooled, and the exposed thiols in the denatured enzyme were titrated.
For the air reoxidized enzyme, the denaturing solution was added
immediately after the bleached mutase recovered its absorption at 450
nm. The resulting mixture was desalted and thiols were similarly titrated.
For assays conducted under reducing conditions, the incubation was
performed in sealed vials in the presence of freshly prepared sodium
dithionite (20 mM). For assessing the effects of the exogenous cofactor
on catalysis, 2 mM of each cofactor being tested was included in the
incubation mixture.
Determination of Kinetic Parameters. To determine the kinetic
parameters of the native enzyme, a series of samples containing the
purified mutase (480 nM) and UDP-Galf (0.10, 0.125, 0.167, 0.25, 0.50,
and 1.0 mM) in a total volume of 30 µL of 100 mM potassium
phosphate (pH 7.6) were prepared. Each sample was incubated at 37
°C for 1.5 min and frozen with liquid nitrogen to terminate the reaction.
The contents of these samples were analyzed by HPLC as described
above, and the ratio of product and substrate peak was used to determine
the activity. The kinetic parameters were deduced by fitting these data
to the Michaelis-Menten equation. The same procedure was also used
to study the reduced enzyme. The incubation mixture in this case
contained 4.8 nM enzyme, 20 mM fresh prepared sodium dithionite,
and appropriate concentration of UDP-Galf (10, 12.5, 16.7, 25, 50, 100,
200, and 500 µM) in 100 µL of 100 mM potassium phosphate buffer
(pH 7.6).
Equilibrium Constant. The equilibrium constant, Keq, was calcu-
lated from the ratio of product to substrate at equilibrium. The product
and substrate ratios were determined for both the forward and reverse
directions by integration of the corresponding peaks from the HPLC
chromatogram. For the reverse reaction, 1 mM UDP-Galf was mixed
with 8 µM purified mutase in 200 µL of 100 mM potassium phosphate
buffer (pH 7.6). The incubation was allowed to reach equilibrium as
Acknowledgment. This work was supported in part by a
grant from the National Institutes of Health (GM54346). We
are grateful to Professor Kevin H. Mayo for allowing us to use
the CD instrument in his laboratory. H.-w. L. also thanks the
National Institute of General Medical Sciences for a MERIT
Award.
Note Added in Proof. A synthesis of 2 was recently reported
(Tsvetkov, Y. E.; Nikolaev, A. V. J. Chem. Soc. Perkin. Trans.
I 2000, 889-891).
JA001333Z
(28) Ashton, W. T.; Brown, R. D.; Tolman, R. L. J. Heterocycl. Chem.
1978, 15, 489-491.