Enzymatic formation and release of a stable glycal intermediate: The mechanism of the reaction catalyzed by UDP-N-acetylglucosamine 2-epimerase

Full text of DOI:10.1021/ja960266z

J. Am. Chem. Soc. 1996, 118, 3033-3034
Enzymatic Formation and Release of a Stable
3033
Glycal Intermediate: The Mechanism of the
Reaction Catalyzed by UDP-N-Acetylglucosamine
2-Epimerase
Rafael F. Sala, Paul M. Morgan, and Martin E. Tanner*
Department of Chemistry
UniVersity of British Columbia, VancouVer
British Columbia V6T 1Z1, Canada
ReceiVed January 25, 1996
The enzyme UDP-N-acetylglucosamine 2-epimerase catalyzes
the interconversion of UDP-N-acetylglucosamine (UDP-GlcNAc)
and UDP-N-acetylmannosamine (UDP-ManNAc) in both Gram-
positive and Gram-negative bacteria (Figure 1).1 This provides
the bacteria with a source of activated ManNAc residues for
use in the biosynthesis of cell wall surface polysaccharides.2
This epimerase is unlike most known racemases and epimerases
in that it must invert a stereogenic center that is not adjacent to
an electron-withdrawing carbonyl or carboxylate group and
therefore cannot employ a simple deprotonation-reprotonation
reaction mechanism.3 A previous report suggested that the
enzyme overcomes this obstacle by transiently oxidizing the
C-3 hydroxyl of the GlcNAc residue to a ketone, thus acidifying
the proton at C-2.4,5 Deprotonation at C-2, followed by
reprotonation on the opposite face and finally reduction of the
ketone, would produce the epimeric sugar nucleotide. In this
communication we report evidence in favor of an alternative
mechanism that proceeds via cleavage of the anomeric C-O
bond, with 2-acetamidoglucal and UDP as enzyme-bound
intermediates (Figure 1).6 We have employed a positional
isotope exchange (PIX) experiment7 in which an 18O label in
the sugar-UDP bridging position (darkened atom in Figure 1)
has been observed to scramble into nonbridging diphosphate
positions during enzymatic epimerization. We have also
demonstrated that the enzyme occasionally releases these
relatively stable intermediates into solution.
Figure 1. Proposed mechanistic scheme for the reactions catalyzed
by UDP-GlcNAc 2-epimerase. Species within brackets are enzyme
bound. U ) uridine; darkened atoms indicate 18O labels.
1
The epimerization of UDP-GlcNAc was monitored by H
NMR spectroscopy and ion-pair reversed-phase HPLC.12 Both
techniques showed the appearance of new signals attributable
to UDP-ManNAc in the proper equilibrium ratio (10:1 in favor
of UDP-GlcNAc).13 Acidic hydrolysis of the equilibrated
mixture produced a 10:1 ratio of GlcNAc:ManNAc identical
with authentic samples.14 The epimerization in D2O was
accompanied by the incorporation of deuterium into the C-2
position of both epimers. This had previously been observed
with enzyme obtained from natural sources and supports a
mechanism that ultimately involves proton transfer at C-2.4
The PIX experiment was performed on UDP-GlcNAc con-
taining an 82% incorporation of 18O label at the GlcNAc
anomeric position.15 This material was enzymatically epimer-
ized in a deuterated phosphate buffer, and the reaction progress
was followed by 1H and 31P NMR spectroscopies. The
epimerization was allowed to proceed well past completion as
indicated by the complete wash-in of deuterium at C-2. This
ensured that each molecule of UDP-GlcNAc had been handled
by the enzyme several times. The 31P chemical shift was then
The gene coding for the Escherichia coli UDP-GlcNAc
2-epimerase, known as rffE, had tentatively been assigned to
an open reading frame, o355, near min 85 on the E. coli
chromosome.8 We have found that rffE is actually located 2.4
kb upstream of this sequence and has been designated nfrC in
other work.9 The nfrC gene product has been overexpressed
and was reported to be a cytoplasmic protein of unknown
activity that is required for bacteriophage N4 adsorption.10 We
have purified this protein to homogeneity and demonstrated that
it is UDP-GlcNAc 2-epimerase.11
(8) Daniels, D. L.; Plunkett, G., III; Burland, V.; Blattner, F. R. Science
1992, 257, 771. This assignment was based on homology arguments and
on previous work that localized the gene onto a 4.9-kb fragment of DNA
containing four open reading frames: Meier-Dieter, U.; Barr, K.; Starman,
R.; Hatch, L.; Rick, P. D. J. Biol. Chem. 1992, 267, 746. The o355 gene
has subsequently been assigned to encode for dTDP-glucose-4,6-dehy-
dratase: Macpherson, D. F.; Manning, P. A.; Morona, R. Mol. Microbiol.
1993, 11, 281. Stevenson, G.; Neal, B.; Liu, D.; Hobbs, M.; Packer, N.
H.; Batley, M.; Redmond, J. W.; Lindquist, L.; Reeves, P. J. Bacteriol.
1994, 176, 4144.
(9) Kiino, D. R.; Rothman-Denes, L. B. J. Bacteriol. 1989, 171, 4595.
(10) Kiino, D. R.; Licudine, R.; Wilt, K.; Yang, D. H. C.; Rothman-
Denes, L. B. J. Bacteriol. 1993, 175, 7074.
(1) (a) Kawamura, T.; Kimura, M.; Yamamori, S.; Ito, E. J. Biol. Chem.
1978, 253, 3595. (b) Kawamura, T.; Ishimoto, N.; Ito, E. J. Biol. Chem.
1979, 254, 8457. (c) Kawamura, T.; Ishimoto, N.; Ito, E. Methods Enzymol.
1982, 83, 515.
(2) (a) Lew, H. C.; Nikaido, H.; Ma¨kela¨, P. H. J. Bacteriol. 1978, 136,
227. (b) Yoneyama, T.; Koike, Y.; Arakawa, H.; Yokoyama, K.; Sasaki,
Y.; Kawamura, T.; Araki, Y.; Ito, E.; Takao, S. J. Bacteriol. 1982, 149,
15. (c) Harrington, C. R.; Baddiley, J. Eur. J. Biochem. 1985, 153, 639.
(d) Kuhn, H.-M.; Meier-Dieter, U.; Mayer, H. FEMS Microbiol. ReV. 1988,
54, 195. (e) Meier-Dieter, U.; Starman, R.; Barr, K.; Mayer, H.; Rick, P.
D. J. Biol. Chem. 1990, 265, 13490.
(3) Adams, E. AdV. Enzymol. Relat. Areas Mol. Biol. 1976, 44, 69.
(4) Salo, W. L. Biochim. Biophys. Acta 1976, 452, 625.
(5) The mechanism necessarily invokes a tightly bound NAD+ cofactor
since no requirements for exogenous cofactors have been detected for this
enzyme.1 We have not been able to detect any bound cofactor using
standard procedures: Palmer, J. L.; Abeles, R. H. J. Biol. Chem. 1979,
254, 1217.
(11) The enzyme was purified by ion-exchange chromatography on a
Waters Protein-Pak Q column in Tris-HCl buffer at pH 7.0 followed by a
second column at pH 8.5. The protein displayed only a single band when
analyzed by SDS-PAGE. Electrospray mass spectrometry results were
consistent with the reported gene sequence: calcd, 42 246 Da; found, 42 254
Da.
(12) Meynial, I.; Paquet, V.; Combes, D. Anal. Chem. 1995, 67, 1627.
(13) UDP-ManNAc was eluted from a C-18 reversed-phase silica column
with a shorter retention time than UDP-GlcNAc under the conditions
described in ref 12. UDP-ManNAc displayed distinctive NMR signals at
5.40 ppm (anomeric proton, 0.06 ppm upfield of that of UDP-GlcNAc)
and 1.99 ppm (methyl protons, 0.04 ppm upfield of those of UDP-GlcNAc).
(14) The hydrolysis conditions (0.06 N HCl, 100 °C, 15 min) are known
not to cause epimerization.1a The sugars were found to have identical
retention times to that of authentic standards when analyzed by HPLC using
a Bio-Rad Aminex HPX-87H column with 13 mM H2SO4 as eluent (see
ref 18).
(15) An 18O label was introduced into the anomeric position of 3,4,6-
tri-O-acetyl-GlcNAc by heating the sugar in 95%-enriched H218O/CH3CN.
The R-dibenzylphosphate was then prepared by phosphitylation and
oxidation: Sim, M. M.; Kondo, H.; Wong, C. H. J. Am. Chem. Soc. 1993,
115, 2260. Complete deprotection of the sugar followed by UMP-
morpholidate coupling using standard conditions (Srivastava, G.; Alton, G.;
Hindsgaul, O. Carbohydr. Res. 1990, 207, 259) provided UDP-GlcNAc
containing an 82% enrichment of 18O in the anomeric position.
(6) A 2-acetamidoglucal intermediate has previously been implicated in
the mechanism of the mammalian enzyme that catalyzes both the inversion
of configuration at C-2 and the loss of UDP to produce free ManNAc from
UDP-GlcNAc: Sommar, K. M.; Ellis, D. B. Biochim. Biophys. Acta 1972,
268, 590.
(7) Raushel, F. M.; Villafranca, J. J. CRC Crit. ReV. Biochem. 1988, 23,
1.
0002-7863/96/1518-3033$12.00/0 © 1996 American Chemical Society

3034 J. Am. Chem. Soc., Vol. 118, No. 12, 1996
Communications to the Editor
upfield from that of the unlabeled material. It results from UDP-
GlcNAc in which the 18O label has scrambled into a nonbridging
position with a P-18O bond order greater than one. The 1:2
ratio of the upfield peaks reflects the statistical distribution
expected for bridging versus nonbridging labels. A similar
pattern is seen in the â-phosphorus signals of the minor UDP-
ManNAc product (not shown in Figure 2).
The observation of deuterium incorporation at C-2 combined
with C-O bond cleavage at the anomeric center leads us to
suspect that an intermediate glycal is formed during the reaction
(Figure 1). Further evidence for this mechanism was obtained
upon extended incubation of UDP-GlcNAc with very high
concentrations of enzyme.17 The epimeric mixture of UDP-
GlcNAc/UDP-ManNAc initially produced was observed to
slowly convert to a 1:1 mixture of free 2-acetamidoglucal and
UDP when monitored by HPLC. The glycal was isolated and
gave an identical 1H NMR spectrum to that of an independently
synthesized sample.18 No free GlcNAc or ManNAc was
detected during this process, indicating that water is not added
to the glycal at any significant rate. The external equilibrium
constant between the two epimers and the released intermediates
was shown to be >600 in favor of UDP and glycal as judged
by integration of the HPLC peaks. We estimate that the
intermediates are released approximately 1 in every 400
turnovers.19
The mechanism we propose involves a trans-elimination of
monoprotic UDP to form 2-acetamidoglucal, followed by a syn-
addition (in the UDP-GlcNAc to UDP-ManNAc direction).
Occasionally the enzyme releases these relatively stable inter-
mediates which accumulate in solution upon extended incuba-
tion. Both steps of the reaction could occur in either a concerted
fashion or stepwise via oxonium intermediates. Neighboring
group participation by either the acetamido or diphosphate
functionalities and the formation of glycosyl-enzyme intermedi-
ates could be involved in this process. Further experiments are
underway in order to distinguish between these possibilities.
Note Added in Proof: A recent paper that also identifies
the correct position of the rffE gene has appeared: Marolda, C.
L.; Valvano, M. A. J. Bacteriol. 1995, 177, 5539.
Figure 2. 31P NMR spectra of the â-phosphorus of 18O-labeled
UDP-GlcNAc: (A) before and (B) after treatment with the epimerase.
All peaks appear as doublets due to coupling to the adjacent
R-phosphorus. U ) uridine; darkened atoms indicate 18O labels.
Acknowledgment. We are grateful to Dr. Lucia B. Rothman-Denes
and Dr. Diane R. Kiino for supplying us with the plasmid-containing
nrfC. We thank Dr. S. G. Withers and Dr. E. C. K. Lai for supplying
us with samples of synthetic 2-acetamidoglucal. We thank the
University of British Columbia and the Natural Sciences and Engineer-
ing Research Council of Canada (NSERC) for financial support.
used to deduce the location of the 18O label in the enzyme-
treated UDP-GlcNAc. It is well established that substitution
of 18O-P for 16O-P causes a small upfield chemical shift in the
resulting 31P signal and that the magnitude of this shift will be
dependent on the O-P bond order.16 Figure 2A shows the 31
P
NMR signals for the â-phosphorus of the 82% 18O-labeled UDP-
GlcNAc before epimerization (the signals appear as doublets
due to coupling to the adjacent R-phosphorus nucleus). The
signal of the labeled material is shifted 0.013 ppm upfield from
that of the unlabeled material, indicative of a P-18O single
bond.16 Figure 2B shows the corresponding 31P NMR signals
following epimerization. A new signal has appeared 0.029 ppm
JA960266Z
(17) In a typical experiment, 0.08 mg of epimerase was incubated in 0.1
mL of pH 8.5 phosphate buffer containing 0.6 mM UDP-GlcNAc at 37
°C. No glycal formation (or epimerization) was observed using heat-killed
enzyme.
(18) Lai, E. C. K.; Withers, S. G. Biochemistry 1994, 33, 14743.
(19) This was based on the ratio of the reported1c specific activity of 7
units/mg for the formation of UDP-ManNAc from UDP-GlcNAc and a
measured specific activity of 0.018 units/mg for the formation of UDP and
2-acetamidoglucal from an equilibrating pool of UDP-ManNAc/UDP-
GlcNAc (as determined by HPLC analysis of the initial rate of formation
of UDP).
(16) (a) Cohn, M.; Hu, A. Proc. Natl. Acad. Sci. U.S.A. 1978, 75, 200.
(b) Singh, A. N.; Hester, L. S.; Raushel, F. M. J. Biol. Chem. 1987, 262,
2554.