Eliminations in the reactions catalyzed by UDP-N-acetylglucosamine 2-epimerase

Article abstract of DOI:10.1021/ja971718q

Mechanistic studies have been carried out on the bacterial enzyme UDP-N-acetylglucosamine 2-epimerase, which catalyzes the interconversion of UDP-N-acetylglucosamine (UDP-GlcNAc) and UDP-N-acetylmannosamine (UDP-ManNAc). This enzyme is interesting because it epimerizes a stereocenter that does not bear an acidic proton, and therefore it cannot utilize a simple deprotonation/reprotonation mechanism. A coupled enzyme assay employing UDP-ManNAc dehydrogenase has been developed. The epimerization in D₂O is found to be accompanied by the incorporation of deuterium into the C-2'' position of both epimers, supporting a mechanism that ultimately involves a proton transfer at this position. The epimerization of [2''-²H]UDP-GlcNAc is slowed by a primary kinetic isotope effect indicating that C-H bond cleavage is occurring during a rate-determining step of the reaction. A positional isotope exchange (PIX) experiment shows that an ¹⁸O label in the sugar-UDP bridging position will scramble into nonbridging diphosphate positions during enzymatic epimerization. These observations are consistent with a mechanism that proceeds via cleavage of the anomeric C-O bond, with 2-acetamidoglucal and UDP as enzyme-bound intermediates. Additional evidence for this mechanism is found in the unusual observation that during extended incubations, the intermediates are gradually released from the enzyme and accumulate in solution. These intermediates are formed by an anti elimination of UDP from UDP-GlcNAc and a syn elimination of UDP from UDP-ManNAc. It is likely that E1-like eliminations via oxocarbenium intermediates are involved in the reaction. Further experiments show that 3''-deoxy-UDP-GlcNAc is not a substrate for the enzyme and that the enzyme does not contain a tightly bound NAD⁺ cofactor.

Full text of DOI:10.1021/ja971718q

J. Am. Chem. Soc. 1997, 119, 10269-10277
10269
Eliminations in the Reactions Catalyzed by
UDP-N-Acetylglucosamine 2-Epimerase
Paul M. Morgan, Rafael F. Sala, and Martin E. Tanner*
Contribution from the Department of Chemistry, UniVersity of British Columbia,
VancouVer, British Columbia V6T 1Z1, Canada
ReceiVed May 27, 1997X
Abstract: Mechanistic studies have been carried out on the bacterial enzyme UDP-N-acetylglucosamine 2-epimerase,
which catalyzes the interconversion of UDP-N-acetylglucosamine (UDP-GlcNAc) and UDP-N-acetylmannosamine
(UDP-ManNAc). This enzyme is interesting because it epimerizes a stereocenter that does not bear an acidic proton,
and therefore it cannot utilize a simple deprotonation/reprotonation mechanism. A coupled enzyme assay employing
UDP-ManNAc dehydrogenase has been developed. The epimerization in D2O is found to be accompanied by the
incorporation of deuterium into the C-2′′ position of both epimers, supporting a mechanism that ultimately involves
a proton transfer at this position. The epimerization of [2′′-2H]UDP-GlcNAc is slowed by a primary kinetic isotope
effect indicating that C-H bond cleavage is occurring during a rate-determining step of the reaction. A positional
isotope exchange (PIX) experiment shows that an 18O label in the sugar-UDP bridging position will scramble into
nonbridging diphosphate positions during enzymatic epimerization. These observations are consistent with a
mechanism that proceeds via cleavage of the anomeric C-O bond, with 2-acetamidoglucal and UDP as enzyme-
bound intermediates. Additional evidence for this mechanism is found in the unusual observation that during extended
incubations, the intermediates are gradually released from the enzyme and accumulate in solution. These intermediates
are formed by an anti elimination of UDP from UDP-GlcNAc and a syn elimination of UDP from UDP-ManNAc.
It is likely that E1-like eliminations via oxocarbenium intermediates are involved in the reaction. Further experiments
show that 3′′-deoxy-UDP-GlcNAc is not a substrate for the enzyme and that the enzyme does not contain a tightly
bound NAD+ cofactor.
Introduction
ManNAc) (Scheme 1).10-12 Very little is known about the
mechanism of this process or how the active site is able to
accommodate two epimers that differ by the position of a
relatively bulky acetamido group. In this paper we report studies
that indicate this enzyme uses an unusual â-elimination/
readdition strategy to carry out the inversion process.13
Most enzymatic racemizations and epimerizations take place
at stereogenic carbon centers that are adjacent to electron-
withdrawing groups such as carbonyl or carboxylate function-
alities.1,2 These reactions ultimately involve a deprotonation
at the stereolabile center followed by a reprotonation in the
opposite stereochemical sense. A few epimerases are known
that act on “unactivated centers” lacking an acidic proton and
must therefore use alternative reaction mechanisms. Perhaps
the best-understood example is found in UDP-galactose 4-epi-
merase, an enzyme that employs hydride transfers instead of
proton transfers.3-8 This enzyme contains a tightly bound
NAD+ cofactor and transiently oxidizes the alcohol at the
galactose C-4 position to a ketone. Reduction of the ketone in
the opposite stereochemical sense produces UDP-glucose and
completes the epimerization process. Another example of an
epimerase that acts at an “unactivated center” is UDP-N-
acetylglucosamine 2-epimerase (UDP-GlcNAc 2-epimerase),9
which catalyzes the interconversion of UDP-N-acetylglu-
cosamine (UDP-GlcNAc) and UDP-N-acetylmannosamine (UDP-
The bacterial UDP-GlcNAc 2-epimerase provides bacteria
with a source of activated ManNAc residues for use in the
biosynthesis of a variety of surface polysaccharides. In Gram-
positive bacteria such as Staphylococcus aureus H and Bacillus
subtilis, a ManNAc residue is a component of the “linkage unit”
that serves to attach teichoic acids to the peptidoglycan.14-17
Teichoic acids are polyol phosphate polymers that can account
for as much as 60% of the dry weight of the Gram-positive
bacterial cell wall and are thought to be important both in
(9) The epimerase acts at the C-2′′ position of UDP-GlcNAc (the C-2
position of the GlcNAc moiety). Throughout the remainder of this text the
notation C-X′′ or H-X′′ will be employed in reference to atoms in the
GlcNAc or ManNAc portions of the sugar nucleotides.
(10) Kawamura, T.; Ishimoto, N.; Ito, E. Methods Enzymol. 1982, 83,
515-519.
(11) Kawamura, T.; Ishimoto, N.; Ito, E. J. Biol. Chem. 1979, 254, 8457-
8465.
X Abstract published in AdVance ACS Abstracts, October 15, 1997.
(1) Adams, E. AdV. Enzymol. Relat. Areas Mol. Biol. 1976, 44, 69-
138.
(2) Gerlt, J. A.; Kenyon, G. L.; Kozarich, J. W.; Neidhart, D. J.; Petsko,
G. A.; Powers, V. M. Curr. Opin. Struct. Biol. 1992, 2, 736-742.
(3) Thoden, J. B.; Frey, P. A.; Holden, H. M. Biochemistry 1996, 35,
2557-2566.
(4) Liu, Y. J.; Vanhooke, J. L.; Frey, P. A. Biochemistry 1996, 35, 7615-
7620.
(5) Thoden, J. B.; Frey, P. A.; Holden, H. M. Biochemistry 1996, 35,
5137-5144.
(6) Vanhooke, J. L.; Frey, P. A. J. Biol. Chem. 1994, 269, 31496-31504.
(7) Burke, J. R.; Frey, P. A. Biochemistry 1993, 32, 13220-13230.
(8) Frey, P. A. In Pyridine Nucleotide Coenzymes; Dolphin, D., Poulson,
R., Avramovic, O., Eds.; Wiley: New York, 1987; Vol. 2B, pp 461-477.
(12) Kawamura, T.; Kimura, M.; Yamamori, S.; Ito, E. J. Biol. Chem.
1978, 253, 3595-3601.
(13) A preliminary communication describing portions of this work has
appeared: Sala, R. F.; Morgan, P. M.; Tanner, M. E. J. Am. Chem. Soc.
1996, 118, 3033-3034.
(14) Yokoyama, K.; Mizuguchi, H.; Araki, Y.; Kaya, S.; Ito, E. J.
Bacteriol. 1989, 171, 940-946.
(15) Yokoyama, K.; Miyashita, T.; Araki, Y.; Ito, E. Eur. J. Biochem.
1986, 161, 479-489.
(16) Harrington, C. R.; Baddiley, J. Eur. J. Biochem. 1985, 153, 639-
645.
(17) Yoneyama, T.; Koike, Y.; Arakawa, H.; Yokoyama, K.; Sasaki, Y.;
Kawamura, T.; Araki, Y.; Ito, E.; Takao, S. J. Bacteriol. 1982, 149, 15-
21.
S0002-7863(97)01718-6 CCC: $14.00 © 1997 American Chemical Society

10270 J. Am. Chem. Soc., Vol. 119, No. 43, 1997
Morgan et al.
Scheme 1. Reactions Catalyzed by UDP-GlcNAc
2-Epimerase and UDP-ManNAc 6-Dehydrogenase
Scheme 2. Salo’s Proposed Mechanism for the Reaction
Catalyzed by UDP-GlcNAc 2-Epimerase34
maintaining proper surface cation concentrations and in cell
division.18,19 In Enterobacteriaceae (Gram-negative), such as
Escherichia coli, N-acetylmannosaminuronic acid (ManNAcUA)
residues are present in the enterobacterial common surface
antigen (ECA).20 These ManNAcUA residues are derived from
a pool of UDP-ManNAcUA that is generated by the sequential
action of UDP-GlcNAc 2-epimerase and UDP-ManNAc 6-de-
hydrogenase on UDP-GlcNAc (Scheme 1).21-23 N-Acetylm-
annosamine is also found as a component of the polysaccharide
capsule in several strains of encapsulated bacteria (both Gram-
positive and Gram-negative).24-26 In the case of Streptococcus
pneumoniae types 19F and 19A, this capsule allows the bacteria
to evade the immune system of the host and is thereby
responsible for the virulence of these strains.24,27 It has recently
been shown that the capsular polysaccharide gene cluster of S.
pneumoniae type 19F contains a gene that encodes for a
functional UDP-GlcNAc 2-epimerase.28
In mammals an enzyme that is also known as UDP-N-
acetylglucosamine 2-epimerase has been identified.29-33 The
classification of this enzyme as an epimerase is not entirely
rigorous since the mammalian enzyme catalyzes the formation
of free ManNAc and UDP from UDP-GlcNAc in an essentially
irreversible manner. The ManNAc formed in this reaction is
ultimately used in the synthesis of sialic acid. Studies on this
enzyme have been somewhat hampered by its instability;
however, it appears likely that certain mechanistic elements will
be common between the two types of enzymes.29
about 38 000 Da.10,11 Its activity was unaffected by the addition
of either NAD+ or NADP+ indicating that if one of these
cofactors were required it must be bound tightly enough to
survive the purification procedure. Changes in the rate of
epimerization as a function of UDP-GlcNAc concentration
showed allosteric behavior with a Hill coefficient of 2.0. In
addition, the reaction in the reverse direction (UDP-ManNAc
f UDP-GlcNAc) was shown to have an absolute requirement
for the presence of UDP-GlcNAc. Together these observations
were taken to indicate that the enzyme contains a modulator
site which is distinct from the active site and is highly specific
for UDP-GlcNAc.
Previous mechanistic studies were limited to the observation
that solvent-derived tritium was incorporated into the C-2′′
position during epimerization in tritium-enriched water.34 This
supports a mechanism that ultimately involves the removal of
a proton at C-2′′ and argues against a direct hydride-transfer
process such as that employed by UDP-glucose 4-epimerase.
An alternative mechanism was proposed by Salo that involves
transient oxidation at C-3′′ by a tightly bound NAD+ cofactor
(Scheme 2).34 This serves to acidify the proton at C-2′′ and
permits the stereochemical inversion to occur via a deprotona-
tion/reprotonation process. Reduction of the ketone at C-3′′
regenerates the NAD+ cofactor and produces the epimeric sugar
nucleotide.
The E. coli UDP-GlcNAc 2-epimerase has been purified and
was reported to be a dimer with a subunit molecular weight of
In this paper we report the identification of the gene encoding
for E. coli UDP-GlcNAc 2-epimerase and the purification of
the overexpressed gene product. We present experimental
evidence that supports an alternative elimination/readdition
mechanism (Scheme 3). In the UDP-GlcNAc-to-UDP-ManNAc
direction, an anti elimination produces the intermediates 2-ac-
etamidoglucal and UDP. A subsequent syn addition generates
the epimeric product. The proposed mechanism is further
supported by the observation that the intermediates are gradually
released into solution and can be isolated and identified.
(18) Heptinstall, S.; Archibald, A. R.; Baddiley, J. Nature 1970, 225,
519-521.
(19) Baddiley, J. Acc. Chem. Res. 1970, 3, 98-105.
(20) Kuhn, H.-M.; Meier-Dieter, U.; Mayer, H. FEMS Microbiol. ReV.
1988, 54, 195-222.
(21) Meier-Dieter, U.; Barr, K.; Starman, R.; Hatch, L.; Rick, P. D. J.
Biol. Chem. 1992, 267, 746-753.
(22) Meier-Dieter, U.; Starman, R.; Barr, K.; Mayer, H.; Rick, P. D. J.
Biol. Chem. 1990, 265, 13490-13497.
(23) Lew, H. C.; Nikaido, H.; Ma¨kela¨, H. J. Bacteriol. 1978, 136, 227-
233.
(24) Lee, C.-J.; Banks, S. D.; Li, J. P. Crit. ReV. Microbiol. 1991, 18,
89-114.
(25) Jann, B.; Jann, K. Curr. Top. Microbiol. Immunol. 1990, 150, 19-
42.
Experimental Section
General Methods. Uridine 5′-(2′′-acetamido-2′′-deoxy-R-D-man-
nopyranosyl diphosphate) was prepared from 2-acetamido-2-deoxy-R-
D-mannopyranosyl monophosphate by the method of Salo and Fletch-
er.35 2-Acetamido-2-deoxy-R-D-mannopyranosyl monophosphate was
prepared by the method of Yamazaki et al.36 The synthesis of 3′′-
deoxy-UDP-GlcNAc was achieved via a slight modification of the route
described by Hindsgaul et al.37,38 Protein concentrations were deter-
(26) Moxon, E. R.; Kroll, J. S. Curr. Top. Microbiol. Immunol. 1990,
150, 65-86.
(27) Lee, C.-J.; Fraser, B. A. J. Biol. Chem. 1980, 255, 6847-6851.
(28) Morona, J. K.; Morona, R.; Paton, J. C. Mol. Microbiol. 1997, 23,
751-763.
(29) Sommar, K. M.; Ellis, D. B. Biochim. Biophys. Acta 1972, 268,
590-595.
(30) Sommar, K. M.; Ellis, D. B. Biochim. Biophys. Acta 1972, 268,
581-589.
(34) Salo, W. Biochim. Biophys. Acta 1976, 452, 625-628.
(35) Salo, W. L.; Fletcher, H. G., Jr. Biochemistry 1970, 9, 878-881.
(36) Yamazaki, T.; Warren, C. D.; Herscovics, A.; Jeanloz, R. W. Can.
J. Chem. 1981, 59, 2247-2252.
(31) Salo, W. L.; Fletcher, H. G. Biochemistry 1970, 9, 882-885.
(32) Spivak, C. T.; Roseman, S. Methods Enzymol. 1966, 9, 612-615.
(33) Glaser, L. Biochim. Biophys. Acta 1960, 41, 534-536.

Eliminations in Reactions by UDP-GlcNAc
J. Am. Chem. Soc., Vol. 119, No. 43, 1997 10271
dehydrogenase reaction). Kinetic parameters were determined from
initial velocities by a direct fit of the data to a Hill equation using the
computer program GraFit.38
Scheme 3. Proposed Glycal Mechanism for the Reactions
Catalyzed by UDP-GlcNAc 2-Epimerasea
The requirement for UDP-GlcNAc as a specific activator of the
epimerase was investigated in the following manner. A solution of
UDP-ManNAc (1.1 mM) in potassium phosphate buffer (40 µL, 50
mM, pH 8.1, containing 2 mM dithiothreitol) was incubated for 10
min at 37 °C with the epimerase (2.6 × 10-4 units). An analogous
solution that also contained UDP-GlcNAc (0.35 mM) was treated in
the same fashion. The reactions were analyzed by ion-paired reversed-
phase HPLC.42
Solvent Deuterium Isotope Incorporation. A sample of UDP-
GlcNAc (15.4 mM) in Trien-HCl buffer (50 mM, pH 8.5) was
lyophilized and resuspended in an equal volume of D2O two times. To
450 µL of this “deuterated” buffer was added epimerase in storage
buffer (10 µL, 4 mg/mL, 0.27 unit). The sample was kept at 25 °C,
1
and H NMR spectra (400 MHz) were collected at timed intervals.
Kinetic Isotope Effect Determination. A 20-mL sample of Tris-
HCl buffer (50 mM, pH 8.8, containing 2 mM dithiothreitol) was
lyophilized and resuspended in an equal volume of D2O two times. To
this “deuterated” buffer were added UDP-GlcNAc (100 mg, 7.7 mM)
and epimerase in the same buffer (400 µL, 3.5 mg/mL, 9.5 units). The
reaction was kept at 37 °C for 2 h, and then it was quenched by heating
at 50 °C for 1 h and filtering the mixture through a centricon
concentrator (Amicon). To the resulting filtrate were added NAD+
(15 mg) and UDP-ManNAc dehydrogenase (1 mg, 12 units). The
solution was incubated at 37 °C for 12 h and analyzed by ion-paired
reversed-phase HPLC to ensure that all of the UDP-ManNAc was
consumed. The product [2H-2′′]UDP-GlcNAc was purified using a
column of Dowex AG1 X8 (20 mL, formate form, 100-200 mesh)
and eluting with a linear gradient of 0-2 M NEt3H2CO3 (400-mL total
volume). Fractions containing the product were lyophilized, resus-
pended in water, and lyophilized again. The product was passed
through a column of Amberlite IR-120(plus) resin (20 mL, Na+ form,
eluted with water) followed by a column of Bio-Gel P-2 (2.5 × 45
cm, eluted with water). 1H NMR spectral and mass spectral analyses
indicated that the extent of deuterium incorporation was >97%:
a Species within the brackets are enzyme-bound.
mined by the method of Bradford,39 using bovine serum albumin as
the standard. A unit is defined as the amount of epimerase that produces
1 µmol of UDP-ManNAc/min under standard assay conditions with 4
mM UDP-GlcNAc.
Purification and Identification of UDP-GlcNAc 2-Epimerase.
The rffE (or nfrC) gene product was overexpressed in E. coli JM109
(DE3, pKI86).40 The resulting protein was purified to homogeneity
(as indicated by a single band on a SDS-PAGE gel) by ion-exchange
chromatography over a short plug of DE-52 followed by an HPLC
separation using a Waters AP-1 Protein-Pak Q column (50 mM Trien-
HCl, pH 8.5, containing 10% glycerol and 2 mM dithiothreitol, 0-0.2
M NaCl gradient). See Supporting Information for full details.
Products of the epimerization reaction were characterized by heating
the epimers (3 mM total concentration) in 0.06 N HCl at 100 °C for
15 min.12 The resulting free sugars were analyzed41 using a Bio-Rad
Aminex HPX-87H column (13 mM H2SO4 as eluent) and compared
with authentic standards. Control reactions confirmed that epimeriza-
tion did not occur during this analysis.
Kinetics and Activation of the Epimerase Reaction. The kinetic
parameters of the overexpressed protein were obtained in the UDP-
GlcNAc-to-UDP-ManNAc direction with a continuous coupled assay
that employs UDP-ManNAc dehydrogenase.48 Very high concentra-
tions of the dehydrogenase were required in order to ensure complete
coupling and to eliminate a “lag phase” that was observed at lower
enzyme concentrations. Rates were determined by following the
increase in absorbance at 340 nm (NADH formation) in assay mixtures
(600-µL total volume) that contained 50 mM Tris-HCl (pH 8.8), 2 mM
dithiothreitol, 4.0 mM NAD+, 5.8 × 10-3 units of epimerase, and 60
units of UDP-ManNAc dehydrogenase (5 mg of protein in each cuvette)
at 37 °C. A background rate that was due to minor impurities in the
dehydrogenase preparation was observed and was found to be
independent of substrate concentration. In each kinetic run the rates
were determined before and after initiation with UDP-GlcNAc (0.07-
3.1 mM). The background rate was then subtracted from the final rate
(it typically amounted to 15% of Vmax), and the resulting value was
divided by a factor of 2 (to account for the stoichiometry of the
2
-LSIMS (thioglycerol) m/z 629 (M - H+(monosodium salt), H-2′′,
100).
The rates of epimerization were determined in triplicate for the
labeled and unlabeled substrate (both at 2.5 mM) using the coupled
assay described above. Substrate concentrations were determined using
A262 with ꢀ ) 8700 M-1 cm-1
.
Positional Isotope-Exchange (PIX) Experiment. (i) Synthesis of
Uridine 5′-(2′′-Acetamido-2′′-deoxy-[1′′-18O]-r-D-glucopyranosyl
diphosphate), 18O-Labeled UDP-GlcNAc. 2-Acetamido-3,4,6-tri-
O-acetyl-2-deoxy-[1-18O]-R-D-glucopyranose. 2-Acetamido-3,4,6-tri-
O-acetyl-2-deoxy-R-D-glucopyranose43 (620 mg, 1.8 mmol) was dis-
solved in 1.0 mL of dry CH3CN containing 0.4 mL of 95% enriched
H218O. The solution was placed in a sealed vial and heated at 90-95
°C for 17 h. The solvent was removed in vacuo, and the resulting
brown syrup was chromatographed on silica gel (acetone-hexane, 1:1)
to give the labeled compound (445 mg, 71%). 1H and 13C NMR spectra
were identical to those reported for the unlabeled compound.43 The
extent of 18O incorporation was determined to be 89% by mass spectral
analysis: +DCIMS (NH3) m/z 350 (M + H+,18O, 100), 348 (M + H+,
16O, 11.3), 330 (M + H+ - H216/18O, 71.3).
Disodium Uridine 5′-(2′′-Acetamido-2′′-deoxy-[1′′-18O]-r-D-glu-
copyranosyl-1′′-18O diphosphate), 18O-Labeled UDP-GlcNAc. 2-Ac-
etamido-3,4,6-tri-O-acetyl-2-deoxy-[1-18O]-R-D-glucopyranose was diben-
zylphosphorylated using published procedures.44 The resulting
2-acetamido-3,4,6-tri-O-acetyl-2-deoxy-[1-18O]-R-D-glucopyranosyl diben-
zylphosphate was fully deprotected and subjected to uridine mono-
phosphomorpholidate coupling using standard conditions.37 The 18O
1
labeled UDP-GlcNAc prepared in this fashion gave H and 13C NMR
spectra that were identical with unlabeled material. The extent of 18
O
(37) Srivastava, G.; Alton, G.; Hindsgaul, O. Carbohydr. Res. 1990, 207,
259-276.
incorporation was determined to be 81% by mass spectral analysis:
(38) See Supporting Information.
(42) Meynial, I.; Paquet, V.; Combes, D. Anal. Chem. 1995, 67, 1627-
1631.
(43) Sebesan, S.; Neira, S. Carbohydr. Res. 1992, 223, 169-185.
(44) Sim, M. M.; Kondo, H.; Wong, C.-H. J. Am. Chem. Soc. 1993,
115, 2260-2267.
(39) Bradford, M. M. Anal. Biochem. 1976, 72, 248-254.
(40) Kiino, D. R.; Licudine, R.; Wilt, K.; Yang, D. H. C.; Rothman-
Denes, L. B. J. Bacteriol. 1993, 175, 7074-7080.
(41) Lai, E. C. K.; Withers, S. G. Biochemistry 1994, 33, 14743-14749.

10272 J. Am. Chem. Soc., Vol. 119, No. 43, 1997
Morgan et al.
+LSIMS (thioglycerol) m/z 654 (M + H+,18O, 100), 652 (M + H+,16O,
23.3). The small decrease in isotopic enrichment observed during this
procedure appears to occur during the dibenzylphosphorylation step.
The location of the 18O label was confirmed by 31P NMR spectros-
copy: 31P NMR (D2O) δ -12.726 (d, JP-P ) 21.0 Hz, 0.2P, â-P-16O),
-12.739 (d, JP-P ) 21.0 Hz, 0.8P, â-P-18O), -11.014 (d, JP-P ) 21.0
Hz, 1P, R-P).
dehydrogenase gene, rffD, are found in the gene cluster that
controls the biosynthesis of the enterobacterial common antigen
(ECA).21,22 The location of the epimerase gene within this
cluster had been tentatively assigned45 based on homology
arguments; however, we13 and others46 began to suspect that
this assignment was incorrect. Ultimately the literature38
indicated that the epimerase was likely one and the same as a
cytosolic protein of unknown function that is required for
bacteriophage N4 adsorption.40,47 The gene encoding for this
protein had been given the name nfrC, and it had been cloned
and overexpressed in E. coli. Drs. Diane R. Kiino and Lucia
B. Rothman-Denes were kind enough to send us a sample of
their pET11a expression vector containing the nfrC gene. We
have purified the gene product and demonstrated that the NfrC
protein is actually an active UDP-GlcNAc 2-epimerase (RffE).
The relationship between this protein and the bacteriophage
adsorption process is likely due to the requirement of UDP-
ManNAc in the ECA biosynthetic pathway. Additional studies
in our laboratory have also unambiguously identified the gene
that encodes for a functional UDP-ManNAc dehydrogenase
(RffD).48
(ii) Scrambling Experiment. A solution of 18O-labeled UDP-
GlcNAc (450 µL, 17 mM) in deuterated phosphate buffer (200 mM,
pD 8.0-8.2) was prepared. The sample was placed in an NMR tube,
Chelex-100 resin was added (20 mg of 200-400 mesh, Na+ form,
1
previously rinsed with D2O), and both H and 31P NMR spectra were
collected. A solution of UDP-GlcNAc 2-epimerase (1.3 units in 200
µL of the same buffer containing 20 mM dithiothreitol) was added,
and the solution was incubated for 8 h at 25 °C. 1H and 31P NMR
spectra were collected. The high-resolution, proton-decoupled, 31P
NMR spectra were obtained on a Bruker AC-200E spectrometer
operating at a frequency of 81 MHz. Acquisition parameters were
sweep width ) 1620 Hz, acquisition time ) 10 s, delay between pulses
) 0.4 s, and pulse width ) 3 µs.
Enzymatic Production of 2-Acetamidoglucal and UDP. Enzyme-
produced 2-acetamidoglucal was identified by incubating a solution of
UDP-GlcNAc (38 mM) in potassium phosphate buffer (50 mM, 1 mL,
pH 8.8) containing 2 mM dithiothreitol and 8.8 units of epimerase at
37 °C for 12 h. The resulting sample was applied to a column of
Dowex AG1 X8 (20 mL, formate form, 100-200 mesh), eluted with
water (200 mL), and then lyophilized to dryness. The sample was
Purification and Identification of UDP-GlcNAc 2-Epime-
rase. The epimerase gene was overexpressed in E. coli as
outlined previously.40 The resulting protein was purified to
homogeneity by ion-exchange chromatography, and its mass
was shown (electrospray mass spectrometry) to be consistent
with that expected from the gene sequence: calcd, 42 246 Da;
found, 42 254 Da. Enzyme prepared in this fashion therefore
retains its N-terminal methionine residue. In initial studies, ion-
paired reversed-phase HPLC was used to detect the activity of
the epimerase.42 Treatment of UDP-GlcNAc with the epimerase
led to the generation of a new peak that was assigned to the
product UDP-ManNAc (see Figure 3, traces A and B). Integra-
tion of the peaks gave a 10:1 ratio of UDP-GlcNAc:UDP-
ManNAc, consistent with the reported equilibrium constant.11
In order to confirm that the product was the expected epimer,
the mixture was subjected to acidic hydrolysis under conditions
known to cleave the glycosidic bond but not to cause further
epimerization.12 The resulting hydrolysate was shown to contain
a 10:1 ratio of GlcNAc:ManNAc by HPLC analysis41 with
comparison to known standards. In subsequent studies authentic
UDP-ManNAc was synthesized according to the method of Salo
and Fletcher.35 This allowed for the direct identification of the
enzyme product as UDP-ManNAc by comparison of the HPLC
retention times. These experiments clearly demonstrated that
our preparation contained substantial epimerase activity.
1
redissolved in D2O, and a H NMR spectrum was collected.38 This
was compared to an authentic sample of 2-acetamidoglucal in D2O.
The rate of UDP/2-acetamidoglucal formation was determined by
incubating solutions (1.0 mL) containing 50 mM Tris-HCl buffer, pH
8.8, 2 mM dithiothreitol, 9.2 mM UDP-GlcNAc, and 0.88 unit of
epimerase at 37 °C. Aliquots (50 µL) were removed at timed intervals
over the course of 120 min and analyzed by ion-paired reversed-phase
HPLC.42 A linear gradient of CH3CN (0-10%) in 50 mM potassium
phosphate buffer, pH 7.0, containing 2.5 mM tetrabutylammonium
hydrogen sulfate was used to separate the components. The concentra-
tion of UDP was determined by integration of the corresponding peak
and comparison to a previously obtained linear calibration curve. Linear
kinetics was obtained during this analysis, and it was found that 10%
of the epimeric UDP-sugars had converted to UDP and 2-acetamido-
glucal.
An attempt to determine the external equilibrium constant was made
by incubating a solution of UDP-GlcNAc (17 mM, 1 mL) in Tris-HCl
buffer (50 mM, pH 8.8, containing 2 mM dithiothreitol) with the
epimerase (51 units) at 37 °C. Aliquots were removed at timed intervals
over the course of 48 h and analyzed by HPLC as described above.
The concentration of 2-acetamidoglucal was taken to be equivalent to
that of UDP. The equilibrium position was calculated using the ratio
of the appropriate peak areas with the assumption that all the nucleotides
had an identical A262. Control reactions containing heat-killed epimerase
were run to ensure that significant amounts of UDP were not
spontaneously formed under these conditions.
Studies with 3′′-Deoxy-UDP-GlcNAc. A solution (0.6 mL) of 3′′-
deoxy-UDP-GlcNAc (9.6 mM) and the epimerase (20 units) in a
deuterated potassium phosphate buffer (200 mM, pD 8.2, containing 2
mM dithiothreitol) was incubated at 37 °C and monitored by 1H NMR
spectroscopy over a period of 12 h. Analogous solutions that also
contained either 8.0 or 0.5 mM UDP-GlcNAc were analyzed in the
same manner.
The requirement for 3′′-deoxy-UDP-GlcNAc as a specific activator
of the epimerase was investigated in the following manner. A solution
of UDP-ManNAc (1.1 mM) and 3′′-deoxy-UDP-GlcNAc (0.35 mM)
in potassium phosphate buffer (40 µL, 50 mM, pH 8.1, containing 2
mM dithiothreitol) was incubated for 10 min at 37 °C with the
epimerase (2.6 × 10-4 unit). The reaction mixture was analyzed by
ion-paired reversed-phase HPLC.42
The enzyme UDP-ManNAc dehydrogenase catalyzes the
effectively irreversible 2-fold oxidation of UDP-ManNAc to
UDP-ManNAcUA with the production of 2 equiv of NADH
(Scheme 1).11 We have overexpressed this enzyme in E. coli48
and used it to develop a continuous spectrophotometric assay
of UDP-GlcNAc 2-epimerase activity. This assay allowed us
(45) Daniels, D. L.; Plunkett, G., III; Burland, V.; Blattner, F. R. Science
1992, 257, 771-777.
(46) Marolda, C. L.; Valvano, M. A. J. Bacteriol. 1995, 177, 5539-
5546.
(47) Kiino, D. R.; Rothman-Denes, L. B. J. Bacteriol. 1989, 171, 4595-
4602.
(48) We have subcloned the rffD gene encoding the dehydrogenase21
(o379 in ref 45) into a pET11a expression vector giving the plasmid pUS01.
High levels of active dehydrogenase are expressed in E. coli JM109 (DE3,
pUS01) following induction by IPTG. The dehydrogenase used in this study
was purified by ion-exchange chromatography and determined to be 90%
homogeneous (as analyzed by SDS-PAGE). It should be noted that a
sequencing error45 led to the incorrect prediction that the rffD gene contains
379 codons. The correct sequence40 indicates that it contains 420 codons.
Full details will be reported elsewhere (Morgan, P. M., Tanner, M. E.,
manuscript in preparation).
Results
Identification of the rffE and rffD Genes. In E. coli the
UDP-GlcNAc 2-epimerase gene, rffE, and the UDP-ManNAc

Eliminations in Reactions by UDP-GlcNAc
J. Am. Chem. Soc., Vol. 119, No. 43, 1997 10273
to determine that the specific activity of our purified enzyme
was 6.8 units/mg. This agrees quite well with the specific
activity reported for the enzyme isolated from natural sources
(7.08 units/mg)11 and indicates that the previous observations
are not simply due to a contamination of endogenous epimerase.
Kinetics and Activation of the Epimerase Reaction. The
initial velocity of the UDP-GlcNAc epimerization reaction was
measured as a function of UDP-GlcNAc concentration using
the coupled assay. The data gave a sigmoidal curve with a Hill
coefficient of 2.29, an apparent KM value of 0.73 mM, and a
kcat value of 4.8 s-1 38
Similar kinetic constants and positive
.
cooperativity were also observed with the enzyme obtained from
natural sources.11
In order to re-examine the report11 of an absolute requirement
for UDP-GlcNAc in the epimerization of UDP-ManNAc, we
used synthetic UDP-ManNAc that contained no traces of the
epimeric sugar nucleotide. Solutions containing 1.1 mM UDP-
ManNAc were incubated in the presence and absence of 0.35
mM UDP-GlcNAc as an activator (0.6 mM UDP-GlcNAc was
reported to give half-maximal activation). In the sample lacking
UDP-GlcNAc, no evidence of epimerization could be detected,
whereas in a comparable sample containing the activator 28%
of the UDP-ManNAc had been converted to UDP-GlcNAc.
Even when the incubation time of the pure UDP-ManNAc
sample was increased 10-fold, less than 1% conversion was
observed. This study confirms the previous report that the
epimerase is tightly regulated by its substrate. We estimate that
the rate of UDP-ManNAc epimerization was reduced by at least
300-fold in the absence of the activator. The specificity of the
activation process suggests that the enzyme possesses a modula-
tor site which is distinct from the active site and is specific for
UDP-GlcNAc; however, one cannot rule out a model in which
one subunit active site must be occupied by UDP-GlcNAc for
the other subunit to function.
Figure 1. 1H NMR spectra obtained during the enzymatic epimerization
of UDP-GlcNAc in D2O. Signals shown are from H-1′′ of the sugar
nucleotides: species A is UDP-GlcNAc (R1 ) H, R2 ) NHAc); species
A′ is deuterated UDP-GlcNAc (R1 ) D, R2 ) NHAc); species B′ is
deuterated UDP-ManNAc (R1 ) NHAc, R2 ) D).
effect kH/kD was found to be 1.8 ( 0.1. There is clearly a
primary isotope effect that slows the epimerization of the
deuterated substrate, indicating that the C-2′′-H bond is broken
during a rate-determining step of the reaction.49
Positional Isotope Exchange (PIX) Experiment. A key
difference between Salo’s mechanism (Scheme 2) and the glycal
mechanism (Scheme 3) is that the bond between the â-phosphate
of the UDP moiety and the C-1 of the hexosamine residue is
transiently broken in the latter process only. In order to probe
for the intermediacy of enzyme-bound UDP, we have employed
a PIX experiment.50 In this experiment UDP-GlcNAc contain-
ing an 18O label at the sugar-UDP bridging position is incubated
with the epimerase, and the epimeric products are examined
for evidence of isotopic scrambling (Scheme 4). If UDP is
formed and has a lifetime comparable to, or greater than, that
required for bond rotation to occur in the terminal phosphate,
then scrambling of the isotope into the nonbridging phosphate
positions should be observed in both of the recovered epimers.
The synthesis of the labeled UDP-sugar began with the
preparation of 3,4,6-triacetyl-2-acetamido-2-deoxy-D-glucose
containing 18O at the anomeric position. The 18O label was
introduced directly into the compound by heating it in H218O/
acetonitrile. The labeled material was converted to the R-diben-
zylphosphate derivative44 and then fully deprotected and coupled
to UMP under known conditions.37 The product UDP-GlcNAc
was shown to contain an 81% incorporation of 18O isotope, and
the position of the label was confirmed using the 31P NMR
signals of the R-phosphorus nuclei (Figure 2A; the signals
appear as a doublet due to coupling with the â-phosphorus).
An isotopic substitution of 18O for 16O is known to cause a
small upfield shift in the 31P NMR signal of phosphates.51,52
Solvent Deuterium Isotope Incorporation. Salo reported
that the enzymatic epimerization of UDP-GlcNAc in tritium-
enriched water was accompanied by incorporation of tritium at
the C-2′′ position.34 In order to confirm this observation with
the recombinant epimerase, we epimerized UDP-GlcNAc in
D2O and followed the process by 1H NMR spectroscopy (Figure
1). The signal corresponding to the proton at the C-2′′ position
of UDP-GlcNAc is somewhat obscured by signals from the
ribose moiety. It is therefore easier to detect the deuterium
incorporation process by observing changes in the coupling
pattern of the anomeric proton (at C-1′′). The anomeric proton
appears as a doublet of doublets due to coupling both to the
â-phosphorus of UDP and to the proton at C-2′′. Deuterium
incorporation at C-2′′ will cause this signal to collapse to a
1
doublet. This is clearly seen in the H NMR spectra and is
accompanied by the appearance of a new doublet due to the
minor, C-2′′-deuterated epimer, UDP-ManNAc. The observa-
tion of deuterium incorporation supports a mechanism involving
breakage of the carbon-hydrogen bond at C-2′′ via removal of
a proton, as opposed to a hydride.
Primary Kinetic Isotope Effect Measurement. The solvent
isotope incorporation at C-2′′ allowed us to prepare [2′′-2H]-
UDP-GlcNAc and measure the deuterium kinetic isotope effect
on the rate of UDP-GlcNAc epimerization. A sample of UDP-
GlcNAc was enzymatically epimerized in D2O to give a mixture
of the C-2′′-deuterated epimers. The residual [2′′-2H]UDP-
ManNAc was oxidized by treatment with UDP-ManNAc
dehydrogenase and NAD+, and the labeled substrate (>97%
isotopic enrichment) was isolated by ion-exchange chromatog-
raphy. Kinetics for labeled and unlabeled substrates (both at
2.5 mM) were measured using the coupled assay, and the isotope
(49) Klinman, J. P. In Transition States of Biochemical Processes;
Gandour, R. D., Schowen, R. L., Eds.; Plenum Press: New York, 1978;
pp 165-200.
(50) Raushel, F. M.; Villafranca, J. J. CRC Crit. ReV. Biochem. 1988,
23, 1-26.
(51) Hester, L. S.; Raushel, F. M. Biochemistry 1987, 26, 6465-6471.
(52) Cohn, M.; Hu, A. Proc. Natl. Acad. Sci. U.S.A. 1978, 75, 200-
203.

10274 J. Am. Chem. Soc., Vol. 119, No. 43, 1997
Morgan et al.
Figure 2. 31P NMR spectra (D2O) of the â-phosphorus of 18O-labeled sugar nucleotides: (A) UDP-GlcNAc before treatment with the epimerase
and (B) equilibrated mixture of UDP-GlcNAc and UDP-ManNAc after treatment with the epimerase. All peaks appear as doublets due to coupling
to the adjacent R-phosphorus. A prime indicates signals due to UDP-ManNAc. U ) uridine; darkened atoms indicate 18O labels.
Scheme 4. PIX Experiment That Can Test for C-O Bond
signals of the R-phosphorus nuclei. The appearance of a new
doublet in the spectrum of UDP-GlcNAc confirms that PIX did
occur (Figure 2B). This doublet is shifted 0.029 ppm upfield
from that of the unlabeled material and can be assigned to a
positional isomer with a P-18O bond order greater than 1.
Integration of the upfield signals shows the expected 2:1 ratio
that reflects the statistical distribution of the label between the
bridging and nonbridging positions. The analogous 31P NMR
signals of the minor UDP-ManNAc epimer display an identical
pattern as expected. The observation of isotopic scrambling
supports the notion that the reaction proceeds with C-O bond
cleavage and involves the formation of UDP as an intermediate.
Cleavage in the Epimerization Reactiona
Enzymatic Production of 2-Acetamidoglucal and UDP.
Further insights into the nature of the reaction intermediates
could be obtained by monitoring the epimerization during
prolonged incubations with high concentrations of enzyme. An
HPLC analysis of such a reaction showed a gradual decrease
in the peaks due to the epimeric UDP-sugars and a correspond-
ing production of two new peaks that had identical retention
times to those of UDP and synthetic 2-acetamidoglucal42,53
(Figure 3). Control reactions with heat-killed epimerase were
used to confirm that this process was enzyme-catalyzed. In
order to confirm that 2-acetamidoglucal was produced, the
reaction mixture was applied to an anion-exchange column that
retained all phosphorylated sugars but allowed neutral species
to pass through. Lyophilization of the resulting eluent and
a U ) uridine; darkened atoms indicate 18O labels.
The observed upfield shift of 0.013 ppm in the labeled material
is consistent with that expected for a P-18O single bond, and
the integrals of the signals are consistent with those expected
for an 81% incorporation.
The enzymatic epimerization of the labeled substrate was
carried out in a deuterated phosphate buffer and monitored using
1H NMR spectroscopy. The reaction was allowed to proceed
well past completion as indicated by the complete exchange of
the protons at C-2′′ with solvent deuterium. The isotopic
scrambling process was then investigated using the 31P NMR
1
analysis by H NMR spectroscopy clearly showed that 2-ac-
etamidoglucal had been produced.38 No signals attributable to
other neutral carbohydrates such as GlcNAc or ManNAc were
observed indicating that the glycal is not significantly hydrated
(53) Pravdic, N.; Fletcher, H. G., Jr. J. Org. Chem. 1967, 32, 1806-
1810.

Eliminations in Reactions by UDP-GlcNAc
J. Am. Chem. Soc., Vol. 119, No. 43, 1997 10275
Scheme 5. Potential Hybrid Mechanism for the Reaction
Catalyzed by UDP-GlcNAc 2-Epimerasea
a The structure of 3′′-deoxy-UDP-GlcNAc is shown in the inset.
reaction intermediates and that the enzyme is simply “sloppy”
and occasionally releases them into solution (Scheme 3). An
alternative explanation invokes a hybrid mechanism that com-
bines elements of both previously discussed possibilities. The
hybrid mechanism also begins with a transient oxidation at C-3′′
by a tightly bound NAD+ cofactor (Scheme 5). The enzyme
then promotes the elimination of UDP to produce an enone
intermediate. A conjugate addition of UDP combined with
protonation in the opposite stereochemical sense inverts the
stereochemistry at C-2′′, and reduction of the ketone at C-3′′
completes the reaction. The postulate of a transient oxidation
process in combination with an elimination event has ample
precedence in the reactions catalyzed by dehydroquinate syn-
thase54 and S-adenosylhomocysteinase.55 Furthermore, the
hybrid mechanism is consistent with all of the isotopic studies.
The release of UDP and 2-acetamidoglucal could be explained
by a premature reduction of the ketone during the lifetime of
the enone intermediate. This would produce 2-acetamidoglucal
that could not serve as a conjugate-addition acceptor and would
simply be released into solution along with UDP. In order to
distinguish between these two possible mechanisms, we inves-
tigated the reactivity of 3′′-deoxy-UDP-GlcNAc (inset of
Scheme 5) and tested for the presence of a bound NAD+
cofactor.
Studies with 3′′-Deoxy-UDP-GlcNAc. The possibility that
a transient oxidation of the hexosamine residue at C-3 is
involved in the epimerization reaction led us to investigate
whether the deoxygenated analog could serve as an alternate
substrate. If 3′′-deoxy-UDP-GlcNAc were enzymatically epimer-
ized, one could rule out mechanisms involving such an oxidation
step.
When 3′′-deoxy-UDP-GlcNAc was incubated with the epi-
merase in D2O, no epimerization or deuterium incorporation
was detected.38 The lack of epimerization could possibly be
Figure 3. Ion-paired reversed-phase HPLC traces obtained during an
extended incubation of UDP-GlcNAc with UDP-GlcNAc 2-epimerase.
under these reaction conditions. The formation of UDP during
extended incubations was also confirmed using 31P NMR
spectroscopy.
The observation that the elimination products accumulate in
solution indicates that they are thermodynamically more stable
than the equilibrating mixture of UDP-sugar epimers. We
previously reported13 that the equilibrium constant had a value
of greater than 600 in favor of 2-acetamidoglucal and UDP.
We have made further attempts to quantitate this number using
an HPLC analysis of an equilibrated reaction mixture and have
found that the equilibrium constant actually has a value that is
in excess of 25 000 favoring the intermediates. An inability to
accurately integrate peaks of widely differing areas prevented
us from obtaining the true value in this fashion.
The kinetics of UDP formation was also followed by HPLC,
and the enzyme was found to have a specific activity of 0.017
units/mg in the presence of a saturating equilibrium mixture of
the epimers. This compares to the specific activity for UDP-
GlcNAc epimerization of 6.8 units/mg in the presence of
saturating UDP-GlcNAc.
(54) Bender, S. L.; Widlanski, T.; Knowles, J. R. Biochemistry 1989,
28, 7560-7572.
(55) Palmer, J. L.; Abeles, R. H. J. Biol. Chem. 1979, 254, 1217-1226.
One explanation for the appearance of 2-acetamidoglucal and
UDP during extended incubations is that these are the true

10276 J. Am. Chem. Soc., Vol. 119, No. 43, 1997
Morgan et al.
attributed to the inability of the deoxygenated compound to act
as an activator of the enzyme. Indeed when 3′′-deoxy-UDP-
GlcNac was incubated with UDP-ManNAc and the epimerase,
no detectable traces of UDP-GlcNAc were produced. These
experiments showed that the removal of the hydroxyl group at
C-3′′ destroys the ability of the sugar-nucleotide to act as an
activator of the enzyme.
Scheme 6. Three Mechanisms for the â-Elimination of UDP
from UDP-GlcNAc
The incubation in D2O was therefore repeated in the presence
of UDP-GlcNAc (both at saturating and subsaturating concen-
trations). Again no signs of epimerization of, or deuterium
incorporation into, the 3-deoxy-GlcNAc were detected upon
extended incubation. The enzyme was clearly active, however,
since the UDP-GlcNAc was rapidly epimerized/deuterated and
then gradually converted to UDP and 2-acetamidoglucal. These
experiments showed that the 3′′-deoxy compound was not a
substrate for the epimerase; however, they do not allow for the
distinction between the glycal mechanism (Scheme 3) and the
hybrid mechanism (Scheme 5). The lack of reactivity could
be due to either the inability of the enzyme to oxidize the 3′′-
deoxy compound at C-3′′ or the removal of a group that affects
the conformation, electronic properties, and crucial binding
interactions of the modified nucleotide. The compound did,
however, appear to bind to the enzyme since it was found to
act as an inhibitor of the epimerase at millimolar concentra-
tions.56
Tests for Bound NAD+. Our inability to rule out an NAD+-
dependent hybrid mechanism (Scheme 5) led us to specifically
test for the presence of the bound cofactor. A direct UV
spectroscopic examination of the purified epimerase failed to
show any significant absorbances past 310 nm that might be
attributed to a chromophoric cofactor.38 A spectrum taken in
the presence of saturating levels of substrate gave similar results.
In a subsequent experiment, a relatively large sample of the
epimerase was proteolytically digested and then assayed55 for
the presence of released NAD+.38 It was found that less than
5% of the epimerase could have contained the bound cofactor.
This experiment, taken with the observation that the addition
of NAD+ to the purified epimerase does not affect its activity,
indicates that the enzyme does not utilize this cofactor during
epimerization. This conclusion is consistent with an analysis
of the amino acid sequence which fails to show any of the
common NAD+ binding site motifs.57
of free UDP and 2-acetamidoglucal since these species are the
proposed reaction intermediates. It is of further interest to note
that these reaction intermediates are more thermodynamically
stable than the equilibrating substrates (when free in solution).
Precedence for the eliminations can be found in the enzymatic
hydrations/dehydrations of glycals by glycosidases58 and nu-
cleoside 2-deoxyribosyltransferase.59 Anti additions normally
occur in the hydration of glycals by retaining â-glucosidases,58
and a syn addition is observed in the hydration of 2-acetami-
doglucal by â-N-acetylhexosaminidase.42 In nonenzymatic acid-
catalyzed additions of alcohols to glycals, syn products are found
to predominate; however, trans addition products are also
observed.60
Three nonradical mechanisms are possible for â-elimination
reactions as shown in Scheme 6 for the anti elimination of UDP
from UDP-GlcNAc. The first is an E1cb mechanism in which
an initial deprotonation event generates a carbanionic intermedi-
ate prior to expulsion of the leaving group. There is ample
precedent for both syn and anti E1cb mechanisms in enzymatic
reactions; however, in all cases an electron-withdrawing func-
tionality such as a carbonyl or carboxyl group serves to stabilize
the anionic intermediate.61,62 It is highly unlikely that this
mechanism is operative in the epimerase reaction due to the
high pKa value of the C-2′′ proton in the UDP-sugar epimers.
The second possibility is a concerted E2 mechanism that
proceeds without any intermediate. In enzymatic reactions,
convincing evidence for an E2 elimination reaction is limited
to the example of crotonase that has been reported to catalyze
a concerted syn elimination of water.63 In this light, it is of
interest to point out that certain acyl-CoA â-epimerases are
thought to be unique crotonases that catalyze sequential hydra-
tions/dehydrations of opposite stereochemistry.64-66 Perhaps the
Discussion
The observation of solvent deuterium isotope incorporation
and of a primary kinetic isotope effect supports a mechanism
involving removal of the C-2′′ proton during a rate-determining
step of the epimerization reaction. The observation of isotopic
scrambling in the UDP-R-phosphate supports a mechanism
involving cleavage of the anomeric C-O bond during the
reaction. Together these results suggest that an anti elimination
occurs with UDP-GlcNAc as the substrate and a syn elimination
occurs with UDP-ManNAc as the substrate. The absence of
any cofactor requirements or any tightly bound NAD+ indicates
that an oxidation at C-3′′ is not involved in the reaction. This
leaves the glycal mechanism (Scheme 3) as a reasonable
minimal description of the overall reaction pathway. This
mechanism is entirely consistent with the observed production
(58) Legler, G. AdV. Carbohydr. Chem. Biochem. 1990, 48, 349-357
and references therein.
(59) Smar, M.; Short, S. A.; Wolfenden, R. Biochemistry 1991, 30,
7908-7912.
(60) Kaila, N.; Blumenstein, M.; Bielawska, H.; Franck, R. W. J. Org.
Chem. 1992, 57, 4576-4578.
(56) Insufficient amounts of the 3′′-deoxy-UDP-GlcNAc were available
to complete a thorough analysis of the mode of inhibition. The inhibition
was studied under the standard assay conditions at a fixed concentration of
UDP-GlcNAc (0.73 mM ) KM), and the concentration of 3′′-deoxy-GlcNAc
was varied from 0.6 to 6.0 mM. An assumption of competitive inhibition
led to the calculation of KI ) 1 mM.
(61) Anderson, V. E. In Enzyme Mechanism from Isotope Effects; Cook,
P. F., Ed.; CRC Press: Cleveland, OH, 1991; pp 389-417.
(62) Creighton, D. J.; Murthy, N. S. R. K. In The Enzymes; Sigman, D.
S., Boyer, P. D., Eds.; Academic Press: San Diego, CA, 1990; Vol. 19, pp
397-403.
(63) Bahnson, B. J.; Anderson, V. E. Biochemistry 1991, 30, 5894-
5906.
(57) Wierenga, R. K.; Terpstra, P.; Hol, W. G. J. J. Mol. Biol. 1986,
187, 101-107.
(64) Yang, S.-Y.; Elzinga, M. J. Biol. Chem. 1993, 268, 6588-6592.

Eliminations in Reactions by UDP-GlcNAc
J. Am. Chem. Soc., Vol. 119, No. 43, 1997 10277
most likely possibility is that the UDP-GlcNAc 2-epimerase
eliminations are E1-like processes. The combination of a good
leaving group and a relatively stable oxocarbenium intermediate
should permit this type of mechanism to take place. Precedence
for an E1 elimination mechanism can be found in the reaction
catalyzed by imidazoleglycerol-phosphate dehydratase where
the cationic intermediate is thought to be stabilized by conjuga-
tion to an imidazole ring.67
a study has recently been published that strongly implicates the
involvement of 1 as an intermediate in the reaction catalyzed
by N-acetyl-â-hexosaminidase.69
The observation of a primary kinetic isotope effect indicates
that C-H bond cleavage occurs in a rate-determining step of
the reaction. This is consistent with either an E2 mechanism
or an E1 mechanism in which deprotonation of the oxocarbe-
nium intermediate is partially rate-determining.
Acknowledgment. We are grateful to Drs. Lucia B. Roth-
man-Denes and Diane R. Kiino for supplying the plasmid
containing nfrC and to Dr. P. D. Rick for supplying a plasmid
containing rffD. We thank Drs. S. G. Withers and E. C. K. Lai
for supplying samples of synthetic 2-acetamidoglucal. We also
thank the University of British Columbia and the Natural
Sciences and Engineering Research Council of Canada (NSERC)
for financial support.
The simple mechanisms outlined above may be more
complicated by the involvement of covalent catalysis with
enzyme-bound intermediates (as in the glycosidase reactions
described above) or by the possibility of neighboring group
participation by the acetamido functionality. In particular, the
oxazolines 1 and 2 could be intermediates in the anti and syn
eliminations, respectively. Nonenzymatic oxazoline formation
in both GlcNac and ManNac systems often occurs under reaction
conditions that promote oxocarbenium formation.68 In addition,
Supporting Information Available: Further experimental
details on purification and identification of UDP-GlcNAc
2-epimerase and selected results (9 pages). See any current
masthead page for ordering and Internet access instructions.
(65) Smeland, T. E.; Cuebas, D.; Schulz, H. J. Biol. Chem. 1991, 266,
23904-23908.
JA971718Q
(66) Hiltunen, J. K.; Palosaari, P. M.; Kunau, W.-H. J. Biol. Chem. 1989,
264, 13536-13540.
(68) Collins, P.; Ferrier, R. Monosaccharides; Wiley: Chichester, 1995;
pp 236-237.
(69) Knapp, S.; Vocadlo, D.; Gao, Z.; Kirk, B.; Lou, J.; Withers, S. G.
J. Am. Chem. Soc. 1996, 118, 6804-6805.
(67) Parker, A. R.; Moore, J. A.; Schwab, J. M.; Davisson, V. J. J. Am.
Chem. Soc. 1995, 117, 10605-10613.