Structure-activity studies of glucose transfer: Determination of the spontaneous rates of hydrolysis of uridine 5′-diphospho-α-D-glucose (UDPG) and uridine 5′-diphospho-α-D-glucuronic acid (UDPGA)

Article abstract of DOI:10.1016/S0968-0896(03)00065-8

The pH-rate profiles for the hydrolysis of uridine 5′-diphospho-α-D-glucose (UDPG) and uridine 5′-diphospho-α-D-glucuronic acid (UDPGA) in aqueous solution have been measured. The results obtained and a comparison with other data suggests that the mechanism of hydrolysis of each activated glycosyl-donor at pH 1-4 probably involves the slow ionisation, via an S_N1 process, of the neutral molecule to a glycosyl ion and UDP. From these data, the catalytic power (k_cat/k_uncat) of the glycosyltransferases has been estimated for the first time to be in the order of 10^11-13.

Full text of DOI:10.1016/S0968-0896(03)00065-8

Bioorganic & Medicinal Chemistry 11 (2003) 2339–2345
Structure–Activity Studies of Glucose Transfer:
Determination of the Spontaneous Rates of Hydrolysis of
Uridine 50-Diphospho-ꢀ-D-glucose (UDPG) and Uridine
50-Diphospho-ꢀ-D-glucuronic Acid (UDPGA)
Colin T. Bedford,a,b,* Alan D. Hickmanb and Christopher J. Loganb,y
aSchool of Biosciences, University of Westminster, 115 New Cavendish Street, London W1W 6UW, UK
bShell Research Ltd, Sittingbourne Research Centre, Sittingbourne, Kent ME9 8AG, UK
Received 20 December 2001; accepted 23 January 2003
Abstract—The pH-rate profiles for the hydrolysis of uridine 50-diphospho-a-d-glucose (UDPG) and uridine 50-diphospho-a-d-glu-
curonic acid (UDPGA) in aqueous solution have been measured. The results obtained and a comparison with other data suggests
that the mechanism of hydrolysis of each activated glycosyl-donor at pH 1–4 probably involves the slow ionisation, via an SN1
process, of the neutral molecule to a glycosyl ion and UDP. From these data, the catalytic power (kcat/kuncat) of the glycosyl-
transferases has been estimated for the first time to be in the order of 1011ꢀ13
.
# 2003 Elsevier Science Ltd. All rights reserved.
Introduction
acid transfer by determining the spontaneous (non-
enzymic) glucuronylating reactivity of UDPGA. This
was achieved by determining its pH-rate profile, and
hence its glucuronylating reactivity towards water.
Because it lacked the ionizable carboxyl group of
UDPGA, the prototype of the series, uridine 50-diphos-
pho-a-d-glucose (UDPG), was also studied. This was of
interest in its own right, of course, since UDPG is the
activated glucosyl-donor employed in the biosynthesis
of glycogen and in many other polysaccharides, glyco-
lipids, and glycoproteins. More mundanely, UDPG is
also utilised in the biotransformation in insects and
plants of natural, and xenobiotic, alcohols and phenols
into ‘glucose conjugates’ (analogous to the role of
UDPGA in the formation of ‘glucuronide conjugates’ in
mammals).
‘Active’ glucose has a pyrophosphate leaving group
Biological glycosylation is largely accomplished by
enzymes that utilise as activated glycosyl-donors the
glycosyl derivatives of either nucleoside pyrophosphates
or polyprenol pyrophosphates.1 The former class of
activated glycosyl-donors form the largest group, for
one of several nucleosides (e.g., uridine, cytidine, ade-
nine) may be linked to one of a whole range of glycosyl
groups (e.g., glucosyl, galactosyl, mannosyl, glucuronyl,
ribosyl). Our interest in this area stemmed from our
work with uridine 50-diphospho-a-d-glucuronic acid
(UDPGA), which can be used routinely to biosynthesise
glucuronide conjugates of drug metabolites (and other
foreign compounds). It is well known that incubation of
an aglycone with boiled enzyme preparations plus
UDPGA produces no glucuronides,2 and we became
interested in investigating the catalytic power of the
enzymes involved, the glucuronyltransferases. To this
end, we began a structure–activity study of glucuronic
In the event, because it lacked the complicating ion-
isable carboxyl group, we conducted a comprehensive
study of UDPG, and collected only a few data for
UDPGA.
Mechanism of biological glycosylation
*Corresponding author at present address: Department of Chemistry,
University College, London WC1H 0AJ, UK. E-mail: c.t.bedford@
ucl.ac.uk
yPresent address: AstraZeneca, Bakewell Road, Loughborough,
Leicestershire LE11 ORH, UK.
In general, these UDPGly-dependent processes effect,
in the main, either O-glycosylation or N-glycosylation
of a substrate. In all cases, whether the substrate be a
0968-0896/03/$ - see front matter # 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0968-0896(03)00065-8

2340
C. T. Bedford et al. / Bioorg. Med. Chem. 11 (2003) 2339–2345
large or small (bio)molecule, the sites (e.g., hydroxyl or
amino groups) of glycosylation are, by definition,
nucleophilic groupings. Since O- and N-glycosylation
are mechanistically congeneric, the prototype of bio-
logical glycosylation reactions can be taken as the glu-
cosyltransferase-mediated glucosylation of an alcohol,
ROH, by UDPG. Since, moreover, all of the nucleotide
glycosyl-donors are a-XDP-glycosides (none is b-), the
prototypical formation from an alcohol (ROH) of its
b-glucoside can, as Axelrod and co-workers3 (in their
work on mammalian steroidal b-glucuronides) recog-
nised more than 40 years ago, be viewed as a glucosyl-
transferase-mediated inverting nucleophilic displacement
at C-1 of UDPG with expulsion of uridine 50-diphosphate
(UDP) (Scheme 1). (Clearly, when a-glucosides are
formed, a double displacement reaction of UDPG is
required, presumably involving the intermediacy of a
glucosylated enzyme. That type of process will not con-
cern us here.)
see ref 4). Subsequently, Schowen and Coward and co-
workers5 determined the spontaneous methylating reac-
tivity of S-adenosylmethionine (SAM) and estimated
the catalytic power (kcat/kuncat
)
of the methyl-
transferases to be in the order of 1016. No attempt has
been made to evaluate the catalytic power of the pre-
nyltransferases, although Tidd’s determination of the
spontaneous alkylating reactivity of dimethylallyl pyr-
ophosphate (DMAPP)6—which is a good model for
GPP and FPP—provides the necessary information.
Recently we determined the spontaneous sulfating reac-
tivity of PAPS and estimated7 the catalytic power (kcat
/
kuncat) of the sulfotransferases to be in the order of
1010ꢀ2—in the event, identical to the range estimated for
the phosphotransferases.4 However, there have been, to
our knowledge, no studies of the measurement of the
spontaneous glycosylating reactivity of any glycosyl
derivative of a nucleoside pyrophosphate which would
yield similar data pertaining to the magnitude of the
catalytic role of the glycosyltransferases. We have now
accomplished this task, and report rates of hydrolysis of
UDPG and UDPGA in the pH range 1–10 and the
magnitude of the spontaneous (non-enzymic) glycosyl-
ating reactivity (towards water) of UDPG and UDPGA
and the pH range at which these reactivities prevail. A
preliminary report of the results of the UDPG study has
been made.8
Related biological transfer processes—methylation,
prenylation, acetylation, phosphorylation and sulfation
Biological glycosylation can be grouped with the two
other major biological alkylation processes, methylation
and prenylation. Moreover, mechanistically we can
group these three biological alkylation reactions with
three important biological ‘acylation’ processes, acetyl-
ation, phosphorylation and sulfation. For each effects
the alkylation or acylation of a nucleophilic group in a
biomolecule by means of a coenzyme—or an activated
donor-molecule (ADM)—in which the group to be
transferred is attached to a good leaving group. The
combinations of coenzyme/enzyme (or ADM/enzyme)
involved are S-adenosylmethionine (SAM)/methyl-
transferases, geranyl or farnesyl pyrophosphate (GPP or
FPP)/prenyltransferases, acetyl coenzyme A (acetyl
CoA)/acetyltransferases, adenosine 50-triphosphate
(ATP)/phosphotransferases and 30-phosphoadenosine-
50-phosphosulfate (PAPS)/sulfotransferases.
Results
Hydrolysis experiments
In order to follow the hydrolysis of UDPG and of
UDPGA by quantitative HPLC, using spectro-
photometric detection of the uridine chromophore, a
mixed ion-pair system capable of resolving UDPG and
UDPGA from their likely uridine-containing hydrolysis
products, UDP, uridine 50-phosphate (UMP), and
uridine 30,50-cyclic-phosphate (cUMP) was developed
(Fig. 1).9
The catalytic power of these transferases has long been
of interest to mechanistic bioorganic chemists, and more
than 30 years ago that of the phosphotransferases
(kinases) came under close scrutiny. By elucidating the
weak spontaneous (non-enzymic) phosphorylating
activity towards water of adenosine 50-triphosphate
(ATP) at pH 7, it was possible to calculate from known
enzymic rates that the catalytic power (kcat/kuncat) of the
phosphotransferases was in the order of 1010–1012 (e.g.,
From the results of pilot experiments, a temperature of
60 ꢁC was chosen for the determination of the pH rate
profile of UDPG and UDPGA. The rate data that were
obtained are shown in Tables 1 and 2 and plotted in
Figure 2. Some rate data of the hydrolysis of the com-
mon hydrolysis product, UDP, were also obtained at
60 ꢁC at pH 1.5–6 and these are shown in Table 3 and
plotted in Figure 2.
Scheme 1.

C. T. Bedford et al. / Bioorg. Med. Chem. 11 (2003) 2339–2345
2341
Table 2. Observed rate constants for the hydrolysis of UDPGA in
aqueous solution at 60 ꢁC (m=1.0)
Buffer
Molarity
(M)
pH
105kobs
(sꢀ1
)
HCl
HCl
HCl
HCl
Succinate
Tris
1.0
0.03
1.10
1.51
2.53
5.63
6.96
403.3
0.10
0.04
0.0063
0.33
0.5
12.71
4.245
1.060
0.0472
0.3814
fragment at m/z 323 due to the corresponding UMP
monoanion. Peak ratio measurements for each sample
were therefore carried out on m/z 403, 404 and 405,
which should indicate any enrichment, and on m/z 323,
324, and 325, which provided an internal check since it
would not be expected to contain any 18O. (No
enhancement of peak area was expected of m/z 404 or of
m/z 324, and this provided an additional internal check.)
Comparison of the mass spectral data of duplicate
samples of UDP obtained by hydrolysis of UDPGA at
pH 1.5 at 60 ꢁC for 0.5 h in 18O water (50% 18O) with
those of authentic UDP and of a blank obtained by
heating UDP under similar conditions revealed that no
incorporation of label had occurred.
Figure 1. Mixed ion-pair reversed-phase HPLC trace of a mixture of
UMP, UDPG, UDP, cUMP, and UDPGA. (Conditions: 10 mL from
an aqueous solution of about 1 mg/mL injected by valve onto a
200ꢂ4.5 mm column packed with Partisil 10 ODS 2. The mobile phase
was 0.002375 M tetrabutylammonium hydroxide, 0.002375 M tetra-
ethylammonium hydroxide, 0.0475 M ammonium dihydrogen ortho-
phosphate, adjusted to pH 6 with KOH in aqueous 10% methanol.)
Discussion
The pH-rate profiles of UDPG and UDPGA
Clearly, amongst the several possible C–O and P–O
hydrolytic cleavage reactions of UDPG, there was a
need to identify and quantify the hydrolytic process
yielding uridine 50-diphosphate (UDP) by C–O fission
(Scheme 1, HOH for ROH). The quantitative HPLC
method9 that was used for the assay of UDPG yielded,
simultaneously, an assay of both UMP and UDP (see
Fig. 1), and this permitted an estimate to be made of the
amounts of each of these reaction products formed
during each run at a given pH value.
Table 1. Observed rate constants for the hydrolysis of UDPG in
aqueous solution at 60 and 95 ꢁC (m=1.0)
Buffer
Molarity
(M)
pH
Temp
(ꢁC)
105kobs
(sꢀ1
)
HCl
HCl
HCl
HCl
HCl
HCl
Formate
Formate
Acetate
Acetate
Succinate
Tris
Tris
Carbonate
NaOH
0.10
0.05
0.027
0.0063
0.0063
0.006
1.0
1.01
1.50
1.84
2.28
2.39
2.61
2.97
3.61
3.76
3.80
5.64
6.97
8.07
9.4260
11.78
60
60
60
60
60
60
60
60
60
60
60
60
60
1
168.8
49.24
19.52
7.652
5.994
3.912
1.581
0.4002
0.3242
0.2415
0.0400
0.2788
1.365
At pH 1–3, the major hydrolysis product of UDPG was
found to be UDP (ca. 95%), with minor amounts of
UMP. Moreover the rate of formation of UDP in this
pH range was virtually identical to the rate of dis-
appearance of UDPG, once allowance had been made
for the rate of hydrolysis of UDP (to UMP), which we
also measured. The rate profile at 60 ꢁC for UDP in the
pH range 1–6 is, for ease of comparison, superimposed
on the UDPG rate profile in Figure 2, and was very
similar to that of adenosine 50-diphosphate (ADP)
determined by Miller and Westheimer11 at 95 ꢁC. [We
made two determinations of the hydrolysis rate of UDP
at 95 ꢁC as a check on our methodology (see Table 3)
and we obtained essentially identical results to those
obtained by Miller and Westheimer.11]
0.10
1.0
0.10
0.33
0.05
0.10
0.33
0.10
19.2
60
1982
HCl
Acetate
0.016
1.0
2.08
3.90
95
95
934.5
14.75
Isotope experiments
The positions of bond fission of UDPGA were deter-
mined from the results of FAB mass spectral analysis of
UDP isolated from reactions run in water enriched with
18O. The negative ion FAB mass spectrum of authentic
UDP has been reported10 to show an intense ion at m/z
403 due to the UDP monoanion with an equally intense
At pH 4, the major hydrolysis product was found to be
UMP, with minor amounts of UDP. As is evident from
Figure 2, the rate of hydrolysis of UDPG to UDP is

2342
C. T. Bedford et al. / Bioorg. Med. Chem. 11 (2003) 2339–2345
Figure 2. The pH-rate profiles at 60 ꢁC of the hydrolysis of UDPG, UDPGA and UDP.
imate proportion to the hydrogen-ion concentration.
Table 3. Observed rate constants for the hydrolysis of UDP in
aqueous solution at 60 and 95 ꢁC (m=1.0)
UDP could have been formed from either substrate via
P–O or C–O fission, but we have confirmed that C–O
fission was occurring by means of a labelling experiment
with UDPGA using H218O (in view of their closely
related structures, we have assumed—though not con-
ducted the experiment— that UDPG would have yiel-
ded a similar result).
Buffer
Molarity
(M)
pH
Temp
(ꢁC)
105kobs
(sꢀ1
)
HCl
HCl
Acetate
Succinate
0.05
1.50
2.39
3.80
5.64
60
60
60
60
0.692
0.380
0.263
0.245
0.0063
0.10
0.33
HCl
0.016
2.08
95
17.0
In the pH range 6–10, UDP, which is stable under mild
alkaline conditions (like ADP11) and would have been
expected to have persisted if it had been formed, was
not a hydrolysis product of UDPG or of UDPGA.
Consequently, only a few rate measurements in this pH
range were made and no product studies were under-
taken. The rate profile for UDPG in the pH range 7–10
(Fig. 2) is probably best explained by hydroxide ion-
catalysed formation of glucose-1,2-cyclic monophos-
phate and uridine 50-monophosphate (UMP), as
observed by Leloir and co-workers12 during their struc-
tural studies of UDPG. This reaction, as would be
expected, increases in rate in approximate proportion to
the hydroxide-ion concentration. The remoteness from
the reaction centre of the carboxylate group of UDPGA
has no effect, predictably, on the rate of the analogous
reaction of UDPGA, the rates of both substrates being
identical at pH 5.5 and 7 (see Fig. 2).
about equal to that of the hydrolysis of UDP to UMP
at pH 4, and this means that UDP at this pH would be
hydrolysed as fast as it was formed. Clearly, although
all, or a part of, the UMP could have been formed
directly from UDPG, it seems more likely that all of it is
formed via the hydrolysis of initially-formed UDP.
Less data were obtained for UDPGA, but as can be
seen from Figure 2, the rates of hydrolysis of UDPGA
and UDPG differ by a factor of about 10 in the range
pH 1–2.5, UDPGA being the less reactive. At low pH
(<3), the carboxylic acid group of UDPGA will be fully
protonated, and this electron-withdrawing group is
exhibiting its expected rate-decreasing destabilising
effect upon the incipient glycosyl cation (vis a vis that
formed from UDPG).
Comparison of UDPG rate data with that of other
phosphates
We can safely conclude, therefore, that the nucleotide
hydrolysis product at pH 1–4 of UDPG and UDPGA is
predominantly, if not solely, UDP, and that the rate of
these reactions, as Figure 2 shows, increases in approx-
For comparison, our rate data for UDPG in the pH
range 1–4 and those for a-d-glucose-1-phosphate,13a

C. T. Bedford et al. / Bioorg. Med. Chem. 11 (2003) 2339–2345
2343
methyl phosphate,13b dimethylallyl phosphate,6 and
dimethylallyl pyrophosphate6 in the pH range 1–8 (all
extrapolated to 60 ꢁC) appear in Figure 3.
deployed for biological glycosylation and prenylation,
UDPG and DMAPP respectively, possess alkylating
reactivities (here measured towards water) that differ by
about two orders of magnitude, DMAPP being the
more reactive.
Gratifyingly, the rectilinear profile that we obtained for
UDPG at pH 1–4 parallels that for a-d-glucose-1-
phosphate (a-GP) at pH 1–4 determined by Bunton and
co-workers,13a and our demonstration of C-O fission in
that pH region reproduces the finding, in experiments
using H218O, that the hydrolysis of a-GP at pH 1–4 also
occurred wholly via C–O fission.13a
It is noteworthy that for dimethylallyl esters,
monophosphate appears to be about as good a leaving
group as pyrophosphate, but for glucosyl esters mono-
phosphate is about 5-fold less good as a leaving group
than is (nucleotidyl) pyrophosphate. The pH-rate profile
of methyl phosphate (see Fig. 3) does not show any fea-
tures at pH 1–4 indicating C–O fission, and labelling
studies have shown that P–O fission is the exclusive
mode in this pH range.13b Put another way, methyl
phosphate is not a methylating agent in this pH range.
Indeed, in contrast to biological glycosylation and pre-
nylation where pyrophosphates are deployed as acti-
vated donor-molecules in SN1-type processes, biological
methylation, most commonly, deploys as coenzyme
The near-identical slopes of the profiles for UDPG and
a-GP are closely similar to the slopes of the profiles for
dimethylallyl phosphate and dimethylallyl pyrophos-
phate (DMAPP) in the pH range 1–7, which Tidd has
shown6 to be due to C–O fission of each. However,
DMAPP is, at pH 4, about a 100-fold more labile
hydrolytically than UDPG. Put another way, this shows
that the prototypes of the activated donor-molecules
Figure 3. The pH-rate profiles at 60 ꢁC of the hydrolysis of methyl phosphate, dimethylallyl phosphate, dimethylallyl pyrophosphate, a-d-glucose-1-
phosphate and UDPG.

2344
C. T. Bedford et al. / Bioorg. Med. Chem. 11 (2003) 2339–2345
S-adenosylmethionine (SAM),
a
methylsulphonium
obtained as aqueous solutions from BDH Ltd. All other
reagents were of AR grade.
compound, MeR2S+, with neutral R2S as a leaving
group in an SN2process.
High-performance liquid chromatograph
Mechanism of hydrolysis of UDPG and UDPGA
The analytical HPLC system was composed of a
Laboratory Data Control (LDC) gradient solvent-
delivery system, a Rheodyne injection valve fitted with a
20 mL loop, a Pye LC3 spectrophotometric detector
operating at 262 nm and a Hewlett Packard 338S Inte-
grator. The columns (200ꢂ4.5 mm) were either Whatman
Partisil 10 SAX for ion-exchange or Partisil 10 ODS2for
paired-ion chromatography. Ion-exchange analyses were
run at room temperature, but paired-ion analyses were
run with the column maintained at 30 ꢁC by means of a
Magnus Scientific water jacket and a Churchill water
pump and heater. The flow rate was 2mL/min.
The mechanism of the hydrolysis of UDPG and
UDPGA at pH 1–4 would be expected to be analogous
to that proposed13a for a-GP and therefore probably
involves the slow ionisation via an SN1(C) process of the
neutral molecule (Scheme 2), or less likely of the mono-
anion, to a glycosyl ion and UDP.
The catalytic power of the glycosyltransferases
In summary, these results reveal that UDPG is a gluco-
sylating agent only at pH 1–4 with a reactivity only
about five-fold greater than that of a-d-glucose-l-phos-
phate. Extrapolation to pH 7 and to 37 ꢁC of the rate
data and conversion to a second-order rate constant
gives a value of about 5ꢂ10ꢀ12 Mꢀ1 sꢀ1 for the gluco-
sylating reactivity (towards water) of UDPG. Expres-
sing this in a way suitable for comparison with an
enzymic reaction and converting to units familiar to
enzymologists, we can say that the spontaneous (non-
enzymic) glucosylation of water by UDPG occurs at a
rate of about 2ꢂ10ꢀ11 mM ꢀ1 minꢀ1. Enzymic rate data
of glycosyltransferases are sparse,14 but are in the range
5–500 mMꢀ1 minꢀ1. By assuming that the magnitude of
the glucosylating reactivity of UDPG towards a typical
hydroxyl-containing substrate, ROH, is similar to that
Ion-exchange chromatography
For isocratic elution, optimal resolution was obtained
using as a mobile phase: 0.05 M KCl and 0.05 M
KH2PO4, adjusted to pH 6.0 with 1 M KOH.
Paired-ion chromatography
Using a single ion-pairing reagent. Optimum isocratic
conditions for separating UDPG, UDP and UMP from
each other and also c-UMP were found to be a mobile
phase of: 0.0475 M NH4PO4, 0.00475 M tetrabutyl
ammonium hydroxide, adjusted to pH 6.0 with 1 M HCl
in aqueous 5% methanol.
of water, we can estimate that the catalytic power (kcat/
kuncat) of the glucosyltransferases—and, by analogy, of
all of the glycosyltransferases—is in the order of
1011ꢀ13
Using two ion-pairing reagents. The optimal isocratic
conditions for separating all of the phosphates was
found to be a mobile phase of: 0.002375 M tetra-
.
ethylammonium
hydroxide,
0.002375 M
tetra-
butylammonium hydroxide, 0.0475 M NH4PO4,
adjusted to pH 6.0 with 1 M HCl in aqueous 5%
methanol.
Experimental
Materials
Uridine 50-diphospho-a-d-glucose (UDPG), uridine 50-
diphospho-a-d-glucuronic acid (UDPGA), uridine 50-
diphosphate (UDP), uridine 30,50-cyclic-phosphate (c-
UMP) and uridine 50-phosphate (UMP) were obtained
as sodium salts from Sigma Ltd. Tetrabutylammonium
hydroxide and tetraethyl-ammonium hydroxide were
Kinetic studies
The kinetics of the hydrolysis of UDPG or UDPGA
were followed by quantitative HPLC either by periodi-
cally measuring the decrease in peak area of UDPG or
UDPGA or the increase in peak area of UDP and/or
Scheme 2.

C. T. Bedford et al. / Bioorg. Med. Chem. 11 (2003) 2339–2345
2345
UMP (the only detectable products). All reactions were
References and Notes
studied at either 60 or 95 ꢁC and were initiated by add-
ing a known volume of pre-heated buffer to a known
amount of substrate. In all cases the substrate dissolved
immediately. Buffers employed for the hydrolyses were
HCl (pH<3.5), formate (0.1–1.0 M), acetate (0.1–
1.0 M), succinate (0.33 M), Tris (0.5–1.0 M), carbonate
(0.33 M), and NaOH (pH>11), with ionic strengths of
1.0 made up with potassium chloride. No corrections of
the rate data for buffer catalysis were necessary, since
representative checks for buffer catalysis at pH 3.6 (for-
mate), pH 3.8 (acetate) and pH 8.0 (Tris) showed there
to be none. All pHs were measured using a Russel glass
electrode and a Corning pH meter. Pseudo-first-order
rate constants were calculated from data acquired over
at least four half-lives.
1. Davies, G.; Sinnott, M. L.; Withers, S. G. In Comprehen-
sive Biological Catalysis; Sinnott, M. L., Ed., Academic: New
York, 1998; Chapter 3, p 119.
2. Mulder, G. J.; Coughtrie, M. W. H.; Burchell, B. In Con-
jugation Reactions in Drug Metabolism; Mulder, G. J., Ed.;
Taylor and Francis: London, 1990; Chapter 4, p 51.
3. Axelrod, J.; Inscoe, J. K.; Tompkins, G. M. J. Biol. Chem.
1958, 232, 835.
4. Benkovic, S. J.; Schray, K. J. In The Enzymes, 3rd ed.;
Boyer, P. D., Ed.; Academic: New York, 1970; Vol. VIII,
Chapter 6, p 201.
5. Mihel, I.; Knipe, J. O.; Coward, J. K.; Schowen, R. L. J.
Amer. Chem. Soc. 1979, 101, 4349.
6. Tidd, B. K. J. Chem. Soc. (B) 1971, 1168.
7. Bedford, C. T.; Kirby, A. J.; Logan, C. J.; Drummond,
J. N. Bioorg. Med. Chem. 1995, 3, 167.
8. Bedford, C. T.; Hickman, A.; Logan, C. J. Bioorg. Med.
Chem. Lett. 1992, 2, 1513.
9. Bedford, C. T.; Hickman, A.; Logan, C. J. J. Chromatogr.
A 1995, 695, 123.
10. Fenselau, C.; Feng, P. C. C.; Chen, T.; Johnson, L. P.
Drug Metab. Dispos. 1982, 10, 316.
11. Miller, D. L.; Westheimer, F. H. J. Am. Chem. Soc. 1966,
88, 1507.
Mass spectrometry
Analyses by fast-atom bombardment (FAB) were car-
ried out in a VG 7070 spectrometer in negative ion
mode using glycerol as the matrix material. A duplicate
measurement of each sample of UDP was recorded and
the mean result determined.
12. Caputto, R.; Leloir, L. F.; Cardini, C. E.; Paladini, A. C.
J. Biol. Chem. 1950, 184, 333.
13. (a) Bunton, C. A.; Llewellyn, D. R.; Oldham, K. G.; Ver-
non, C. A. J. Chem. Soc. 1958, 3588. (b) Bunton, C. A.;
Llewellyn, D. R.; Oldham, K. G.; Vernon, C. A. J. Chem. Soc.
1958, 3574.
14. Tvaroska, I.; Andre, I.; Carver, J. P. J. Am. Chem. Soc.
2000, 122, 8762.
Acknowledgements
We thank Dr. F. H. Cottee and Dr. J. A. Page, Analy-
tical Chemistry Division, Shell Research Ltd for the
mass spectral measurements.