Experimental and kinetic studies of the Escherichia coli glucuronylsynthase: An engineered enzyme for the synthesis of glucuronide conjugates

Article abstract of DOI:10.1021/jo101914s

The detection and study of glucuronide metabolites is essential in many fields including pharmaceutical development, sports drug testing, and the detection of agricultural residues. Therefore, the development of improved methods for the synthesis of glucuronide conjugates is an important aim. The glycosynthase derived from E. coli β-glucuronidase provides an efficient, scalable, single-step synthesis of β-glucuronides under mild conditions. In this article we report on experimental and kinetic studies of the E. coli glucuronylsynthase, including the influence of acceptor substrate, pH, temperature, cosolvents, and detergents, leading to optimized conditions for glucuronide synthesis. Enzyme kinetics also reveals that both substrate and product inhibition may occur in glucuronylsynthase reactions but that these effects can be ameliorated through the judicious choice of acceptor and donor substrate concentrations. An investigation of temporary polar substituents was conducted leading to improved aqueous solubility of hydrophobic steroidal acceptors. In this way the synthesis of the steroidal metabolite dehydroepiandrosterone 3-β-d-glucuronide was achieved in three steps and 86% overall yield from dehydroepiandrosterone.

Full text of DOI:10.1021/jo101914s

ARTICLE
pubs.acs.org/joc
Experimental and Kinetic Studies of the Escherichia coli
Glucuronylsynthase: An Engineered Enzyme for the Synthesis of
Glucuronide Conjugates
Shane M. Wilkinson, Morgan A. Watson, Anthony C. Willis, and Malcolm D. McLeod*
Research School of Chemistry, Australian National University, Canberra, ACT 0200, Australia
S Supporting Information
b
ABSTRACT: The detection and study of glucuronide metabolites is
essential in many fields including pharmaceutical development,
sports drug testing, and the detection of agricultural residues.
Therefore, the development of improved methods for the synthesis
of glucuronide conjugates is an important aim. The glycosynthase
derived from E. coli β-glucuronidase provides an efficient, scalable,
single-step synthesis of β-glucuronides under mild conditions. In
this article we report on experimental and kinetic studies of the E. coli glucuronylsynthase, including the influence of acceptor
substrate, pH, temperature, cosolvents, and detergents, leading to optimized conditions for glucuronide synthesis. Enzyme kinetics
also reveals that both substrate and product inhibition may occur in glucuronylsynthase reactions but that these effects can be
ameliorated through the judicious choice of acceptor and donor substrate concentrations. An investigation of temporary polar
substituents was conducted leading to improved aqueous solubility of hydrophobic steroidal acceptors. In this way the synthesis of
the steroidal metabolite dehydroepiandrosterone 3-β-D-glucuronide was achieved in three steps and 86% overall yield from
dehydroepiandrosterone.
onjugation with glucuronic acid during phase II metabolism
is a major pathway for the elimination of hydrophobic
family 2 and catalyzes the hydrolytic cleavage of terminal β-
glucuronide residues. The enzyme displays substrate promiscuity
and has been widely exploited in the field of analytical chemistry
for the deconjugation of glucuronide metabolites.8 The enzyme
active site contains two key catalytic residues. The side chain of
glutamic acid 504 (E504)9 acts as a nucleophile, and glutamic
acid 413 (E413) is responsible for general acid/base catalysis in a
double displacement mechanism (Scheme 2a). As demonstrated
for a range of other retaining β-glycosidase enzymes,10,11 muta-
tion of the E. coli β-glucuronidase catalytic nucleophile to a non-
nucleophilic glycine (E504G), alanine (E504A), or serine
(E504S) residue disables the hydrolytic pathway (Scheme 2b).
However, the resulting glycosynthase enzyme, or glucuronyl-
synthase, can catalyze the formation of glucuronide product 1
from R-D-glucuronyl fluoride 2 and an acceptor alcohol substrate.
Our previous report7 detailed the application of a spectrophoto-
metric screening protocol using the wild-type E. coli β-glucur-
onidase to identify likely acceptor alcohol substrates for the
glycosynthase mediated reaction, expression of the enzyme, and
confirmation of these screening hits with the E. coli glucuronyl-
synthase mutants. In this paper we report details of the glucur-
onylsynthase development together with a kinetic study of this
enzyme mediated transformation and practical improvements to
the process for the synthesis of steroidal glucuronides.
C
xenobiotic and endogenous compounds in mammalian systems.1
The identification, quantification, and pharmacological evalua-
tion of these metabolites is essential for many fields including
drug development,1 sports drug testing,2 and the detection of
agricultural residues.3 This creates a significant demand for the
synthesis of glucuronide conjugates 1 as standards.
The preparation of glucuronides provides a significant chal-
lenge for existing methods of glucuronylation (Scheme 1).4
Chemical methods5 of glucuronylation based on the Koenigs-
Knorr reaction or more recent procedures often suffer from poor
yields and side reactions due to the low reactivity of glucuronic
acid derived glycosyl donors4,5 and require one or more depro-
tection steps to liberate free glucuronide. Enzymatic methods6
of synthesis employ uridine 50-diphosphoglucuronosyl trans-
ferases (UGTs), a superfamily of enzymes responsible for glucur-
onylation in the body.1 This procedure provides a mild and
stereospecific synthesis in a single step. However, UGTs are
substrate-specific to the acceptor alcohol, and practical considera-
tions often limit this procedure to small scale syntheses. Given the
limitations associated with existing methods, the development of
improved glucuronylation protocols is an important goal.
Recently we reported a distinct strategy for the synthesis
of glucuronides based on the glycosynthase derived from
Escherichia coli β-glucuronidase (EC 3.2.1.31).7 Escherichia coli
β-glucuronidase is a member of the retaining β-glycosidase
Received: October 7, 2010
Published: February 24, 2011
r
2011 American Chemical Society
1992
dx.doi.org/10.1021/jo101914s J. Org. Chem. 2011, 76, 1992–2000
|

The Journal of Organic Chemistry
ARTICLE
Scheme 1. Preparation of Glucuronide Conjugates 1
Scheme 2. Proposed Mechanism of Action of (a) E. coli Wild-
Type β-Glucuronidase and (b) E. coli E504G
Glucuronylsynthase
’ RESULTS AND DISCUSSION
The development of the E. coli glucuronylsynthase system
required reliable access to both the mutant enzymes and the
R-D-glucuronyl fluoride 2 donor sugar. The wild-type and three
putative glycosynthase mutant enzymes, E504G, E504A, and
E504S, were overexpressed in a glucuronidase deficient strain of
E. coli (GMS407(DE3)) as previously described.7 A minor
modification of replacing the method of cell lysis from freeze-
thaw cycles to French pressure cell press improved the yield of
purified protein from 18 to 25 mg L-1 of culture. The enzymes
produced by this protocol are histidine tagged (His6), allowing
for purification by nickel affinity chromatography. To assess if
this modification has a deleterious effect on enzyme activity, the
kinetic competence of the wild-type enzyme was investigated by
colorimetric assay using p-nitrophenyl β-D-glucuronide as sub-
strate and monitoring enzyme mediated hydrolysis at 410 nm.12
The observed Km of 0.17 ( 0.02 mM and kcat of 38 ( 1 s-1
(21 °C, 50 mM phosphate buffer, pH 7.5) compare well with
reported literature values determined under similar conditions
for hydrolysis by a similar histidine tagged glucuronidase con-
struct (Km = 0.20 mM, kcat = 68 s-1, Tris-HCl buffer, pH 7.6)13
and also purified native enzyme (Km = 0.22 mM, 37 °C, 76 mM
phosphate buffer, pH 7.0).14
The second key component required for the glucuronyl-
synthase system is the R-D-glucuronyl fluoride 2 donor sugar.
At the outset of this study one reported synthesis15 provided the
potassium salt of this sugar in 65% yield and employed the
TEMPO-mediated oxidation of the primary hydroxyl group of
R-D-glucosyl fluoride 3.16 Purification was achieved by ion
exchange, followed by recrystallization to remove potassium
acetate buffer salt. In our hands the recrystallization of the sugar
from significant amounts of potassium acetate salt proved
challenging. This difficulty was circumvented by the use of
ammonium hydrogen carbonate as eluting buffer in the anion
exchange chromatography, which was readily removed on lyo-
philization to afford the ammonium salt as an off-white powder.
This was further purified by recrystallization from aqueous
ethanol to give the R-D-glucuronyl fluoride 2 as the ammonium
salt7 in 87% yield on a multigram scale (Scheme 3). The identity
of sugar 2 was confirmed by single crystal X-ray analysis from
crystals grown in ethanol/water/acetone, which clearly showed
the presence of the expected ammonium counterion.17 Analysis
of the material by 13C NMR in D2O solvent confirmed the
absence of residual hydrogen carbonate buffer (δ ∼160).
In our previous investigations the aglycone specificity of the
E. coli β-glucuronidase and of the putative glycosynthase en-
zymes was determined using an established spectrophotometric
screening protocol against a panel of 123 alcohol acceptors.7,11
The 13 alcohols identified by this screen included a range of
primary and cyclic secondary aliphatic alcohols, substituted
benzyl alcohols, and isomeric naphthalene methanols. None of
the 54 carbohydrates tested in the screen were identified as
acceptors. Preparative scale synthesis of glucuronide conjugates
was conducted from each of the 13 alcohol hits shown in
Scheme 4. Of the three glucuronylsynthase mutants, the
E504G proved most effective. The E504A mutant provided
lower yields in the cases examined, and the E504S mutant did
not promote glucuronide synthesis.
Reactions conducted on a small selection of alcohols not
identified by the screen also hinted at a broader substrate scope.
Phenol, an acceptor present in the initial screen but not identified
as a hit, was glucuronylated in a low 13% yield (Scheme 4).
Ethanol, an acceptor not included in the original screen, failed to
afford detectable amounts of the glucuronide product. Of great
1993
dx.doi.org/10.1021/jo101914s |J. Org. Chem. 2011, 76, 1992–2000

The Journal of Organic Chemistry
ARTICLE
Scheme 3. Synthesis of R-D-Glucuronyl Fluoride 2
Scheme 4. Product Yields for E504G Glucuronylsynthase
Mediated Synthesis
Figure 1. Plot of initial rate versus concentration for varying R-D-
glucuronyl fluoride 2 donor with fixed (97 mM) 2-phenylethanol 4
acceptor (21 °C, 100 mM phosphate buffer, pH 7.5). The line represents
the data fitted (least-squares) to the Michaelis-Menten equation.
substrate concentration led to a decrease in the isolated yield of
glucuronide product from 84% to 64% based on the acceptor
alcohol. In order to understand the factors governing this process
and to optimize glucuronide synthesis, we embarked on a study
of the kinetics of this enzyme catalyst.
The enzyme kinetics of the glucuronylsynthase reaction was
conveniently investigated by an automatically sampled reverse-
phase HPLC-UV assay monitored at 211 nm. Both alcohol
acceptor and glucuronide product concentrations were deter-
mined relative to an internal standard over time to give a measure
of initial rate (ν0). The first kinetic measurements were con-
ducted at a fixed concentration of 2-phenylethanol 4 acceptor
(97 mM), due to limited solubility of this substrate in the
aqueous buffer, and a varying concentration of R-D-glucuronyl
fluoride 2. At this concentration of alcohol acceptor, the
app
R-D-glucuronyl fluoride 2 donor gave apparent values of Km
= 15.0 ( 1 μM and kcatapp = 0.024 ( 0.01 s-1 by least-squares
fitting to the Michaelis-Menten equation (Figure 1). Because of
the kinetically subsaturating concentration of the acceptor
alcohol, the donor Kmapp and kcatapp values provide an estimate
of the kinetic parameters for this substrate. Nevertheless, the
donor Kmapp is low relative to other glycosynthase enzymes18 and
well below the concentration required for preparative enzyme-
mediated synthesis. The kcatapp or apparent turnover number for
this donor, although an underestimate of kcat, is also low relative
to other glycosynthases, leaving considerable scope for improv-
ing the catalytic efficiency of the enzyme.18
The kinetic parameters for 2-phenylethanol 4 donor were also
investigated at a fixed saturating concentration of R-D-glucuronyl
fluoride 2 donor (1 mM). Under these conditions the acceptor
alcohol did not kinetically saturate the enzyme due to the limited
aqueous solubility of the acceptor alcohol. An estimate of the
kinetic parameters for 2-phenylethanol 4 was obtained by fitting
the available data to the Michaelis-Menten equation. This
indicated a Km of 140 ( 10 mM, significantly higher than the
highest acceptor substrate concentration tested (107 mM). An
estimate of the specificityconstant kcat/Km of 0.30 ( 0.02 M-1 s-1
was obtained from the slope of the plot of initial rate against
acceptor concentration at low acceptor concentration (Figure 2,
Table 1).
interest, the steroid dehydroepiandrosterone (DHEA), an ac-
ceptor not included in the original screen, was observed to give a
modest 26% yield of the glucuronide product.
The isolated yields from these early experiments showed that
the enzyme provided high yields of glucuronide products for
some substrates, but this was not universal (Scheme 4). The
yields varied considerably for some closely related acceptors such
as 3-methoxybenzyl alcohol (84%) and 4-methoxybenzyl alcohol
(42%). In a number of instances cosolvents or detergents were
used to boost reaction yield for substrates with low aqueous
solubility. However, the beneficial effects of cosolvents and
detergents were not universally observed. Finally the variation
of reaction variables such as substrate concentration provided
some surprising results. In the case of 3-methoxybenzyl alcohol
doubling the substrate concentration was expected to increase
enzyme activity and therefore chemical yield. In fact doubling the
The significant difference in estimated Michaelis constants
(Km) for the donor and acceptor substrates is noteworthy. This
may reflect the binding interactions for glucuronide substrates at
the wild-type β-glucuronidase and the role of this enzyme in
1994
dx.doi.org/10.1021/jo101914s |J. Org. Chem. 2011, 76, 1992–2000

The Journal of Organic Chemistry
ARTICLE
Figure 3. Dependence of initial rate (ν0, solid bars) and HPLC yield
after 5 d (%, lined bars) on pH for reaction of 2-phenylethanol 4
acceptor (107 mM) and R-D-glucuronyl fluoride 2 donor (1 mM) (100
mM phosphate buffer, 21 °C).
Figure 2. Plot of initial rate versus concentration for varying 2-phenyl-
ethanol 4 acceptor with fixed saturating (1 mM) R-D-glucuronyl fluoride
2 donor (21 °C, 100 mM phosphate buffer, pH 7.5). Line represents data
fitted (least-squares) to the Michaelis-Menten equation.
at high acceptor concentrations, consistent with substrate inhibi-
tion. Because of substrate inhibition the data could not be fitted
to the Michaelis-Menten model. Kinetic parameters for these
reactions (Table 1) were estimated in two ways. An estimate of
acceptor specificity kcat/Km was obtained from the slope of the
plot of initial rate against acceptor concentration at low acceptor
concentration. A second measure of enzyme activity, the max-
imum observed kcat (kcatmax), was calculated from the initial
velocity at the concentration of the local maximum (concnmax).12
It was not possible to reliably estimate Km from this data.
The kinetic parameters for acceptors 4-7 give some insight
into the yields observed in previous synthetic work (Scheme 4).7
The observed yields parallel the maximum kcat observed with the
exception of 2-phenylethanol 4 (kcatmax = 0.020 s-1, 96%), which
Table 1. Estimated Kinetic Parameters for Alcohol Acceptors
4-8 with Constant Saturating (1 mM) R-D-Glucuronyl
Fluoride 2 Donora
afforded a yield greater than that of 3-methoxybenzyl alcohol 5
(kcat
= 0.023 s-1, 84%). Comparing estimated specificity
max
constants (kcat/Km) for acceptor alcohols 4-7 gives 3-methoxy-
benzyl alcohol 5 (1.3 ( 0.1 M-1 s-1) highest, followed by
4-fluorobenzyl alcohol 6 (0.62 ( 0.04 M-1 s-1), 2-phenyletha-
nol 4 (0.30 ( 0.02 M-1 s-1), and then phenol 7 (0.22 ( 0.01
M-1 s-1) as lowest. Interestingly 2-phenylethanol 4 substrate
that provided the highest yield of product (96%) in previous
work has among the lowest specificity constant (kcat/Km). The
higher yield observed for 2-phenylethanol 4 can be attributed to
the absence of substrate inhibition, which was relevant for all
other acceptor alcohols under the high acceptor concentrations
(∼100 mM) used in preparative work. The kinetic observations
also explain the reduction in yield based on acceptor alcohol that
was observed on doubling the substrate concentration of accep-
tor 3-methoxybenzyl alcohol from 50 mM (84%) to 100 mM
(64%). A higher yield is afforded by the reaction performed with
substrate concentration close to the local maximum of enzyme
activity (concnmax = 55 mM). The lowest yielding acceptor
acceptor
kcat/Km (M-1 s-1
)
kcatmax (s-1
)
concnmax (mM)
4
5
6
7
8
0.30 ( 0.02
1.3 ( 0.1
0.020
107b
55
0.023
0.62 ( 0.04
0.22 ( 0.01
4.3 ( 0.6
0.0081
0.0029
0.0014
37
30
0.55
a 100 mM phosphate buffer, pH 7.5, 21 °C. b The solubility limit of
2-phenylethanol 4 in aqueous buffer at 21 °C.
E. coli, the producing organism. Escherichia coli is an enteric
organism that employs a glucuronide transporter/β-glucuroni-
dase enzyme system to harvest the glucuronide residue as a
carbon source from biliary metabolites excreted in the gut.19
These glucuronide conjugates are formed from a wide variety of
xenobiotic and some endogenous compounds so the primary
recognition of the glucuronide conjugate by wild-type β-glucur-
onidase is expected for the target carbohydrate residue over the
variable aglycone portion. For the glucuronylsynthase enzyme,
the function of the wild-type enzyme is reflected in a low
max
phenol 7 (13%) gave the lowest values for kcat and kcat/Km
of the four simple alcohol acceptors investigated. This study
highlights the potential for substrate inhibition under the high
acceptor alcohol concentrations usually employed for prepara-
tive purposes. It indicates that operating at lower substrate
concentrations or employing slow addition of substrate may be
beneficial for some glucuronylsynthase reactions.
app
apparent Km for the R-D-glucuronyl fluoride 2 donor (Km
15.0 ( 1 μM) and a much higher estimated Km for the
2-phenylethanol acceptor 4 (Km 140 ( 10 mM).
Kinetic investigations on four additional acceptors 5-8
(Table 1) were conducted at a saturating concentration of R-D-
glucuronyl fluoride 2 donor (1 mM). Each of these substrates
showed a negative slope of initial rate against acceptor concentration
To further improve the glucuronylsynthase reaction, variables
such as pH, temperature, and the effect of additives were also
investigated. The pH optimum of the enzyme was determined at
a fixed concentration of R-D-glucuronyl fluoride 2 donor (1 mM)
1995
dx.doi.org/10.1021/jo101914s |J. Org. Chem. 2011, 76, 1992–2000

The Journal of Organic Chemistry
ARTICLE
Figure 4. Dependence of initial rate (ν0, solid bars) and HPLC yield
after 13 d (%, lined bars) on temperature for reaction of 3-methoxy-
benzyl alcohol 5 acceptor (89 mM) and R-D-glucuronyl fluoride 2 donor
(89 mM) (100 mM phosphate buffer, pH 7.5).
and 2-phenylethanol 4 acceptor (107 mM). Two measurements
of enzyme activity were taken at each pH value. First, the initial
rate (ν0) of reaction was measured at low conversion. Second,
the reactions were incubated for 5 d, and the yield of product
formed at end point was determined by HPLC. High yields were
obtained at pH values of 5-9 with highest initial rates across a
broad plateau between pH 6 and 9 (Figure 3). The enzyme
showed an attenuated initial rate at pH 5 but maintained activity
at this pH to afford a high yield of product after 5 d. Studies on the
hydrolytic activity of wild-type β-glucuronidase shows a similar
broad pH optimum.8,20
Figure 5. Effect of additives on the initial rate (ν0 as a percentage of the
initial velocity for the control) for reaction of 2-phenylethanol 4 acceptor
(94 mM) and R-D-glucuronyl fluoride 2 donor (1 mM) (100 mM buffer,
pH 7.5, 21 °C). Brij-56, poly(ethylene glycol) hexadecyl ether; BB,
commercial protein extraction reagent; TX-100, triton X-100; DDM,
dodecyl β-D-maltoside.
Scheme 5. Synthesis of Phenyl β-D-Glucuronide 9
Next, the enzyme activity was evaluated at several tempera-
tures (4, 11, 18, 24, 30, 37 °C), at a fixed concentration of R-D-
glucuronyl fluoride 2 donor (89 mM) and 3-methoxybenzyl
alcohol 5 acceptor (89 mM). The use of 3-methoxybenzyl
alcohol as acceptor alcohol and a reduced enzyme concentration
(0.02 mg mL-1) relative to typical synthesis experiments was
also used to allow the observation of initial rates at higher
temperatures and more keenly assess the influence of tempera-
ture on overall yield. Two measurements of enzyme activity were
taken at each temperature value. First, the initial rate (ν0) of
reaction was determined, and as expected there was an increase of
initial rate with temperature (Figure 4). Second, the reactions
were incubated for 13 d, and the yield of product formed was
determined by HPLC. The reactions maintained at 30 and 37 °C
gave similar yields after 13 d, suggesting that the higher initial
reaction rate observed at the higher temperatures is attenuated
over time by a higher rate of enzyme deactivation or R-D-
glucuronly fluoride 2 hydrolysis.21,22 The study indicated a
temperature of 30-37 °C was preferred for the glucuronyl-
synthase reactions.
Next the effect of different cosolvent and detergent additives
was investigated. One limitation of the glucuronylsynthase
reaction is the low solubility of many of the hydrophobic acceptor
alcohols relative to the glucuronide product. Additives such as
detergents or cosolvents generally offer enhancement of sub-
strate solubility but could also potentially have a deleterious
effect on enzyme activity and stability.
The effect of different additives on the initial velocity of the
enzyme was determined at a saturating concentration of R-D-
glucuronyl fluoride 2 donor (1 mM) and a fixed concentration of
2-phenylethanol 4 acceptor (94 mM). All reactions with addi-
tives demonstrated a decrease in enzyme activity (Figure 5).
Detergents maintained the highest activity with no velocities
falling below 70% of the control (no additives). Organic cosol-
vents were more detrimental to the enzyme, albeit 5% tert-
butanol and glycerol demonstrated enzymatic activities greater
than 80% of the control.23
It should be noted that in the cases where the higher acceptor
concentrations can be achieved with the aid of cosolvents, this may
occur at the cost of lower enzyme activity. Furthermore, high
substrate concentrations may not always be advantageous given
the observation of substrate inhibition. In summary, the optimal
conditions of enzyme mediated synthesis of glucuronides occur at
pH 6-9 and temperatures of 30-37 °C. Because of the influence of
substrate inhibition, the kinetic behavior of individual acceptors
must be investigated in order to optimize additional variables such as
acceptor concentration or cosolvent composition.
The reaction of phenol 7 as acceptor alcohol was investigated
to gauge the combined effect of the optimized conditions on the
1996
dx.doi.org/10.1021/jo101914s |J. Org. Chem. 2011, 76, 1992–2000

The Journal of Organic Chemistry
ARTICLE
Scheme 6. Synthesis of DHEA 3-β-Glucuronide 13
Figure 6. Reaction yield against time for reaction of CMO-DHEA 8
acceptor (1.4 mM) and varying equivalents of R-D-glucuronyl fluoride 2
donor (100 mM phosphate buffer, pH 7.5, 37 °C). (blue circles, 1 equiv;
green squares, 2 equiv; yellow diamonds, 3 equiv; orange circles, 4 equiv;
red squares, 5 equiv).
of the simple acceptor alcohols, a plot of initial rate against
acceptor concentration showed a negative slope at high acceptor
concentrations, consistent with substrate inhibition.12 Relative to
the simple alcohols this substrate gave the highest estimated kcat/
Km (4.3 ( 0.6 M-1 s-1) and smallest kcatmax (0.0014 s-1) at a
concentration of 0.55 mM, with this local maximum occurring
below the solubility limit (2 mM).
The CMO-DHEA 8 glucuronylation reaction was followed
by HPLC. Using 1 equiv of R-D-glucuronyl fluoride 2 donor gave
a reaction profile that reached a 48% HPLC yield (Figure 6).
Increasing equivalents of R-D-glucuronyl fluoride 2 afforded
higher yields of product with 5 equiv giving 98% HPLC yield.
These observations suggested the operation of product inhibi-
tion at higher conversion, a phenomenon that could be overcome
with increasing concentrations of R-D-glucuronyl fluoride 2
donor.26
reaction outcome. This substrate served as a useful benchmark, as
it was not identified by the spectrophotometric screening pro-
tocol used to identify potential acceptors and the previous
synthetic procedure had provided phenyl glucuronide in a low
13% yield. Repeating this reaction with 40 mM substrate con-
centration, just above the local maximum of enzyme activity, and
at 30 °C in pH 7.5 buffer afforded a phenyl β-D-glucuronide 9 in
an improved yield of 43% (Scheme 5).
To further test the scope of the glucuronylsynthase reaction,
we turned our attention to the synthesis of steroid glucuronides.
In previous work the glucuronylation of dehydroepiandrosterone
(10, DHEA) was achieved in 5% in phosphate buffer, 17% with
25% v/v DMSO cosolvent and 26% with 2.5% w/v dodecyl
maltoside (DDM) nonionic detergent additive.7 However, in
each case the reaction was performed as a suspension of DHEA
10 in the reaction mixture due to the low aqueous solubility of the
hydrophobic steroid. The solubility of DHEA 10 has been
estimated as 90 μM in aqueous buffer.24
To address the challenge of working with hydrophobic
acceptor alcohols, the temporary introduction of polar substitu-
ents was investigated as a strategy to boost solubility into a
suitable range without the need for organic cosolvents or
detergent additives. A three-step process was envisaged involving
(1) introduction of a polar substituent, (2) enzyme catalyzed
glucuronylation, and (3) cleavage of the polar substituent to give
the steroid glucuronide. A range of derivatives were explored
with the O-(carboxymethyl)oxime derivative appearing most
suitable for this purpose.25 The oxime unit is readily generated
by condensation of commercially available O-(carboxymethyl)
hydroxylamine reagent with a carbonyl compound promoted by
pyrrolidine (Scheme 6). The charged derivative CMO-DHEA 8
thus formed in 88% had significantly greater solubility than the
parent steroid and could be fully dissolved in aqueous buffer up
to 2 mM at 21 °C.
To assess this product inhibition, the influence of increasing
concentrations of CMO-DHEA 3-β-D-glucuronide 11 as in-
hibitor on the glucuronylsynthase reaction between 2-phenyl-
ethanol 4 (88 mM) and R-D-glucuronyl fluoride 2 at several
concentrations was investigated (Scheme 7). The analysis of
reciprocal plots indicated mixed, predominantly competitive
inhibition (Kic = 71 μM, Kiu = 180 μM) for CMO-DHEA
3-β-D-glucuronide 11.27,28 The observed competitive inhibition
app
constant is higher than the apparent Michaelis constant (Km
=
15 ( 1 μM) for R-D-glucuronyl fluoride 2. The implication of this
finding is that higher concentrations of R-D-glucuronyl fluoride 2
should successfully compete for occupancy of the enzyme active
site and maintain enzyme activity as product concentration in-
creases. This analysis is borne out in the results presented in
Figure 6 with 5 equiv (7 mM) of R-D-glucuronyl fluoride giving a
98% HPLC yield of CMO-DHEA 3-β-D-glucuronide 11.
To further assess the potential for product inhibition in the
glucuronylsynthase reaction, the reverse experiment was con-
ducted. The inhibition derived from increasing concentrations of
2-phenylethyl β-D-glucuronide 12 on the glucuronylsynthase
reaction involving CMO-DHEA 8 (0.6 mM) and R-D-glucur-
onyl fluoride 2 at several concentrations was investigated. In this
instance no inhibition was observed for 2-phenylethyl β-D-
glucuronide 12 between 0 and 75 mM.29 Interestingly for the
reaction of 2-phenylethanol 4 the absence of observable substrate
and product inhibition correlates to a high yield of the 2-phenylethyl
The higher aqueous solubility allowed for completion of an
enzyme kinetic study with a varying concentration of CMO-
DHEA 8 and a saturating concentration of R-D-glucuronyl
fluoride 2 donor (1 mM) (Table 1). In common with a number
1997
dx.doi.org/10.1021/jo101914s |J. Org. Chem. 2011, 76, 1992–2000

The Journal of Organic Chemistry
ARTICLE
Scheme 7. Study of Glucuronylsynthase Reaction Inhibition
by CMO-DHEA 3-β-D-Glucuronide 11
previously reported single-step glucuronylsynthase mediated
synthesis (26%).7 The reaction also compares favorably with
the two-step Koenigs-Knorr glycosylation followed by base-
promoted deprotection that is reported to afford DHEA 3-β-D-
glucuronide 13 in 20% yield.31 The UGT catalyzed synthesis of
DHEA 3-β-D-glucuronide 13 has not been reported.
The synthesis of testosterone 17-β-D-glucuronide 14 pro-
ceeded in a similar manner (Scheme 8). Testosterone 15 was
converted to the oxime derivative 16 in 91% yield as a 1:1.7
mixture of Z and E (major) isomers. This charged derivative 16
had a solubility significantly greater than that of the parent steroid
15 and could be fully dissolved in aqueous buffer up to 10 mM at
21 °C. This smoothly underwent glucuronylation under opti-
mized conditions and using 3 equiv of R-D-glucuronyl fluoride 2
to give the glucuronide product 17 in 72% yield. Partial cleavage
of the oxime was achieved using titanium(III) chloride in
aqueous ammonium acetate buffer to regenerate the ketone
and afford testosterone 17-β-D-glucuronide 14 in 38% yield.
The mild reaction conditions allowed recovery of unreacted
oxime 17 (53%). No cleavage of the glycosidic bond was
observed under the mild reaction conditions and no byproduct
was observed from over-reduction of the unsaturated ketone or
imine functionality.32 This three-step sequence afforded the
target glucuronide in 25% yield. This contrasts with the two-
step Koenigs-Knorr glycosylation followed by base-promoted
deprotection, which is reported to afford testosterone 17-β-D-
glucuronide 14 in 17% yield.31 The method can also be com-
pared to a recent report that described the UGT-mediated
synthesis of 11 steroid glucuronides.6a The steroid glucuronides
were afforded in 13-77% yield to provide 1.1-6.5 mg of
product. Pertinent to this study was the synthesis of testosterone
17-β-D-glucuronide 14 in a yield of 77% (6.5 mg). However,
drawbacks of this method include the high levels (1 mg mL-1) of
rat liver microsomal enzyme that is obtained through animal
sacrifice. By contrast the glucuronylsynthase affords testosterone
17-β-D-glucuronide in a yield of 24% (16 mg) over three steps.
Scheme 8. Synthesis of Testosterone 17-β-D-Glucuronide 14
Furthermore the method uses soluble enzyme (0.2 mg mL-1
)
that is readily obtained on a large scale by standard methods of
protein expression and nickel affinity purification.
’ CONCLUSION
β-D-glucuronide 12 product (96%) using 1.2 equiv of R-D-
glucuronyl fluoride donor. For the reaction of CMO-DHEA 8
both substrate and product inhibition are observed in this study.
In the former case this can be overcome by selection of a suitable
substrate concentration and in the latter case by maintaining a
high concentration of the R-D-glucuronyl fluoride donor 2.
Performing the glucuronylation of CMO-DHEA 8 on a
preparative scale under optimized conditions and using 5 equiv
of R-D-glucuronyl fluoride donor and 1.9 mM acceptor alcohol
gave the glucuronide product 11 in 98% yield (Scheme 6). A
single crystal X-ray structure of the disodium salt of the glucur-
onide 11, grown from ethyl acetate/methanol/water solution
confirmed the expected structure.30 The oxime exists solely as the
E-isomer and the glycosidic linkage presents as the β anomer.
Cleavage of the oxime unit was readily achieved using titanium-
(III) chloride in aqueous ammonium acetate buffer to regenerate
the ketone and afford DHEA 3-β-D-glucuronide 13 in quantita-
tive yield. This three-step sequence afforded the target glucur-
onide (45 mg) in 86% yield, which compares favorably with the
Improved procedures for the glucuronylsynthase catalyzed
synthesis of glucuronide conjugates have been developed. The
influence of acceptor substrate, pH, temperature, cosolvents, and
detergents on the enzyme activity has been investigated by
HPLC/UV. Substrate inhibition was observed for the majority
of acceptor alcohols investigated, indicating that the substrate
concentration is an important consideration for optimizing
enzyme activity. The operation of mixed, predominantly com-
petitive inhibition was observed for the product CMO-DHEA
3-β-D-glucuronide 11 in the glucuronylsynthase reaction of
2-phenylethanol 4 as acceptor. This inhibition could be over-
come through the use of higher concentrations of R-D-glucuronyl
fluoride 2 donor. The temporary introduction of polar substit-
uents was investigated as a way to increase the solubility of
hydrophobic steroidal substrates. In this way the synthesis of
DHEA 3-β-D-glucuronide 13 was achieved in three steps and in
86% overall yield from DHEA 10. The glucuronylsynthase
reaction provides a practical alternative to existing methods for
the synthesis of glucuronide metabolites. Further engineering
1998
dx.doi.org/10.1021/jo101914s |J. Org. Chem. 2011, 76, 1992–2000

The Journal of Organic Chemistry
ARTICLE
will be directed to increasing the substrate scope and improving
the catalytic efficiency of the glucuronylsynthase enzyme.
added. The aqueous titanium trichloride was added via cannula to the
stirred CMO-DHEA 3-β-D-glucuronide solution at room temperature.
The reaction instantly turned black-violet upon addition and gradually
changed to a white-gray as the reaction proceeded. After 2 h, the reaction
was deemed complete so the solution was acidified to pH 2 and dried
onto silica. The dry residue was subjected to flash chromatography
(7:2:1 ethyl acetate/methanol/water þ0.1% formic acid) to isolate
DHEA 3-β-D-glucuronide 13 (44.8 mg, 100%) as a colorless solid;
[R]20D -29 (c 0.70, MeOH), (lit.31 [R]25D -35.5 (c 1, EtOH)); Rf 0.25
(7:2:1 ethyl acetate/methanol/water). IR (NaCl): 3369 (O-H), 2938,
’ EXPERIMENTAL SECTION
17-Carboxymethoximino-dehydroepiandrosterone (CMO-
DHEA) 8.33 Pyrrolidine (600 μL, 7.19 mmol) was added to DHEA 10
(1.00 g, 3.47 mmol) dissolved in dry methanol (40 mL) at 4 °C. After 1 h
the solution had turned yellow. Carboxymethoxylamine hemihydrochloride
(800 mg, 7.32 mmol) was dissolved in a dry solution of methanol (10 mL)
and pyrrolidine (600 μL, 7.19 mmol) and transferred to the DHEA mixture
via cannula with methanol washing (5 mL). The solution immediately
cleared and was heated to reflux. The reaction was complete after 6 h. The
solvent was removed under reduced pressure, and water (100 mL) was
added to the residue. The pH was adjusted to 2 with aqueous hydrochloric
acid (2 M), and the white precipitate was extracted with ethyl acetate until
all precipitate had dissolved (5 ꢀ 150 mL with sonication required). The
organic extracts were combined, washed with water (200 mL), dried over
magnesium sulfate, and then evaporated to dryness. Cold chloroform
(5 mL) was added to the resulting yellow-white residue (1.15 g). The
white precipitate was filtered and washed with cold chloroform (2 ꢀ 3 mL)
to afford 17-carboxymethoximino-dehydroepiandrosterone 8 (1.10 g,
1
2903 (C-H), 1733 (CdO), 1636. H NMR (800 MHz, MeOD): δ
5.42 (1H, d, J = 4.9 Hz), 4.45 (1H, d, J = 7.8 Hz), 3.80 (1H, d, J = 9.7 Hz)
3.58-3.47 (2H, m), 3.38 (1H, t, J = 9.1 Hz), 3.20 (1H, t, J = 8.4 Hz),
2.48-1.43 (2H, m), 2.29 (1H, m), 2.16-2.07 (3H, m), 2.00-1.87 (3H,
m), 1.78 (1H, m), 1.74-1.52 (5H, m), 1.40-1.23 (2H, m), 1.11-1.04
(2H, m), 1.07 (3H, s), 0.90 (3H, s). 13C NMR (75 MHz, MeOD/D2O):
δ 225.5, 169.8, 142.1, 122.2, 102.1, 79.7, 77.6, 74.7, 73.5, 72.5, 52.9, 51.6,
39.5, 38.3, 37.9, 36.8, 32.7, 32.5, 31.8, 30.4, 22.7, 21.4, 19.8, 13.9, one
carbon overlapping or obscured. LRMS (-ESI) m/z: 927 ([2M - H]-,
43%), 463 ([M - H]-, 100). HRMS (-ESI) calcd for C25H35O8
([M - H]-) 463.2338, found 463.2342.
88%); mp 215-217 °C (decomp); [R]20 -36 (c 1.0, DMSO) {lit.33
’ ASSOCIATED CONTENT
D
[R]24D -37.9 (c 1, EtOH)}; Rf 0.38 (7:2:1 ethyl acetate/methanol/water).
IR (KBr): 3351 (broad, O-H), 2946, 2910, 2865, 2505 (C-H), 1680
(CdO). 1H NMR (800 MHz, d6-DMSO): δ 5.28 (1H, s), 4.61 (1H, s,
broad), 4.42 (2H, s), 3.26 (1H, obscured, m), 2.46 (1H, dd, J = 18.9, 9.0
Hz), 2.38 (1H, m), 2.16 (1H, dd, J = 12.9, 2.4 Hz), 2.09 (1H, t, J= 12.1 Hz),
1.98 (1H, m), 1.82-1.70 (3H, m), 1.67 (1H, d, J = 12.2 Hz), 1.58 (2H, m),
1.51 (1H, m), 1.42 (1H, m), 1.38-1.28 (3H, m), 1.12 (1H, m), 1.02-0.91
(2H, m), 0.96 (3H, s), 0.85 (3H, s), COOH not observed. 13C NMR (200
MHz, d6-DMSO): δ 171.3, 170.6, 141.4, 120.1, 70.0, 69.8, 53.5, 49.8, 43.5,
42.2, 36.9, 36.2, 33.8, 31.4, 30.82, 30.78, 25.7, 22.8, 20.2, 19.2, 16.8. LRMS
(þESI) m/z: 384 ([M þ Na]þ, 100%). LRMS (-ESI) m/z: 360 ([M -
H]-, 100%).; HRMS (þESI) calcd for C21H31NO4Naþ ([M þ Na]þ)
384.2151, found 384.2151.
S
Supporting Information. CIF files for the X-ray crystal
b
structures of R-D-glucuronyl fluoride 2 and CMO-DHEA
3-β-D-glucuronide 11, protein expression procedures, experimen-
tal procedures, spectroscopic data and 1H and 13C NMR spectra
for compounds 2, 8, 9, 11, 13, 14, 16 and 17. This material is
available free of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
*E-mail: malcolm.mcleod@anu.edu.au.
CMO-DHEA 3-β-D-Glucuronide 11. Glucuronylsynthase
(4.14 mL, 1.45 mg/mL) was added to a solution containing CMO-
DHEA 8 (20 mg, 0.055 mmol, final concentration 1.9 mM) and
R-D-glucuronyl fluoride 2 (57.5 mg, 0.270 mmol) in 100 mM sodium
phosphate buffer pH 7.5 (25 mL). The reaction was incubated at 37 °C
without agitation for 3 days and then dried onto reverse-phase silica. The
dried residue was subjected to reverse-phase flash chromatography (25%
aqueous acetonitrile þ 0.1% formic acid) to isolate CMO-DHEA 3-β-
D-glucuronide 11 (29 mg, 98%) as a colorless solid; [R]20D -62 (c 1.0,
DMSO); Rf 0.02 (7:2:1 ethyl acetate/methanol/water). IR (KBr): 3423
’ ACKNOWLEDGMENT
We thank the Australian Government Department of Health
and Ageing Anti-Doping Research Program for financial support.
’ REFERENCES
(1) (a) Lucaciu, R.; Ionescu, C. Farmacia 2005, 53, 10–19. (b)
Shipkova, M.; Wieland, E. Clin. Chim. Acta 2005, 358, 2–23. (c) Wells,
P. G.; Mackenzie, P. I.; Chowdhury, J. R.; Guillemette, C.; Gregory,
P. A.; Ishii, Y.; Hansen, A. J.; Kessler, F. K.; Kim, P. M.; Chowdhury,
N. R.; Ritter, J. K. Drug Metab. Dispos. 2004, 32, 281–290. (d) King,
C. D.; Rios, G. R.; Green, M. D.; Tephly, T. R. Curr. Drug Metab. 2000,
1, 143–161. (e) Mulder, G. J. Annu. Rev. Pharmacol. Toxicol. 1992,
32, 25–49.
1
(O-H), 2942 (C-H), 1745 (CdO). H NMR (800 MHz, D2O): δ
5.51 (1H, s), 4.59 (1H, d, J = 7.3 Hz), 4.38 (2H, s), 3.72-3.69 (2H, m),
3.52-3.50 (2H, m), 3.27 (1H, t, J = 7.1 Hz), 2.63 (1H, dd, J = 19.2, 8.3
Hz), 2.52 (1H, m), 2.47 (1H, d, J = 12.6 Hz), 2.29 (1H, t, J = 12.0 Hz),
2.10 (1H, d, J = 11.3 Hz), 2.00-1.92 (2H, m), 1.90-1.84 (2H, m),
1.71-1.60 (4H, m), 1.55 (1H, m), 1.46-1.41 (2H, m), 1.27 (1H, m),
1.10 (1H, t, J = 13.0 Hz), 1.06 (3H, s), 1.06-1.02 (1H, m), 0.95 (3H, s).
13C NMR (200 MHz, D2O): δ 178.9, 177.3, 177.2, 143.3, 123.4, 101.7,
81.0, 77.7, 77.7, 74.4, 73.3, 72.9, 55.0, 51.2, 45.6, 39.5, 38.1, 38.0, 34.9,
32.29, 32.27, 30.3, 27.8, 24.2, 21.6, 20.3, 17.7. LRMS (-ESI) m/z: 536
([M - H]-, 100%), 558 ([M - 2H þ Na]-, 30). HRMS (-ESI) calcd
for C27H38NO10 ([M - H]-) 536.2496, found 536.2495; calcd for
C27H37NO10Na ([M - 2H þ Na]-) 558.2315, found 558.2302.
DHEA 3-β-D-Glucuronide 13.31 CMO-DHEA 3-β-D-glucuro-
nide 11 (50 mg, 0.093 mmol) and ammonium acetate (96 mg, 1.2 mmol)
were purged with nitrogen, and then dioxane (0.5 mL) and aqueous acetic
acid (50%, 38 μL) were added. In a second flask, titanium trichloride (36
mg, 0.23 mmol) was purged with nitrogen before water (6 mL) was
€
(2) (a) Th€orngren, J.-O.; Oestervall, F.; Garle, M. J. Mass Spectrom.
2008, 43, 980–992. (b) Schulze, J. J.; Lundmark, J.; Garle, M.; Skilving,
I.; Ekstr€om, L.; Rane, A. J. Clin. Endocrinol. Metab. 2008, 93, 2500–2506.
(c) Van Eenoo, P.; Lootens, L.; Spaerkeer, A.; Van Thuyne, W.;
Deventer, K.; Delbeke, F. T. J. Anal. Toxicol. 2007, 31, 543–548. (d)
Avois, L.; Mangin, P.; Saugy, M. J. Pharm. Biomed. Anal. 2007,
44, 173–179. (e) Spyridaki, M.-H.; Kiousi, P.; Vonaparti, A.; Valavani,
P.; Zonaras, V.; Zahariou, M.; Sianos, E.; Tsoupras, G.; Georgakopoulos,
C. Anal. Chim. Acta 2006, 573-574, 242–249. (f) Jimꢀenez, C.; Ventura,
R.; Williams, J.; Segura, J.; de la Torre, R. Analyst 2004, 129, 449–455.
(3) (a) Kim, H.-J.; Ahn, K. C.; Ma, S. J.; Gee, S. J.; Hammock, B. D.
J. Agric. Food Chem. 2007, 55, 3750–3757. (b) Hebestreit, M.; Flenker,
U.; Buisson, C.; Andre, F.; Le Bizec, B.; Fry, H.; Lang, M.; Weigert, A. P.;
Heinrich, K.; Hird, S.; Sch€anzer, W. J. Agric. Food Chem. 2006,
1999
dx.doi.org/10.1021/jo101914s |J. Org. Chem. 2011, 76, 1992–2000

The Journal of Organic Chemistry
ARTICLE
54, 2850–2858. (c) Hubert, T. D.; Bernardy, J. A.; Vue, C.; Dawson,
V. K.; Boogaard, M. A.; Schreier, T. M.; Gingerich, W. H. J. Agric. Food
Chem. 2005, 53, 5342–5346. (d) Llorach, R.; Tomꢀas-Barberꢀan, F. A.;
Ferreres, F. Eur. Food Res. Technol. 2005, 220, 31–36.
(4) (a) Stachulski, A. V.; Harding, J. R.; Lindon, J. C.; Maggs, J. L.;
Park, B. K.; Wilson, I. D. J. Med. Chem. 2006, 49, 6931–6945. (b)
Stachulski, A. V.; Jenkins, G. N. Nat. Prod. Rep 1998, 15, 173–186. (c)
Kaspersen, F. M.; Van Boeckel, C. A. A. Xenobiotica 1987,
17, 1451–1471.
Int. Ed. 2007, 46, 3885–3888. (i) Jahn, M.; Stoll, D.; Warren, R. A. J.;
Szabꢀo, L.; Singh, P.; Gilbert, H. J.; Ducros, V. M.-A.; Davies, G. J.;
Withers, S. G. Chem. Commun. 2003, 1327–1329.
(19) Liang, W.-J.; Wilson, K. J.; Xie, H.; Knol, J.; Suzuki, S.;
Rutherford, N. G.; Henderson, P. J. F.; Jefferson, R. A. J. Bacteriol.
2005, 187, 2377–2385.
(20) (a) Goto, J.; Sato, A.; Suzaki, K.; Nambara, T. Chem. Pharm.
Bull. 1981, 29, 1975–1980. (b) Bowers, L. D.; Johnson, P. R. Biochim.
Biophys. Acta 1981, 661, 100–105.
(5) For recent chemical synthesis, see: (a) Engstrom, K. M.; Daanen,
J. F.; Wagaw, S.; Stewart, A. O. J. Org. Chem. 2006, 71, 8378–8383. (b)
Harding, J. R.; King, C. D.; Perrie, J. A.; Sinnott, D.; Stachulski, A. V. Org.
Biomol. Chem. 2005, 3, 1501–1057. (c) Polꢀakovꢀa, M.; Pitt, N.; Tosin, M.;
Murphy, P. V. Angew. Chem., Int. Ed. 2004, 43, 2518–2521. (d)
Ferguson, J. R.; Harding, J. R.; Killick, D. A.; Lumbard, K. W.;
Scheinmann, F.; Stachulski, A. V. J. Chem. Soc., Perkin Trans. 1
2001, 3037–3041.
(6) For recent enzymic synthesis, see: (a) Hintikka, L.; Kuuranne,
T.; Aitio, O.; Thevis, M.; Sch€anzer, W.; Kostiainen, R. Steroids 2008,
73, 257–265. (b) J€antti, S. E.; Kiriazis, A.; Reinil€a, R. R.; Kostiainen,
R. K.; Ketola, R. A. Steroids 2007, 72, 287–296. (c) Khymenets, O.;
Joglar, J.; Clapꢀes, P.; Parella, T.; Covas, M.-I.; de la Torre, R. Adv. Synth.
Catal. 2006, 348, 2155–2162. (d) Kuuranne, T.; Aitio, O.; Vahermo, M.;
Elovaara, E.; Kostiainen, R. Bioconjugate Chem. 2002, 13, 194–199. (e)
Stevenson, D. E.; Hubl, U. Enzyme Microb. Technol. 1999, 24, 388–396.
(f) Werschkun, B.; Wendt, A.; Thiem, J. J. Chem. Soc., Perkin Trans. 1
1998, 3021–3024.
(7) Wilkinson, S. M.; Liew, C. W.; Mackay, J. P.; Salleh, H. M.;
Withers, S. G.; McLeod, M. D. Org. Lett. 2008, 10, 1585–1588.
(8) (a) Gomes, R. L.; Meredith, W.; Snape, C. E.; Sephton, M. A.
J. Pharm. Biomed. Anal. 2009, 49, 1133–1140. (b) Wakabayashi, M.;
Fishman, W. H. J. Biol. Chem. 1961, 236, 996–1001.
(9) Wong, A. W.; He, S.; Withers, S. G. Can. J. Chem. 2001,
79, 510–518.
(10) Mackenzie, L. F.; Wang, Q.; Warren, R. A. J.; Withers, S. G.
J. Am. Chem. Soc. 1998, 120, 5583–5584.
(21) The first order rate constant for hydrolysis (kH) of 1.8 ꢀ 10-6
s-1 was determined for R-D-glucuronyl fluoride 2 by 1H NMR analysis
(unbuffered D2O, 37 °C). Details are reported in Supporting Informa-
tion. This hydrolysis rate constant (kH) is of equivalent magnitude to
other reported nonenzymatic glycosyl fluoride hydrolysis: (a) 3-O-
cellobiosyl-R-D-glucosyl fluoride kH = 2.57 ꢀ 10-6 s-1 (50 mM maleate
buffer, pH 7.0, 0.1 mM CaCl2 35 °C) Faijes, M.; Pꢀerez, X.; Pꢀerez, O.;
Planas, A. Biochemistry 2003, 42, 13304–13318. (b) R-D-glucopyranosyl
fluoride kH = 8.42 ꢀ 10-6 s-1 (50 mM sodium acetate buffer, pH 6.0,
40 °C) Zhang, Y.; Bommuswamy, J.; Sinnott, M. L. J. Am. Chem. Soc.
1994, 116, 7557–7563. (c) R-D-glucopyranosyl fluoride kH = 1.5 ꢀ 10-6
s-1 (5 mM potassium phosphate buffer, 30 °C) Banait, N. S.; Jencks,
W. P. J. Am. Chem. Soc. 1991, 113, 7951–7958. The hydrolysis of R-D-
glucopyranosyl fluoride is catalyzed by phosphate buffer and is subject to
isotope effects. (d) Banait, N. S.; Jencks, W. P. J. Am. Chem. Soc. 1991,
113, 7958–7963. We are currently investigating the influence of buffer
composition on the hydrolysis of R-D-glucuronyl fluoride 2.
(22) The HPLC reaction yields increased on average 3.1 times from
day 4 to day 13 for each temperature regime. The HPLC yield of the
reaction held at 37 °C for 13 d was 2.9 times greater than that
determined at 4 d. Details are reported in Supporting Information.
(23) A more comprehensive study of wild-type E. coli β-glucuroni-
dase activity and stability in a wide range of cosolvents and detergents is
reported in Supporting Information. Similar results are observed for the
influence of cosolvent and detergent activity on both the wild-type
glucuronidase and the glucuronylsynthase enzymes. For the effects of
cosolvents on immobilized E. coli β-glucuronidase activity see ref 20b.
(24) Laneri, S.; Sacchi, A.; di Frassello, E. A.; Luraschi, E.; Colombo,
P.; Santi, P. Pharm. Res. 1999, 16, 1818–1824.
(11) For reviews on glycosynthases, see: (a) Rakiꢀc, B; Withers, S. G.
Aust. J. Chem. 2009, 62, 510–520. (b) Hancock, S. M.; Vaughan, M. D.;
Withers, S. G. Curr. Opin. Chem. Biol. 2006, 10, 509–519. (c) Williams,
S. J.; Withers, S. G. Aust. J. Chem. 2002, 55, 3–12.
(25) Derivatives investigated included: hydrazone, 2-(2-pyri-
dyl)hydrazone, oxime, O-(carboxymethyl)oxime.
(26) Nonenzymatic hydrolysis of R-D-glucuronyl fluoride was ex-
cluded as a significant factor in the incomplete reactions on kinetic
grounds. See ref 21.
(27) (a) Cortꢀes, A.; Cascante, M.; Cꢀardenas, M. L.; Cornish-
Bowden, A. Biochem. J. 2001, 357, 263–268. (b) Cornish-Bowden, A.
Biochem. J. 1974, 137, 143–144. (c) Dixon, M. Biochem. J. 1953,
55, 170–171.
(28) Plots of 1/ν0 against [CMO-DHEA 3-β-D-glucuronide]
(Dixon plot), [R-D-glucuronyl fluoride]/v0 against [CMO-DHEA 3-β-
D-glucuronide], and 1/i0.5 against ν0/Vmax are reported in Supporting
Information.
(29) A plot of initial rate against concentration of 2-phenylethyl β-D-
glucuronide 12 (0-75 mM) for the glucuronylsynthase reaction of
CMO-DHEA 8 and R-D-glucuronyl fluoride 2 is reported in Supporting
Information.
(30) The ORTEP diagram and CIF file of the CMO-DHEA 3-β-D-
glucuronide 11 crystal structure (CCDC 793480) is provided in
Supporting Information.
(31) Wotiz, H. H.; Smakula, E.; Lichtin, N. N.; Leftin, J. H. J. Am.
Chem. Soc. 1959, 81, 1704–1708.
(32) Blaszczak, L. C; McMurry, J. E. J. Org. Chem. 1974,
39, 258–259.
(33) Rosenfeld, R. S.; Rosenberg, B.; Kream, J.; Hellman, L. Steroids
(12) Plots of initial rate against substrate concentration are reported
in Supporting Information.
(13) Geddie, M. L.; Matsumura, I. J. Biol. Chem. 2004,
279, 26462–26468.
(14) Kim, D.-H; Jin, Y.-H.; Jung, E.-A.; Han, M.-J.; Kobashi, K. Biol.
Pharm. Bull. 1995, 18, 1184–1188.
(15) Heeres, A.; van Doren, H. A.; Gotlieb, K. F.; Bleeker, I. P.
Carbohydr. Res. 1997, 299, 221–227.
(16) Kitahata, S.; Brewer, C. F.; Genghof, D. S.; Sawai, T.; Hehre,
E. J. J. Biol. Chem. 1981, 256, 6017–6026.
(17) The ORTEP diagram and CIF file for the R-D-glucuronyl
fluoride 2 crystal structure (CCDC 793479) is provided in Supporting
Information.
(18) Glycosynthases derived from exo-glycosidases: (a) Ben-David,
A.; Bravman, T.; Balazs, Y. S.; Czjzek, M.; Schomburg, D.; Shoham, G.;
Shoham, Y. ChemBioChem 2007, 8, 2145–2151. (b) Hommalai, G.;
Withers, S. G.; Chuenchor, W.; Ketudat Cairns, J. R.; Svasti, J.
Glycobiology 2007, 17, 744–753. (c) Faijes, M.; Saura-Valls, M.; Pꢀerez,
X.; Conti, M.; Planas, A. Carbohydr. Res. 2006, 341, 2055–2065. (d)
Nashiru, O.; Zechel, D. L.; Stoll, D.; Mohammadzadeh, T.; Warren,
R. A. J.; Withers, S. G. Angew. Chem., Int. Ed. 2001, 40, 417–420. (e)
Mayer, C.; Zechel, D. L.; Reid, S. P.; Warren, R. A. J.; Withers, S. G. FEBS
Lett. 2000, 466, 40–44. Glycosynthases derived from endo-glycosidases:
(f) Hancock, S. M.; Rich, J. R.; Caines, M. E. C.; Strynadka, N. C. J.;
Withers, S. G. Nat. Chem. Biol. 2009, 5, 508–514. (g) Piens, K;
Henriksson, A.-M.; Gullfot, F.; Lopez, M.; Faurꢀe, R.; Ibatullin, F. M.;
Teeri, T. T.; Driguez, H.; Brumer, H. Org. Biomol. Chem. 2007,
5, 3971–3978. (h) Yang, M.; Davies, G. J.; Davis, B. G. Angew. Chem.,
1973, 21, 723–733.
2000
dx.doi.org/10.1021/jo101914s |J. Org. Chem. 2011, 76, 1992–2000