Probing synergy between two catalytic strategies in the glycoside hydrolase O-GlcNAcase using multiple linear free energy relationships

Article abstract of DOI:10.1021/ja904506u

Human O-GlcNAcase plays an important role in regulating the post-translational modification of serine and threonine residues with β-O-linked N-acetylglucosamine monosaccharide unit (O-GlcNAc). The mechanism of O-GlcNAcase involves nucleophilic participation of the 2-acetamido group of the substrate to displace a glycosidically linked leaving group. The tolerance of this enzyme for variation in substrate structure has enabled us to characterize O-GlcNAcase transition states using several series of substrates to generate multiple simultaneous free-energy relationships. Patterns revealing changes in mechanism, transition state, and rate-determining step upon concomitant variation of both nucleophilic strength and leaving group abilities are observed. The observed changes in mechanism reflect the roles played by the enzymic general acid and the catalytic nucleophile. Significantly, these results illustrate how the enzyme synergistically harnesses both modes of catalysis; a feature that eludes many small molecule models of catalysis. These studies also suggest the kinetic significance of an oxocarbenium ion intermediate in the O-GlcNAcase-catalyzed hydrolysis of glucosaminides, probing the limits of what may be learned using nonatomistic investigations of enzymic transition-state structure and offering general insights into how the superfamily of retaining glycoside hydrolases act as efficient catalysts.

Full text of DOI:10.1021/ja904506u

Published on Web 08/28/2009
Probing Synergy between Two Catalytic Strategies in the
Glycoside Hydrolase O-GlcNAcase Using Multiple Linear Free
Energy Relationships
Ian R. Greig,†,‡ Matthew S. Macauley,† Ian H. Williams,‡ and David J. Vocadlo*,†
Department of Chemistry, Simon Fraser UniVersity, 8888 UniVersity DriVe, Burnaby, British
Columbia, Canada, V5A 1S6, and Department of Chemistry, UniVersity of Bath, Bath,
United Kingdom, BA2 6AS
Received June 3, 2009; E-mail: dvocadlo@sfu.ca
Abstract: Human O-GlcNAcase plays an important role in regulating the post-translational modification of
serine and threonine residues with ꢀ-O-linked N-acetylglucosamine monosaccharide unit (O-GlcNAc). The
mechanism of O-GlcNAcase involves nucleophilic participation of the 2-acetamido group of the sub-
strate to displace a glycosidically linked leaving group. The tolerance of this enzyme for variation in substrate
structure has enabled us to characterize O-GlcNAcase transition states using several series of substrates
to generate multiple simultaneous free-energy relationships. Patterns revealing changes in mechanism,
transition state, and rate-determining step upon concomitant variation of both nucleophilic strength and
leaving group abilities are observed. The observed changes in mechanism reflect the roles played by the
enzymic general acid and the catalytic nucleophile. Significantly, these results illustrate how the enzyme
synergistically harnesses both modes of catalysis; a feature that eludes many small molecule models of
catalysis. These studies also suggest the kinetic significance of an oxocarbenium ion intermediate in the
O-GlcNAcase-catalyzed hydrolysis of glucosaminides, probing the limits of what may be learned using
nonatomistic investigations of enzymic transition-state structure and offering general insights into how the
superfamily of retaining glycoside hydrolases act as efficient catalysts.
to over 107-fold.2,4 Of course, such analyses are hindered in
their accuracy since deletion of the nucleophilic side chain
Introduction
General acid/base and nucleophilic catalysis are among the
most common mechanisms used by enzymes to increase rates
of reaction above those of uncatalyzed reactions. Many enzymes
catalyzing phosphoryl, acyl, or glycosyl group transfer reactions
use these two catalytic strategies concomitantly.1 Structure-
reactivity studies of enzymes making use of altered substrates
and mutant enzymes have firmly established the independent
importance of both of these catalytic strategies. Experimental
studies of enzymes that probe the interaction and synergy of
these two mechanistic strategies, however, are notably lacking.
Glycoside hydrolases are a vast superfamily of enzymes that
can comprise up to 1% of the genome of species.2 The
pioneering determination of the X-ray structure of hen egg white
lysozyme3 has spurred sustained speculation and study into the
catalytic and mechanistic roles of enzymic active site nucleo-
philes and general acids/bases in glycoside hydrolases.4 Studies
evaluating the contribution to catalysis by nucleophiles of
glycosidases commonly use mutagenesis and the resulting
decreases in kcat/Km values have been found to range from 10-
results in a dramatic perturbation of the electrostatic character
of an enzyme active site by changing, for example, the pKa
values of other active-site residues5 and may even alter the
stereochemical outcome of glycosyl transfer.6 Owing to limita-
tions in the number of naturally occurring amino acids, probing
the role of the catalytic nucleophile in a more delicate manner,
such as by attenuating nucleophilicity through a series of
systematic changes, has not been realized. We were intrigued
by this problem and, owing to our interest in the catalytic
mechanism of a glycoside hydrolase known as O-GlcNAcase,
recognized the feasibility of investigating this fundamental issue
using this enzyme as a representative test case for the large
superfamily of glycoside hydrolases.2,4
O-GlcNAcase (OGA) is a ꢀ-glucosaminidase from family
GH84 responsible for removing 2-acetamido-2-deoxy-ꢀ-D-
glucopyranose moieties from post-translationally modified serine
and threonine residues of nucleocytoplasmic proteins.7 This form
of glycosylation has been implicated in various disease states,8
† Simon Fraser University.
(5) McIntosh, L. P.; Hand, G.; Johnson, P. E.; Joshi, M. D.; Ko¨rner, M.;
Plesniak, L. A.; Ziser, L.; Wakarchuk, W. W.; Withers, S. G.
Biochemistry 1996, 35, 9958–9966.
‡ University of Bath.
(1) Fersht, A. Structure and Mechanism in Protein Science; W. H. Freeman
& Co. Ltd.: New York, 1999.
(6) Wang, Q. P.; Graham, R. W.; Trimbur, D.; Warren, R. A. J.; Withers,
S. G. J. Am. Chem. Soc. 1994, 116, 11594–11595.
(7) Gao, Y.; Wells, L.; Comer, F. I.; Parker, G. J.; Hart, G. W. J. Biol.
Chem. 2001, 276, 9838–9845.
(2) Davies, G. J.; Henrissat, B. Biochem. Soc. Trans. 2002, 30, 291–297.
(3) Blake, C. C. F.; Koenig, D. F.; Mair, G. A.; North, A. C. T.; Phillips,
D. C.; Sarma, V. R. Nature 1965, 206, 757–761.
(4) Vocadlo, D. J.; Davies, G. J. Curr. Opin. Chem. Biol. 2008, 12, 539–
555.
(8) Hart, G. W.; Housley, M. P.; Slawson, C. Nature 2007, 446, 1017–
1022.
9
10.1021/ja904506u CCC: $40.75  2009 American Chemical Society
J. AM. CHEM. SOC. 2009, 131, 13415–13422 13415

A R T I C L E S
Greig et al.
Figure 1. (A) O-GlcNAcase hydrolyses 2-acetamido-2-deoxy-ꢀ-D-glucopyranosides with retention of stereochemistry at the anomeric center using a substrate-
assisted catalytic mechanism. (B) The series of aryl substrates synthesized and studied: R ) CH3, CH2F, CHF2, CF3, and n-Bu and X ) 3,4-DNP, 3-F-4-
NO2, 4-NO2, 4-MU, 4-Cl-3-NO2, 4-CN, 3-NO2, 3,4-DF, 4-Cl, H, and 4-OMe. (C) The series of methyl glycosides synthesized and studied; R ) CH3, CH2F,
CHF2, CF3.
and an understanding of the catalytic mechanism of O-
GlcNAcase9 has led to the discovery of inhibitors that are
proving useful in determining the role of the O-GlcNAc
modification within a cellular context.10-12 O-GlcNAcase uses
a substrate-assisted catalytic mechanism in which the 2-aceta-
mido group of the substrate acts as the catalytic nucleophile to
form an oxazoline intermediate (Figure 1A), in contrast to the
large majority of glycoside hydrolases that use an enzymic
nucleophile to cleave their substrates with retention of config-
uration at the anomeric center. Notably, the structure of two
bacterial homologues of human O-GlcNAcase have been solved
and reveal a spacious active site which tolerates larger groups
appended to the 2-acetamido group.13,14 The biological sub-
strates of O-GlcNAcase are hundreds of different O-GlcNAc-
modified proteins,8 and O-GlcNAcase must therefore tolerate
many different leaving group structures. Consistent with this
observation, O-GlcNAcase tolerates glucosaminides possessing
various substituted phenolic leaving groups.9,15,16
We recognized that these tolerances of the active site to
variation in these substituents of O-GlcNAcase substrates would
enable us to concomitantly vary the electronic nature of both
the nucleophile and leaving group in a quantifiable manner. In
this way a series of linear free energy relationships could be
generated that correlate different thermodynamic parameters of
the substrates in solution, with the kinetic parameters governing
their enzyme-catalyzed hydrolysis. Such linear free energy
relationships are a widely used tool to study both enzyme
catalysis and small molecule reactions in solution since they
provide insight into how changes in the electronic properties
of the substrate affect the sensitivity and, therefore, structure
of the transition state of the reaction of interest. We speculated
that, by using concomitant variation in two substituents of the
substrate, the enzymic transition state could be investigated
through multiple simultaneous free-energy relationships using
an approach favored by Jencks to study small molecule
systems.17 Here we investigate the O-GlcNAcase-catalyzed
hydrolysis of a large series of substrates (Figure 1B,C) using
such multiple free energy relationships. These results allow us
to evaluate the synergistic interaction between nucleophilic and
acid catalysis and discuss the kinetic significance of an enzyme-
sequestered oxocarbenium ion intermediate in glycosyl group
transfer.
Materials and Methods
Substrate and Enzyme Preparation. Substrates were prepared
according to previously reported synthetic schemes.16 Spectroscopic
characterization data for novel compounds may be found in the
Supporting Information. Recombinant wild-type O-GlcNAcase was
prepared according to previously reported procedures.16
Steady-State Kinetic Analysis. For the majority of substrates,
enzyme specificity constants (kcat/Km) were determined by following
substrate depletion using substrate concentrations at least 10 times
lower than previously determined Km values.15 Stock solutions
containing 100 µM substrate and 1% DMSO (v/v) in reaction buffer
(pH 7.4 phosphate-buffered saline) were diluted 4-fold into the
reaction buffer for kinetic assays. Reactions were followed spec-
trophotometrically at wavelengths reported in the Supporting
Information (Table S1). First-order rate constants were determined
from a nonlinear least-squares fit to the observed change in
absorbance with time and the values reported represent the mean
of at least three separate reactions. Where prior data were available,
kinetic parameters determined using these DMSO-containing reac-
tion volumes were identical (within error) to those previously
determined in the absence of DMSO. For substrates having
trifluorinated and difluorinated nucleophiles and for slow-reacting
substrates possessing poor leaving groups with small absorbance
changes occurring at wavelengths at which significant protein
absorbance occurs, values of kcat/Km were evaluated by the initial
rates method. The values of kcat/Km determined by substrate
depletion and initial rates methods for the difluorinated compounds
were identical within experimental error.
Investigation of the enzyme-catalyzed hydrolysis of methyl
glycosides Me-GlcNAc-F0 and Me-GlcNAc-F1 were carried out as
follows. Reactions were carried out in a total reaction volume of
200 µL using PBS (pH 7.4) as a buffer. Reactions contained a final
concentration of 0, 10, 20, or 40 µM substrate and were initiated
by the addition of enzyme (100 nM), after which the reaction
mixture was incubated at 37 °C for 1 h. Reactions were terminated
by the addition of 800 µL of 95% cold ethanol and stored at -20
°C for 30 min to allow protein to precipitate. Upon termination of
(9) Macauley, M. S.; Whitworth, G. E.; Debowski, A. W.; Chin, D.;
Vocadlo, D. J. J. Biol. Chem. 2005, 280, 25313–25322.
(10) Macauley, M. S.; Bubb, A. K.; Martinez-Fleites, C.; Davies, G. J.;
Vocadlo, D. J. J. Biol. Chem. 2008, 283, 34687–34695.
(11) Dorfmueller, H. C.; Borodkin, V. S.; Schimpl, M.; van Aalten, D. M. F.
Biochem. J. 2009, 420, 221–227.
(12) Yuzwa, S. A.; Macauley, M. S.; Heinonen, J. E.; Shan, X.; Dennis,
R. J.; He, S.; Whitworth, G. E.; Stubbs, K. A.; McEachem, E. J.;
Davies, G. J.; Vocadlo, D. J. Nat. Chem. Biol. 2008, 4, 483–490.
(13) Dennis, R. J.; Taylor, E. J.; Macauley, M. S.; Stubbs, K. A.;
Turkenburg, J. P.; Hart, S. J.; Black, G. N.; Vocadlo, D. J.; Davies,
G. J. Nat. Struct. Mol. Biol. 2006, 13, 1079–1085.
(14) Rao, F. V.; Dorfmueller, H. C.; Villa, F.; Allwood, M.; Eggleston,
I. M.; van Aalten, D. M. EMBO J. 2006, 25, 1569–1578.
(15) C¸ etinbas¸, N.; Macauley, M. S.; Stubbs, K. A.; Drapala, R.; Vocadlo,
D. J. Biochemistry 2006, 45, 3835–3844.
(16) Macauley, M. S.; Stubbs, K. A.; Vocadlo, D. J. J. Am. Chem. Soc.
2005, 127, 17202–17203.
(17) Jencks, W. P. Chem. ReV. 1985, 85, 511–527.
9
13416 J. AM. CHEM. SOC. VOL. 131, NO. 37, 2009

Synergy in Catalysis by O-GlcNAcase
A R T I C L E S
the reaction, 2 µL of 1 mM fucose was added to each reaction as
an internal standard. The suspension was centrifuged at 17 000 rpm
(Eppendorf 5415C) for 20 min and the supernatant was collected
and then dried by vacuum centrifugation. The residues were
dissolved in 150 µL of ddH2O, vortexed, and centrifuged at 17 000
rpm to remove residual insoluble debris. Carbohydrates were
separated by high performance anion exchange chromatography
(ASI 100 automated sample injector, Carbopack PA20 column, and
ICS 3000; Dionex) and detected using an electrochemical detector
(ED50, Dionex) using a gold working electrode and an Ag/AgCl
reference electrode. An isocratic elution of 20 mM NaOH was used,
which afforded optimal separation of fucose (3 min) and the
hemiacetal products of the enzymatic reaction, GlcNAc (GlcNAc-
F0, 8 min) and GlcNAc-F1 (12 min). Substrates (F0 and F1 methyl
glycosides) eluted within the first 2 min. The amounts of the
products produced in the reactions were determined by integration
of the peak corresponding to GlcNAc-F0 or GlcNAc-F1 and
corrected by comparison to the internal fucose standard. The
absolute amount of each product formed was determined using
standard curves constructed for GlcNAc-F0 or GlcNAc-F1 (the
standard curve for GlcNAc-F1 was generated by complete enzymatic
hydrolysis of 3-fluoro-4-nitrophenyl-GlcNAc-F1). All reactions were
carried out in triplicate and the values of kcat/Km were determined
by plotting initial rates obtained for each of the concentrations and
taking the slope.
(data not shown), indicating no gross change in mechanism,
such as change to an inverting mechanism, which would lead
to formation of alternative products. Anomerization to a
mixture containing predominantly the R-anomer was observed
over time, and analysis of chemical shifts supports assign-
ment of these products as GlcNAc-F0 or GlcNAc-F3 hemiac-
etals.
Free Energy Relationships. Variations in the log(kcat/Km)
values obtained are illustrated as a function of leaving group
ability, which are defined by the pKa value of the conjugate
acid of the leaving group aglycon, (pKa)1g. Plotting these two
free energy terms against each other yields a series of
Brønsted slopes that provide a measure of the relative
negative charge build up on the leaving group oxygen.4
Similarly, plotting the Taft σ* parameter against log(kcat/Km)
provides quantitative insight into N-acyl group nucleophilic in-
volvement at the transition state.4,9,19
The quantitative description of the regimens of different
gradients found in the Brønsted plots of Figure 2A may be found
in Table 1, and the fits to these data are described by equations
1 and 2 outlined below. For substrates possessing the naturally
occurring nucleophiles, reactivities are correlated by a single
Brønsted slope as previously described.15 Free energy relation-
ships comprising a series of compounds that all react via the
same mechanism and kinetically significant, rate-determining,
transition state are linear. Equation 1 shows the correlation for
such a typical Brønsted plot.
Solvent kinetic isotope effects were determined in phosphate-
buffered saline at pD ) 7.4. The stereochemistry of the first formed
product of O-GlcNAcase-catalyzed N-acetylglucosaminide hydroly-
sis was determined in D2O (pD ) 7.4, phosphate buffered saline).
Enzyme inhibition constants (Ki) were determined using previously
reported procedures.10
log(kcat/Km) ) ꢀ1(pKa)lg + c1
(1)
Results
The reactivities of fluorinated N-acylglucosaminides display
multiphasic Brønsted plots with regimens (I and III) of negative
gradient at high and low leaving group pKa values, separated
by a regimen (II) of positive gradient at intermediate leaving
group pKa values. The Brønsted plots for the reactivity of Ar-
GlcNAc-F2 and Ar-GlcNAc-F3 series clearly display both
upward (change in mechanism) and downward (change in rate-
determining step) deflections (see Discussion). The minimal
kinetic model used to correlate these data is given in Sch-
eme 1.
We prepared 52 substrates (Ar-GlcNAc-Rs, Figure 1B) in
which both steric (R ) n-butyl) and electronic [R ) methyl
(F0), fluoromethyl (F1), difluoromethyl (F2), and trifluoromethyl
(F3)] perturbations are made to the 2-acetamido group of the
substrate, while a variety of substituted phenols (Ar) were
installed at the leaving group position (see Supporting Informa-
tion for details of their synthesis and characterization). For all
aryl glycosides prepared we avoided the use of ortho-substituted
phenols, since these have been found to have effects on binding
of substrates to glycoside hydrolases.18 To investigate substrates
having leaving groups with pKa values closer to those of the
natural substrates of O-GlcNAcase, we also prepared a series
of methyl glycosides with corresponding changes to the aceta-
mido group: Me-GlcNAc-F0, Me-GlcNAc-F1, Me-GlcNAc-F2,
Me-GlcNAc-F3. We then investigated the O-GlcNAcase-
catalyzed hydrolysis of these series of substrates (See Supporting
Information for the complete data).
Scheme 1. Minimal Kinetic Scheme for O-GlcNAcase Processing
a
Ar-GlcNAc-F2 and Ar-GlcNAc-F3
O-GlcNAcase-Catalyzed Hydrolysis of N-Acetylglucosaminides
Occurs with Retention of Stereochemistry. Previous studies have
shown that deletion of the catalytic nucleophile in glycoside
hydrolases can change the catalytic mechanism, resulting in
an altered stereochemical outcome for the catalyzed reaction.
a A change in kinetically significant mechanism and kinetically significant
rate-determining step are clearly demarked in the free energy relationships.
R represents the O-GlcNAcase substrate, I represents a kinetically significant
intermediate, and P represents the product of the reaction, in this case the
oxazoline intermediate.
1
To address this possibility, we carried out studies using H
NMR spectroscopy and monitored the time-dependent O-
GlcNAcase-catalyzed hydrolysis of four representative Ar-
GlcNAc-Rs in D2O possessing different leaving groups (Ar
) 3-fluoro-4-nitrophenyl or 3,4-difluorophenyl) and either
good or poor nucleophiles (R ) F0 or F3). The results show
that for all substrates the first formed product is the ꢀ-anomer
The observed rate of reaction (kcat/Km) may therefore be
expressed in terms of individual rate constants according to
equation 2.
k2
kcat/Km ) k1 +
(2)
1 + k /k
(
)
3
-2
(18) Street, I. P.; Kempton, J. B.; Withers, S. G. Biochemistry 1992, 31,
9970–9978.
(19) Vocadlo, D. J.; Withers, S. G. Biochemistry 2005, 44, 12809–12818.
9
J. AM. CHEM. SOC. VOL. 131, NO. 37, 2009 13417

A R T I C L E S
Greig et al.
Figure 2. Effects of simultaneous variation of nucleophile and leaving group on the second-order rate constant of the O-GlcNAcase-catalyzed reaction,
binding effects of fluorine substitution, and the dependence of ꢀlg values on σ* values. (A) Brønsted plots showing the changes in kcat/Km with
variation in (pKa)lg as the nucleophile structure is varied. Regimen I (shaded dark gray) represents the rate-limiting departure of the leaving group as
a phenolate anion (ANDN). The upward break passing from regimen I to II represents a change in mechanism to acid catalyzed leaving group departure.
Positive slopes in regimen II (shaded light gray) reveal that the formation of an intermediate is rate-determining (DN‡*AN). The subsequent downward
break between regimens II and III represents a change in rate-determining step such that in regimen III (unshaded) breakdown of an intermediate is
rate-determining (DN*AN‡). (B) Observed rates for n-Bu (]) and CH3 hydrolysis (b) as well as the extrapolated rates for the chemical step of CH3
hydrolysis (upper line lacking symbols). (C) Fluorine substitution has little effect on binding of methyl N-fluoroacetylglucosaminides to O-GlcNAcase.
(D) ꢀlg values for hydrolysis of Ar-GlcNAc-Fn glycosides in regimens II and III vary depending on the strength of the N-acyl nucleophile (reflected
by the σ* value), revealing the synergistic interaction of general acid and nucleophilic catalysis. Fits shown are nonlinear fits determined as described
in the text. The ꢀlg value for the CH2F nucleophile in regimen II has been excluded due to limited data in this region.
The (Brønsted) free energy relationship associated with
log(kcat/Km) is therefore described by eq 3.
Table 1. Brønsted Data for a Variety of Substrates Possessing
Various Nucleophilesa
nucleophile
regimen
ꢀlg
c
r
n
10ꢀ +c
1 + 10∆ꢀ+∆c
2
2
log(kcat/Km) ) log 10ꢀ +c
+
(3)
1
1
13
CH3
-
III
I
II
III
I
II
III
I
-0.11 ( 0.01
-0.32 ( 0.03
-0.92 ( 0.09
+1.38 ( 0.34
-0.71 ( 0.46
-0.93 ( 0.11
+1.64 ( 0.48
-1.17 ( 0.62
-0.95 ( 0.04
-0.23 ( 0.01
+4.89 ( 0.06
+6.11 ( 0.29
+9.10 ( 0.54
-7.46 ( 2.60
+9.30 ( 3.50
+8.66 ( 0.68
-10.08 ( 3.69
+12.90 ( 4.82
+9.10 ( 0.27
+4.11 ( 0.07
0.984
0.980
0.997
11
8
11
[
]
CH2F
CHF2
In which ꢀ1 and c1 are the gradients and offsets, respectively,
associated with the linear free energy relationship of rate
constant k1 (similarly ꢀ2 and c2). ∆ꢀ (equal to ꢀ-2 - ꢀ3) and
∆c (equal to c-2 - c3) report on differences in gradients and
offsets of linear free energy relationships associated with the
relevant rate constants.
CF3
0.994
9
n-Bu
0.999
0.998
3
5
III
To probe the chemical step for a nucleophile whose intrinsic
strength is similar to that of the 2-acetamido group, the reactivity
of a series of substrates possessing the bulkier n-butyl nucleo-
phile was investigated. The Brønsted plot for these substrates
was biphasic with two regimens (I and III) of negative gradient.
For this series of compounds, which displays a break in the
free energy relationship, linear free energy relationships are used
a Where sufficient data exists and regimens I, II, and III are
clearly defined (i.e., for Ar-GlcNAc-F2 and Ar-GlcNAc-F3),
least-squares fit to the appropriate kinetic expression (eq 3) was
carried out. Where particular regimens are poorly delineated (e.g.,
regimens I and II for Ar-GlcNAc-F1), linear fits to compounds falling
within clearly defined regimens are carried out (e.g., regimen III for
Ar-GlcNAc-F1).
a
9
13418 J. AM. CHEM. SOC. VOL. 131, NO. 37, 2009

Synergy in Catalysis by O-GlcNAcase
A R T I C L E S
to correlate data within the difference regimens. This approach
is used because regions in which mechanism and rate-determin-
ing step change are not well-defined and a least-squares fit to a
kinetic scheme representing solely a change in mechanism
(upward break) is inappropriate. Since the data for the Ar-
GlcNAc-F1 are suggestive of, yet do not clearly define, a region
having positive gradient, we treat the data for this series of
compounds in the same manner.
O-GlcNAcase Cleavage of Methyl Glycosides. We were able
to determine the second order rate constants (kcat/Km) for the
O-GlcNAcase-catalyzed hydrolysis of Me-GlcNAc-F0 and
Me-GlcNAc-F1 (see Supporting Information for the complete
data) using an HPLC assay in which the formation of the
products are monitored; however, due to limitations in
experimental design and assay sensitivity, we were unable
to obtain kinetic data for Me-GlcNAc-F2 and Me-GlcNAc-
F3. The rate constants obtained for the Me-GlcNAc-F0 and
Me-GlcNAc-F1 substrates correlate well with the data
obtained for the aryl glycosides in regimen III (Figure 4,
see Conclusion).
Solvent Kinetic Isotope Effects (KIEs) Are Sensitive to
Nucleophilic Strength. Small and normal solvent KIEs, H(kcat/
Km)/D(kcat/Km), were measured with several representative Ar-
GlcNAc-Rs possessing different leaving groups (Ar )
3-fluoro-4-nitrophenyl or 3,4-difluorophenyl) and either good
or poor nucleophiles (R ) F0 or F3). Values of 1.48 ( 0.05
and 1.28 ( 0.06 were determined for 3-fluoro-4-nitrophenyl-
GlcNAc-F0 (pKa ) 6.4, regimen I) and 3,4-difluorophenyl-
GlcNAc-F0 (pKa ) 9.1, regimen III), respectively. Larger
solvent KIEs were measured for the enzyme-catalyzed
hydrolysis of 3-fluoro-4-nitrophenyl-GlcNAc-F3 (2.2 ( 0.1,
pKa ) 6.4, regimen I) and 3,4-difluorophenyl-GlcNAc-F3 (2.8
( 0.2, pKa ) 9.1, regimen III).
Substrate Binding Is Weakly Sensitive to the Nature of
the Nucleophile. Because kcat/Km values reflect the free energy
change on going from free enzyme and substrate to the
transition state of the first irreversible step of the reaction,
the reactivity parameters we attribute to on-enzyme chemical
steps may be complicated by effects associated with substrate
binding.20 To address whether addition of fluorine to the
2-acetamido group caused significant perturbations to binding,
a series of Me-GlcNAc-Fns were tested as competitive
inhibitors. As shown in Figure 2C, binding is well-correlated
but very weakly sensitive to increasing fluorine substitution
of the N-acyl group:
single positive inflection for substrates with the natural good
nucleophile are entirely unprecedented. Given their apparent
complexity, we will discuss our interpretations for each regimen
separately.
Substrates in Regimen I. Before embarking on an interpreta-
tion of the data in this regimen, several prior observations require
consideration. Earlier mutagenesis studies of O-GlcNAcase have
identified Asp175 as the general acid catalyst responsible for
assisting departure of the aglycon leaving group,15 and the X-ray
structure of Bacteriodes thetaiotaomicrometer ꢀ-glucosamini-
dase (BtGH84), a GH84 enzyme having conserved active site
residues with human O-GlcNAcase, supports this assignment.13
Key to the ensuing discussions, a kinetic pKa value of ap-
proximately 7.8 has been assigned to Asp175.15,16 Substrates
possessing phenolic leaving groups having pKa values below
that of the general acid cannot benefit significantly from acid
catalysis, as proton transfer is thermodynamically unfavorable.21
Thus, the reactivity of compounds in regimen I (Figure 2A,B)
must arise from a mechanism involving departure of phenolate
anion without the benefit of acid catalysis, here termed
“spontaneous departure”.
The magnitude of the slope of plots of log(kcat/Km) against
the (pKa)lg indicates the extent of negative charge accumulation
on the glycosidic oxygen in the transition state. The above
analysis suggests that we should observe a large negative ꢀlg
value for substrates having (pKa)lg values below that of the
general acid catalyst (pKa ≈ 7.8), since these must depart as
the phenolate anion. The reactivities of all Ar-GlcNAc-F0s
(which span the pKa of the enzymic general acid), however,
are well-correlated by a single line, implying that no variation
in mechanism or rate-determining step occurs with (pKa)lg. This
observation, however, is inconsistent with the lack of general
acid catalysis for substrates having low (pKa)lg. We speculated
that this pattern of reactivity associated with Ar-GlcNAc-F0s
(Figure 2A) may be readily explained by a kinetically significant
step (e.g., an enzyme conformational change) that precedes
glycosidic bond cleavage, which is the first irreversible chemical
step of the reaction. One piece of evidence supporting this
interpretation is the observed solvent KIEs. For the O-GlcNA-
case-catalyzed hydrolysis, significantly smaller KIEs were
observed for substrates possessing the 2-acetamido nucleophile
than for those KIEs found for substrates possessing the
2-trifluoroacetamido nucleophile, suggesting the KIE is partly
masked by a significant contribution from a nonchemical step
to the overall rate constant. Further support for this proposal is
that the biphasic Brønsted plot of log(kcat/Km), previously
determined for the Asp175Ala mutant-catalyzed hydrolysis of
Ar-GlcNAc-F0s, reveals a negligible gradient for substrates
having (pKa)lg values less than the pKa of the general acid
catalyst (regimen I), whereas for those substrates having (pKa)lg
values greater than 7.8, a significant negative slope is observed.
These data suggest that, for the Asp175Ala mutant, a kinetically
significant step precedes cleavage of the glycosidic linkage for
good substrates, whereas for worse leaving groups cleavage of
the glycosidic bond is the kinetically significant step.15
log(1/Ki) ) (0.07 ( 0.01)(σ*)nuc + (1.91 ( 0.01)
(r ) 0.999, n ) 4)
On the basis of this finding, our subsequent analysis of the
influence of different nucleophiles on reactivity can be
interpreted without requiring corrections for steric or elec-
tronic effects on binding.
Discussion
The complex multiple free energy relationship measured here
for O-GlcNAcase offers unique insight into the catalytic
mechanism of glycoside hydrolases, enabling us to gain an
understanding of the interplay between general acid catalysis
and nucleophilic catalysis. The observed pattern of positive and
negative inflections for substrates with poor nucleophiles and a
To test this hypothesis, we aimed to make the chemical step
rate-determining while minimizing changes to the structure of
the transition states associated with the chemical steps. Indeed,
this strategy has been used with great success in other enzyme
classes.22 We therefore prepared a series of substrates having a
(21) Jencks, W. P. Acc. Chem. Res. 1980, 13, 161–169.
(20) Schramm, V. L. Curr. Opin. Chem. Biol. 2007, 11, 529–536.
(22) Kline, P. C.; Schramm, V. L. Biochemistry 1993, 32, 13212–13219.
9
J. AM. CHEM. SOC. VOL. 131, NO. 37, 2009 13419

A R T I C L E S
Greig et al.
bulky n-butyl alkyl chain, Ar-GlcNAc-n-Bu, which possess a
nucleophile of similar intrinsic strength [σ*(n-Bu) ) -0.15] to
the 2-acetamido group [σ*(Me) ) 0.00]. On the basis of
previous studies with O-GlcNAcase,23 the increased bulk of the
n-butyl alkyl group raises by similar amounts the energy of both
the rate-determining transition state and that of the Michaelis
complex relative to free enzyme and substrate. The net result
is that the chemical steps become rate-determining (Figure 2B).
Consistent with our prediction, we find for this series of Ar-
GlcNAc-n-Bu a ꢀlg value of -0.95 for those substrates having
leaving groups that cannot benefit from general acid catalysis
(regimen I) and a ꢀlg value of -0.23 for substrates with worse
leaving groups that can benefit from general acid catalysis
(regimen III). Notably, the steep negative slope (ꢀlg ) -0.95)
in regimen I for the Ar-GlcNAc-n-Bu series is comparable to
that observed for the Asp175Ala mutant (ꢀlg ) -1.0),15
consistent with this process reflecting expulsion of the leaving
group as a phenolate. This ꢀlg value also closely resembles those
found for the Ar-GlcNAc-F2 (-0.92) and Ar-GlcNAc-F3
(-0.93) in regimen I. Thus, when values of ꢀlg report on the
chemical step for spontaneous glycoside hydrolysis, rather than
some preceding nonchemical rate-determining step, the extent
of negative charge buildup on the leaving group is similar
regardless of nucleophilic strength. These results indicate that
the transition states for the spontaneous expulsion of phenoxides
from glucosaminides occurring in aqueous and this enzymic
environment are broadly similar: they both involve significant
negative charge accumulation on the leaving group and signifi-
cant nucleophilic participation.24,25
Using this data, and on the basis of prior studies investigating
the influence of the N-acyl substituent on transition state and
transition state analogue inhibitor binding, we estimate values of
kcat/Km that would be observed for the hydrolysis of Ar-GlcNAc-
F0 if the chemical step were kinetically significant (Figure 2B and
Supporting Information).23 Throughout the paper we use these
extrapolated kcat/Km values for the Ar-GlcNAc-F0 series.
The significance of nucleophilic participation at the transition
state for the chemical step is quantified by the Taft-reaction
sensitivity (F*). Obtaining F* involves correlating the Taft
parameter of each substrate (σ*), which reflect the differing
strengths of the N-acyl nucleophile of each substrate, against
the log(kcat/Km) values obtained for each corresponding substrate
in each series. For substrates in the Ar-GlcNAc-F0, Ar-GlcNAc-
F2, and Ar-GlcNAc-F3 series having leaving groups with (pKa)lg
values that place them in regimen I, the Taft-parameter is
negative, suggesting significant nucleophilic participation that
is fairly similar in all series [(F*)kcat/Km ) -0.85 ( 0.04; data
not shown]. Collectively, the linear and parallel Brønsted (Figure
2A,B) and Taft plots obtained for substrates in regimen I, for
which the chemical step is clearly rate-determining, are con-
sistent with a concerted (SN2-like, more clearly described as an
ANDN mechanism26) transition state that remains relatively
invariant with changing leaving group and nucleophile structure.
a change in rate-determining step, and positive deviations denote
a change in mechanism. The positive deviations that we
consistently observe as leaving group pKa values increase
beyond that of 4-nitrophenol (pKa ) 7.2) (regimen I passing to
regimen II), therefore, reflect an enforced change in mechanism.
More specifically, a change from spontaneous to acid-catalyzed
glucosaminide cleavage seems most likely as proton transfer
to the leaving group becomes thermodynamically favorable.
Single positive deviations, as observed for the Ar-GlcNAc-n-
Bu series, have previously been reported for a 1,3-1,4-ꢀ-D-
glucan hydrolase.27 However, the large positive ꢀlg values of
regimen II, following the inflection, have not been observed in
glycoside hydrolase-catalyzed reactions. Furthermore, the pat-
terns of subsequent negative deviations on passing from regimen
II to III are unprecedented in free energy relationships for
glycosyl hydrolases, perhaps because attenuating the nucleo-
philicity of an enzymic group has been difficult or perhaps
because such a downward break has not yet fallen into a region
that is convenient to observe experimentally. This negative
inflection, however, suggests a change in rate-determining step
arising from the existence of a kinetically significant intermedi-
ate state (SN1-like, more clearly described as an DN*AN
mechanism26). The microscopic nature of this intermediate state
is of obvious interest and there is a compelling mechanistic
rationale offering insight into its nature, as well as providing
the basis for the trends in the Brønsted plots.
For glycoside hydrolases4 and many ribosyl transferases28 that
possess preorganized active site nucleophiles, a mechanism in
which motion of the anomeric carbon defines the reaction
coordinate (Figure 1A) is invoked. This process may occur
through either a concerted transition state (ANDN) or through a
dissociative pathway possessing a kinetically significant, oxo-
carbenium ion-like intermediate (DN*AN). For glycosidases, the
anomeric center is generally thought to move from bonded
contact with the leaving group to bonded contact with the
enzymic nucleophile, which both remain stationary.4,29 Accord-
ingly, if a kinetically significant oxocarbenium ion intermediate
lies along the reaction coordinate, then the overall rate of
‡
reaction may be limited by either its rate of formation (DN *AN)
‡
or its collapse (DN*AN ). The character of this intermediate
would therefore be highly sensitive to the nature of the flanking
nucleophile; decreasing nucleophile strength would increase the
ionic character of the intermediate, contributing to its kinetic
significance and giving shape to the Brønsted plots for poor
nucleophiles. Electronically, this is consistent with decreasing
´
(27) Planas, A.; Abel, M.; Millet, O.; Palas´ı, J.; Plallare´s, C.; Viladot, J.-
L. Carbohydr. Res. 1998, 310, 53–64.
(28) Schramm, V. L. Acc. Chem. Res. 2003, 36, 588–596.
(29) Davies, G. J.; Ducros, V. M.-A.; Varrot, A.; Zechel, D. L. Biochem.
Soc. Trans. 2003, 31, 523–527.
(30) Ballardie, F. W.; Capon, B.; Dearie, W. M.; Foster, R. L. Carbohydr.
Res. 1976, 49, 79–92.
(31) Alternative interpretations of the data that might account for the observed
reactivity trends observed can be envisioned. For example, one alternative
proposal is that regimen II might reflect rate-determining formation of a
cationic intermediate and regimen III might reflect rate-determining
diffusion of the phenol away from the cationic intermediate within the
enzyme active site (rather than attack of the nucleophile). Although such
a mechanism is not inconsistent with the data presented in this paper, we
do not favor this alternative interpretation. These (SNi-type) mechanisms
may operate for certain glycosyl transferases (and possibly even some
mutant glycosyl hydrolases);32 however, for O-GlcNAcase the presence
of a preorganized nucleophilic species of reasonable strength on the
opposing face of the incipient cationic intermediate to the leaving group
suggests that an intermediate possessing significant covalent character
will be formed.
Substrates with Poor Nucleophiles in Regimens II and III.
Inflections in linear free energy relationships, such as a Brønsted
plot, can stem from two phenomena:1 negative deviations reflect
(23) Whitworth, G. E.; Macaulay, M. S.; Stubbs, K. A.; Dennis, R. J.;
Taylor, E. J.; Davies, G. J.; Greig, I. R.; Vocadlo, D. J. J. Am. Chem.
Soc. 2007, 129, 635–644.
(24) Cocker, D.; Sinnott, M. L. J. Chem. Soc. Perkin Trans. 2 1976, 2,
618–620.
(25) Dunn, B. M.; Bruice, T. C. AdV. Enzymol. 1973, 37, 1–60.
(26) Guthrie, R. D.; Jencks, W. P. Acc. Chem. Res. 1989, 22, 343–349.
9
13420 J. AM. CHEM. SOC. VOL. 131, NO. 37, 2009

Synergy in Catalysis by O-GlcNAcase
A R T I C L E S
Therefore, the decreases in ꢀlg values observed for substrates
having poorer nucleophiles (Figure 2D), which have a greater
partial positive charge on the glycosyl moiety of the intermedi-
ate, is consistent with progressively tighter association between
the glycosyl moiety and protonated leaving groups.31
Substrates with Better Nucleophiles and Poor Leaving
Groups. The natural substrates of O-GlcNAcase possess a
2-acetamido group that is a significantly stronger nucleophile
than fluorinated 2-N-acyl groups. O-GlcNAcase uses this
nucleophile to displace serine or threonine leaving groups of
polypeptide chains having (pKa)lg values exceeding that of
Asp175. Therefore, characterization of transition states falling
within regimen III for substrates possessing the 2-acetamido
group provide greatest insight into the biologically significant
O-GlcNAcase mechanism.
Perhaps the most striking feature of the free energy relation-
ships of regimens II and III for substrates bearing good
nucleophiles is the absence of a clear negative deviation in the
Brønsted relationship (Figure 2A,B). There are two possible
explanations for the absence of such a break: the oxocarbenium
ion-like intermediate state is no longer kinetically significant
and a concerted ANDN mechanism is followed instead of a
dissociative DN*AN pathway; in this scenario the break does
not exist. Alternatively, the break does exist but now falls within
regimen I, for which an entirely different mechanism operates.
Although it is difficult to draw firm conclusions as to whether
a change in mechanism from dissociative to concerted pathways
has actually occurred or not in the specific case of the
2-acetamido nucleophile, trends of decreasingly positive values
of ꢀlg in regimen II and decreasingly negative values of ꢀlg in
regimen III are relatively clear (Figure 2D). The convergence
of these values with increasing nucleophilic strength indicates
that a change in mechanism from dissociative to concerted
pathways very likely occurs for an appropriately strong nucleo-
phile, perhaps, for example, for the acetamido group of the
normal substrate of O-GlcNAcase.33 Such a change in mech-
anism is consistent with the expectation that a positively charged
glycosyl group will collapse much more rapidly with ap-
propriately positioned strong nucleophiles. In any event, further
investigations using comprehensive kinetic isotope effect studies
are being undertaken to definitively clarify these and other
detailed mechanistic proposals.36
Figure 3. Proposal describing the formation and fate of intermediates
in O-GlcNAcase-catalyzed glycosyl hydrolysis. The rate-determining step
depends on the nature of the leaving group and nucleophile. For the
best leaving groups, regimen I reflects a concerted (ANDN) pathway with
departure of the leaving group as a phenolate being operative. For poor
nucleophiles, regimen II reflects formation of an oxocarbenium ion being
rate determining and regimen III reflects collapse of this intermediate
(DN*AN) pathways. For acid-assisted departure of the leaving group,
either increasing nucleophile strength or decreasing leaving group ability
will ultimately change the pathway followed from DN*AN to ANDN. For
the different strength nucleophiles used in this study n ) 0-3 and m )
3 - n, with increasing numbers of fluorine atoms decreasing nucleo-
philicity.
nucleophilic strength resulting in greater partial positive charge
on the glycosyl moiety of the intermediate.
The reaction followed by compounds having poor nucleo-
philes in regimen II is represented in Figure 3 (poor nucleophile
and good leaving group). The transition state that lies en route
to the cationic intermediate is expected to be rate-determining.
The rationale for the positive ꢀlg values is that protonation of
the leaving group significantly leads glycosidic bond fission for
all substrates in this regimen. Substrates with worse nucleophiles
have more positive ꢀlg values (Figure 2D) and this can be
explained by the progressively greater extent of partial positive
charge on the glycosyl moiety of the intermediate. As the charge
on the glycosyl moiety increases, the leaving group oxygen bears
more relative positive charge, and this leads to consequent
increases in ꢀlg values.
The reaction followed by compounds in regimen III is
represented in Figure 3 (poor nucleophile and poor leaving
group). This regimen defines a rate-determining step leading to
collapse of the oxocarbenium ion intermediate. The N-acetyl
group has been shown to be a nucleophilic participant in
substitution reactions occurring even in solution, a condition
where it must be disordered in the ground state.30 This strongly
suggests that ultimately a covalent oxazoline or oxazolinium
ion intermediate will be formed in the enzyme active site. A
DN*AN mechanism rationalizes observed reactivity trends: more
basic phenols interact more strongly with the cationic intermedi-
ate in the active site. As a consequence, the back reaction
becomes more favorable with more basic phenols and so
formation of the cationic intermediate likely becomes reversible.
Also, the rates of trapping of the cationic intermediate by the
N-acyl nucleophile decrease with increasing pKa (ꢀlg is negative).
Conclusion
The preorganized nature of the substrate within the active
site of O-GlcNAcase involves strategic positioning of both
nucleophile and general acid catalyst, a feature likely enabling
concomitant acidic and nucleophilic catalysis. The potential
synergy between these two modes of catalysis may be inves-
tigated by extrapolation of our data to a pKa of 16, the
approximate pKa of the natural substrate of O-GlcNAcase
(32) Lairson, L. L.; Henrissat, B.; Davies, G. J.; Withers, S. G. Annu. ReV.
Biochem. 2008, 77, 521–555.
(33) Symmetry consideration (i.e., the nucleophilic species in the forward
direction is equivalent to the leaving group in the reverse direction
and vice versa) suggests that for sufficiently poor leaving groups
(strong nucleophiles in the reverse direction) a change in mechanism
from dissociative (DN*AN) to concerted (ANDN) pathways may also
occur. This change in mechanism may be used to explain unusual
upward breaks in the Brønsted plots observed for the acid-catalyzed
hydrolysis of some acetals in solution.34,35
(34) Jensen, J. L.; Herold, L. R.; Lenz, P. A.; Trusty, S.; Sergi, V.; Bell,
K.; Rogers, P. J. Am. Chem. Soc. 1979, 101, 4672–4677.
(35) Capon, B.; Nimmo, K. J. Chem. Soc. Perkin Trans. 2 1975, 1113–
1118.
9
J. AM. CHEM. SOC. VOL. 131, NO. 37, 2009 13421

A R T I C L E S
Greig et al.
the case of O-GlcNAcase acting on substrates with good
nucleophiles and poor leaving groups, for which this enzyme
has evolved, effective synergy of these catalytic strategies likely
alter the reaction coordinate, such that the oxocarbenium ion
intermediate ceases to be kinetically significant and the reaction
possibly becomes concerted (Figure 3). Further investigation
of this mechanistic proposal, using detailed kinetic isotope effect
studies, are being undertaken to further assess these microscopic
interpretations and will be reported in due course. More
significantly, here we provide compelling evidence for synergy
between nucleophilic and general acid catalysis, which, although
widely speculated upon in many enzyme systems, has proven
remarkably difficult to capture in small molecule models of
enzymic catalysis,40 including N-acetylglucosaminides,30,41 or
to experimentally demonstrate in enzyme systems. Such syn-
chronization, however, is undoubtedly a key factor contributing
to the proficient nature of enzyme catalysts, including glycoside
hydrolases, as we reveal here experimentally for O-GlcNAcase.
Figure 4. Evidence that O-GlcNAcase synergistically harnesses nucleo-
philic and general acid catalysis. The lines show rates of hydrolysis
extrapolated to a leaving group pKa of 16 for a mechanism involving (A)
spontaneous departure of leaving group using a crippled nucleophile (Ar-
GlcNAc-F3s in regimen I, 9), (B) spontaneous departure of leaving group
using the naturally occurring nucleophile (Ar-GlcNAc-F0s in regimen I,
b), (C) acid-catalyzed leaving group departure using a crippled nucleophile
(Ar-GlcNAc-F3s in regimen III, 0), (D) acid-catalyzed leaving group
departure using the naturally occurring nucleophile (Ar-GlcNAc-F0s,
regimen III, O). The extrapolated rates are likely reasonable since rates of
hydrolysis of the methyl 2-N-acyl-ꢀ-glucopyranosides Me-GlcNAc-F0 (bold
line, D, O) and Me-GlcNAc-F1 (dashed line, ]) correlate well with linear
regressions for the processing of aryl substrates having either the acetamido
or the monofluoracetamido nucleophile, respectively, that define regimen
III.
Acknowledgment. We thank David Shen for generous expert
assistance in carrying out kinetic assays using HPLC. Dr. Bennet
is thanked for enjoyable discussions. For support of this research
we thank the Natural Sciences and Engineering Research Council
of Canada (NSERC) for a grant to D.J.V., the Mizutani Foundation
for Glycoscience for a grant to I.R.G. and I.H.W. I.R.G. is a fellow
of the Leverhulme Trust. D.J.V. is the Canada Research Chair in
Chemical Glycobiolgy and a scholar of the Michael Smith Founda-
tion for Health Research (MSFHR). M.S.M. is a holder of
scholarships from NSERC and the MSFHR.
Supporting Information Available: Synthesis and character-
ization of compounds used in this study, measured rate data,
procedure for NMR monitoring of reaction stereochemistry and
the extrapolation of rates of chemical step for Ar-GlcNAc-F0s.
This material is available free of charge via the Internet at http://
pubs.acs.org.
(Figure 4). Notably, the observation that the data for the methyl
glycosides (Me-GlcNAc-F0 and Me-GlcNAc-F1) correlate well
with the data obtained for aryl glycosides of regimen III supports
the validity of this extrapolation. As would be expected, when
both forms of catalysis are impaired, the predicted rate of
hydrolysis is slowest. It is notable, however, that the rate
enhancement stemming from general acid or nucleophilic
catalysis, on their own, do not come close to being additive to
the rate enhancement observed when both modes of catalysis
are operative. This difference highlights how the enzyme
employs both general acid and nucleophilic catalysis in a
synergistic manner and seems most likely to arise from the
preorganized nature of the enzyme active site, a feature that
makes this pathway favored by the enzyme whereas in solution
its observation would likely prove impossible due to competing
mechanistic pathways.
The linear free energy relationships we describe here for
O-GlcNAcase are most readily interpreted as reflecting the
kinetic significance of an oxocarbenium ion intermediate for
substrates having poor nucleophiles. There is a long history of
investigations into the existence of such carbocation-like
intermediates in the hydrolysis of glycosides, with solution
studies pointing to a lifetime dependent on subtle factors.21 In
enzymes, the existence of such a species has also long been
contemplated,38 and recent studies of a mutant glycosidase
lacking the general acid catalyst support its viable existence.39
Detailed microscopic interpretations of the Brønsted data are
not essential for evaluating the extent of synergy. However, in
JA904506U
(36) Kinetic isotope effects, and in particular the values of the anomeric
carbon and endocyclic ring oxygen KIEs, are perhaps the most widely
used means of differentiating between concerted (ANDN) and dis-
sociative (DN*AN) mechanisms of group transfer.37 Previous studies
for dissociative mechanisms using models for the transition states that
lack either the nucleophilic species (DN‡*AN) or the leaving group
(DN*AN‡) have revealed good matching of experimental and computed
KIEs, suggesting that these groups may be relatively insignificant in
determining the vibrational environment of the transition state. Our
results, however, show that values of kcat/Km determined in both
regimens II (DN‡*AN) and III (DN*AN‡) are sensitive to both the nature
of the leaving group and the nucleophile, indicating that these groups
both contribute to the electrostatic environment of the transition state.
Vibrational models of these transition states constructed using a
comprehensive series of KIEs along with computational models should
clarify details of the mechanistic proposal advanced here for human
O-GlcNAcase.
(37) Berti, P. J.; Tanaka, K. S. E. AdV. Phys. Org. Chem. 2002, 37, 239–
314.
(38) Sinnott, M. L.; Souchard, I. J. Biochem. J. 1973, 133, 89–98.
(39) Richard, J. P.; Huber, R. E.; Heo, C.; Amyes, T. L.; Lin, S.
Biochemistry 1996, 35, 12387–12401.
(40) Kirby, A. J. Angew. Chem., Int. Ed. Engl. 1996, 35, 707–724.
(41) Piszkiewicz, D.; Bruice, T. C. J. Am. Chem. Soc. 1968, 90, 5844–
5848.
9
13422 J. AM. CHEM. SOC. VOL. 131, NO. 37, 2009