Solvent and α-secondary kinetic isotope effects on β-glucosidase

Article abstract of DOI:10.1016/j.bbapap.2015.02.015

β-Glucosidase from sweet almond is a retaining, family 1, glycohydrolase. It is known that glycosylation of the enzyme by aryl glucosides occurs with little, if any, acid catalysis. For this reaction both the solvent and α-secondary kinetic isotope effects are 1.0. However, for the deglucosylation reaction (e.g., k_cat for 2,4-dinitrophenyl-β-D-glucopyranoside) there is a small solvent deuterium isotope effect of 1.50 (± 0.06) and an α-secondary kinetic isotope effect of 1.12 (± 0.03). For aryl glucosides, k_cat/K_M is very sensitive to the pK_a of the phenol leaving group [β_lg - 1; Dale et al., Biochemistry 25 (1986) 2522-2529]. With alkyl glucosides the β_lg is smaller (between - 0.2 and - 0.3) but still negative. This, coupled with the small solvent isotope effect on the pH-independent second-order rate constant for the glucosylation of the enzyme with 2,2,2-trifluoroethyl-β-glucoside [^D2O(k_cat/K_M) = 1.23 (± 0.04)] suggests that there is more glycone-aglycone bond fission than aglycone oxygen protonation in the transition state for alkyl glycoside hydrolysis. The kinetics constants for the partitioning (between water and various alcohols) of the glucosyl-enzyme intermediate, coupled with the rate constants for the forward (hydrolysis) reaction provide an estimate of the stability of the glucosyl-enzyme intermediate. This is a relatively stable species with an energy about 2 to 4 kcal/mol higher than that of the ES complex. This article is part of a Special Issue entitled: Enzyme Transition States from Theory and Experiment.

Full text of DOI:10.1016/j.bbapap.2015.02.015

BBAPAP-39539; No. of pages: 6; 4C:
Biochimica et Biophysica Acta xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Biochimica et Biophysica Acta
journal homepage: www.elsevier.com/locate/bbapap
☆
Solvent and α-secondary kinetic isotope effects on β-glucosidase
⁎
Miaomiao Xie, Larry D. Byers
Department of Chemistry, Tulane University, New Orleans, LA 70118, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
β-Glucosidase from sweet almond is a retaining, family 1, glycohydrolase. It is known that glycosylation of
the enzyme by aryl glucosides occurs with little, if any, acid catalysis. For this reaction both the solvent
and α-secondary kinetic isotope effects are 1.0. However, for the deglucosylation reaction (e.g., kcat for
2,4-dinitrophenyl-β-D-glucopyranoside) there is a small solvent deuterium isotope effect of 1.50
Received 7 February 2015
Accepted 23 February 2015
Available online xxxx
(
0.06) and an α-secondary kinetic isotope effect of 1.12 ( 0.03). For aryl glucosides, kcat/KM is very sen-
Keywords:
sitive to the pKa of the phenol leaving group [βlg ≈ −1; Dale et al., Biochemistry 25 (1986) 2522–2529].
With alkyl glucosides the βlg is smaller (between −0.2 and −0.3) but still negative. This, coupled with
the small solvent isotope effect on the pH-independent second-order rate constant for the glucosylation
of the enzyme with 2,2,2-trifluoroethyl-β-glucoside [D2O(kcat/KM) = 1.23 ( 0.04)] suggests that there is
more glycone–aglycone bond fission than aglycone oxygen protonation in the transition state for alkyl gly-
coside hydrolysis. The kinetics constants for the partitioning (between water and various alcohols) of the
glucosyl-enzyme intermediate, coupled with the rate constants for the forward (hydrolysis) reaction pro-
vide an estimate of the stability of the glucosyl-enzyme intermediate. This is a relatively stable species
with an energy about 2 to 4 kcal/mol higher than that of the ES complex. This article is part of a Special
Issue entitled: Enzyme Transition States from Theory and Experiment.
β-glucosidase
2,4-Dinitrophenyl β-D-glucopyranoside
Secondary kinetic isotope effect
Glucosyl-enzyme energy
© 2015 Elsevier B.V. All rights reserved.
1. Introduction
occurs with very little oxocarbenium ion character. This is a fairly
common feature for a variety of β-glucosidases [6–9] which often
β-glucosidase (EC 3.2.1.21) from sweet almonds is a retaining
glycohydrolase [1] and a member of the GH1 family of sequence-
related glycosyl hydrolases [2]. The enzyme catalyzes the highly effi-
cient hydrolysis of a variety of aryl and alkyl glycosides [3] via a double
displacement mechanism (see Scheme 1).
demonstrate larger secondary KIEs in the deglycosylation reaction
than in the glycosylation reaction. In order to see if this is also the sit-
uation with the almond enzyme we examined the α-secondary deu-
terium KIE on kcat for 2,4-dinitrophenyl β-glucoside hydrolysis, for
which deglucosylation is the rate-limiting step [3].
The reaction utilizes two carboxylic (glutamic) acids with one of
these functioning as a nucleophile leading to a glycosyl-enzyme inter-
mediate (an acylal) and the other as a BrØnsted acid catalyst in the
formation, and as a base catalyst in the breakdown, of this intermedi-
ate [4]. The mechanism of action of this enzyme has been extensively
investigated and has revealed some intriguing features. Sweet almond
β-glucosidase was one of the first enzymes examined for a secondary
kinetic isotope effect. Dahlquist et al. [5] found an α-secondary deute-
rium KIE of D(kcat/KM) = 1.01 for the enzyme-catalyzed hydrolysis of
phenyl β-glucoside. This is significantly less than the value expected
for a SN1 (DN + AN) mechanism, or for that found for the acid-
catalyzed hydrolysis (Dk = 1.13) of this substrate or for the
lysozyme-catalyzed hydrolysis of its substrates [5]. This suggests a
transition-state structure for the glycosylation of β-glucosidase
The almond enzyme shows no evidence of protonation of the leaving
group in the transition state for the hydrolysis of aryl glucosides. We
found no solvent KIE [D2O(V/K) ≈ 1.0] on the pH-independent second-
order rate constant for the enzyme-catalyzed hydrolysis of 4-
nitrophenyl β-glucoside [3]. Rosenberg and Kirsch [10] found a large
18O-kinetic isotope effect [18(V/K) = 1.038] for the β-glucosidase-
catalyzed hydrolysis of this substrate. Also, there is a large negative
Brønsted coefficient (βlg ≈ −1) for the dependence of the enzyme
glucosylation rate constant on the leaving group acidity (for phenols
with pKa ≥7) [3]. While a transition state structure with substantial
C–O cleavage and little protonation of the leaving group is plausible
for glucosides with “good” leaving groups such as phenols with pKas
≤10, it is difficult to imagine such a structure for the enzyme-
catalyzed hydrolysis of unactivated substrates, such as methyl gluco-
side. Nevertheless, we found that there is no solvent kinetic isotope
effect for the enzyme-catalyzed hydrolysis of methyl β-glucosidaside
[11]. This was attributed to a fortuitous cancellation of a normal equilib-
rium isotope effect for the protonation of the enzyme-bound glucoside
and an inverse kinetic isotope effect accompanying the cleavage of the
☆
This article is part of a Special Issue entitled: Enzyme Transition States from Theory
and Experiment.
Corresponding author. Tel.: +1 504 861 7044.
E-mail address: byers@tulane.edu (L.D. Byers).
⁎
http://dx.doi.org/10.1016/j.bbapap.2015.02.015
1570-9639/© 2015 Elsevier B.V. All rights reserved.
Please cite this article as: M. Xie, L.D. Byers, Solvent and α-secondary kinetic isotope effects on β-glucosidase, Biochim. Biophys. Acta (2015),
http://dx.doi.org/10.1016/j.bbapap.2015.02.015

2
M. Xie, L.D. Byers / Biochimica et Biophysica Acta xxx (2015) xxx–xxx
Scheme 1. Double displacement mechanism of ß-glucosidase.
glycosidic bond. In order to confirm this hypothesis, and to get an idea of
the extent of proton transfer to the anomeric oxygen in the transition
state we decided to examine the hydrolysis reactions with more acidic
(fluorinated) alkyl glucosides.
containing 0.01 M NaCl at pH 6.3, and the reactions were initiated by ad-
dition of enzyme. With pNPG1 as substrate, product production was
monitored spectrophotometrically by measuring the absorbance at
400 nm (ε = 2.9 × 103 M−1 cm−1 at pH 6.3). With DNPG as substrate
the reaction was monitored at 400 nm (ε = 1.09 × 104 M−1 cm−1
2. Materials and methods
for 2,4-dinitrophenolate) under first order conditions ([S] ≪
KM ≈ 0.8 mM). When larger substrate concentrations were used
(i.e., in the determination of the Michaelis–Menten parameters) higher
wavelengths were used (e.g., at 480 nm, ε = 760 M−1 cm−1). These pa-
rameters were determined by monitoring the initial velocities at sub-
2.1. Reagents
Almond β-glucosidase, specific activity ≈ 25 units/mg (salicin, 37°,
pH 5.0), was obtained from Sigma-Aldrich Chemical Co., St. Louis, MO
and further purified (to a specific activity ≈ 35 units/mg) by ion ex-
change chromatography on DEAE Sepharose (eluting with increasing
NaCl concentrations at pH 6.0). SDS-PAGE (4–12% gradient cross-
linked NuPage™ gel, Invitrogen Life Technologies) gave a single band
corresponding to a molecular weight of ~65,000, consistent with this
being the subunit of the homodimeric isozyme A [12]. Protein concen-
tration was determined by absorbance at 278 nm based on a value of
E1% = 18.8 [11,12]. D2O (99.9%) was obtained from Cambridge Isotope
Laboratories. 4-Nitrophenyl β-D-glucopyranoside (pNPG1), DCl, NaOD,
buffers and other reagents were obtained from Sigma-Aldrich.
strate concentrations ranging from ~0.3 to
9 mM. (Substrate
concentrations were determined from the absorbance of
dinitrophenolate following complete enzymatic hydrolysis of a sample
from the stock substrate solutions.) The reactions were initiated by the
addition of enzyme solution (final concentration ≈ 0.03 units/ml =
12 nM) and the initial velocities were fit to the Henri–Michaelis–Menten
equation via non-linear regression. The α-secondary kinetic isotope ef-
fect on kcat for DNPG hydrolysis was determined at pH 6.3 by measuring
a series of (10–15) initial velocities with each substrate (the 1-deutero-
and the 1-proteo-DNPG) at high ([S] ~12KM) concentrations. The Vmax
values were calculated from these initial velocity measurements by
multiplying them by the saturation correction factor (1 + KM/[S]),
which is Vmax/v from the Henri–Michaelis–Menten equation.
2.2. Synthesis
For the reaction with alkyl glucosides, the reactions were monitored
by the rate of glucose production by removal of an aliquot of the reaction
mixture and then determination of the glucose concentration using the
coupled enzyme assay of hexokinase (yeast) and glucose-6-phosphate
dehydrogenase (L. mesenteroides), containing 10 mM ATP, 5 mM
MgSO4, 1 mM NAD+, pH 7.5, and measuring the resulting absorbance
at 340 nm due to the formation of NADH (ε = 6.32 × 103 M−1 cm−1
[20]). The KM values for these substrates were determined by measuring
the inhibition of pNPG1 hydrolysis under first-order conditions. For the
reverse reactions (Glc + ROH → GlcOR), the rate constant for
the alcoholysis of the glycosyl-enzyme was obtained by monitoring the
product partitioning when p-nitrophenol is produced from pNPG in the
presence of various concentrations of the alcohol. The product ratio
([Glc]/[GlcOR]) is equal to the rate-constant ratio (kROH[ROH/kW[H2O])
where kW is the rate constant for the reaction of water with the
glucosyl-enzyme and kROH is the rate constant for the reaction of the
alcohol with the glucosyl-enzyme. In determining the product ratio,
the glucose concentration (measured as described above and corrected
for the absorbance of pNPG and of p-nitrophenol/phenolate at 340 nm)
and the p-nitrophenol/phenolate concentration (= [Glc] + [GlcOR])
were used to calculate the concentrations.
All compounds were characterized by 1H-NMR (400 MHz, Varian
Unity Inova 400 spectrometer).
2,4-Dinitrophenyl β-D-glucopyranoside (DNPG1) was prepared by
the method of Koeners et al. [13] via the reaction of 2,3,4,6-tetra-O-ace-
tyl-β-D-glucopyranose with 2,4-dinitrofluorobenzene. The 1-deutero
analog of the tetraacetate was prepared by the method of Berven
and Withers [14] via reduction of corresponding lactone with
NaBD4 (Sigma-Aldrich, 98 atom % D). 1H-NMR analysis of deuterated
substrate indicated an extent of isotopic incorporation greater than
95%. The 1-proteo compound was prepared in an identical manner,
using NaBH4.
Ethyl β-D-glucopyranoside was prepared by the method of Koenigs
and Knorr from α-bromo-2,3,4,6-tetra-O-acetylglucose [15] with a flash
column chromatography step (EtOAc/hexane 2:3) for the purification of
the ethyl glucoside tetraacetate, followed by Zemplen deacetylation
(NaOMe in MeOH), neutralization with Dowex 50W-X8)*, concentra-
tion under reduced pressure and purification by column chromatogra-
phy (silica gel, CH2Cl2/CH3OH, 5:1), yielding a solid which was
crystallized from EtOAc/MeOH, mp 97–100°, lit. 98–100° [16].
2,2,2-Trifluoroethyl β-D-glucopyranoside was prepared by the
method of Xue et al. [17]. Isopropyl β-D-glucopyranoside was prepared
enzymatically (800 units of β-glucosidase in 10 ml of 90% 2-propanol
containing 3 mmol glucose) as described by Lu et al. [18]. The resulting
syrup was subjected to flash column chromatography (silica gel,
CH2Cl2/MeOH, 5:1) and collected as a white solid. The 1,1,1,3,3,3-
hexafluoro analog of isopropyl β-D-glucopyranoside was prepared by
the method of Gueyrard et al. [19].
2.4. Solvent kinetic isotope effect
The Vmax/KM values for hydrolysis of 2,2,2-trifluoroethyl β-D-
glucopyranoside were determined under pseudo zero-order conditions
by adding enzyme (final concentration = 30 nM) to an appropriately
buffered solution containing 10 mM substrate (≪ KM ≈ 120 mM at
pH 5) and measuring the initial velocity (by removing aliquots and
determining the glucose concentration at various times). The initial ve-
locity was divided by the substrate concentration to yield Vmax/KM. The
buffer solutions consisted of 0.01 M buffer [formate, pL(=pH/pD) 3.0–
4.0; acetate, pL 3.5–5.8; MES1, pL 5.5–7.0; PIPES1, pL 6.1–7.6], prepared
in deionized water (or D2O) containing 0.01 M NaCl. pH measurements
were made using a glass combination electrode (Accumet pH meter).
2.3. Kinetics
The reactions were monitored on a Hewlett-Packard model
8452A diode array spectrophotometer equipped with a circulating
water bath (T = 25.0
0.1 °C). Concentrated enzyme solutions
(~10 mg/ml ≈ 150 μM subunits) were prepared in 0.01 M MES1 buffer
Please cite this article as: M. Xie, L.D. Byers, Solvent and α-secondary kinetic isotope effects on β-glucosidase, Biochim. Biophys. Acta (2015),
http://dx.doi.org/10.1016/j.bbapap.2015.02.015

M. Xie, L.D. Byers / Biochimica et Biophysica Acta xxx (2015) xxx–xxx
3
Table 1
pD values were estimated by adding 0.41 to the pH meter reading [21].
The data were fit to the following equation:
Alkyl β-glucoside hydrolysis, pH 5.5.
ROH
pKa
kcat/KM, M−1 s−1
KM, mM
kcat, s−1
h
ꢀ
ꢁi
CH3CH2OH
CF3CH2OH
(CH3)2CHOH
(CF3)2CHOH
15.9
12.4
17.1
9.3
48 ( 2)
1620 ( 58)
130 ( 8)
230 ( 12)
110 ( 6)
210 ( 9)
79 ( 4)
11 ( 1)
178 ( 11)
27 ( 2)
log kobs ¼ log klim= 1 þ 10ðpK1−pLÞ þ 10ðpL−pK2Þ
ð1Þ
8090 ( 18)
640 ( 30)
3. Results
3.2.2. On the glucosylation step — 2,2,2-trifluoroethyl β-D-glucopyranoside
hydrolysis
3.1. Secondary kinetic isotope effect
Fig. 1 shows the pL (pH/pD) dependence of kcat/KM for the hydrolysis
of TFEG1 in H2O and in D2O. The data show typical increases in the ap-
parent pKas when the solvent is changed from H2O to D2O as well as a
small, but significant, solvent kinetic isotope effect on the limiting
second-order rate constant of D2O(kcat/KM) = 1.23 ( 0.04).
The α-deuterium secondary KIE for the hydrolysis of 2,4-DNPG was
measured at pH 6.3 (0.01 M MES, 0.01 M NaCl, 25 °C). Estimates of the
kinetic parameters, obtained under initial velocity conditions, for the
enzymatic hydrolysis of [1H]- and [2H]-glucosides of 2,4-dinitrophenol
are: KhM = 0.79 ( 0.08) mM, KMd = 0.7 ( 0.1) mM and kHcat = 510
(
16) s−1, kD = 450 ( 23) s−1. Under pseudo first-order condi-
cat
3.3. Alkyl glucoside hydrolysis
tions, the measured rate constants, and thus the kcat/KM values, are
identical [6.46 ( 0.09) × 105 M−1 s−1 for the protio substrate and
6.43 ( 0.08) × 105 M−1 s−1 for the deuterio substrate] yielding a sec-
ondary KIE of D(kcat/KM) = 1.00. This is not surprising since the second-
order rate constant for this substrate is largely diffusion-controlled [3].
The cleavage of aryl glucosides catalyzed by almond β-glucosidase
proceeds with little, if any, proton transfer [3,10]. The BrØnsted coeffi-
cient, based on kcat/KM for the hydrolysis of a series of 13 substituted
phenyl glucosides is large and negative (βlg = −0.97) [3]. In order to
obtain a qualitative comparison for the enzyme-catalyzed hydrolysis
of alky glucosides with that of aryl glucosides we compared the kinetics
for the hydrolysis of ethyl and isopropyl glucosides with their fluorinat-
ed analogs. The results are shown in Table 1. It is clear that within each
structurally analogous pair, the more acidic alcohol is the better leaving
group. Furthermore, the sensitivity of either kcat/KM or kcat to the ali-
phatic alcohol leaving group pKa (βlg ~ −0.3) is not nearly as large as
it is for aryl leaving groups (with pKa ≥ 7).
In order to obtain a more precise estimate of the isotope effect on kcat
,
a series of initial velocity measurement were made with each substrate
at high concentrations (9.2 mM ≈ 13.1 KM for the deuterio-DNPG and
9.8 mM ≈ 12.4 KM for the protio-DNPG). These initial velocities convert-
ed to Vmax values by multiplying by the corresponding saturation cor-
rection factor, (1 + KM/[S]) = 1.076 for [2H]-substrate and 1.081 for
the [1H]-DNPG. The ratio of these maximal velocities yields a value for
the α-secondary KIE of Dkcat = 1.12 ( 0.03).
3.2. Solvent kinetic isotope effects
3.4. Enzymatic synthesis of alkyl glucosides
3.2.1. On the deglucosylation step — dinitrophenyl glucoside hydrolysis
The value of kcat for the hydrolysis of DNPG was found to vary by b3%
between pH 5 and 6. When measured at the pH optimum (pH 5.2) this
value (determined from 12 initial velocity measurements at [S] =
20 mM) is 990 ( 28) s−1. In D2O (pD = 5.6) this value is 660
The reactions illustrating the competition between trans-
glucosylation, in the presence of added alcohols, and hydrolysis are
summarized in Scheme 2:
The kinetics of product partitioning of pNPG (1 mM) between glu-
cose and alkyl glucoside were determined in the presence of added alco-
hols at pH 5.5. Assuming that the reaction of the glucosyl-enzyme
intermediate with water (i.e., formation of Glc) is first order in [H2O]
and its reaction with the accepting alcohol (i.e., formation of GlcOR) is
first-order in [ROH] (vide infra) then the second-order rate constant
ratio for these two reactions, kROH/kW is readily obtained (e.g., see
[22]) from the ratio of the initial velocities for p-nitrophenol production
(vpNP) and for glucose production (vglc):
(
(
21) s−1, corresponding to a solvent kinetic isotope effect of 1.50
0.06) on the deglucosylation reaction.
vpNP=vglc ¼ 1 þ kROH½ROH=kWH2Oꢀ
ð2Þ
Fig. 2 shows the results obtained using 2-propanol and its
hexafluoro analog. The results are linear at least up to the highest
alcohol concentrations tested [ca. 20% v/v with (CF3)2CHOH
(=1.9 M) and 35% v/v with (CH3)2CHOH (=4.6 M)]. Similar results
were obtained with ethanol and 2,2,2-trifluoroethanol. This indicates
that the reactions are, indeed, first-order in alcohol concentration and
show no evidence of enzyme denaturation during the time course
(several minutes) for which the reactions are followed. (This stability
is not surprising. The 2-propyl glucoside was prepared by incubating
the enzyme in 90% 2-propanol for 2 days at 50° [18].) The partitioning
of the glucosyl-enzyme intermediate between the reaction with the
alcohol and with water determined by this method is kROH/kW = 7.5
for (CF3)2CHOH and 4.0 for (CH3)2CHOH. With CF3CH2OH and
CH3CH2OH the partitioning ratios are 9.2 and 5.9.
Fig. 1. Solvent isotope effect on the enzymatic hydrolysis of 2,2,2-trifluoroethyl β-D-
glucopyranoside (25 °C). The lines are the fit to Eq. (1) with the following values: in
lim
H2O (open circles); (kcat/KM
)
= 1.85( 0.03) × 103 M−1 s−1, pKa1 ~2.2 ( 0.2),
lim
pKa2
=
6.31
(
0.05) and in D2O (filled circles); (kcat/KM
)
=
1.51( 0.05) × 103 M−1 s−1, pKa1 ~2.8 ( 0.1), pKa2 = 7.14 ( 0.09), yielding a solvent
isotope effect on the pH-independent second-order rate constant of D2O(kcat/KM
1.23 ( 0.04).
)
=
lim
Please cite this article as: M. Xie, L.D. Byers, Solvent and α-secondary kinetic isotope effects on β-glucosidase, Biochim. Biophys. Acta (2015),
http://dx.doi.org/10.1016/j.bbapap.2015.02.015

4
M. Xie, L.D. Byers / Biochimica et Biophysica Acta xxx (2015) xxx–xxx
Scheme 2. Partitioning of intermediate.
4. Discussion
glucosylation (k2) transition state. 2) The BrØnsted coefficient for
this step is large and negative [3]. Indeed, this value (βlg = −1) com-
pared with the equilibrium β value for complete glycosidic bond
cleavage (yielding the unprotonated phenoxide, ArO−), determined
by Richard et al. to be −1.56 for β-galactoside hydrolysis [22] and as-
suming that the βeq for aryl glucoside hydrolysis is similar to this,
4.1. Secondary α-deuterium KIE
The mechanism of action of retaining glycohydrolases, such as al-
mond β-glucosidase, is a double-displacement reaction involving a
covalent glycosyl-enzyme intermediate (Scheme 1). Consistent with
this is the biphasic nature of the BrØnsted plot obtained when the
log kcat for hydrolysis of phenyl glucosides is plotted against the pKa
of the substituted phenol leaving groups (where the slope, βlg, chang-
es from −1 to 0 as the pKa of the leaving group decreases below ~8)
[3]. Thus, for aryl glucosides with good leaving groups, such as 2,4-
DNP-Glc (leaving group pKa = 4.1), the rate-limiting step is
deglucosylation (i.e., kcat = k3). For this substrate, kcat ~990 s−1 at
pH 5.5 and 510 s−1 at pH 6.3. When kcat was measured with both
the protio and deuterio substrates, the α-deuterium secondary KIE
was found to be 1.12 ( 0.03), suggesting substantial sp2- character
of the anomeric carbon in the transition state for the deglucosylation
reaction. This is in contrast to the enzyme-catalyzed hydrolysis of
less reactive substrates (for which k3 ≫ k2), such as phenyl gluco-
side (leaving group pKa = 10), for which Dahlquist et al. [5] mea-
sured an α-deuterium secondary KIE of 1.01. It is known that there
is substantial glycosidic bond cleavage in the transition state for
the glycosylation reaction (k2): 1) Rosenberg and Kirsch [10] found
a large 18O kinetic isotope effect on kcat/KM for the hydrolysis of
suggests a transition state where bond cleavage is ~64% (100% βlg/βeq
)
complete. This data indicates that the lack of a secondary kinetic isotope
for phenyl glucoside hydrolysis [5] is not the result of an early transition
state with little, if any, bond cleavage. It is consistent with an almost SN2-
like transition state for the glucosylation reaction. There is a considerable
body of evidence for an SN2 (ANDN) type mechanism in the glycosylation
step for many retaining glycohydrolases. For example, Berti's group [6],
using natural abundance 13C-NMR, has measured a large 13C-KIE for al-
mond β-glucosidase-catalyzed hydrolysis of methyl glucoside, indicative
of an ANDN mechanism.
The secondary KIE (Dkcat = 1.12) that we find with the reactive sub-
strate, DNPG, (i.e., the KIE on k3, the deglucosylation step) is of typical
magnitude compared to other reactions involving formation of the
glucopyranosyl cation (e.g., Dk = 1.13 for the acid-catalyzed hydrolysis
of phenyl β-glucoside [23]) but it is lower than that (Dkcat = 1.25) ob-
served for the hydrolysis of 2,4-dintrophenyl β-galactoside catalyzed
by β-galactosidase [24]. But even with β-galactosidase, which shows a
substantially larger KIE on the deglycosylation reaction (k3 in
Scheme 1) than is seen here, or with other glycohydrolases [7,9,25]
the isotope effect on the glycosylation reaction k2 is unity [24]. Indeed,
for most of the glycohydrolases examined (i.e., all of which we are
aware) the secondary α-deuterium KIE on k3 is significantly larger
than that on k2 (≈1.00) [7,9,24,25] suggesting an SN2-like (ANDN) tran-
sition state for the glycosylation reaction and an SN1-like (DN + AN)
transition state for the deglycosylation reaction. Our results with the
almond β-glucosidase are consistent with this.
pNPG suggesting nearly complete (89
14%) C–O cleavage in the
4.2. Alkyl glucoside hydrolysis
For the reaction of aryl glucosides with the enzyme (kcat/KM), the
more acidic the leaving group (up to pKa ≈ 7), the faster the reaction,
corresponding to a βlg ≈ −1 [3]. Here we did not examine enough
alkyl glucosides to obtain a reliable value of βlg. However, if we look at
the two pairs of structurally similar substrates shown in Table 1, we
see that kcat (=k2) increases 16-fold when trifluoroethyl glucoside re-
places ethyl glucoside. Fluorination of the of the leaving group decreases
its pKa by 3.5 units. This corresponds to a βlg ~ −0.3. With the isopropyl
and hexafluoroisopropyl glucosides it is also apparent that the more
acidic alcohol is more reactive. This gives an apparent “β” value
of(βlg)k2 ~ −0.2. Of course, neither of these “β” values are quantitatively
reliable, being based on only two points, and the apparent difference be-
tween them is not statistically reliable. What is clear, however, is that
Fig. 2. Effect of increasing concentration of alcohol on the ratio of initial velocities of forma-
tion of p-nitrophenoxide and glucose (1 mM pNPG, 9 μg/ml β-glucosidase, .01 M MES,
.01 M NaCl, pH 5.5, 25°).
Please cite this article as: M. Xie, L.D. Byers, Solvent and α-secondary kinetic isotope effects on β-glucosidase, Biochim. Biophys. Acta (2015),
http://dx.doi.org/10.1016/j.bbapap.2015.02.015

M. Xie, L.D. Byers / Biochimica et Biophysica Acta xxx (2015) xxx–xxx
5
the more acidic the alkyl alcohol leaving group the more reactive the
glucoside will be. A similar phenomenon was observed by Richard
et al. [26] who, working with β-galactosidase, clearly demonstrated in-
creasing reactivity with galactosides of more acidic alkyl alcohols
[(βlg)k2 = −0.5( 0.1), determined with a set of 7 structurally homoge-
neous galactosides].
kinetic isotope effect [D2O(kcat/KM)lim = 1.23 ( 0.04)], unlike the reac-
tion of phenyl glucosides (D2O(kcat/KM lim ≈ 1.0 [3]). Although this a
)
small isotope effect compared with those typically seen in general
acid catalyzed reactions or with other hydrolytic enzymes, it is sig-
nificant, particularly in view of the fact that no solvent isotope effect
was detected on the pH-independent second-order rate constant for
the β-glucosidase-catalyzed hydrolysis of methyl glucoside
4.3. Alkyl glucoside synthesis
[ )
D2O(kcat/KM lim = 1.05( 0.08)] [11]. Since no solvent isotope effect
is seen on the binding of substrates (e.g., methyl or phenyl glucoside)
it was concluded that the lack of a solvent isotope effect on the hydro-
lysis of methyl glucoside was a fortuitous cancellation of isotope effects
(i.e., a normal isotope effect on a pre-equilibrium protonation of the
bound substrate followed by an inverse isotope effect on the cleavage
of the protonated glucosidic bond). Because the trifluoroethyl oxygen
in TFEG is even less basic than that in methyl glucoside (the pKa of
trifluoroethanol is 3.5 units lower than that of ethanol), a pre-
equilibrium protonation step is not very likely in the reaction. In order
to avoid such an unstable intermediate (the enzyme-bound protonated
glucoside) a concerted proton transfer from the acidic group (Glu) on
the enzyme coupled to the glycosidic bond cleavage is expected, accord-
ing to the principles of Jencks' libido rule [28]. A normal (albeit small)
solvent isotope effect is consistent with this. While we are unable to
quantitate the progression along the reaction coordinate where the
transition state occurs, we can at least eliminate the two extremes. If
protonation occurs before any glycosidic bond is cleaved in the transi-
tion state we would expect a larger solvent isotope effect and a
This is simply the reverse of the glycosylation step and provides a di-
rect measure of k−2 (Scheme 1). The partitioning of the glucosyl-
enzyme intermediate (between reaction with ROH and H2O) provides
a measure of the relative values of the rate constants for the reverse
step (k−2) and the forward step (k3). The observation that the
partitioning ratios (kROH/kW = k−2/k3) are greater with the more acidic
(fluorinated) alcohols than the corresponding alkyl alcohols is consis-
tent with base catalysis of addition of the alcohol to the intermediate.
The “βnuc” value based on the two pairs of alcohols examined is ~
−0.05 (“βnuc” = −0.04 for hexafluoro-isopropanol/isopropanol and
−0.06 for trifluoroethanol/ethanol) and suggests a small negative
charge on the alcohol nucleophile in the transition state. This is analo-
gous to the situation with β-galatosidase discussed by Richard et al.
[22]. The partitioning ratio also allows us to construct a reaction coordi-
nate diagram (Fig. 3) and reveals the relative stability of the glucosyl-
enzyme intermediate. This intermediate (generated from ethyl gluco-
side) is higher in energy than the enzyme–substrate complex by
3.7 kcal/mol. Calculations based on the data (k2/k−2) obtained with the
other glucosides (trifluoroethyl-, isopropyl- and hexafluoroisopropyl-)
yield similar values (ΔΔGo′ = 2 to 3 kcal/mol) for the energy difference
between the covalent glucosylated enzyme and the non-covalent ES
complex.
βlg N 0. If the transition state occurred when the cleavage of the proton-
ated glycosidic bond was nearly complete we would expect a βlg value
near −0.56 (βeq for glycoside hydrolysis [22]) and a solvent isotope ef-
fect b1 [due to the higher pKa (12.4) of the alcohol than of the enzymic
acid group (pKa2 ≈ 6.3)]. Qualitatively, our results suggest a transition
state structure where there is more glycosidic bond cleavage than pro-
tonation. Because of the stability of the glucosyl-enzyme intermediate
(only about 2–4 kcal higher energy than the Michaelis complex, see
Fig. 3), a late transition state (i.e., one resembling the product) for the
glycosylation step is unlikely. It is, however, clear that leaving group
protonation is more strongly coupled to C–O cleavage in alkyl glucoside
hydrolysis than in aryl glucoside hydrolysis.
4.4. Solvent isotope effect on trifluoroethyl-β-glucoside hydrolysis
The smaller sensitivity, on the leaving group pKa, of the rate constant
(k2) for the glucosylation of the enzyme with alkyl glucosides than with
aryl glucosides could reflect a transition state structure with a smaller
extent of glycosidic bond cleavage and/or a larger degree of proton
transfer from the enzyme to the oxygen (of the incipient alcohol prod-
uct). The hydrolysis of trifluoroethyl glucoside (TFEG1) shows a solvent
Abbreviations
Glc
D-glucose
HEPES
MES
DNPG
PIPES
pNPG
TFEG
N-(2-hydroxyethyl)-piperazine-N′-(2-ethansulfonate)
2-(N-morpholino)ethane-sulfonate
2,4-dinitrophenyl β-D-glucopyranoside
piperazine-N,N′-bis(2-ethansulfonate)
p-nitrophenyl β-D-glucopyranoside
2,2,2-trifluoroethyl β-D-glucopyranoside
References
[1] a) C.S. Hudson, H.S. Paine, The hydrolysis of salicin by the enzyme emulsin, J. Am.
Chem. Soc. 31 (1909) 1242–1248;
b) D.E. Eveleighand, A.S. Perlin, A proton magnetic resonance study of the anomeric
species produced by D-glycosidases, Carbohydr. Res. 10 (1969) 87–95.
[2] S. He, S.G. Withers, Assignment of sweet almond β-glucosidase as a family 1 glyco-
sidase and identification of its active site nucleophile, J. Biol. Chem. 272 (1997)
24864–24867.
[3] M.P. Dale, W.P. Kopfler, I. Chait, L.D. Byers, β-glucosidase: substrate, solvent and vis-
cosity variation as probes of the rate-limiting steps, Biochemistry 25 (1986)
2522–2529.
[4] a) M.L. Sinnott, Catalytic mechanisms of enzymic glycosyl transfer, Chem. Rev. 90
(1990) 1171–1202;
Fig. 3. Reaction coordinate diagram for the hydrolysis of ethyl β-D-glucopyranoside (pH
5.5, 25 °C, 55.6
M H2O). The first step is assumed to be diffusion-controlled
(k1 ≈ 106 M−1 s−1 for aryl glucosides [3]). The equilibrium constant for the overall reac-
tion, Keq′ ≈ 15 M, is from the value of Szekeres and Tettamanti [27]. The KM (=KS) for
ethyl glucoside is 0.23 M, corresponding to a binding energy of −0.9 kcal/mol. For this
substrate, kcat = k2 = 11 s−1 corresponding to an activation energy of 16 kcal/mol from
the ES (15.1 kcal/mol from E + S). The value of k−2 (=6000 s−1) corresponds to an acti-
vation energy of 12.3 kcal/mol from the glucosyl-enzyme intermediate. Thus, the energy
of this covalent intermediate is 2.8 kcal/mol higher than that of the free
enzyme + substrate and 3.7 kcal/mol higher than that of the enzyme–substrate complex
(k2/k−2 = 1.8 × 10−3).
b) J.D. McCarter, S.G. Withers, Mechanisms of enzymatic glycoside hydrolysis, Curr.
Opin. Struct. Biol. 4 (1994) 885–892.
[5] F.W. Dahlquist, T. Rand-Meir, M.A. Raftery, Application of secondary α-deuterium
kinetic isotope effects to studies of enzyme catalysis. Glycoside hydrolysis by lyso-
zyme and β-glucosidase, Biochemistry 8 (1969) 4214–4221.
[6] J.K. Lee, A.D. Bain, P.J. Berti, Probing the transition states of four glucoside hydrolyses
with 13C kinetic isotope effects measured at natural abundance by NMR spectroscopy,
J. Am. Chem. Soc. 126 (2004) 3769–3776.
Please cite this article as: M. Xie, L.D. Byers, Solvent and α-secondary kinetic isotope effects on β-glucosidase, Biochim. Biophys. Acta (2015),
http://dx.doi.org/10.1016/j.bbapap.2015.02.015

6
M. Xie, L.D. Byers / Biochimica et Biophysica Acta xxx (2015) xxx–xxx
[7] Y.-K. Li, J. Chir, F.-Y. Chen, Catalytic mechanism of a family 3 β-glucosidase and mu-
[18] W. Lu, G. Lin, H. Yu, A. Tong, J. Xu, Facile synthesis of alkyl β-D-glucopyranosides
from D-glucose and the corresponding alcohols using fruit seed meals, J. Mol.
Catal. B Enzym. 44 (2007) 72–77.
[19] D. Gueyrard, P. Rollin, T. Nga, M. Ourevitch, J. Begue, D. Bonnet-Delpon, A convenient
synthesis of fluoroalkyl and fluoroaryl glycosides using Mitsunobu conditions,
Carbohydr. Res. 318 (1999) 171–179.
tagenesis on residue Asp-247, Biochem. J. 355 (2001) 835–840.
[8] G. Legler, M.L. Sinnott, S.G. Withers, Catalysis by β-glucosidase A3 of Aspergillus
wentii, J. Chem. Soc. Perkin Trans. 2 (1980) 1376–1383.
[9] J.B. Kempton, S.G. Withers, Mechanism of Argobacterium β-glucosidase: kinetic
studies, Biochemistry 31 (1992) 9961–9969.
[10] S. Rosenberg, J.F. Kirsch, Oxygen-18 leaving group kinetic isotope effects on the
hydrolysis of nitrophenyl glycosides. 2. Lysozyme and β-glucosidase: acid and alka-
line hydrolysis, Biochemistry 20 (1981) 3196–3204.
[11] E.B. Golden, L.D. Byers, Methyl glucoside hydrolysis catalyzed by β-glucosidase,
Biochim. Biophys. Acta 1794 (2009) 1643–1647.
[20] R.B. McComb, L.W. Bond, R.W. Burnett, R.C. Keech, G.N. Bowers Jr., Determination of
the molar absorptivity of NADH, Clin. Chem. 22 (1976) 141–150.
[21] A.K. Covington, M. Paabo, R.A. Robinson, R.G. Bates, Use of the glass electrode in deu-
terium oxide and the relation between the standardized pD scale and the operation-
al pH in heavy water, Anal. Chem. 40 (1968) 700–706.
[12] T. Kleinschmidt, H. Glossmann, J. Horst, Sweet almond emulsin. Further character-
ization of components A and B, Hoppe-Seyer's, Z. Physiol. Chem. 351 (1970)
349–358.
[22] J.P. Richard, J.G. Westerfeld, S. Lin, J. Beard, Structure–reactivity relationships for β-
galactosidase (E. coli, lac Z). 2. Reactions of the galactosyl-enzyme intermediate with
alcohols and azide ion, Biochemistry 34 (1995) 11713–11724.
[13] H.J. Koeners, A.J. de Kok, C. Romers, J.H. van Boom, The use of the 2,4-dinitrophenyl
group in sugar chemistry re-examined, Recl. Trav. Chim. Pays-Bas 99 (1980)
355–362.
[14] L.A. Berven, S.G. Withers, A convenient and high yielding synthesis of 1,2,3,4,6-
penta-O-acetyl 2,3,4,6-β-D-[2H]-glucopyranose, Carbohydr. Res. 156 (1986)
282–285.
[15] W. Koenigs, E. Knorr, Over some derivatives of grape sugar and galactose, Ber 34
(1901) 957–981.
[16] J.H. Ferguson, Relations between rotatoty power and structure in the sugar group.
XXXI. The alpha and beta forms of ethyl-D-glucoside and their tetraacetates, J. Am.
Chem. Soc. 54 (1932) 4086–4090.
[23] F.W. Dahlquist, T. Rand-Meir, M.A. Raftery, Demonstration of carbonium ion inter-
mediate during lysozyme catalysis, PNAS 61 (1968) 1194–1198.
[24] M.L. Sinnott, I.J.L. Souchard, The mechanism of action of β-galactosidase, Biochem. J.
133 (1973) 89–98.
[25] D. Tull, S.G. Withers, Mechanisms of cellulases and xylanases: a detailed kinetic
study of the exo-β-1,4-glycanase from Cellulomonas fimi, Biochemistry 33 (1994)
6363–6370.
[26] J.P. Richard, J.G. Westerfeld, S. Lin, Structure–reactivity relationships for β-
galactosidase (E. coli, lac Z). 1. Br Ønsted parameters for cleavage of alkyl β-D-
galactopyranosides, Biochemistry 34 (1995) 11703–11712.
[27] L. Szekeres, K. Tettamanti, Some new data regarding the formation of
alkylglycosides, Microchem. J. 17 (1972) 148–150.
[17] J.L. Xue, S. Cecioni, H. He, S. Vidal, J.-P. Praly, Variations on the SnCl4 and CF3CO2Ag-
promoted glycosidation of sugar acetates: a direct, versatile and apparently simple
method with either α or β stereocontrol, Carbohydr. Res. 344 (2009) 1646–1653.
[28] W.P. Jencks, Requirements for general acid–base catalysis of complex reactions, J.
Am. Chem. Soc. 94 (1972) 4731–4732.
Please cite this article as: M. Xie, L.D. Byers, Solvent and α-secondary kinetic isotope effects on β-glucosidase, Biochim. Biophys. Acta (2015),
http://dx.doi.org/10.1016/j.bbapap.2015.02.015