N-phenylglucosylamine hydrolysis: A mechanistic probe of β-glucosidase

Article abstract of DOI:10.1016/j.bioorg.2011.02.003

The spontaneous hydrolysis of glycosylamines, where the aglycone is either a primary amine or ammonia, is over a hundred million-times faster than that of O- or S-glycosides. The reason for this (as pointed out by Capon and Connett in 1965) is that, in contrast to the mechanism for O- or S-glycoside hydrolysis, hydrolysis of these N-glycosides (e.g., glc-NHR) involves an endocyclic C-O bond cleavage resulting in formation of an imine (iminium ion) which then reacts with water. Since ring-opening is kinetically favored with glycosylamines, compounds such as phenylglucosylamine can be a useful probes of enzymes that have been suggested to possibly follow this mechanism. With β-glucosidase from sweet almonds, the enzyme is highly efficient in catalyzing the hydrolysis of phenyl glucoside (k_cat/k_non ～ 10¹⁴) and phenyl thioglucoside (k_cat/k_non ～ 10¹⁰) while with either the almond or the Aspergillus niger enzyme or with yeast α-glucosidase, there is no detectable catalysis of phenylglucosylamine hydrolysis (k_cat/k_non < 20). These results are consistent with the generally accepted mechanism involving exocyclic bond cleavage by these enzymes.

Full text of DOI:10.1016/j.bioorg.2011.02.003

Bioorganic Chemistry 39 (2011) 111–113
Contents lists available at ScienceDirect
Bioorganic Chemistry
journal homepage: www.elsevier.com/locate/bioorg
N-phenylglucosylamine hydrolysis: A mechanistic probe of b-glucosidase
⇑
Ying Na, Hong Shen, Larry D. Byers
Department of Chemistry, Tulane University, New Orleans, LA 70118, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 31 December 2010
Available online 27 February 2011
The spontaneous hydrolysis of glycosylamines, where the aglycone is either a primary amine or ammo-
nia, is over a hundred million-times faster than that of O- or S-glycosides. The reason for this (as pointed
out by Capon and Connett in 1965) is that, in contrast to the mechanism for O- or S-glycoside hydrolysis,
hydrolysis of these N-glycosides (e.g., glc-NHR) involves an endocyclic C–O bond cleavage resulting in for-
mation of an imine (iminium ion) which then reacts with water. Since ring-opening is kinetically favored
with glycosylamines, compounds such as phenylglucosylamine can be a useful probes of enzymes that
have been suggested to possibly follow this mechanism. With b-glucosidase from sweet almonds, the
Keywords:
b-Glucosidase
Phenylglucosylamine
Endocyclic bond cleavage
14
enzyme is highly efficient in catalyzing the hydrolysis of phenyl glucoside (kcat/knon ꢀ 10 ) and phenyl
1
0
thioglucoside (kcat/knon ꢀ 10 ) while with either the almond or the Aspergillus niger enzyme or with yeast
a-glucosidase, there is no detectable catalysis of phenylglucosylamine hydrolysis (kcat/knon < 20). These
results are consistent with the generally accepted mechanism involving exocyclic bond cleavage by these
enzymes.
Ó 2011 Elsevier Inc. All rights reserved.
1
. Introduction
The non-enzymic hydrolysis of glycosylamines (glycosylated
nitrogen, resulting in a ring-opening mechanism [2] yielding a
Schiff base (imine) intermediate (Mech. I), in contrast to the ‘‘clas-
sical’’ mechanism for nucleophilic substitution at the anomeric
ammonia or primary amines) is a much more rapid process than
center (Mech. II) usually seen with O-glycosides:
hydrolysis of glycosides or thioglycosides [1]. The reason for this
is due to the effective ‘‘push’’ by the non-bonding electrons on
For some glycosides (especially conformationally strained ones)
there is direct experimental evidence for endocyclic C–O bond
cleavage in the acid-catalyzed hydrolysis reactions (e.g., [3,4]).
However, for the majority of acid, and enzyme, catalyzed hydroly-
sis of glycosides, the overwhelming body of evidence is consistent
⇑
Corresponding author.
E-mail address: byers@tulane.edu (L.D. Byers).
0
045-2068/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.bioorg.2011.02.003

1
12
Y. Na et al. / Bioorganic Chemistry 39 (2011) 111–113
zyme. For example, with a 6 mM solution of phenylglucosylamine,
there was no detectable difference in the amount of glucose re-
leased (over an 8 h time period) in the presence or absence of
5
.2 mg/ml (130 units/ml) of almond b-glucoside (=80 lM). Fig. 1
shows the initial velocity data. The linear regression fit of the data
in this range yield the following equations:
with enzyme : ½Glcꢁ ¼ 0:240 ðꢂ:003Þ mM=h þ 0:200 ðꢂ0:002Þ mM
ð1Þ
without enzyme : ½Glcꢁ ¼ 0:234 ðꢂ:004Þ mM=h þ 0:200 ðꢂ0:004Þ mM
ð2Þ
ꢄ4
The spontaneous rate of glucose production was 2 ꢃ 10 M/h.
Thus, the enzymic reaction must occur at a rate that is less than
ꢄ5
1
0% of this value (i.e., <2 ꢃ 10 M/h). Based on a substrate
concentration (6 mM = 0.6K
m
), this indicates the Vmax must be less
ꢄ5
ꢄ8
Fig. 1. Rate of hydrolysis of N-phenylglucosylamine (6 mM, pH 6.3, 25 °C) in the
absence of enzyme (s) and the presence of 80 M almond b-glucosidase ( ). The
than 5.3 ꢃ 10 M/h (=1.5 ꢃ 10 M/s) and, therefore, kcat < 2 ꢃ
ꢄ4 ꢄ1
l
1
0
s . The kinetic parameters for this ‘‘substrate’’ and those
line is the linear regression fit of the (separate) data through the initial velocity
range (<8% completion).
of its O and S analogs are summarized in Table 1. Clearly the rate
of the enzyme-catalyzed hydrolysis of phenylglucosylamine (if it
occurs at all) is much less than that of the oxygen analog. While
the upper limit of kcat for phenylglucosylamine hydrolysis is com-
parable to the measured value of kcat for phenyl thioglucoside
hydrolysis, this corresponds to a very poor catalytic proficiency
with a process analogous to Mech. II involving varying extents of
nucleophilic participation in the transition state [5–8]. Neverthe-
less, there have been occasional proposals of enzymic mechanisms
for O-glycoside hydrolysis involving initial cleavage of the endocyc-
lic, rather than the exocyclic, C–O bond (analogous to Mech. I).
These suggestions were based mainly inhibition analysis [9] and/
or structural and computational approaches [10–12]. While for
most enzymes involving glycosyl transfer (i.e., glycohydrolases
and glycotransferases), physical organic studies are more consis-
tent with mechanism II than with mechanism I (e.g., multiple atom
kinetic isotope effects [13]), proposals occasionally reappear, rais-
ing the possibility of endocyclic cleavage, especially for b-glycosi-
dases [12]. To test this possibility, we examined the substrate
10
(
2
k
cat/k
n
< 20) compared to that for it’s thiol analog (kcat/k
5 °C. This catalytic ‘‘proficiency’’ for phenylglucosylamine hydro-
lysis is even more negligible when compared to that for that for
n
10 ) at
1
4
n
phenyl glucoside hydrolysis (kcat/k 10 ). Similarly, no enhance-
ment could be detected in the hydrolysis rate of 40 mM phenylg-
lucosylamine (pH 6.3) when incubated in the presence of
b-glucosidase from either almonds or A. niger (15 units/ml, K
phenylglucosylamine = 9.3 mM) or -glucosidase from yeast (100
units/ml, K for phenylglucosylamine = 8.6 mM). If the catalytic
i
for
a
i
mechanism of these glucosidases did involve an endocyclic C–O
cleavage step, phenylglucosylamine would be expected to be a very
good substrate for the enzyme. Our results indicate that this is not
the case. Contrary to the proposal of Frank [12] there appear to be at
least two b-glucosidases, one from sweet almonds (a family 1 gly-
cohydrolase) and one from A. niger (a family 3 glycohydrolase), that
do not catalyze the hydrolysis of b-glucosides via a ring-opening
mechanism. Considering the relative ease in which phenylglycosyl-
amines can be prepared, these compounds can serve as useful
probes of possible endocyclic cleavage with other glycohydrolases.
potential of N-phenyl-
sweet almonds (a family 1 glycohydrolase) and from Aspergillus ni-
ger (a family 3 glycohydrolase) as well as with yeast -glucosidase.
D-glucosylamine with b-glucosidases from
a
The b-glucosidases have a fairly relaxed specificity for the leaving
group, cleaving a wide variety of O-glycosides (e.g., [14]) and the
almond enzyme has been shown to cleave S-glycosides [15]. If
these enzymes operate via an endocyclic cleavage mechanism, than
phenylglucosylamine should be an even better substrate for the
enzyme than is phenylglucoside.
3
. Experimental
2
. Results and discussion
3.1. Materials
While phenyl b-glucoside and phenyl b-thioglucoside are stable
in aqueous solution (t1/2 half a million years for spontaneous
hydrolysis at 25 °C [16]), phenyl glucosylamine is relatively unsta-
ble. The half-life for spontaneous hydrolysis of a 10 mM solution
The chromatographically purified almond b-glucosidase (spe-
cific activity 25 units/mg) was obtained from Sigma–Aldrich
Chemical Co., St. Louis, MO. Enzyme concentration was determined
by absorbance at 278 nm using a molar extinction coefficient of
(
(
1
pH = 6.3) of an anomeric equilibrium mixture of PhNHGlc is 21
±2) h, corresponding to a hydrolysis rate constant which is about
5
ꢄ1
ꢄ1
8
e
= 1.22 ꢃ 10 M cm , based on the percent (1 g/100 ml) extinc-
tion coefficient (=18.8) of Legler [17] and a subunit molecular
weight of 65,000 [18,19]. Yeast (Saccharomyces cerevisiae) -gluco-
0 -times larger than that for either the corresponding O or S glu-
coside [16]. We were unable to increase the rate of hydrolysis of
phenylglucosylamine, even in the presence of large amounts of en-
a
sidase (specific activity = 170 units/mg) was also obtained from
Sigma–Aldrich. The b-glucosidase from A. niger was prepared from
crude ‘‘cellulase’’ powder, obtained from Fluka (Sigma–Aldrich)
essentially as described by Seidle et al. [20]. The substrates, both
Table 1
Hydrolysisa of phenyl b-glucopyranoside derivatives, Ph-X-Glc.
ꢄ1
ꢄ1 ꢄ1
X
kcat (s
)
K
m
(mM)
kcat/K
m
, (M
s
)
a and b p-nitrophenyl-D-glucopyranoside (pNPG) and phenyl b-D-
glucopyranoside (GlcOPh), were obtained from Sigma–Aldrich as
were buffers and all other reagents.
O
S
NH
1.2 (±0.1)
32.0 (±0.7)
8.7 (±0.5)
39 (±3)
ꢄ4
ꢄ2
3.2 (±0.5) ꢃ 10
3.7 (±0.6) ꢃ 10
b
ꢄ4
10.0 (±0.2)c
ꢄ2
<2 ꢃ 10
<2 ꢃ 10
a
b
c
3.1.1. Phenyl b-
GlcSPh was prepared by reaction of 2,3,4,6-tetra-O-acetyl-
glucopyranosyl bromide with thiophenol in acetone (containing
D-thioglucopyranoside
Almond b-glucosidase, pH 6.3, 25 °C.
Anomeric mixture.
K value.
i
a-D-

Y. Na et al. / Bioorganic Chemistry 39 (2011) 111–113
113
potassium carbonate). The acetylated glucoside was deacetylated
in dry methanol containing catalytic amounts of sodium methox-
pNPG, as previously described for methyl glucoside hydrolysis
[19].
ide. Crystallization from ethanol yielded phenyl b-thioglucoside
1
(
400 MHz H NMR: dH1 = 4.63 ppm, J1,2 = 9.9 Hz, mp = 130–131°,
References
lit mp = 135° [21]).
[
1] W. Pigman, E.A. Cleveland, D.H. Couch, J.H. Cleveland, J. Am. Chem. Soc. 73
1951) 1976–1979.
(
[
[
2] B. Capon, B.E. Connett, J. Chem. Soc. (1965) 4497–4502.
3] A.J. Bennet, M.L. Sinnott, W.S.S. Wijesundera, J. Chem. Soc. Perkin II (1985)
3
.1.2. Phenyl glucosylamine
GlcNHPh was prepared essentially as described by Capon and
1233–1236.
1
Connett [22]. H NMR (CD
3
OD) revealed a 3:2 mixture of
a:b ano-
[4] P. Deslongchamps, S. Li, Y.L. Dory, Org. Lett. 6 (2004) 505–508.
[
[
5] M.L. Sinnott, Carbohydrate Chemistry and Biochemistry, RSC Publishing,
Cambridge, UK, 2007.
6] B. Capon, Chem. Rev. 69 (1969) 407–498.
mers. Since the rate of anomerization in aqueous solution is faster
than the hydrolysis rate [2], no attempt was made to separate the
anomers.
[7] D.L. Zechel, S.G. Withers, Acc. Chem. Res. 33 (2000) 11–18.
[
[
8] D.J. Vocadlo, G.J. Davies, Curr. Opin. Chem. Biol. 12 (2008) 539–555.
9] G.W.J. Fleet, Tetrahedron Lett. 26 (1985) 5073–5076.
3
.2. Kinetics
[10] C.B. Post, M. Karplus, J. Am. Chem. Soc. 108 (1986) 1317–1319.
[
11] A.K. Mazur, R. Haser, F. Payan, Biochem. Biophys. Res. Commun. 204 (1994)
97–302.
12] R.W. Franck, Bioorg. Chem. 20 (1992) 77–88.
2
Reactions were carried out in buffer solutions [0.01 M MES
[
(
morpholino-ethansulfonate), 0.01 M NaCl, pH 6.3] at 25 °C. Stock
[13] J.K. Lee, A.D. Bain, P.J. Berti, J. Am. Chem. Soc. 126 (2004) 3769–3776.
[
14] M.P. Dale, W.P. Kopfler, I. Chait, L.D. Byers, Biochemistry 25 (1986)
522–2529.
15] H. Shen, L.D. Byers, Biochem. Biophys. Res. Commun. 362 (2007) 717–720.
enzyme solutions were prepared by dissolving ꢀ10 mg/ml of
protein in the MES buffer. The reaction was monitored by removal
of an aliquot of the reaction mixture and then determining the
glucose concentration using the coupled enzyme assay of
hexokinase (yeast) and glucose-6-phosphate dehydrogenase
2
[
[16] M.E. Rountree, L. Byers, R. Wolfenden, Y. Yuan, Abstracts of Papers, 239th ACS
National Meeting, San Francisco, CA, United States, March 21–25, 2010.
[
[
17] G. Legler, Z. Physiol. Chem. 351 (1970) 25–31.
18] T. Kleinschmidt, H. Glossmann, J. Horst, Hoppe-Seyer’s Z. Physiol. Chem. 351
(1970) 349–358.
(
Leuconostoc mesenteroides), in a buffered solution (0.2 M trietha-
nolamine, pH 7.5), containing 10 mM ATP, 5 mM MgSO , 1 mM
NAD and measuring the increase in absorbance at 340 nm due
[
19] E.B. Golden, L.D. Byers, Biochim. Biophys. Acta 1794 (2009) 1643–1647.
4
+
[20] H.F. Seidle, I. Marten, O. Shoseyov, R.E. Huber, Protein J. 23 (2004) 11–23.
21] E. Fischer, K. Delbriick, Ber. Dtsch. Chem. Ges. 42 (1909) 1476–1482.
[22] B. Capon, B.E. Connett, J. Chem. Soc. (1965) 4492–4497.
23] R.B. McComb, L.W. Bond, R.W. Burnett, R.C. Keech, G.N. Bowers Jr., Clin. Chem.
2 (1976) 141–150.
[
3
ꢄ1
ꢄ1
to the formation of NADH (
or K ) values for the compounds PhXGlc (X = O, S, NH) were deter-
mined by the inhibition of the enzyme-catalyzed hydrolysis of
e
= 6.32 ꢃ 10 M cm [23]). The K
m
[
(
i
2