Influence of solvent viscosity on the rate of hydrolysis of dipeptides by carboxypeptidase Y

Article abstract of DOI:10.1002/poc.752

The influence of solvent viscosity on the rate of enzymatic hydrolysis of a series of dipeptides (Z-Phe-Gly, Z-Phe-Sar, Z-Phe-Ala, Z-Phe-NMeAla, Z-Phe-Aib and Z-Phe-Pro) by carboxypeptidase Y was investigated. The effect of solvent viscosity on the enzymatic hydrolysis revealed that whereas all K_cat values decreased with viscosity, those of the N-alkyl peptides decreased more than those of the N-H peptides. The kinetic behaviour implies the involvement of conformational changes of the enzyme in terms of the 'induced-fit' process. Copyright

Full text of DOI:10.1002/poc.752

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY
J. Phys. Org. Chem. 2004; 17: 448–457
Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/poc.752
Influence of solvent viscosity on the rate of hydrolysis
of dipeptides by carboxypeptidase Y
Yoshifumi Kanosue,1 Satoshi Kojima1,2 and Katsuo Ohkata1*
1Department of Chemistry, Graduate School of Science, Hiroshima University, 1-3-1 Kagamiyama, Higashi-Hiroshima 739-8526, Japan
2Center for Quantum Life Sciences, Hiroshima University, 1-3-1 Kagamiyama, Higashi-Hiroshima 739-8526, Japan
Received 18 October 2003; revised 12 December 2003; accepted 23 January 2004
ABSTRACT: The influence of solvent viscosity on the rate of enzymatic hydrolysis of a series of dipeptides (Z-Phe-
Gly, Z-Phe-Sar, Z-Phe-Ala, Z-Phe-NMeAla, Z-Phe-Aib and Z-Phe-Pro) by carboxypeptidase Y was investigated. The
effect of solvent viscosity on the enzymatic hydrolysis revealed that whereas all kcat values decreased with viscosity,
those of the N-alkyl peptides decreased more than those of the N-H peptides. The kinetic behaviour implies the
involvement of conformational changes of the enzyme in terms of the ‘induced-fit’ process. Copyright # 2004 John
Wiley & Sons, Ltd.
KEYWORDS: viscosity; induced fit process; carboxypeptidase-Y; dipeptides; hydrolysis
INTRODUCTION
viscosity was suggested by Doolitle,6 who intuitively
pointed out that the translational motion of a molecule
in a liquid is only possible when sufficient free volume
(Vf) is available, i.e. when Vf is larger than some ‘critical’
value V0. The fluidity (ꢀꢀ1) is proportional to the prob-
ability factor [exp(ꢀV0 /Vf)] for the translational motion
of an ensemble of molecules. Therefore, the free-volume
dependence of the viscosity may be expressed by Eqn (1),
where A is a proportionality factor:
Solvent viscosity constitutes an informative medium
variable, whose utility is of prime importance for under-
standing the nature of solvent–solute interactions.1 Con-
sequently, the diversity of physical models developed to
rationalize the viscosity dependence of molecular trans-
formations has reached an impressive level. Furthermore,
viscosity has played a definitive role in the detection of
reaction intermediates and the elucidation of their che-
mical behaviour. The classical cage-effect studies con-
stitute an excellent example of the use of solvent
viscosity as a powerful mechanistic tool. In pioneering
work, Kramers2 proposed a chemical reaction model by
considering the Brownian motion of a particle along the
reaction coordinate in the presence of a potential-energy
barrier. There have been many investigations on the
influence of solvent viscosity on rate in various chemical
reactions, such as stereochemical inversion upon nitrogen
ꢀ ¼ A expðV0=VfÞ
ð1Þ
In contrast to translational diffusion, enzymatic confor-
mational changes involve only a portion of the molecule.
Thus, only a fraction ꢁV0 (ꢁ < 1) of the critical volume V0
is required to execute the conformational motion, for
which the catalytic rate constant kcat of the reaction from
enzyme–substrate complex to product in the enzymatic
pocket is given by Eqn (2). Substitution of Eqn (1) into
Eqn (2) results in Eqn (3). From Eqn (3), the higher the
solvent viscosity, the smaller the reaction rate becomes.
extrusion from cyclic azoalkanes,3 electron transfer
3ꢀ/4ꢀ
between a platinum electrode and Fe(CN)6
or
ferrocene0/þ,4 and the isomerization reactions of stil-
kcat ¼ k0 expðꢀꢁV0=VfÞ
ð2Þ
ð3Þ
bene,5 to name just a few.
ꢁ
kcat ¼ k0ðA=ꢀÞ
Our approach is based on the conceptually simple free-
volume model of viscosity (ꢀ), in which the fractional-
power viscosity dependence (ꢀꢀꢁ) provides the effective
volume available for the enzymatic conformational
change under investigation. The free-volume model of
Carboxypeptidase-Y (CPD-Y) (EC 3.4.16.5) from ba-
ker’s yeast is a lysosomal serine exopeptidase, which
catalyses the hydrolysis of the C-terminal peptide bond of
proteins. It has been suggested that a Ser-His-Asp cata-
lytic triad is involved in the process as found in the
trypsin and chymotrypsin families of endopeptidases.7
An interesting feature is that a peptide bond consisting
of a proline residue can be hydrolysed by catalysis of
*Correspondence to: K. Ohkata, Department of Chemistry, Graduate
School of Science, Hiroshima University, 1-3-1 Kagamiyama, Higashi-
Hiroshima 739-8526, Japan. E-mail: ohkata@sci.hiroshima-u.ac.jp
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 448–457

INFLUENCE OF SOLVENT VISCOSITY ON HYDROLYSIS RATES OF DIPEPTIDES
449
CPD-Y, whereas this bond is resistant to cleavage by
ꢁ-chymotrypsin. CPD-Y is synthesized in the nucleus as
a preproenzyme and transported to vacuoles. CPD-Y is a
popular model protein for studying protein sorting in
yeast.8 Initial studies have focused on applications such
as C-terminal sequence analysis. Recently, the structure
of monomeric CPD-Y has been determined by x-ray
diffraction studies.9
In general, enzyme catalysis is initiated by interaction
between the enzyme and the substrate prior to reaction,
which gives rise to what is known as the enzyme–
substrate (E–S) complex or Michaelis complex. The
‘induced-fit’ process has been suggested to account for
the reactivity in enzymatic reactions.10–12 According to
this assumption, the enzyme reaction proceeds by cata-
lytic conformations induced by interaction between the
enzyme and the substrate. The flexible active site opens
up to allow the incorporation of the substrate, but once
the substrate is incorporated, the amino acid residues of
the active site become fixed around the substrate.
Increasing solvent viscosity generally results in strong-
er friction forces between solvent and solute. Likewise,
structural fluctuations of enzymes are also expected to be
diminished.13 Therefore, the rate constant of a step where
the conformational effect of an enzyme is significant
should be strongly affected by solvent viscosity. In fact,
there are several examples where the rate of an enzymatic
reaction or isomerization of a protein depends on the
solvent viscosity.14–17 For instance, it has been reported
that an increase in solvent viscosity causes a decrease in
the rate in the photocycle of bacteriorhodopsin,18 in the
diffusion of atom groups CO or O2 in myoglobin19 and in
enzymatic reactions by carboxypeptidase A.16 In the case
of the diffusion of CO or O2, the relation between
viscosity and rate could be correlated by the value of
0.4 < ꢁ < 0.8. In another example, from the dependence
on solvent viscosity of kinetic parameters for the hydro-
lysis of p-nitroanilides catalyzed by ꢁ-chymotrypsin, it
was concluded that the formation of a tetrahedral inter-
mediate in the acylation of the enzyme is significantly
more susceptible to conformational effects of the enzyme
than the breakdown of the intermediate.20
Scheme 1
Kinetics
All kinetic experiments were performed in pH 6.5 at
37 ꢁC. The progress of the reactions was determined by
monitoring spectrophotomerically (230–240 nm) the re-
lative amount of the released products. Kinetic para-
meters as designated in Scheme 2 for the hydrolysis of
these dipeptides by CPD-Y are summarized in Table 1.
The listed values are the average of 3–5 runs.
In addition, the enzymatic reaction of six substrates
was carried out in glycerol or sucrose at four or five
different concentrations and the kinetic data were calcu-
lated by non-linear fitting of the Michaelis–Menten
equation to plotted experimental data.
As a step towards gaining an understanding of the
mechanism of the hydrolysis of peptides by enzymes,
we have pursued this issue in terms of solvent viscosity
by comparing the difference in hydrolytic behaviour of N-
alkyl and N-H peptides catalysed by carboxypeptidase Y.
RESULTS AND DISCUSSION
Preparation and hydrolysis of
N-benzyloxycarbonyl-protected dipeptides
A series of dipeptides (Z-Phe-Gly, Z-Phe-Sar, Z-Phe-Ala,
Z-Phe-NMeAla, Z-Phe-Aib and Z-Phe-Pro) as shown in
Scheme 1 were prepared by typical procedures.
Scheme 2
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 448–457

450
Y. KANOSUE, S. KOJIMA AND K. OHKATA
Table 1. Kinetic parameters for hydrolysis of dipeptides by
CPD-Y
making the E–S complex relative to those of Ala, Sar, Aib
and Gly substrates. It was found that there is a high
correlation between hydrophobicity of the dipeptide sub-
strates and Km, suggesting that the hydrophobicity of the
C-terminal amino acid is a major factor in governing the
stability of the E–S complex.21 Therefore, the higher
reactivity (kcat/Km) of the Pro substrate relative to the
Sar, Aib and Gly substrates could be attributed to its
stronger affinity induced by hydrophobic interactions and
to the difference in steric demands to which it is subjected
in the transition state compared with the other substrates.
Examination of the x-ray structure of CPD-Y9 led to
the proposal that in CPD-Y the segment of the enzyme
bearing the amino acid residues Ser146, Asp338 and
His397, which could constitute the catalytic triad or
charge relay system, is the active site or catalytic cavity,
as illustrated in Scheme 3.
Dipeptide
Z-Phe-Xaa
Km
(mM)
kcat
(sꢀ1
kcat/Km
)
(mMꢀ1 sꢀ1
)
-NMeAla
-L-Pro
-L-Ala
-Sar
-Gly
-Aib
0.273 ꢃ 0.033
0.338 ꢃ 0.018
0.939 ꢃ 0.099
1.94 ꢃ 0.058
3.36 ꢃ 0.12
0.657 ꢃ 0.027 2.40 ꢃ 0.30
45.0 ꢃ 0.8
187.7 ꢃ 11.4
25.9 ꢃ 0.55
189 ꢃ 5.5
133 ꢃ 7
200 ꢃ 24
13.3 ꢃ 0.49
56.4 ꢃ 2.6
51 ꢃ 4.9
2.86 ꢃ 0.15
147 ꢃ 9.66
Hydrolysis of dipeptides by carboxypeptidase Y
The order of the overall reactivity (kcat/Km) of the C-
terminal residue in the dipeptides was Ala > Pro >
GlyꢂAib > Sar > NMeAla, as shown in Table 1.
On the basis of the Km values (NMeAla <
Pro < Ala < Sar < Aib < Gly), the affinity of the Pro
substrate towards the CPD-Y enzyme is favourable for
Near these catalytic residues are two hydrophobic
pockets consisting of the residues Thr60, Phe64, Glu65,
Tyr256, Tyr269, Leu272, Met398 and disulfide 56–298
Scheme 3. Acylation step in the hydrolysis by serine proteinase
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 448–457

INFLUENCE OF SOLVENT VISCOSITY ON HYDROLYSIS RATES OF DIPEPTIDES
451
designated the S10 binding site, and Tyr147, Leu178,
Influence of solvent viscosity on the kinetic
parameters Km and kcat
Tyr185, Tyr188, Leu245, Trp312, Ile340 and Cys341
designated the S1 binding site. It would be possible for
these pockets to accommodate hydrophobic side-chains
at the P1 (corresponding to the N-terminal amino acid
moiety) and P10 (corresponding to the C-terminal amino
acid moiety) positions, respectively. Another important
region at the active site, which could be assumed to be the
oxyanion hole consisting of Gly53 and Tyr147, is also
present.
Since more conformational change of the enzyme is
expected to be necessary in the enzymatic hydrolysis of
N-alkyl peptides relative to the N-H substrates, we
expected that the kcat of N-alkyl substrates would be
more sensitive than N-H substrates to changes in solvent
viscosity. Thus, the relation between solvent viscosity
and parameters Km and kcat was next examined. In order
to ascertain that viscosity did not affect the structure of
the enzyme, circular dichroism (CD) spectra in the
wavelength range 200–300 nm were recorded for the
enzyme dissolved in Tris–HCl buffer containing 50%
glycerol (6 M), and compared with spectra of a reference
solution containing just Tris–HCl buffer and of a solution
containing 8 M urea with buffer. As shown in Fig. 1, since
the spectral profiles of the 50% glycerol and reference
solutions were essentially identical, it was assumed that
the presence of glycerol did not greatly affect the enzyme
secondary structure.
The mechanism of bond cleavage can be rationalized
as proceeding via tetrahedral intermediates formed by
the attack of Ser146 upon the dipeptide substrates, which
go on to collapse to acyl-enzyme intermediates by
extrusion of the amino group of the cleaved peptide
bond following inversion upon the amide nitrogen atom
and protonation, as illustrated in Scheme 3. According to
the x-ray structure, there seems to be higher steric
hindrance at the tetrahedral intermediate (I) in the
enzymatic hydrolysis by ꢁ-chymotrypsin than carbox-
ypeptidase Y.22,23 When the C-terminal residue is
NMeAla and Sar, which have acyclic structures, differ-
ent steric interactions are expected to arise upon con-
formational change of the enzyme as compared with the
Pro substrate in the tetrahedral intermediate (II). This is
based on the principle of microscopic reversibility and
the stereoelectronic control theory.22 The role of the
His397 and Asp338 residues is thought to be to maintain
the Ser146 residue in a state capable of reacting with the
incoming peptide chain, and also to stabilize the tetra-
hedral intermediate formed during catalysis. The imida-
zole group of His397, which is essential in the enzymatic
catalysis,7b assists as a strong general base; the func-
tional group abstracts the alcoholic proton of Ser146 and
shuttles it to the amino leaving group. The negatively
charged Asp338 stabilizes the positive charge developed
on His397.
The order of the catalytic rate constants (kcat) in
the present work was Gly > Ala > Aib > Pro > Sar >
NMeAla and happens to be approximately the same as
the order of pKa values of the corresponding amino acids
[Gly (9.60), Ala (9.69), Aib (10.21), Pro (10.60) and Sar
(10.01)].24 Since the electron-donating ability of the
amino acid nitrogen atom is expected to correlate with
the pKa of the amino acid, the experimental results can be
rationalized as the more basic and thus the more electron-
donating amino acid being less capable of being attacked
by the oxide anion of the Ser146, thus leading to lower
kcat values.
In the hydrolysis of dipeptides by carboxypeptidase-Y,
Km increased only slightly with viscosity and the max-
imum variation was within 1.28 times the value with no
viscogen (Fig. 2, Table 2).
Generally, the effects of viscogen on enzymatic reac-
tions have been explained by excluded volume effects,25
secondary and tertiary structual change,26 changes in
bulk water concentration27 and inhibition by the added
viscogen.26 As the additives (glycerol and sucrose) are
very hydrophilic hydroxy compounds, these compounds
have no excluded volume effect but produce a stabiliza-
tion effect of proteins in aqueous solution.28 The compar-
ison of the CD spectra (Fig. 1) indicates that secondary
and tertiary structure changes by excluded volume effects
did not occur substantially in the enzyme itself. Changes
in bulk water concentration and inhibition by the additive
is concerned with the changes in Km values. On the other
hand, it is considered that the catalytic rate constants kcat
from the E–S complex to the product are not directly
affected by the additive but are indirectly influenced
through the enzymatic conformational changes since
the active site of CPD-Y is isolated from the surrounding
solvent.9
Since Km was little influenced by the presence of
viscogens, we could assume that the association of
the enzyme and a substrate does not require appreciable
conformational changes of the enzyme in the present
series of reactions. The small increase in the Km values
could be attributed to the loss of entropy upon the
incorporation of the substrate into the enzyme interior
by the partial destruction of the water cluster structure
caused by the intervention of viscogen molecules
and the competitive inhibitor by the viscogen. When
substrates are more difficult to incorporate into the
enzyme by competitive inhibitor, the values of Km
increase.
Since N-H substrates are similarly hydrolysed by both
ꢁ-chymotrypsin and carboxypeptidase-Y, whereas N-
alkyl substrates are only readily hydrolysed by the latter,
it can be assumed that whereas the sizes of the cavity of
both enzymes are similar, carboxypeptidase-Y is much
more flexible and large conformational changes facilitate
the formation of E–S complexes even for N-alkyl sub-
strates in the hydrolysis reactions.
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 448–457

452
Y. KANOSUE, S. KOJIMA AND K. OHKATA
Figure 1. CD spectra of carboxypeptidase-Y in Tris–HCl buffer at pH 6.5 (no additive), in 50 wt% glycerol in Tris–HCl buffer at
pH 6.5 and in 8 M urea in Tris–HCl buffer
adenosine deaminase-catalysed hydrolysis of adenosine
by transition-state analogue inhibitors,15 where sucrose
or Ficoll (Ficoll 400 is a neutral, highly branched,
hydrophilic polymer of sucrose which dissolves readily
in aqueous solution) was used as the viscogen. The
association and dissociation rates of the inhibitor showed
a decrease in the presence of the small viscogen sucrose
whereas the large viscogen Ficoll had no effect.15
Furthermore, in the hydrolysis of benzoylglycyl-L-
phenyllactate by carboxypeptidase A, where sucrose
and glycerol were used as viscogens,16 the reaction rate
constant kcat in the glycerol–water system was found to
be smaller than that in the sucrose–water system at the
same viscosity, revealing that glycerol affects the reaction
more than sucrose.
Here, the reaction rate constant kcat decreased as the
solvent viscosity increased for all the substrates studied.
It can be assumed that the decreases are due to the loss of
conformational mobility of the enzyme and can be
associated with the conformational change of enzymes
known as the ‘induced-fit’ process, induced by the inter-
action between an enzyme and the substrate. Hence it is
suggested that the kb step involving the formation of
Figure 2. (a) Dependence of the Km of CPD-Y-catalysed
tetrahedral intermediate (I) is the rate-determining step.20
hydrolysis of dipeptides Z-Phe-Xaa on viscosity. (b) Depen-
Furthermore, as described above, the dependence of the
dence of Km of the dipeptide Z-Pha-Xaa hydrolysis on
electron-donating ability of the amino acid moiety at the
C-terminus on the kcat values is consistent with this step
being rate determining.
Strajbl et al. have demonstrated that in the event of
nucleophilic attack of ROH such as serine protease on an
amide carbonyl group, the alcohol is assisted by base
by deprotonation.29 From the principle of microscopic
reversibility, the tetrahedral intermediate must therefore
eliminate ROꢀ instead of ROH. A comparison of the
pKa values of methanol and ethanol (15.6 and 16,
viscosity with sucrose or glycerol as viscogen
Figure 3(a) shows that the kcat values of the N-alkyl
peptides were more strongly affected by viscosity than
those of the N-H peptides. Figure 3(b) indicates that the
influence of viscosity on the kcat value in the sucrose–
water system is smaller than that in the glycerol–water
system.
The tendency for the smaller viscogen to be more
influential has also been observed in the inhibition of
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 448–457

INFLUENCE OF SOLVENT VISCOSITY ON HYDROLYSIS RATES OF DIPEPTIDES
453
Table 2. Kinetic parameters for the hydrolysis of the dipeptides Z-Phe-Xaa by CPD-Y in glycerol–water or sucrose–water mixed
solvent
Viscogenic
agent
Concentration
(wt%)
Viscosity
(mPa s)
kcat
(sꢀ1
Km
(mM)
)
Z-Phe-Ala
Glycerol
0
0.7197
0.7776
1.107
1.370
1.766
0.7197
0.7514
0.9791
1.329
1.888
0.7197
0.7665
0.9195
1.311
1.903
0.7197
0.8128
1.036
1.727
2.146
0.7197
1.026
1.376
1.77
2.053
0.7197
1.046
1.416
1.809
2.201
0.7197
0.8955
1.219
1.666
2.236
0.7197
0.9075
1.186
1.600
2.228
187.7 ꢃ 11.4
181.4 ꢃ 9.3
0.9386 ꢃ 0.099
1.020 ꢃ 0.054
1.055 ꢃ 0.025
1.176 ꢃ 0.023
1.180 ꢃ 0.391
0.3387 ꢃ 0.0180
0.3359 ꢃ 0.0039
0.3483 ꢃ 0.0088
0.3752 ꢃ 0.0113
0.4062 ꢃ 0.0149
3.358 ꢃ 0.121
3.971 ꢃ 0.337
4.091 ꢃ 0.121
4.298 ꢃ 0.198
4.303 ꢃ 0.216
1.944 ꢃ 0.058
2.134 ꢃ 0.084
2.375 ꢃ 0.182
2.418 ꢃ 0.104
2.429 ꢃ 0.078
0.9386 ꢃ 0.099
0.9745 ꢃ 0.0057
1.028 ꢃ 0.089
1.049 ꢃ 0.086
1.055 ꢃ 0.028
0.3387 ꢃ 0.0180
0.3392 ꢃ 0.0028
0.3469 ꢃ 0.0052
0.3537 ꢃ 0.0040
0.3586 ꢃ 0.0134
0.2740 ꢃ 0.0325
0.2759 ꢃ 0.0026
0.2979 ꢃ 0.0041
0.3095 ꢃ 0.0073
0.3341 ꢃ 0.0102
2.864 ꢃ 0.147
3.182 ꢃ 0.107
3.470 ꢃ 0.104
3.518 ꢃ 0.136
3.549 ꢃ 0.363
5.008
18.11
25.29
33.73
0
3.236
14.16
23.39
34.38
0
4.514
13.18
24.13
34.51
0
7.048
16.47
31.81
38.36
0
14.08
23.13
28.69
32.13
0
13.47
21.82
27.55
31.00
0
9.51
20.47
30.19
39.56
0
9.554
20.01
29.74
39.61
169.9 ꢃ 3.9
164.8 ꢃ 3.2
154.8 ꢃ 5.1
Z-Phe-Pro
Glycerol
44.97 ꢃ 0.78
41.09 ꢃ 0.36
36.16 ꢃ 0.72
33.96 ꢃ 0.86
30.84 ꢃ 1.00
189.4 ꢃ 5.5
Z-Phe-Gly
Glycerol
190.8 ꢃ 16.1
184.8 ꢃ 5.4
171.3 ꢃ 7.9
162.1 ꢃ 7.4
Z-Phe-Sar
Glycerol
25.93 ꢃ 0.55
23.76 ꢃ 0.92
21.36 ꢃ 1.62
19.69 ꢃ 0.84
18.33 ꢃ 0.59
187.7 ꢃ 11.4
182.6 ꢃ 1.0
Z-Phe-Ala
Sucrose
173.5 ꢃ 14.5
166.8 ꢃ 13.3
163.5 ꢃ 4.2
Z-Phe-Pro
Sucrose
44.97 ꢃ 0.78
40.01 ꢃ 0.26
35.45 ꢃ 0.42
34.38 ꢃ 0.34
34.12 ꢃ 1.07
0.6569 ꢃ 0.0270
0.5937 ꢃ 0.0039
0.5406 ꢃ 0.0058
0.5147 ꢃ 0.0099
0.4803 ꢃ 0.0126
147.3 ꢃ 9.7
Z-Phe-NMe-Ala
Glycerol
Z-Phe-Aib
Glycerol
135.8 ꢃ 4.6
129.2 ꢃ 3.9
118.2 ꢃ 4.6
117.2 ꢃ 2.4
respectively)30 with those of the ammonium group of
amino acids also indicates that the ROꢀ moiety has a
poorer leaving group ability from the tetrahedral inter-
mediate (I) than the counterpart protonated amino acid.
Here, the increase in the pKa values of the corresponding
amino acid makes the addition of ROꢀ even harder. In
other words, the addition of ROH assisted by base is more
difficult than the elimination of the corresponding amino
acid moiety. Since the activation energies of the N-
protonation and the N-inversion steps are generally
much smaller than those of the addition of alkoxide ion
and the elimination of amino acid steps, the rates of N-
protonation (ka, kd) and N-inversion (kc) should be faster
than those of addition (kb) and elimination (ke). If the step
involving C-terminal amino acid product release (kf)
were rate determining, a relationship between kcat and
the hydrophobicity of the corresponding amino acid
moieties would be expected since there are two hydro-
phobic pockets in CPD-Y. However, for the six substrates
in this study, no relationship (correlation factor r ¼ 0.12)
between kcat and hydrophobic parameter ÁGret,21 which
is correlated with the hydrophobicity of the correspond-
ing amino acid moiety, was observed. As the N-terminal
residues of all the substrates are the same and kcat differs
for all the substrates examined, the final kg step could not
be rate determining. This would leave the step involving
the formation (kb) of the tetrahedral intermediate (I)
and the step involving breakdown (ke) of the protonated
tetrahedral intermediate as candidates for the rate-
determining step. The fact that kcat diminished with
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 448–457

454
Y. KANOSUE, S. KOJIMA AND K. OHKATA
In order to evaluate the ꢁ values in Eqn (3), substitution
0
of kcat and ꢀ0, which are the values for no viscogen,
into Eqn (3) results in Eqn (4). Equation (3) divided by
Eqn (4) produces Eqn (5).
ꢁ
kc0at ¼ k0ðA=ꢀ0Þ
ð4Þ
ð5Þ
kcat=kc0at ¼ ðꢀ=ꢀ0Þꢀꢁ
The ꢁ values for the substrates employed here were
calculated from normalized kcat/k0cat and normalized
ꢀ/ꢀ0 with Eqn (5) by non-linear fitting (Fig. 4, Table 3).
The ꢁ values correspond to the degree of dependence
of the rate on viscosity and here may be regarded as the
degree of solvation around the active site which is located
on the inside of the surface of the enzyme; ꢁ ¼ 1
corresponds to a case where the variations in the solvent
mobility are fully transferred to the reaction site of the
Figure 3. (a) Dependence of the kcat of CPD-Y-catalysed
hydrolysis of dipeptides Z-Phe-Xaa on viscosity in glycerol–
Tris–HCl buffer mixed solvent. (b) Dependence of kcat of the
dipeptide Z-Phe-Xaa hydrolysis on viscosity with sucrose or
glycerol as viscogen
higher amino acid basicity and higher solvent viscosity
implies that the former formation step is the overall rate-
determining step.
Support for this conclusion in seen in ab initio calcula-
tions of the base-catalysed methanolysis of formamide,
which can be considered a simplified model system of an
enzymatic hydrolysis reaction. The nucleophilic attack of
an alkoxide anion is indicated to be the rate-determining
step and the calculated activation energy of this step is
consistent with experimental values for the general base-
catalysed hydrolysis of several amides.29
Further support for the rate-determining step was
provided by Fersht and Jencks, who showed that the
rate constant k2 (Mꢀ1 sꢀ1) for the nucleophilic attack
of R00Oꢀ on amide groups increases with decreasing
electron-donating ability of the amine moiety according
to the relationship log k2 ¼ 7.2 ꢀ 0.67pKa(R0NH3þ),
where k2 represents the rate constant for the step invol-
ving acylation of the enzyme in the hydrolysis by serine
proteases.31 Furthermore, it is feasible for the catalytic
His to abstract a proton from the catalytic Ser simulta-
neously as the latter attacks the peptide bond to be
cleaved, because CPD-Y is known to be highly active
even at low pH where other serine endopeptidases
are inactive.8 In the alcoholysis of amides at low pH,
the rate-determining step is generally the formation of a
tetrahedral intermediate.31
Figure 4. (a) Normalized kcat value against normalized
viscosity of the dipeptide Z-Phe-Xaa hydrolysis in glycerol–
Tris–HCl buffer mixed solution. (b) Normalized kcat value
against normalized viscosity of the dipeptide Z-Phe-Xaa
hydrolysis with sucrose or glycerol as viscogen
Table 3. ꢁ Values calculated from Eqn (5)
Glycerol
Sucrose
Z-Phe-L-Ala
Z-Phe-Gly
Z-Phe-Aib
Z-Phe-L-Pro
Z-Phe-NMeAla
Z-Phe-Sar
0.21 ꢃ 0.01
0.16 ꢃ 0.01
0.24 ꢃ 0.02
0.45 ꢃ 0.06
0.30 ꢃ 0.02
0.34 ꢃ 0.04
0.12 ꢃ 0.01
0.29 ꢃ 0.02
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 448–457

INFLUENCE OF SOLVENT VISCOSITY ON HYDROLYSIS RATES OF DIPEPTIDES
455
enzyme. From the ꢁ values in Table 3, the hydrolysis
reaction of the N-alkyl peptides was more sensitively
affected than the N-H peptides by variations of viscosity.
According to the free-volume model of viscosity,6 as
viscosity increases, the stronger an effect free volume has,
resulting in a larger ꢁ value. In other words, the hydrolysis
reaction of the N-alkyl peptides accompanied larger struc-
tural fluctuations (larger free volume) than the N-H pep-
tides. These results are consistent with the ‘induced-fit’
process model in the enzymatic reactions, induced by the
interaction between the enzyme and the substrate.
Examination of the x-ray structure9 of the main chain
of CPD-Y [Fig. 5(a)] reveals that the catalytic His397 is
near the C-terminal helix domain (Phe401–Ile415). This
domain is not bound by a surrounding peptide chain and
not located within a ꢂ-strand structure [Fig. 5(c)] and
may therefore fluctuate with ease. On the other hand, the
catalytic His of ꢁ-chymotrypsin is located within a ꢂ-
strand structure (Trp51–Thr54 and Val65–Ala68) and is
fixed to a surrounding peptide chain [Fig. 5(b)]. For this
reason, we believe that the hydrolysis reaction of the N-
alkyl peptides accompanied more fluctuation of the
domain including His397.
during the reaction, and therefore is more susceptible to
solvent effects. The decrease in k with increasing ꢀ
implies that the increase in friction forces exerted by
the viscous solvent leads to retardation of the fluctuation
of the enzyme.
It has also been reported that most proteins are effec-
tively hydrated on the addition of glycerol, this being the
origin of structure stabilization of proteins in aqueous
glycerol.28 As shown in Fig. 3, although there is a slight
difference in the degree of change, the profiles were
similarly shaped for the two different viscogens. This
supports our conclusion that viscosity is directly related
to the speed of conformational change and that the
phenomenon is independent of the nature of the viscogen.
EXPERIMENTAL
Instrumentation. CD spectra were recorded on Jasco J-
500CH instrument. A Shimadzu UV-240 UV–visible
recording spectrophotometer was used for kinetic mea-
surements. Viscosity was measured by using a stalagm-
ometer at 310 K.
Scheme 4 illustrates the effect of viscosity on the
enzymatic reaction by the free-volume model.
General synthetic method. Treatment of L-phenylalanine
(Phe) with benzyloxycarbonyl chloride (Z-Cl) and L-
alanine (Ala) with di-tert-butyl dicarbonate (Boc2O)
afforded Z-Phe and Boc-Ala, respectively. Methylation
of Boc-Ala with methyl iodide gave the N-methylalanine
derivative (Boc-MeAla). Reaction of ꢁ-amino acid
(Xaa ¼ L-Ala, Gly, NMeAla, L-Pro, Aib and Sar) with
thionyl chloride followed by MeOH gave the correspond-
ing methyl ester (HClꢄXaa-OMe). The coupling reactions
With the conformational change of the substrate during
the conversion from the Michaelis complex to a tetra-
hedral intermediate as depicted, the geometry of the
amide bond nitrogen changes from planar to pyramidal
with the accompaniment of a conformational change of
the enzyme. As in Scheme 4(b), it is expected that a larger
conformational change of the enzyme (larger free vo-
lume) is required in the case of the N-alkyl substrates
Figure 5. The main chain of CPD-Y x-ray structure (PDB code: 1YSC).9 (b), (c) Sequence and secondary structure of ꢁ-
chymotrypsin [(PDB code: 4CHA),23 Leu33–Lys82, (b)] and carboxypeptidase-Y [Ala371–Leu421, (c)]. The sequence is shown as
a single letter code and H is the catalytic histidine of each enzyme. The assignments of secondary structure are: H ¼ helix;
B ¼ residue in the isolated beta bridge; E ¼ extended ꢂ-strand; T ¼ hydrogen-bonded turn; S ¼ bend
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 448–457

456
Y. KANOSUE, S. KOJIMA AND K. OHKATA
Scheme 4. The viscosity effect on the enzymatic reaction
were mediated by dicyclohexylcarbodiimide to give the
corresponding N- and O-protected dipeptides (Z-Phe-
Xaa-OMe). Deprotection of the C-terminus of the dipep-
tides was carried out by treatment with an equivalent
volume of 1 M NaOH for 0.5–3 h at room temperature to
give the dipeptide substrates Z-Phe-Xaa.
(c ¼ 0.100, MeOH); 1H NMR (500 MHz, CDCl3), ꢃ
7.33–7.16 (10H, m), 6.23 (1H, d, J ¼ 9.0 Hz), 5.17 (1H,
q, J ¼ 7.5 Hz), 5.12 (1H, d, J ¼ 8.5 Hz), 5.09 (1H, d,
J ¼ 8.5 Hz), 4.95 (1H, quintet, J ¼ 7.4 Hz), 3.09–2.94
(2H, m), 2.86 (3H, s), 1.39 (3H, d, J ¼ 7.5 Hz). Found:
C, 65.47; H, 6.46; N, 7.19%. Calculated for C21H24N2O5:
C, 65.63; H, 6.25; N, 7.29%.
N-Benzyloxycarbonyl-L-phenylalanyl-2-methylalanine (Z-
20
Phe-Aib). M.p. 158 ꢁC (lit.32 161 ꢁC), ½ꢁꢅD ¼ ꢀ9.7
N-Benzyloxycarbonyl-L-phenylalanyl-L-proline (Z-Phe-L-
20
20
(c ¼ 0.10, MeOH) (lit.32 ½ꢁꢅD ¼ ꢀ9.5, c ¼ 0.1, MeOH);
Pro). M.p. 107–108 ꢁC (lit.35 106.5 ꢁC), ½ꢁꢅD ¼ ꢀ43.7
1H NMR (500 MHz, CDCl3), ꢃ 7.47–7.05 (10H, m), 6.48
(1H, s), 5.69 (1H, s), 5.08 (2H, s), 4.48 (1H, m), 3.05 (2H,
m), 1.55 (3H, s), 1.42 (3H, s).
(c ¼ 0.130, MeOH); 1H NMR (500 MHz, CDCl3) ꢃ
7.37–7.120 (10H, m), 5.57 (1H, d, J ¼ 9.0 Hz), 5.11
(1H, d, J ¼ 12.5 Hz), 5.06 (1H, d, J ¼ 12.5 Hz), 4.69
(1H, dd, J ¼ 8.0, 15.1 Hz), 4.58 (1H, d, J ¼ 5.5 Hz),
3.51 (1H, dd, J ¼ 8.5, 15.9 Hz), 3.06–3.03 (2H, m),
2.85–2.80 (1H, m), 2.34–2.30 (1H, m), 2.01–1.93 (1H,
m), 1.89–1.83 (1H, m), 1.78–1.71 (1H, m). Found: C,
66.64; H, 6.13; N 6.90%. Calculated for C22H24N2O5: C,
66.65; H, 6.10; N, 7.06%.
N-Benzyloxycarbonyl-L-phenylalanyl-L-alanine (Z-Phe-
20
L-Ala). M.p. 160–162 ꢁC (lit.33 165 ꢁC), ½ꢁꢅD ¼ ꢀ10.8,
20
(c ¼ 0.0198, MeOH) (lit.33 ½ꢁꢅD ¼ ꢀ11.0, c ¼ 2, alco-
hol);1H NMR (500 MHz, CDCl3) ꢃ 7.47–7.05 (10H, m),
6.47 (1H, s), 5.41 (1H, s), 5.08 (2H, s), 4.52 (1H, q,
J ¼ 7.5 Hz), 4.50 (1H, t, J ¼ 7.5 Hz), 3.10 (1H, q,
J ¼ 7.5 Hz), 3.06 (1H, q, J ¼ 7.5 Hz), 1.38 (3H, d,
J ¼ 7.5 Hz).
N-Benzyloxycarbonyl-L-phenylalanylsarcosine (Z-Phe-
20
Sar). M.p. 61–63 ꢁC (lit.36 46–50 ꢁC), ½ꢁꢅD ¼ ꢀ9.4
20
(c ¼ 0.40, EtOH) (lit.36 ½ꢁꢅD ¼ ꢀ9.5, c ¼ 4.0, EtOH);
N-Benzyloxycarbonyl-L-phenylalanylglycine (Z-Phe-Gly).
1H NMR (500 MHz, CDCl3), ꢃ 7.34–7.15 (10H, m),
5.91 (1H, d, J ¼ 9.0 Hz), 5.08 (1H, d, J ¼ 12.5 Hz), 5.02
(1H, d, J ¼ 12.5 Hz), 4.95 (1H, q, J ¼ 8.2 Hz), 4.20 (1H,
d, J ¼ 17.5 Hz), 3.94 (1H, d, J ¼ 17.5 Hz), 3.07–2.94 (2H,
m), 2.88 (3H, s). Found: C, 64.71; H, 6.24; N, 7.45%.
Calculated for C20H22N2O5: C, 64.85; H, 5.99; N, 7.56%.
FABMS: m/z 371.1639. Calculated for C20H23N2O5:
371.1608.
20
M.p. 144–147 ꢁC (lit.34 154 ꢁC), ½ꢁꢅD ¼ ꢀ9.9
18
(c ¼ 0.025, AcOH) (lit.34 ½ꢁꢅD ¼ ꢀ10.2, c ¼ 2.73,
AcOH); 1H NMR (500 MHz, CDCl3), ꢃ 7.47–7.05
(10H, m), 6.52 (1H, s), 5.39 (1H, s), 5.07 (2H, s), 4.54
(1H, dd, J ¼ 6.5, 12.5 Hz), 4.06 (1H, d, J ¼ 17.9 Hz), 3.92
(1H, d, J ¼ 17.9 Hz), 3.96–3.92 (1H, m), 3.08 (2H, m).
Found: C, 63.94; H, 5.57; N, 7.92%. Calculated for
C19H20N2O5: C, 64.04; H,5.66; N; 7.68%.
Solvent viscosity. The viscosity of the reaction solvent
was controlled by adding an appropriate amount of
glycerol or sucrose to Tris–HCl buffer (50 mM, pH 6.5).
N-Benzyloxycarbonyl-L-phenylalanyl-N-methyl-L-alanine
20
(Z-Phe-NMeAla).
M.p.
39–40 ꢁC,
½ꢁꢅD ¼ ꢀ33.2
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 448–457

INFLUENCE OF SOLVENT VISCOSITY ON HYDROLYSIS RATES OF DIPEPTIDES
457
When an equal volume of viscous fluid is measured by
using the same thin tube, the viscosity (ꢀ) is given by
REFERENCES
1. (a) Schroeder A, Troe J. In The Barrier Crossing Problem,
Fleming GR, Hanggi P (eds). World Scientific: Singapore,
1993; 206–240; (b) Voth GA, Hochstrasser RM. J. Phys. Chem.
1996; 100: 13034–13049; (c) Sun Y-P, Saltiel J, Park NS, Hoburg
EA, Waldeck DH. J. Phys. Chem. 1991; 95: 10336–10344.
2. Kramers HA. Physica. 1940; 7: 284–304.
3. (a) Adam W, Grune M, Diedering M, Trofimov AV. J. Am. Chem.
Soc. 2001; 123: 7109–7112; (b) Adam W, Marti V, Sahin C,
Trofimov AV. Chem. Phys. Lett. 2001; 340: 26–32; (c) Adam W,
Diedering M, Trofimov AV. Chem. Phys. Lett. 2001; 350: 453–
458; (d) Adam W, Diedering M, Trofimov AV. Phys. Chem. Chem.
Phys. 2002; 4: 1036–1039; (e) Adam W, Diedering M, Trofimov
AV. J. Am. Chem. Soc. 2002; 124: 5427–5430; (f) Adam W,
Diedering M, Trofimov AV, Sajimon MC. Phys. Chem. Chem.
Phys. 2003; 5: 329–332; (g) Adam W, Corley DA, Trofimov AV,
White RC. Org. Lett. 2002; 4: 4277–4280; (h) Adam W, Trofimov
AV. Acc. Chem. Res. 2003; 36: 571–579.
ꢀ ¼ ðꢄtꢀ0Þ=ðꢄ0t0Þ
ð6Þ
where t is the time required for passage through the tube,
ꢄ is the solution density and ꢀ0, t0 and ꢄ0 are the values for
water as the standard fluid. The time for passage of 5 ml
of solutions and water was measured five times with the
stalagmometer. The average passage times were used to
calculate the solution viscosities relative to water.
CD spectra. The effect of solvent viscosity on the sec-
ondary structure of the enzyme was investigated by CD
spectroscopy in the wavelength range 200–300 nm. The
enzyme was dissolved in solutions controlled at pH 6.5
with Tris–HCl buffer (50% glycerol, 8 M urea and a
reference solution without either additive).
4. Zhang X, Leddy J, Bard AJ. J. Am. Chem. Soc. 1985; 107: 3719–
3721.
5. Park NS, Waldeck DH. J. Chem. Phys. 1989; 91: 943–952.
6. Doolittle AK. J. Appl. Phys. 1951; 22: 1471–1475.
7. (a) Isolation: Zuber H. Nature 1964; 201: 613; (b) Hayashi R, Bai
Y, Hata T. J. Biol. Chem. 1975; 250: 5221–5226.
8. Jung G, Ueno H, Hayashi R. J. Biochem. 1999; 126: 1–6.
9. Endrizzi JA, Breddam K, Remington SJ. Biochemistry 1994; 33:
11106–11120.
10. Schanstra JP, Janssen DB. Biochemistry 1996; 35: 5624–5632.
11. Krooshof GH, Ridder IS, Tepper AWJW, Vos GJ, Rozeboom HJ,
Kalk KH, Dijkstra BW, Janssen DB. Biochemistry 1998; 43:
15013–15023.
12. (a) Koshland JDE. Proc. Natl Acad. Sci. USA 1958; 44: 98–104;
(b) Koshland JDE, Neet KE. Annu. Rev. Biochem. 1968; 37: 359–
410.
13. Gavish B, Werber MM. Biochemistry 1979; 18: 1269–1275.
14. Rayfield W. Photochem. Photobiol. 1986; 43: 171–174.
15. Kurz LC, Weitkamp E, Frieden C. Biochemistry 1987; 26: 3027–
3032.
Enzyme CPD-Y. CPD-Y prepared by Oriental Yeast was
purchased from Wako Pure Chemical Industries and used
without purification.
Kinetic measurements. A spectrophotometer was used for
kinetic measurements. Kinetics were measured by fol-
lowing the decrease in absorbance at 230 nm. The reac-
tion conditions were kept constant over all kinetic
measurements: temperature 310 K, concentration of the
buffer 50 mM, pH 6.5.
16. Goguadze NG, Hammerstad-Pedersen JM, Khoshtariya DE,
Ulstrump J. Eur. J. Biochem. 1991; 200: 423–429.
17. Kawai Y, Matsuo T, Ohno A. Bull. Chem. Soc. Jpn. 1998; 71:
2187–2196.
18. Beece D, Boune SF, Czege J, Eisenstein L, Fraufelder H, Marden
MC, Ormos J, Reinisch L, Yue KT. Photochem. Photobiol. 1981;
33: 517–522.
19. Beece D, Eisenstein L, Frauenfelder H, Good D, Marden MC,
Reinisch L, Reynolds AH, Sorensen LB, Yue KT. Biochemistry
1980; 23: 5147–5157.
20. Kawai Y, Matsuo T, Ohno A. J. Chem. Soc., Perkin Trans. 2 2000;
887–891.
21. Kanosue Y, Hiraga Y, Kojima S, Ohkata K. Chem. Lett. 2001;
418–419.
22. (a) Bizzozero SA, Zweifel BO. FEBS Lett. 1975; 59: 105–108; (b)
Dugas H, Bioorganic Chemistry, a Chemical Approach to Enzyme
Action (3rd edn). Springer: Berlin, 1996; 222.
23. Tsukada H, Blow DM. J. Mol. Biol. 1985; 184: 703–711.
24. Dawson RMC. Data for Biochemical Research. Clarendon Press:
Oxford, 1959.
25. Ralston GB. J. Chem. Educ. 1990; 67: 857–860.
26. Mozhaev VV, Khmelnitsky YL, Sergeeva MV, Belova AB,
Klyachko NL, Levashov AV, Martinek K. Eur. J. Biochem.
1989; 184: 597–602.
As a typical procedure of these experiments, carbox-
ypeptidase-Y (4.68 mg, ca 7.8 ꢆ 10ꢀ8 mol) was dissolved
in Tris–HCl buffer (1 ml) and the solution was diluted
appropriately. A stock solution of substrate was prepared
(ca 1 mM) by dissolving the substrate in Tris–HCl buffer
and diluting appropriately. The solution of substrate
(3 ml) was placed in a thermostated compartment of the
spectrophotometer and incubated for 5 min. After 20 ml of
enzyme solution had been added, the mixture was shaken
to make it homogeneous. The absolute absorbance of this
solution was then measured. One set of measurements
was composed of data from five different substrate
concentrations and the initial velocity of each substrate
concentration was calculated as the average of 3–5 runs.
The dipeptide and the hydrolysis product of the dipeptide
Z-Phe-Xaa of carboxypeptidase-Y were checked by re-
versed-phase HPLC and no diastereomer was detected.
The mobile phase (0.5 ml minꢀ1) consisted of acetoni-
trile–water (3:7) containing 60 mM AcONa buffer
(pH 6.0).
27. Pocker Y, Janjic N. Biochemistry 1988; 27: 4114–4120.
28. Gekko K, Timasheff SN. Biochemistry 1981; 20: 4667–4686.
29. StrajblM, FlorianJ, WarshelA. J. Am. Chem. Soc. 2000;122:5354–
5366.
30. Ballinger P, Long FA. J. Am. Chem. Soc. 1960; 82: 795–798.
31. Fersht AR, Jencks WP. J. Am. Chem. Soc. 1970; 92: 5442–5452.
32. Jelokhani-Niaraki M, Yoshioka K, Takahashi H, Kato F, Kondo M.
J. Chem. Soc., Perkin. Trans. 2 1992; 1187–1193.
33. Grassman W, Wunsch E, Riedel A. Chem. Ber. 1958; 91: 455–
462.
34. Weygand F, Steglich W. Chem. Ber. 1960; 93: 2983–3005.
35. Sakakibara S, Itoh M. Bull. Chem. Soc. Jpn. 1967; 40: 656–661.
36. Gerber C, Seebach D. Helv. Chim. Acta 1991; 74: 1373–1385.
Acknowledgements
We are indebted to the Instrument Center for chemical
analysis, Hiroshima University and Hiroshima Prefec-
tural Institute of Industrial Science and Technology for
machine time on the NMR and MS instruments.
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 448–457