Protein Engineering vol.14 no.9 pp.683–689, 2001
The chemical modification of α-chymotrypsin with both
hydrophobic and hydrophilic compounds stabilizes the enzyme
against denaturation in water–organic media
A.A.Vinogradov1,2, E.V.Kudryashova3, V.Ya.Grinberg4,
N.V.Grinberg4, T.V.Burova4 and A.V.Levashov3
1A.N.
Nesmeyanov Institute of Organoelement Compounds, 28 Vavilov str.,
119991 Moscow, 3Department of Chemical Enzymology, M.V. Lomonosov
Moscow State University, Vorobievy gory, 117899 Moscow and
4N.M. Emmanuel Institute of Biochemical Physics, 4 Kosygina str.,
117977 Moscow, Russia
2To
whom correspondence should be addressed.
E-mail: [email protected].
We considered α-chymotrypsin (CT) in homogeneous
water–organic media as a model system to examine the
influence of enzyme chemical modification with hydrophilic
and hydrophobic substances on its stability, activity and
structure. Both types of modifying agents may lead to
considerable stabilization of the enzyme in water–ethanol
and water–DMF mixtures: (i) the range of organic cosolvent
concentration at which enzyme activity (Vm) is at least
100% of its initial value is broadened and (ii) the range
of organic cosolvent concentration at which the residual
enzyme activity is observed is increased. We found that for
both types of modification the stabilization effect can be
correlated with the changes in protein surface hydrophobicity/hydrophilicity brought about by the modification.
Circular dichroism studies indicated that the effects of
these two types of modification on CT structure and its
behavior in water–ethanol mixtures are different. Differential scanning calorimetry studies revealed that after modification two or three fractions or domains, differing in
their stability, can be resolved. The least stable fractions
(or domains) have properties similar to native CT.
Keywords: α-chymotrypsin/chemical modification/kinetics/
stabilization/water–organic mixtures
Introduction
Recently there has been impressive progress in biocatalyst
design tools, such as directed evolution (Arnold et al., 1999).
Nevertheless, understanding structure–function relationships is
still one of the challenging tasks in protein science. It is well
known that even minor structural changes may cause dramatic
changes of protein properties. The protein surface is primarily
responsible for interaction with the surrounding solvent. Thus,
the properties of the enzyme surface may be crucial for
maintaining its catalytically active conformation, particularly
in non-natural environments (non-water and water–organic
solvents). The possibility of enzymatic reactions in such
media has greatly increased the industrial use of enzymes
(Khmelnitsky and Rich, 1999). Therefore, it is interesting from
both theoretical and practical points of view to determine the
relationships between enzyme surface properties and enzyme
stability in water–organic solvent media.
Biocatalysis in non-aqeous media has been the subject of
numerous studies (Oldfield, 1994; Tuena de Gomez-Puyou and
Gomez-Puyou, 1998; Halling, 2000; Klibanov, 2001). Some
© Oxford University Press
enzyme suspensions in non-polar water-immiscible organic
solvents were shown be to very stable (Zaks and Klibanov,
1985). It has been suggested that in more polar solvents, water
molecules, essential for stabilizing the native conformation,
are stripped from the protein surface (Zaks and Klibanov,
1984). Several approaches have been proposed to make water
molecules bind more tightly or to immobilize the protein in
hydrophilic matrices (Khmelnitsky and Rich, 1999).
Enzymes dissolved in polar organic solvents or in homogeneous water–organic media usually lose their native conformation and catalytic activity. It is generally believed that the
introduction of additional hydrophilic groups on the protein
surface may improve its stability in such systems owing to the
enhanced ability of the hydrophilized surface to keep the
hydration shell and form additional electrostatic interactions,
hydrogen bonds or salt bridges (Khmelnitsky et al., 1991;
Mozhaev et al., 1996). On the other hand, hydrophobic
groups, if introduced near the hydrophobic region on protein
surface, may also contribute to protein stabilization (Lee
and Richards, 1971; Chotia, 1984). Arnold and co-workers
suggested that the replacement of disordered, uncompensated
surface charge with hydrophobic residues using site-directed
mutagenesis may dramatically improve enzyme stability
(Arnold, 1988; Martinez and Arnold, 1991). Crambin, a small
hydrophobic protein, is an amazing demonstration of natural
protein design for non-aqueous media: it is soluble and retains
its native structure at very high concentrations of polar organic
solvents (De Marco, 1981). Crambin is not soluble in aqueous
solutions, but has water-soluble homologs with similarly folded
structure (Teeter et al., 1981). In comparison with these
homologs, crambin possesses amino acid substitutions, removing hydrophilic side chains: lysines, arginines and asparagines
are almost excluded in crambin.
In recent work, we have performed hydrophobization/
hydrophilization via chemical modification of α-chymotrypsin
surface amino groups. The aim was to understand how both
types of modification affect enzyme activity, structure and
stability.
Materials and methods
Covalent modification of α-chymotrypsin
Acetylation of α-chymotrypsin (CT) (Sigma, EC 3.4.21.1) was
carried out with pyromellitic anhydride (Aldrich) and succinic
anhydride (Reanal) as described previously (Mozhaev et al.,
1988).
Reductive alkylation with glyceric, pentylic, isobutyric and
acetic aldehydes (Reanal) was carried out according to
Mozhaev et al. (Mozhaev et al., 1992). Modification with a
mixture of glyceric and pentyl aldehydes was made in exactly
the same way (a 100 molar excess of each aldehyde with
respect to the enzyme was taken).
Determination of the degree of modification
The degree of CT modification was calculated from the number
of amino groups in it which reacted with trinitrobenzene683
A.A.Vinogradov et al.
sulfonic acid (TNBS) as compared with the unmodified enzyme
(Fields, 1971).
Titration of α-chymotrypsin active sites
The concentration of active sites in all CT samples was
determined with N-trans-cinnamoylimidazole (Sigma) as titration reagent (Schonbaum et al., 1961).
Other procedures
Inhibition with phenylmethanesulfonyl fluoride (PMSF) was
performed as described (Fahrney and Gold, 1963).
The enzymatic activity of α-chymotrypsin in binary
water–organic mixtures was determined with N-benzoylL-tyrosine-p-nitroanilide (BTNA) (Sigma) and with 4-methylumbelliferyl-p-trimethylammonium
cinnamate
chloride
(MUTMAC) as described previously (Kudryashova et al.,
1997).
Calculations of hydrophobicity of modifiers and of change
of hydrophobicity of modified CT were made according to
Mozhaev et al. (Mozhaev et al., 1992).
Differential scanning calorimetry (DSC) measurements were
made essentially as described elsewhere (Grinberg et al.,
2000). Protein solutions for calorimetric measurements were
dialyzed against corresponding media for at least 5 h at 4°C.
Buffer solution was 20 mM MOPS, pH 7.45. The protein
concentration after dialysis was determined spectrophotometrically using the extinction coefficient E280 ⫽ 50 000
M–1 cm–1 (Volini and Tobias, 1969). The concentration of
protein solutions used for calorimetric measurements was
0.5–1.2 mg/ml. DSC measurements were carried out with
a DASM-4 adiabatic differential scanning microcalorimeter
(Biopribor, Puschino, Russia) at a heating rate of 1°C/per min
and at extra pressure 1 atm. Primary data processing was
performed using NAIRTA software (Institute of Biochemical
Physics, Russian Academy of Sciences, Moscow, Russia).
A quantitative thermodynamic analysis should be based on
equilibrium studies. Nevertheless, in some cases one can use
this analysis for practically irreversible processes (Privalov
and Potekhin, 1986). The thermal denaturation of CT is known
to be irreversible. To minimize the interference of irreversible
denaturation, all measurements were made under the same
conditions. We aimed to compare the calorimetric parameters
of modified and native samples of CT.
Circular dichroism experiments were carried out with a
Jobin Yvon Mark V spectrometer at 25°C with quartz cells of
pathlength 1 mm in the far-UV region (200–260 nm). The
protein concentrations were 0.1–0.2 mg/ml. The secondary
structure percentage predictions were made using CDNN
software (http://bioinformatik.biochemtech.uni-halle.de/cdnn).
Polyacrylamide gel electrophoresis was performed under
denaturing conditions according to Laemmli (Laemmli, 1970).
Results
Enzymatic activity
We found that both surface hydrophilization and hydrophobization enhance protein stability against denaturation in water–
organic mixtures. This effect was expressed in (i) broadening
of the range of organic cosolvent concentration at which
enzyme activity (Vm) is at least 100% of its initial value and
(ii) the increase in the organic cosolvent concentration range
at which the residual enzyme activity is observed.
In water–organic solvent mixtures native (non-modified) CT
is relatively unstable (Mozhaev et al., 1989; Kudryashova
684
Fig. 1. Catalytic activity of some CT samples in water–DMF mixtures. r,
Native CT; m, CT modified with pyromellitic anhydride, degree of
modification 11 (pyr11CT); *, CT modified with glyceric aldehyde, degree
of modification 13 (glyc13CT); j, CT modified with pentyl aldehyde,
degree of modification 11 (pent11CT).
Fig. 2. Catalytic activity of some CT samples in water–ethanol mixtures.
Symbols as in Figure 1.
et al., 1994; Gladilin et al., 1995). It is almost inactive at 50%
(v/v) ethanol (EtOH), which is in agreement with data from
other groups (Sato et al., 2000); in 40% (v/v) dimethylformamide (DMF) CT is completely inactive owing to denaturation.
At moderate organic cosolvent concentrations (up to 20%,
v/v) native CT is more active than in a water environment
(Figures 1 and 2). Further, the catalytic activity is gradually
decreased to complete inactivation.
Native CT was modified with several hydrophilic reagents
(degrees of modification are given in parentheses): succinic
anhydride (3, 6), pyromellitic anhydride (2, 8, 11, 15) and
glyceric aldehyde (2, 6, 13) (Table I). A substantial stabilization
effect was achieved for pyromellitic and glyceric anhydrides.
At higher degrees of modification this effect was greater
(Figures 1 and 2). The stabilization effect of succinic anhydride
was less pronounced. Samples modified with glyceric aldehyde
(glycCT) were found to be more resistant to aggregation in
water–ethanol mixtures: precipitation of glycCT was observed
at 80% (v/v) ethanol, whereas native CT precipitated at 40%.
For hydrophobic modification of CT we used pentylic
aldehyde (degree of modification 3 and 11), isobutyric aldehyde
(degree of modification 6 and 7) and acetic aldehyde (degree
of modification 5). Modifications with pentylic (pentCT) and
isobutyric aldehydes (ibutCT) led to substantial stabilization
against denaturation in water–organic mixtures (Figures 1 and
2). This case is particularly interesting, because the replacement
of a surface charged amino group with a hydrophobic chain
Chemical modification of α-chymotrypsin
Table I. The increments of hydrophobicity (∆∆G) introduced by the
modifications
Fig. 3. Correlation of stabilizing effect of modification, expressed by DMF
concentration (C100) up to which the activity of CT is at least 100% of its
value in aqueous solution, with Hansch hydrophobic increments (∆∆G) of
functional groups and atoms introduced by the modification.
Table II. The measured calorimetric parameters of denaturation in aqueous
solution (20 mM MOPS, pH 7.5) for modified and non-modified
chymotrypsin samples
aThe values of (∆∆G) per group were calculated from Hansch hydrophobic
increments of functional groups and atoms as described elsewhere
(Leo et al., 1971; Mozhaev et al., 1992).
bData on catalytic activity of these samples in water–ethanol mixtures have
been published by Kudryashova et al. (Kudryashova et al., 1997).
is not usually expected to stabilize the protein. Moreover, the
resistance of the samples against aggregation in water–ethanol
mixtures was enhanced in comparison with native CT: ibutCT
and pentCT formed true solutions at 80% (v/v) ethanol.
Modification with acetic aldehyde did not have much effect
on the CT behavior in water–organic solvents.
Normally, not more than 20% of CT active sites were lost
in the course of modification. The catalytic properties of
modified samples were not dramatically different from native
enzyme owing to several experimental facts. (1) The hydrolysis
of two substrates differing in the rate-limiting step, BTNA and
MUTMAC, was investigated. The rate-limiting step for BTNA
is acyl-enzyme formation (West et al., 1990) and for MUTMAC
it is deacylation (Maurel, 1978). The catalytic activity of
modified samples in aqueous solution does not change from
native CT by more than 50%, taking into account the percentage
of the active sites. In water–organic media the modified
samples demonstrate similar trends to the native enzyme, but
possess higher stability (Figures 1 and 2). (2) The pH dependence of catalytic activity for modified chymotrypsins is similar
to that for native chymotrypsins (data not shown). (3) All
modified samples were characterized using SDS–PAGE and
no difference from the native form was observed (data not
shown). This also indicates that intermolecular ‘sewing’, which
could be expected in the case of pyromellitic anhydride, does
not take place. (4) The Km values of modified samples are the
same order of magnitude as that for the native enzyme. The
Km values of modified and native CT samples behave similarly:
they increase with organic cosolvent concentration (data not
shown).
We found that the values of the stabilization effect, expressed
by organic cosolvent concentration at which enzyme activity
returns to its initial value (C100), can be correlated with
the increment of hydrophobicity (∆∆G) introduced by the
Sample
T1/2 (°C)
∆Hd (kJ/mol)
Native CT
Pent11CT
Glyc13CT
48
50.3
50.6
690
1148
1165
modification (Figure 3). The values of (∆∆G) were calculated
from Hansch hydrophobic increments of functional groups and
atoms (Mozhaev et al., 1992; Leo et al., 1971) and are given
in Table I. In the case of hydrophilic and hydrophobic
modifications at small and moderate increments (∆∆G), the
stabilization effect is proportional to the amounts of hydrophobicity introduced. It is interesting that left and right
halves of the curve, corresponding to hydrophobization and
hydrophilization of enzyme surface, are almost symmetric
about the y-axis (Figure 3). At higher ∆∆G the stabilization
effect is almost independent of further increases in the
increment.
In addition, we performed the modification of CT with a
mixture of glyceric and pentylic aldehydes (glyc–pentCT)
taken in equal amounts. The total degree of modification of
this sample was determined as 11. From our experience, the
reactivities with glyceric and pentylic aldehydes are identical,
so one might suppose that the degrees of modification for
them are almost equal and hence the ∆∆G value should be
about –40 kJ/mol (Table II). Despite such a small increment,
glyc–pentCT appeared to be the most stable of all modified
samples studied in recent work – it was characterized by
C100 ⫽ 43.5% in DMF.
To reveal the mechanisms by which the two types of
modification affect the enzyme properties, we studied native
and modified CT samples in water–organic mixtures using
circular dichroism (CD) spectroscopy and differential scanning
calorimetry (DSC).
Circular dichroism
The samples analyzed by far-UV CD spectroscopy, glyc13CT
(hydrophilic modification) and ibut7CT (hydrophobic modification), are fairly similar considering the catalytic activity in
water–ethanol media (Figure 2). They are much more stable
685
A.A.Vinogradov et al.
Table III. The percentage of secondary structure elements, estimated from
CD spectra using CDNN software (http://bioinformatik.biochemtech.unihalle.de/cd_spec/cdnn)
Sample
Native CT,
theoretical dataa
Native CT
glyc13ct
ibut7CT-PMSF
aExtracted
EtOH
(%)
α-Helix
β-Sheets
β-Turn
Random
coil
0
0
20
30
50
70
0
20
30
40
45
50
0
10
10.7
10.3
7.5
8.0
7.8
12.9
11.0
9.3
13.6
38.6
57.8
9.1
34
32.6
32.0
38.8
39.0
39.1
28.3
30.4
30.2
24.2
8.2
6.2
32.8
20
22.3
22.6
19.4
18.8
19.0
23.3
23.2
24.2
25.3
24.0
15.3
22.4
36
34.4
35.1
34.3
34.2
34.1
35.4
35.5
36.3
36.9
29.1
20.6
35.7
from reference file of CDNN program.
against inactivation than the native form. However, from the
CD spectra we found that the structural properties of the
analyzed samples in aqueous solution and in water–ethanol
media are very different.
The CD spectrum for native CT (not shown) was similar to
the published spectrum (Gorbunoff, 1971). For comparison,
the values according to theoretical data are also presented.
Upon addition of ethanol up to 20% (v/v) practically no
changes in the CD spectra are observed. The values of
secondary structure contents vary within the estimation error
(Table III). The addition of more ethanol (30%, v/v) to the
media gives rise to considerable changes in the structure: the
β-sheet content is increased mainly at the expense of α-helix
and β-turn. Very similar changes take place at higher ethanol
concentrations (50 and 70%, v/v). The observed increase in
β-sheet content can be attributed to aggregation of the protein,
which is known for CT under the given conditions (Kudryashova et al., 1997). It is worth noting that the decrease in
catalytic activity is the sharpest between 25 and 30% (v/v)
ethanol (Figure 2). Hence the inactivation of CT in water–
ethanol mixtures can be accounted for by aggregation. The
formation of β-structures and loss of activity in 50% ethanol
were reported for native CT also by another group (Sato
et al., 2000).
The CD spectrum of glyc13CT (Figure 4) in aqueous solution
is less regular in comparison with native CT. Deconvolution of
the CD spectrum indicates a lower percentage of β-sheets and
a higher percentage of α-helix and β-turns (Table III). However,
from the values of these variations one can judge that the
secondary structure has not been distorted dramatically by
the modification. Much more pronounced perturbations were
observed upon addition of ethanol to the media (Figure 4).
The corresponding values of secondary structure content
change non-monotonically (Table III). At ethanol concentrations up to 30% (v/v) the α-helix content gradually decreases
from 12.9 to 9.3%, probably leading to the slight increase of
other structures. However, at higher ethanol concentrations the
α-helix content increases about 6-fold, to 57.8% at 50% (v/v)
EtOH. At the same time the β-sheet content drops approximately 5-fold, from 30.2 to 6.2%. The random coil and β-turn
contents also decrease substantially. Instead of aggregation
and complete loss of activity (which is the case for the native
686
Fig. 4. Circular dichroism spectra of glyc13CT in the far-UV region at
various ethanol concentrations (v/v): 0, 0%; 1, 20%; 2, 30%; 3, 40%; 4,
45%; 5, 50%.
Fig. 5. Circular dichroism spectra of ibut7CT in the far-UV region at
various ethanol concentrations (v/v): 0, 0%; 1, 40%; 2, 50%; 3, 70%; 4,
80%.
enzyme), for glyc13CT we observe the refolding to some new,
‘incorrect’ structure, which is able to remain soluble and retains
considerable catalytic activity (Figure 2). A phenomenon of
such ‘incorrect’ refolding to a structure with higher helix
contents has been reported for some proteins on exposure to
water–alcohol mixtures (Buck et al., 1993; Fan et al., 1993).
Probably this is the way in which the protein can ‘adapt’ to a
non-natural environment.
It was surprising to see that the far-UV CD spectrum of
ibut7CT is extremely destructured (Figure 5) under conditions
where the enzyme is fully active. Several computer programs
were tried, but all of them failed to yield a reliable estimation
of secondary structure contents in ibut7CT samples. However,
we found that the CD spectrum is dramatically changed after
inhibition of ibut7CT with PMSF. PMSF is an irreversible
inhibitor of CT and a PMSF-inhibited sample can be considered
as a model of an enzyme–substrate transition state. Figure 6
shows the CD spectrum of PMSF-ibut7CT in comparison with
ibut7CT. The molar ellipticity of the PMSF-inhibited sample
is substantially higher and the corresponding percentage of
secondary structure elements is similar to that of native CT
Chemical modification of α-chymotrypsin
Fig. 7. The experimental excess heat capacity function of pent11CT in 20%
(v/v) ethanol (as an example) can be deconvoluted as a superposition of two
independent components. Solid line, experimental trace; dashed line,
deconvolution.
Fig. 6. Circular dichroism spectra in the far-UV region of ibut7CT (1) and
ibut7CT inhibited with PMSF (2).
(Table III). Apparently, the modification with butyric aldehyde
led to partial reversible unfolding of the protein. However, it
is possible that the core structure of the protein crucial for
catalytic activity is still intact. Unlike native CT, the given
‘partially unfolded structure’ can undergo considerable perturbations, but still remain soluble and retain its catalytic
activity at high ethanol concentrations. On addition of PMSF
the ‘partially unfolded structure’ is reversed to a native-like
form. Perhaps ibut7CT, as well as glyc13CT, is able to ‘adapt’
to a non-natural medium. However, the mechanism of its
‘adaptation’ is apparently different.
Differential scanning calorimetry
For DSC studies we selected glyceric aldehyde-modified CT
with degree of modification 13 (hydrophilic modification,
glyc13CT); and pentylic aldehyde-modified CT with degree
of modification 11 (hydrophobic modification, pent11CT). As
we mentioned, these samples were considerably more stable
than native CT in respect of retention of catalytic activity in
water–DMF mixtures (Figure 1). The parameters obtained by
DSC indicate that in aqueous media both modifications lead
to the formation of CT structure with higher conformational
stability (Table II).
We also followed the behavior of pent11CT and glyc13CT
in water–DMF media using DSC techniques. For both modified
and native CT, the addition of DMF causes an almost linear
decrease in denaturation temperature and a non-linear decrease
in enthalpy of denaturation. For native CT the excess heat
capacity function contains a single peak, which apparently
corresponds to a single cooperative system (Privalov, 1979).
However, for both modified samples, particularly at 10–20%
(v/v) DMF, the experimental excess heat capacity function
shows the presence of several fractions or quasi-independent
structural components (Privalov, 1981). We deconvoluted the
experimental excess heat capacity functions as a superposition
of several independent components and calculated the parameters for each of them (Figure 7).
From such a deconvolution, it can be judged that pent11CT
is more likely to contain two components. The behavior of
the less stable one (component 2) is similar to that of the
native CT, whereas component 1 is considerably more stable.
This can be judged from the changes in the corresponding
∆Hd and T1/2 with increase in DMF concentration (Figures 8
Fig. 8. Dependence of denaturation temperature (T1/2) on DMF
concentration for pent11CT represented as a two-component system. s,
Native CT; n, component 1 (less stable) of pent11CT; u, component 2
(more stable) of pent11CT.
Fig. 9. Dependence of denaturation enthalpy (∆Hd) on DMF concentration
for pent11CT represented as a two-component system. The total height of
the column corresponds to the total entalpy of denaturation. Black, the
contribution of component 1 (less stable); gray, the contribution of
component 2 (more stable).
and 9). Perhaps the modification transforms some part of the
protein globule into a stable domain, which can be resolved
using DSC. It is also possible that the above components 1
and 2 represent two fractions of pent11CT: ‘successfully’ and
‘unsuccessfully’ modified molecules. The presence of residual
activity at high concentrations of organic cosolvent can be
687
A.A.Vinogradov et al.
Fig. 10. Dependence of denaturation temperature (T1/2) on DMF
concentration for glyc13CT represented as a three-component system. s,
Native CT; j, component 1 (the most stable) of pent11CT; m, component 2
of pent11CT; r, component 3 (the least stable).
Fig. 11. Dependence of denaturation enthalpy (∆Hd) on DMF concentration
for glyc13CT represented as a three-component system. The total height of
the column corresponds to the total enthalpy of denaturation. Black, the
contribution of component 3 (the least stable); white, the contribution of
component 2; gray, the contribution of component 1 (the most stable).
accounted for by the functioning of a ‘successfully’ modified
domain or fraction.
The experimental excess heat capacity function of glyc13CT
is well fitted as a three-component system. Component 1 (the
most stable) is obviously more stable in comparison with
native CT, but components 2 and 3 are similar or even less
stable (Figures 10 and 11). Component 1 could represent a
‘successfully’ modified domain or fraction. Component 2 and
especially component 3 are in fact ‘non-successfully’ modified.
Perhaps in the latter case, the modification of some ‘incorrect’
residues took place. For example, owing to the higher degree
of modification, amino groups crucial for maintaining the
hydrophilic interaction network were modified. Also, the
introduced hydrophilicity of glyceric aldehyde has not compensated for the loss of those interactions.
Discussion
The introduction of additional hydrophobicity, as well as
hydrophilicity, on to the CT surface led to the formation of a
more stable protein structure (domains or fractions), less
vulnerable to denaturation in water–organic mixtures. It is
most probable that the mechanisms of stabilization for two
types of modifying agents are different. These conclusions can
688
be drawn from the recent work. While the stabilization effect
was not surprising to us in the case of hydrophilic modification
(Khmelnitsky et al., 1991; Kudryashova et al., 1997), the
mechanism of stabilization upon hydrophobic modification
is unclear.
Replacements of surface-charged amino acids with hydrophobic ones using site-directed mutagenesis was reported to
improve enzyme stability substantially (Arnold, 1988; Martinez
and Arnold, 1991; Martinez et al., 1992; van den Burg et al.,
1994). This approach works if the replaced hydrophilic residue
is not crucial for maintaining salt bridges and a hydrogenbond network. But how could such explanation be extended
to the case of chemical modifications, where the sites of
modification are not known?
From the chemical point of view, the reactivity of amino
functions involved in multiple hydrophilic interactions will be
lower. Hence they would be less susceptible to chemical
modification. If that is so, uncompensated, unsatisfied surface
residues (lysines and N-terminal amino acids) would have the
highest reactivity. If our assumption is correct, all these residues
are the first candidates for chemical modification. Substituted
with a hydrophobic substance, they can be expected to contribute to protein stabilization in water–organic mixtures. However,
if this mechanism was really working in our case, one could
expect the stabilizing effect not only for pentylic and isobutyric
aldehydes, but also for acetic aldehyde. Our data indicates that
this is not so – no effect of the latter was observed. Therefore,
to make the protein more stable it is not sufficient just to
replace the unfavorable surface charge.
In our case, it could be supposed that additional hydrophobic
groups introduced during modification create additional interactions with unfavorable hydrophobic regions on the protein
surface. Apparently, this must contribute to protein stabilization
(Lee and Richards, 1971; Chotia, 1984). The hydrophobic
increment of acetic aldehyde is perhaps not sufficient to build
such interactions.
Another possibility is that in the course of modification the
distribution of modificator molecules around the protein globule is not even: more hydrophobic molecules would preferentially react with amino groups adjacent to uncompensated
hydrophobic regions on the protein surface. If so, we have a
kind of self-adjusting system, where the modifier ‘finds’
the way to attach near the uncompensated region. Similar
suggestions can be put forward for hydrophilic modification.
For the examination of our hypothesis we performed the
modification of CT with a mixture of glyceric and pentylic
aldehydes (glyc–pentCT) taken in equal amounts. As mentioned in the Results section, the total degree of modification
of this sample was determined as 11; also considering the
reactivity with glyceric and pentylic aldehydes, the degrees of
modification for them should be equal. Hence the ∆∆G value
should be about –40 kJ/mol (see Table II). It is amazing that
despite such a small increment, glyc–pentCT appeared to be
the most stable of all modified samples studied in recent work.
This is an indication that the most efficient stabilization is not
achieved by hydrophilization or hydrophobization itself, but
by ‘successful’ modification. Perhaps in glyc–pentCT the
structure has been optimally readjusted, so that both stabilization mechanisms are working.
Acknowledgements
We thank Dr Vadim V.Mozhaev of EnzyMed, Iowa City, IA, USA for his
substantial contribution to this work in the initial stages. We also thank
Chemical modification of α-chymotrypsin
Martina Reddington of Ireland for checking our English. This work was
supported by the Russian program ‘The New Methods in Bioengineering:
Engineering Enzymology’.
References
Arnold,F.H. (1988) Protein Eng., 2, 21–25.
Arnold,F.H. and Volkov,A.A. (1999) Curr. Opin. Chem. Biol., 3, 54–59.
Buck,M., Radford,S.E. and Dobson,C.M. (1993) Biochemistry, 32, 669–678.
Chotia,C. (1984) Annu. Rev. Biochem., 53, 537–572.
De Marco,A., Lecomte,J.T.J. and Llinas,M. (1981) Eur. J. Biochem., 119,
483–490.
Fahrney,D. and Gold,A.M. (1963) J. Am. Chem. Soc., 85, 997–1000.
Fan,P., Bracken,C. and Baum,J. (1993) Biochemistry, 32, 1573–1582.
Fields,R. (1971) Biochem. J., 124, 581–590.
Gladilin,A.K.,
Kudryashova,E.V.,
Vakurov,A.V.,
Izumrudov,V.A.,
Mozhaev,V.V. and Levashov,A.V. (1995) Biotechnol. Lett., 17, 1329–1334.
Gorbunoff,M. (1971) Biochemistry, 10, 250–257.
Grinberg,V.Ya., Burova,T.V., Haertle,T. and Tolstoguzov,V.B. (2000)
J. Biotechnol., 79, 269–280.
Halling,P.J. (2000) Curr. Opin. Chem. Biol., 4, 74–80.
Khmelnitsky,Y.L. and Rich,J.O. (1999) Curr. Opin. Chem. Biol., 3, 47–53.
Khmelnitsky, Yu.L., Belova,A.B., Levashov,A.V. and Mozhaev,V.V. (1991)
FEBS Lett., 284, 267–269.
Klibanov,A.M. (2001) Nature, 409, 241–246.
Kudryashova,E.V., Belova,A.B., Vinogradov,A.A. and Mozhaev,V.V. (1994)
Russ. J. Bioorg. Chem., 20, 274–280.
Kudryashova,E.V., Gladilin,A.K., Vakurov,A.V., Heitz,F., Levashov,A.V. and
Mozhaev,V.V. (1997) Biotechnol. Bioeng., 55, 267–277.
Laemmli,U.K. (1970) Nature, 227, 680–684.
Lee,B. and Richards,F.M. (1971) J. Mol. Biol., 55, 187–211.
Leo,A., Hansch,C. and Elkins,D. (1971) Chem. Rev., 71, 525–643.
Martinez,P. and Arnold,F.H. (1991) J. Am. Chem. Soc., 113, 6336–6337.
Martinez,P., van Dam,M.E., Robinson,A.C., Chen,K. and Arnold,F.H. (1992)
Biotechnol. Bioeng., 39, 141–147.
Maurel,P. (1978) J. Biol. Chem., 253, 1677–1683.
Mozhaev,V.V.,
Silksnis,V.A.,
Melik-Nubarov,N.S.,
Galkantaite,N.Z.,
Denis,G.J., Butkus,E.P., Zaslavsky,B.Yu., Mestechkina,N.M. and Martinek
K. (1988) Eur. J. Biochem., 173, 147–154.
Mozhaev,V.V., Khmelnitsky,Y.L., Sergeeva,M.V., Belova,A.B., Klyachko,N.L.,
Levashov,A.V. and Martinek,K. (1989) Eur. J. Biochem., 184, 597–602.
Mozhaev,V.V., Melik-Nubarov,N.S., Siksnis,V.A., Levitsky,V.Y. and
Martinek,K. (1992) Biotechnol. Bioeng., 40, 650–662.
Mozhaev,V.V., Kudryashova,E.V., Efremova,N.V. and Topchieva,I.N. (1996)
Biotechnol. Tech., 10, 849–854.
Oldfield,C. (1994) Biotechnol. Genet. Eng. Rev., 12, 255–327.
Privalov,P.L. (1979) Adv. Protein Chem., 33, 167–241.
Privalov,P.L. (1981) Adv. Protein Chem., 35, 1–104.
Privalov,P.L. and Potekhin,S.A. (1986) Methods Enzymol., 131, 4–51.
Sato,M., Sasaki,T., Kobayashi,M. and Kise,H. (2000) Biosci. Biotechnol.
Biochem., 64, 2552–2558.
Schonbaum,G.R., Zerner,B. and Bender,M. (1961) J. Biol. Chem., 236,
2930–2938.
Teeter,M.M., Mazer,J.A. and L’Italien,J.J. (1981) Biochemistry, 20, 5437–5443.
Tuena de Gomez-Puyou,M. and Gomez-Puyou,A. (1998) Crit. Rev. Biochem.
Mol. Biol., 33, 53–89.
van den Burg,B., Dijkstra,B.W., Vriend,G., van der Vinne,B., Venema,G. and
Eijsink,V.G. (1994) Eur. J. Biochem., 220, 981–985.
Volini,M. and Tobias,P. (1969) J. Biol. Chem., 244, 5105–5109.
West,J.B., Hennen,W.J., Lalonde,J.L., Bibbs,J.A., Zhong,Z., Meyer,E.F.,Jr. and
Wong,C.H. (1990) J. Am. Chem. Soc., 112, 5313–5320.
Zaks,A. and Klibanov,A.M. (1984) Science, 224, 1249–1251.
Zaks,A. and Klibanov,A.M. (1985) Proc. Natl Acad. Sci. USA, 82, 3192–3196.
Received February 16, 2001; revised June 8, 200l; accepted June 18, 2001
689