On the rational design of substrate mimetics: The function of docking approaches for the prediction of protease specificities

Article abstract of DOI:10.1039/b316641d

The behaviour of substrate mimetics in mediating the acceptance of nonspecific acyl moieties by proteases has been investigated as a direct function of their site-specific ester leaving groups. In this contribution we report on a computational approach to

Full text of DOI:10.1039/b316641d

View Article Online / Journal Homepage / Table of Contents for this issue
On the rational design of substrate mimetics: the function of
docking approaches for the prediction of protease specificities
Robert Günther,*a Christian Elsner,a Stephanie Schmidt,b Hans-Jörg Hofmanna
and Frank Bordusa*b
a University of Leipzig, Faculty of Biosciences, Pharmacy and Psychology,
Institute of Biochemistry, Brüderstr. 34, D-04103 Leipzig, Germany.
E-mail: robguent@uni-leipzig.de; Fax: ϩ49 341 973 6998; Tel: ϩ49 341 973 6737
b Max-Planck Society, Research Unit for Enzymology of Protein Folding, Weinbergweg 22,
D-06120 Halle/Saale, Germany. E-mail: bordusa@enzyme-halle.mpg.de;
Fax: ϩ49 345 551 1972; Tel: ϩ49 345 552 2806
Received 22nd December 2003, Accepted 23rd March 2004
First published as an Advance Article on the web 16th April 2004
The behaviour of substrate mimetics in mediating the acceptance of nonspecific acyl moieties by proteases has
been investigated as a direct function of their site-specific ester leaving groups. In this contribution we report on a
computational approach to rationalise this interplay and to predict the power of a potential ester moiety to act as
a suitable substrate mimetic for a given enzyme by means of an automated docking procedure. Investigations with
seven distinct substrate mimetics and two proteases, subtilisin and chymotrypsin, show a clear correlation between
the theoretically calculated binding energies ∆E and the specificity constants kcat KMϪ1 obtained from parallel
hydrolysis kinetic studies. These results prove the general function of the docking approach as a rational model
not only in predicting the general acceptance of a substrate mimetic in a qualitative manner, but also to provide
reliable information on its individual specificity towards proteases.
substrates must be arranged. Approaches to the screening of
Introduction
potential substrate mimetics used so far, however, are based on
simple empirical comparisons of the structural similarity
between the side chain of preferred amino acid residues and
respective ester leaving groups and allow at best only very
rough estimations of the general suitability of a given ester
moiety.11,12
Enzymes exhibit unique specificities and selectivities compared
to traditional means of chemical catalysis making them
attractive technological tools in applied chemical processes.1
Especially for modifications of complex molecules with
multiple stereocenters or a wide variety of functional groups
they open up a new dimension of reaction control that cannot
be achieved by classical chemical approaches. With these
advantages, however, comes the disadvantage that the
pronounced specificities reduce simultaneously the universality
of enzymes as synthetic catalysts. Efforts to broaden the limited
substrate spectrum of native enzymes are, therefore, of decisive
importance for raising the scope of these catalysts for synthesis.
Related to proteases, substantial improvements in the enzyme’s
synthesis flexibility have been attained by the development of
substrate mimetics used as the acyl donor components.1 Based
on their unique architecture, which is typically characterized
by a shift of the site-specific amino acid moiety from the
C-terminus of the peptide residue to the substrate’s ester
leaving group, substrate mimetics mediate the acceptance of
completely nonspecific acyl moieties. This behaviour enables
the synthesis of a broad spectrum of compounds including
for instance peptides,2 isopeptides,3 all--peptides,4 peptide
isosteres,5 peptidoglycans6 or nonpeptidic carboxylic acid
amides.7 Moreover, a function of the approach to the resolution
of racemates has recently been proposed.8 Furthermore, it has
been shown that suitably adapted substrate mimetics are
well accepted by virtually all serine and cysteine proteases
independent of their individual native substrate preferences
making the approach a rather general one in protease catalysis.1
As suggested by attempts to visualize the structure of the
substrate mimetic–enzyme complex formed during reaction,
a specific binding of the ester leaving group to the specificity-
determining S1 position of the respective protease (nomen-
clature according to ref. 9) can be considered as the key event
for mediating the acceptance of substrate mimetics.10 Moreover,
it became evident that for productive binding a conformation
of the scissile ester bond similar to that of normal-type
This contribution reports on the utility of an automated
docking method to predict the acceptance and specificity of
potential substrate mimetics to proteases of known 3D-
structure. The general function of the approach benefits from
the kinetics of substrate mimetic-mediated reactions that
involve a fast acylation of the enzyme followed by a deacylation
being rate limiting.13 In cases where identical acyl moieties
within the substrate mimetics are used, the rate of the latter can,
therefore, be considered equal for the corresponding enzyme
without having any influence on the specificity of the respective
substrate (cf. Scheme 1). Consequently, the specificity of the
individual substrate mimetic is exclusively encoded in the bind-
ing and acylation behaviour of the ester leaving group. The
former can be calculated directly by the docking approach in
terms of the binding energies while for estimating the efficiency
of the acylation step the productivity of the computed
substrate–enzyme complexes is analyzed.
Scheme
1
Kinetic model of protease-mediated substrate ester
hydrolysis. EH, free enzyme; Ac-X, substrate ester; [E...Ac-X],
Michaelis–Menten complex; HX, leaving group; Ac-E, acyl enzyme
intermediate; Ac-OH, hydrolysis product.
Results and discussion
Theoretical calculations
We initially selected a set of seven individual types of substrate
mimetics with respect to the primary specificity of subtilisin
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2, 1 4 4 2 – 1 4 4 6
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 4
1442

View Article Online
Fig. 1 Protease–ligand complexes of (a) subtilisin (pdb-entries:14 1av7, 1avt, 1cse, 1lw6, 1scn, 2sec, 3sec, 1sib, 2sic, 5sic, 2sni, 1sua and 3vsb) and (b)
chymotrypsin (pdb-entries: 1acb, 1cho, 2cha, 6cha, 1hja and 1mtn). Only the residues of the catalytic triad, the residues forming the oxanion hole
(orange) and the P1-residue of the ligand (grey) are shown. The hydrogen bonds are drawn as dotted lines; the arrow indicates the attack of the
catalytic serine.
and chymotrypsin. Since both proteases are well known to
prefer originally aromatic amino acid moieties,15 leaving
groups derived from benzyl, indoyl and in particular phenyl
esters have been empirically chosen as potential candidates.
The latter were favoured because of the already reported
acceptance of phenyl (OPh) and 4-guanidinophenyl (OGp)
esters by chymotrypsin.16 For selecting a suitable substitution
pattern of the aromatic ring systems, earlier findings were
considered which reveal a role of hydrogen bonds for stabilising
the complex of the two proteases with their native substrates.15
Thus, functionalities were preferred that act either as a
donor or acceptor of hydrogen bonds. As the acyl residue, the
tert-butyloxycarbonyl-alanine moiety has been chosen in all
cases, which can be considered as a neutral and only very
poorly accepted amino acid derivative by both proteases. A
complete list of the ester structures selected is illustrated
in Scheme 2.
Prior to synthesis, the esters were docked against the crystal
structures of subtilisin and chymotrypsin, respectively,
employing the program package AutoDock.17 The calculated
substrate–enzyme complexes showing the lowest binding
energies were subsequently analysed for a catalytically product-
ive orientation of the substrate mimetic within the active site
of the respective protease. The criteria for a productive sub-
strate mimetic–enzyme complex were defined after analysing
3D structures of subtilisin and chymotrypsin complexed with
specific ligands. As illustrated in Fig. 1, all these protein–ligand
arrangements share characteristic features of the orientation of
the P1 residue of the complexed compound with respect to the
active site of the corresponding protease; i.e. the distance
between the substrate’s carbonyl carbon atom of the ester bond
and the γ-oxygen atom of the active serine is less than 3 Å.
Furthermore, the carbonyl oxygen of all ligands points towards
the oxanion hole, which is formed by the backbone amide
moiety of serine 221 and the side chain amide of asparagine 155
in the case of subtilisin and by the backbone amide groups of
serine 195 and glycine 193 in the case of chymotrypsin. For
detailed analysis of binding energies, only those substrate
mimetic–protease arrangements fulfilling the aforementioned
productive orientation have been considered. In most cases
these were the most energetically favourable complexes or the
highest populated complexes in the docking runs, respectively
(cf. Experimental section).
Fig. 2 shows a snapshot of the individual binding of the
substrate mimetics to the two proteases in their productive and
energetically most preferred conformations. Importantly, all the
complexes have in common that the ester leaving groups point
into the S1 binding pocket while the acyl moiety is located in the
SЈ region of the respective protease. In accordance to earlier
studies,10 such an arrangement can be considered typically for
substrate mimetics which already indicates an acceptance of all
substrate esters by the two enzymes.
A more rational and detailed picture becomes apparent from
Table 1 listing the individual binding energies ∆E calculated for
the productively bound substrate mimetics. Considering an
intrinsic error of the energy evaluation method of approx-
imately 1.5 kcal molϪ1, the ∆E values obtained for subtilisin are
almost equivalent without displaying any significant deviations.
This finding let one expect that the variation of the ester leaving
group’s structure should have almost no effect on the specificity
of subtilisin towards the substrate mimetics involved in this
study. This is in good agreement with the well-known broad
specificity of subtilisin. On the contrary, a completely different
situation is found for chymotrypsin. For this enzyme, the bind-
ing energies vary between Ϫ6.5 and Ϫ10.2 kcal molϪ1 indicating
significant differences in the specificity of chymotrypsin
Scheme 2 Structures of substrate mimetics: OBzl: benzyl ester; OCap:
4-amido carboxy phenyl ester; OCp: 4-carboxy phenyl ester; OGp:
4-guanidino phenyl ester; OInd: 2-carboxy-1H-indol-5-yl ester; OPh:
phenyl ester; OPic: 4-picoyl ester.
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2, 1 4 4 2 – 1 4 4 6
1443

View Article Online
Fig. 2 Productive complexes of the investigated substrate mimetics docked to (a) subtilisin and (b) chymotrypsin. Surfaces of the proteins are
coloured according to their electrostatic potential (blue = positive, red = negative) calculated with Grasp.19 The residues of the catalytic triad and the
residues of the oxanion hole are in orange.
Table 1 Binding energies of productive substrate–protease complexes
calculated with AutoDock
Table 2 Specificity constants for the hydrolysis of substrate mimetics
by subtilisin and chymotrypsina
∆E/kcal molϪ1 a
kcat KMϪ1/MϪ1 sϪ1 b
Ester
Subtilisin
Chymotrypsin
Ester
Subtilisin
Chymotrypsin
Boc-Ala-OBzl
Boc-Ala-OCap
Boc-Ala-OCp
Boc-Ala-OGp
Boc-Ala-OInd
Boc-Ala-OPh
Boc-Ala-OPic
Ϫ6.3
Ϫ5.9
Ϫ6.8
Ϫ6.2
Ϫ7.3
Ϫ5.3
Ϫ5.9
Ϫ7.9
Ϫ8.9
Ϫ6.5
Ϫ9.2
Ϫ10.2
Ϫ7.1
Ϫ7.7
Boc-Ala-OBzl
Boc-Ala-OCap
Boc-Ala-OCp
Boc-Ala-OGp
Boc-Ala-OInd
Boc-Ala-OPh
Boc-Ala-OPic
1.7 × 104
1.1 × 104
6.5 × 103
7.2 × 103
2.1 × 103
1.6 × 104
1.2 × 104
1.1 × 102
5.1 × 103
3.7 × 101
3.6 × 103
1.0 × 102
1.9 × 102
6.8 × 102
a All calculated data have an intrinsic error of approximately 1.5 kcal
a Conditions: 0.1 M Hepes, pH 8.0, 0.1 M NaCl, 10 mM CaCl2, 8%
17
molϪ1
.
(v/v) DMF, 25 ЊC. b Errors are less than 15%.
towards the individual substrate esters with the lowest for
Boc-Ala-OCp and the highest for Boc-Ala-OInd. Boc-
Ala-OGp, which is already commonly used as a substrate
mimetic for this enzyme,16,18 shows a binding energy very close
to that of Boc-Ala-OInd. Interestingly, a comparable value of
∆E was also calculated for Boc-Ala-OCap, which only differs
from the worst ester Boc-Ala-OCp by having a neutral amide
group instead of a negatively charged carboxylate at position 4
of the aromatic leaving group. Higher binding energies and,
thus lower specificities were predicted for Boc-Ala-OBzl, Boc-
Ala-OPic, and Boc-Ala-OPh, which themselves show ∆E values
very close to each other.
than a factor of ten. Although less specific in total, chymo-
trypsin shows a more pronounced specificity profile towards the
individual substrates with specificity constants varying by more
than two orders of magnitude. Plotting of the specificity
Ϫ1
constants kcat KM vs. the calculated binding energies ∆E for
subtilisin (Fig. 3) and chymotrypsin (Fig. 4) reflects clearly the
different behaviour of the two proteases. Furthermore, it illus-
trates the correlation of the two values and, thus allows for
Enzymatic hydrolysis
To prove the predictions made by the automated docking
approach, the specificity constants have been experimentally
determined by usual steady-state hydrolysis studies. Plotting
of the initial rates of hydrolysis vs. the substrate ester concen-
tration resulted in straight lines for all enzyme/substrate com-
binations (plots are not shown). From the slopes of the respect-
Ϫ1
ive curves, the apparent second order rate constants kcat KM
have been calculated (Table 2). Analysis of the resulting data
show that all substrate esters were found to be specifically
hydrolysed by the two proteases as it was predicted by the dock-
ing approach. Also in good agreement to the docking calcu-
lations are the similar kcat KM values of the distinct esters in
the case of subtilisin, which only differ from each other by less
Ϫ1
Fig. 3 Plot of specificity constants kcat KM of subtilisin-catalysed
Ϫ1
hydrolysis of substrate mimetics vs. calculated binding energies ∆E of
the corresponding productive complexes.
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2, 1 4 4 2 – 1 4 4 6
1444

View Article Online
used, a good correlation between the calculated binding
energies and the experimentally determined specificity con-
stants has been found in almost all cases. Provided that the
three-dimensional structure of a suitable protease is available,
this computational method therefore enables a rational screen-
ing of de-novo designed substrate mimetics prior to their syn-
thesis, which finally facilitates the identification of novel, and
highly specific substrate mimetics significantly. Already within
this initial study, Boc-Ala-OCap has been identified as an
efficient and even slightly higher specific substrate mimetic
for chymotrypsin than the previously used OGp ester. This is
a result that clearly proves the potential of this approach.
Broader studies with larger substrate mimetic libraries might
lead to further promising candidates with even higher specific-
ities. Studies in this direction including also a larger number of
proteases are presently under investigation. Further efforts,
however, are necessary to optimize the docking calculations in
particular their capability to consider the degree of ester bond
activation. Differences in the latter might be responsible for
the non-correct prediction of the specificity of chymotrypsin
towards Boc-Ala-OInd. At the present stage of development,
best results can be expected for ester moieties of similar basic
structures varying mainly in the substitution pattern.
Fig.
4
Plot of specificity constants of chymotrypsin-catalysed
hydrolysis of substrate mimetics vs. calculated binding energies of
the corresponding productive complexes. The correlation coefficient is
R = 0.93; the confidence interval of 95% is plotted as dotted lines.
Correlation does not include Boc-Ala-OInd (cf. text).
conclusions as to the suitability of the docking approach for
predicting the specificity of a given substrate mimetic in a
quantitative manner. Analysis of Fig. 3 shows that the resulting
data points in the case of subtilisin are almost located within
the same area. Based on similar ∆E values, which would ideally
correlate with similar values for kcat KMϪ1, this finding matches
exactly the prediction made by the docking approach for this
enzyme. In general, a good correlation between the two values
is also found for chymotrypsin (cf. Fig. 4). In this case, an
increase in the binding energies comes along with an increase in
the specificity constants. Linear regression of the data results in
a correlation coefficient R of 0.93 with almost all the data
located within a confidence interval of 95%. An exception was
found, however, for Boc-Ala-OInd whose specificity constant is
slightly lower than expected from the theoretical studies. As
already pointed out, the applied computational docking
approach relies on the calculation of binding energies and, thus,
provides primarily information on the affinity of ligand bind-
ing. Assuming that the binding energy is calculated correctly
and that the rate of deacylation is equal for all esters due to
their identical acyl residues, this discrepancy should be caused
by differences in the rate of acylation as the third remaining
step of the reaction. Following this hypothesis, the lower speci-
ficity of chymotrypsin for hydrolysing Boc-Ala-OInd could be
therefore the result of a reduced rate of acylation compared to
the other ester types. Although this was apparently not the case
for subtilisin, the reason for this different acylation behaviour
might be based on the unique chemical structure of the indoyl
ester. Among the substrate mimetics used, Boc-Ala-OInd is a
bicyclic ring system while all other esters are derived from
monocyclic phenyl or benzyl moieties. Due to the high electron
density of the heterocyclic indoyl moiety a stabilising effect on
the scissile ester bond can be expected which would lower the
degree of ester bond activation and make the ester hydrolysis
more demanding. The exclusion of Boc-Ala-OInd from the
Experimental
General
Boc-Ala-OH, coupling reagents, ester components, enzymes
and buffer components were products of Bachem, Sigma,
Fluka (Switzerland) or Merck (Germany). All reagents were
of the highest purity commercially available. Solvents were
purified and dried by the usual methods.
Molecular modelling and automated docking
The protein ligand docking studies were performed on the basis
of the crystal structures of α-chymotrypsin complexed with
phenylethane boronic acid (pdb entry 6cha20) and subtilisin
(pdb-entry 1svn21). All solvent molecules, ions and the
cocrystalised inhibitor were removed from the structures. Polar
hydrogen atoms were added using the modelling package
Quanta98 (Accelrys Inc., San Diego CA, 1998). After assign-
ment of template charges the protein structures were subjected
to a short geometry relaxation (500 steps steepest descent,
CHARMm23 force field22). The various substrate esters were
modelled in the same way, but with charges derived from their
electrostatic potential, which were obtained from single point
calculations at the AM1 level23 of MO theory using the
program Spartan 5.0 (Wavefunction Inc., Irvine CA, 1998).
For the docking calculations, the program package Auto-
Dock 3.0 was employed. To allow the ligand to search the entire
active site, a cubic box of 30 × 30 × 30 Å was centred on the
γ-oxygen atom of the catalytic active serine (serine 195 for
chymotrypsin and serine 221 for subtilisin). The search for the
possible ligand binding sites was performed within this box at
the default grid point spacing of 0.375 Å. All relevant torsion
angles in the ester ligands were released. Each single docking
experiment consisted of 30 runs employing a Lamarckian
genetic algorithm and took a computing time of about 15 min
on a standard Pentium-4 system to be finished. The docking
runs were initiated with 50 randomly chosen protein–ligand
arrangements and iterated through 5 × 106 energy evaluations.
Up to 50 iterations of local search were applied with a fre-
quency of 0.1. Thus, approximately 150 million protein–ligand
arrangements were evaluated per substrate and enzyme. All
other docking parameters were set to their default values. To
ensure correct calculation of the free binding energy, aromatic
carbon atoms were treated with an additional solvation term, as
it is incorporated in the AutoDock program. The resulting
protease–substrate complexes were clustered employing an
Ϫ1
linear regression of the ∆E and kcat KM values results, how-
ever, in the aforementioned good correlation coefficient
whereas Boc-Ala-OCap shows the highest binding energy and
specificity constant as well.
Conclusions
The results of our studies show that automated docking
methods based on the calculation of binding energies can be
successfully applied to predict the general acceptance of a
given substrate mimetic by proteases. Besides this qualitative
prediction, the approach further allows for a quantitative
estimation of the individual specificity of a potential substrate
mimetic. Independent of the enzyme and substrate mimetic
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2, 1 4 4 2 – 1 4 4 6
1445

View Article Online
rmsd tolerance of 1 Å. Only clusters representing productive
ligand–protein complexes were further analysed, which were in
most cases the highest populated cluster or the cluster with the
lowest binding energy, respectively.
at flow rates of 1.0 ml minϪ1. Detection was at 254 nm. The
reaction rates were calculated from peak areas of the ester sub-
strates using 4-toluenesulfonic acid as an internal standard.
Acknowledgements
Chemical syntheses
This work has been supported by the Deutsche Forschungs-
gemeinschaft (Sonderforschungsbereich 610 “Proteinzustände
mit zellbiologischer und medizinischer Relevanz” A3 and
Z3 and BO 1770/1-1) and Fonds der Chemischen Industrie.
The technical assistance of Regina Reppich is gratefully
acknowledged.
Boc-Ala-OGp was prepared according to our previously
described protocols.10 Similarly, Boc-Ala-OPh and Boc-Ala-
OPic were synthesized by DCC coupling of the appropriate
Boc-protected amino acid with phenol and 4-hydroxymethyl
pyridine, respectively. Boc-Ala-OCap and Boc-Ala-OCp were
prepared by coupling the ester bond using TBTU. The final
amino acid esters were purified by preparative HPLC. The
identity and purity of all final products were checked by
analytical HPLC at 220 nm, NMR, thermospray mass spectro-
scopy, and elemental analysis (analytical data see ref. 24). In
all cases, satisfactory analytical data were found ( 0.4% for
C, H, N).
References
1 F. Bordusa, Chem. Rev., 2002, 102, 4817.
2 F. Bordusa, D. Ullmann, C. Elsner and H.-D. Jakubke, Angew.
Chem., 1997, 109, 1125 (Angew. Chem. Int. Ed. Engl., 1997, 36,
2473).
3 N. Wehofsky, M. Alisch and F. Bordusa, Chem. Commun., 2001,
1602.
4 N. Wehofsky, S. Thust, J. Burmeister, S. Klussmann and F. Bordusa,
Angew. Chem., 2003, 115, 701 (Angew. Chem. Int. Ed., 2003, 42,
677).
5 R. Günther, A. Stein and F. Bordusa, J. Org. Chem., 2000, 65, 1672.
6 N. Wehofsky, R. Löser, A. Buchynskyy, P. Welzel and F. Bordusa,
Angew. Chem., 2002, 114, 2895 (Angew. Chem. Int. Ed., 2002, 41,
2735).
7 R. Günther and F. Bordusa, Chem. Eur. J., 2000, 6, 463.
8 N. Wehofsky, N. Koglin, S. Thust and F. Bordusa, J. Am. Chem.
Soc., 2003, 125, 6126.
9 I. Schechter and A. Berger, Biochem. Biophys. Res. Commun., 1967,
27, 157.
10 M. Thormann, S. Thust, H.-J. Hofmann and F. Bordusa,
Biochemistry, 1999, 38, 6056.
11 N. Wehofsky and F. Bordusa, FEBS Lett., 1999, 443, 220.
12 N. Wehofsky, J.-D. Wissmann, M. Alisch and F. Bordusa, Biochim.
Biophys. Acta, 2000, 1479, 114.
13 H. Sekizaki, K. Itoh, E. Toyota and K. Tanizawa, Chem. Pharm.
Bull., 1996, 44, 1577.
14 H. M. Berman, J. Westbrook, Z. Feng, G. Gilliland, T. N. Bhat,
H. Weissig, I. N. Shindyalov and P. E. Bourne, Nucleic Acids Res.,
2000, 28, 235.
15 J. J. Perona and C. S. Craik, Protein Sci., 1995, 4, 337.
16 V. Cerovsky and F. Bordusa, J. Peptide Res., 2000, 55, 325.
17 G. M. Morris, D. S. Goodsell, R. S. Halliday, R. Huey, W. E. Hart,
R. K. Belew and A. J. Olson, J. Comput. Chem., 1998, 19, 1639.
18 V. Cerovsky, J. Kockskämper, H. G. Glietsch and F. Bordusa,
ChemBioChem, 2000, 2, 126.
Enzyme kinetics
Hydrolysis reactions were performed in a total volume of 450 µl
at 25 ЊC. Stock solutions of Boc-Ala-OY esters were prepared
in 0.1 M Hepes buffer (pH 8.0), 0.1 M NaCl, and 10 mM CaCl2
containing 8% (v/v) DMF as cosolvent. The substrate con-
centrations were between 0.1 and 9.0 mM and the enzyme
concentrations between 6.4 × 10Ϫ8 and 3.2 × 10Ϫ6 M. The active
enzyme concentration was determined by active-site titration
using 4-nitrophenyl acetate.25 After thermal equilibration of
assay mixtures, the reactions were initiated by addition of the
appropriate enzyme stock solutions. Subsequently, the mixtures
were rapidly shaken and transferred into a thermomixer
adjusted to 25 ЊC. The reaction rates were analyzed by
RP-HPLC determining the disappearance of the substrate
esters by at least ten different concentrations. For this purpose,
50 µl aliquots were withdrawn at defined time intervals and
diluted with a quenching solution of 50% methanol containing
5% TFA. As a control for the spontaneous hydrolysis of the
acyl donor esters, parallel reactions without enzyme were
analyzed. On the basis of these control reactions the extent
of spontaneous hydrolysis was found to be less than 5%. The
kinetic parameters were calculated by iterative nonlinear curve
fitting of the untransformed data using the program SigmaPlot
Scientific Graphic System (Version 6.10, SPSS Inc.). The
data reported are the average of at least three independent
experiments.
19 A. Nicholls, K. A. Sharp and B. Honig, Proteins: Struct., Funct.,
Genet., 1991, 11, 281.
20 A. Tulinsky and R. A. Blevins, J. Biol. Chem., 1987, 262, 7737.
21 C. Betzel, S. Klupsch, G. Papendorf, S. Hastrup, S. Branner and
K. S. Wilson, J. Mol. Biol., 1992, 223, 427.
HPLC analyses
22 F. A. Momany and R. Rone, J. Comput. Chem., 1992, 13, 888.
23 M. J. S. Dewar, E. G. Zoebisch, E. F. Healy and J. J. P. Steward,
J. Am. Chem. Soc., 1985, 107, 3902.
24 K. Rall and F. Bordusa, J. Org. Chem., 2002, 67, 9103.
25 H. U. Bergmeyer, M. Graßl and H.-E. Walter, in Methods of
Enzymatic Analysis, ed. H. U. Bergmeyer, Verlag Chemie, Weinheim,
1983, vol 2, p. 319.
HPLC measurements were performed with
a Shimadzu
LC-10A HPLC system using a LiChrospher RP 18 column
(250 mm × 4 mm, 5 µm, Merck, Germany) or a Capcell PAK
C8 column (250 mm × 4 mm, 5 µm, Shiseido, Japan). Samples
were eluted with various mixtures of water and acetonitrile
containing 0.1% TFA under isocratic and gradient conditions
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2, 1 4 4 2 – 1 4 4 6
1446