Probing the abilities of synthetically useful serine proteases to discriminate between the configurations of remote stereocenters using chiral aldehyde inhibitors

Article abstract of DOI:10.1021/ja952835t

The abilities of the synthetically useful serine proteases, subtilisin Carlsberg (SC) and α-chymotrypsin (CT), to discriminate between R- and S-configurations of stereocenters remote from the catalytic site have been explored using chiral aldehyde transition state analog inhibitors as probes. The inhibitors evaluated were (R)- and (S)-3-phenylbutanal and (R)- and (S)-4-phenylpentanal, for which the stereocenters at C-3 and C-4 respectively are distant from the aldehyde functionality that interacts with the catalytic serine residue. The achiral parent compounds, 3-phenylpropanal and 4-phenylbutanal, respectively, were also assessed for reference purposes. Each aldehyde was found to be a competitive inhibitor for both enzymes, with CT being significantly more potently inhibited than SC. Within this series, the presence of an R-center methyl group improved binding significantly over that of the achiral parent aldehyde for both enzymes. In contrast, the effects on binding of S-methyl substituents in the same positions were modest, and generally somewhat deleterious. Furthermore, the greater the separation of the stereocenter from the aldehyde group, the lower the degree of configuration discrimination. The most effective inhibition, and the highest degree of remote stereocenter discrimination, observed was that by CT of (R)-3-phenylbutanal, whose K(I) of 8.4 μM was 61-fold lower than that of its achiral parent 3-phenylpropanal, and 88-fold lower than the K(I) of its S-enantiomer. Molecular mechanics and molecular dynamics calculations were performed to identify each favored aldehyde-enzyme complex and to reveal the binding and orientation differences responsible for the R- and S-enantiomer binding discriminations observed.

Full text of DOI:10.1021/ja952835t

502
J. Am. Chem. Soc. 1996, 118, 502-508
Probing the Abilities of Synthetically Useful Serine Proteases
To Discriminate between the Configurations of Remote
Stereocenters Using Chiral Aldehyde Inhibitors
Taekyu Lee and J. Bryan Jones*
Contribution from the Department of Chemistry, UniVersity of Toronto, 80 St. George Street,
Toronto, Ontario, Canada M5S 1A1
ReceiVed August 17, 1995X
Abstract: The abilities of the synthetically useful serine proteases, subtilisin Carlsberg (SC) and R-chymotrypsin
(CT), to discriminate between R- and S-configurations of stereocenters remote from the catalytic site have been
explored using chiral aldehyde transition state analog inhibitors as probes. The inhibitors evaluated were (R)- and
(S)-3-phenylbutanal and (R)- and (S)-4-phenylpentanal, for which the stereocenters at C-3 and C-4 respectively are
distant from the aldehyde functionality that interacts with the catalytic serine residue. The achiral parent compounds,
3-phenylpropanal and 4-phenylbutanal, respectively, were also assessed for reference purposes. Each aldehyde was
found to be a competitive inhibitor for both enzymes, with CT being significantly more potently inhibited than SC.
Within this series, the presence of an R-center methyl group improved binding significantly over that of the achiral
parent aldehyde for both enzymes. In contrast, the effects on binding of S-methyl substituents in the same positions
were modest, and generally somewhat deleterious. Furthermore, the greater the separation of the stereocenter from
the aldehyde group, the lower the degree of configuration discrimination. The most effective inhibition, and the
highest degree of remote stereocenter discrimination, observed was that by CT of (R)-3-phenylbutanal, whose KI of
8.4 µM was 61-fold lower than that of its achiral parent 3-phenylpropanal, and 88-fold lower than the KI of its
S-enantiomer. Molecular mechanics and molecular dynamics calculations were performed to identify each favored
aldehyde-enzyme complex and to reveal the binding and orientation differences responsible for the R- and
S-enantiomer binding discriminations observed.
Introduction
it is not clear whether the meagerness of the database in this
regard is due to the fact that enzymatic discrimination of a
remote stereocenter is intrinsically difficult or simply that too
few examples have yet been sought. These questions cannot
be answered at this time because our knowledge of the factors
controlling and determining enzyme stereospecificity are still
poorly understood. Accordingly, for full advantage to be taken
of the synthetic potential that enzymes offer, it is important to
identify and to extend our understanding of the factors control-
ling substrate and inhibitor binding, and it is toward this goal
that the present study is targeted.
With synthetic applications of esterases and related hydrolytic
enzymes being of such widespread current interest,1 two serine
proteases, subtilisin Carlsberg (SC, EC 3.4.21.14) and R-chy-
motrypsin (CT, EC 3.4.21.1), were selected as representative
hydrolases for our initial studies on remote stereocenter ste-
reospecificity. SC and CT are commercially available enzymes
that have been applied in a wide range of synthetic transforma-
tions4 and for which high-resolution X-ray crystal structures are
Enzymes have now gained general acceptance as chiral
catalysts in asymmetric synthesis.1 In almost all of these cases,
the stereocenter being introduced or selected is adjacent to the
site of catalysis. Very few examples have been reported where
the stereocenter of interest is three or more bonds removed from
the carbonyl group of the ester function undergoing hydrolysis.2
The paucity of examples of stereocenter control remote from
enzyme’s catalytic sites parallels the situation in non-enzyme
catalyzed asymmetric synthesis, where control of the configura-
tions of stereocenters remote from a chiral auxiliary or catalyst
represents a major problem that has not yet been solved.3
However, in enzymatic catalysis, since the whole of an enzyme’s
active site region that envelopes a substrate in the enzyme-
substrate (ES) complex is chiral, discrimination of any substrate
stereocenter is feasible in principle, no matter how remotely
such a stereocenter is located from the catalytic site. At present,
X Abstract published in AdVance ACS Abstracts, December 15, 1995.
(1) (a) Wong, C.-H., Whitesides, G. M. Enzymes in Synthetic Organic
Chemistry; Pergamon: New York, 1994. (b) PreparatiVe Biotransforma-
tions; Roberts, S. M., Ed.; Wiley: New York, 1993. (c) Biotransformations
in PreparatiVe Organic Chemistry; Faber, K., Ed.; Springer-Verlag:
Heidelberg, 1992.
(2) (a) Hughes, D. L.; Bergan, J. J.; Amato, J. S.; Bhupathy, M.; Leazer,
J. L.; McNamara, J. M.; Sidler, D. R.; Reider, P. J.; Grabowski, E. J. J. J.
Org. Chem. 1990, 55, 6252. (b) Ladner, W. E.; Whitesides, G. M. J. Am.
Chem. Soc. 1984, 106, 7250. (c) Jones, J. B.; Marr, P. W. Tetrahedron
Lett. 1973, 3165. (d) Mohr, P.; Waespe-Sarcevic, N.; Tamm, C.; Gawronska,
K.; Gawronski, J. K. HelV. Chim. Acta 1983, 66, 2501. (e) Cohen, S. G.;
Weinstein, S. Y. J. Am. Chem. Soc. 1964, 86, 725.
(3) (a) Ojima, I. Catalytic Asymmetric Synthesis; VCH: New York, 1993.
(b) Noyori, R. Asymmetric Catalysis in Organic Synthesis; Wiley: New
York, 1994. (c) Koskinen, A. Asymmetric Synthesis of Natural Products;
Wiley: New York, 1993. (d) Hegedus, L. Transition Metals in the Synthesis
of Complex Organic Molecules; S. University Science Books: Mill Valley,
CA, 1994.
(4) (a) Delinck, D. L.; Margolin, A. L. Tetrahedron Lett. 1990, 31, 3093.
(b) Pugniere, M.; San Juan, C.; Previero, A. Tetrahedron Lett. 1990, 31,
4883. (c) Gotor, V.; Garcia, M. J.; Rebelleo, F. Tetrahedron Asymmetry
1990, 1, 277. (d) Margolin, A. L.; Delinck, D. L.; Whalon, M. R. J. Am.
Chem. Soc. 1990, 112, 2849. (e) Chenevert, R.; Desjardins, M.; Gagnon,
R. Chem. Lett. 1990, 33. (f) Chenevert, R.; Letourneau, M.; Thiboutot, S.
Can. J. Chem. 1990, 960. (g) Frigerio, F.; Coda, A.; Pugliese, L.; Lionetti,
C.; Menegatti, E.; Amiconi, G.; Schnebli, H. P.; Ascenzi, P.; Bolognesi,
M. J. Mol. Biol. 1992, 225, 107. (h) Tsukada, H.; Blow, D. M. J. Mol.
Biol. 1985, 184, 703. (i) Birktoft, J. J.; Blow, D. M. J. Mol. Biol. 1972, 68,
187. (j) Blevins, R. A.; Tulinsky, A. J. Biol. Chem. 1985, 260, 4264. (k)
Ricca, J. M.; Crout, D. H. J. Chem. Soc., Perkin Trans. 1 1989, 2126. (l)
Kitaguchi, H.; Fitzpatrick, P. A.; Huber, J. E.; Klibanov, A. M. J. Am. Chem.
Soc. 1989, 111, 3094. (m) Brieva, R.; Rebelleo, F.; Gotor, V. Chem. Comm.
1990, 1386. (n) Gutman, A. L.; Meyer, E.; Kalerin, E.; Polyak, F.; Sterling,
J. Biotechnol. Bioeng. 1992, 40, 760. (o) Roper, J. M.; Bauer, D. P. Synthesis
1983, 1041.
0002-7863/96/1518-0502$12.00/0 © 1996 American Chemical Society

Discrimination of R and S with Serine Proteases
J. Am. Chem. Soc., Vol. 118, No. 3, 1996 503
Table 1. Inhibition of Subtilisin Carlsberg and a-Chymotrypsin by
Aldehydesa
Figure 1. Schematic representation of the EI complex formed when
a transition state analog aldehyde inhibitor such as C6H5CH2CH2CHO
(1) binds at the active site of subtilisin Carlsberg. Nucleophilic attack
of the carbonyl group of 1 by the side chain OH of the serine residue
of the catalytic triad Ser221-His64-Asp32 gives the tetrahedral,
transition state-like, structure shown, whose anion is stabilized by the
amide functions of Asn155 and Ser221 that comprise the “oxyanion
hole”. The phenethyl group is located in the non-polar S1-pocket, in
accord with the dominance of the structural specificity of the enzyme
by such hydrophobic interactions. S2 is where the peptide chain binds
and S1′ is where the alcohol moiety of ester binds.
available.5 These serine proteases have an extended active site
6
binding region composed of several subsites, of which the S1
-pocket dominates, particularly in the binding of hydrophobic
groups.
a KI values for both enzymes were determined12 in duplicate at pH
7.5 at 25 °C in 0.1 M NaH2PO4, 0.5 M NaCl, 0.25 mM of substrate,
5% DMSO in volume, inhibitor concentrations in the range of 5.9-
9.1 mM, and enzyme concentrations 4.5 nM (SC) and 16 nM (CT).
b Data obtained9a at 25 °C and pH 7.7 ( 0.3.
For systematic probing of enzyme specificity, evaluating the
binding affinities of transition state analog competitive inhibi-
tors7 represents a convenient strategy, and kinetic studies on
boronic acid inhibitors of this type, coupled with graphics
analyses and molecular modeling, have been successfully used
for the systematic probing of the structural and electrostatic
specificity of the S1-site of SC.8 This strategy is continued in
this paper, using aldehydes as inhibitors. Aldehydes are well
known to be transition state analog competitive inhibitors of
serine proteases,9 as illustrated in Figure 1. In this study, the
enantiomeric aldehydes (R)- and (S)-2,4, with remote stereo-
centers positioned either â or γ to the aldehyde, together with
the achiral parent structures 1,3 as reference compounds, have
been used to evaluate the remote stereocenter stereoselectivity
potential of the S1 pockets of SC and CT and to further probe
the structural specificity of these sites.
Results and Discussion
The aldehyde structures 1-4 were selected as suitable
structures for S1-site interaction on the basis of graphics analyses
using the X-ray structures of SC and CT.5 The starting material
for the chiral aldehydes was commercially available (()-3-
phenylbutanoic acid, which was resolved into its R- and
S-enantiomers by recrystallizations of its diastereomeric salts
with (R)- and (S)-R-phenethylamines. The methyl esters of the
resolved acids were then reduced and the alcohols obtained
oxidized, using the Dess-Martin reagent,10 to the corresponding
aldehydes (R)- and (S)-2. The homologous aldehydes (R)- and
(S)-4 were obtained similarly from the enantiomers of methyl
4-phenylpentanoate obtained Via Arndt-Eistert homologations11
of the above (R)- and (S)-3-phenylbutanoic acids. The achiral
reference inhibitor 3-phenylpropanal (1) was commercially
available, while 4-phenylbutanal (3) was prepared from pur-
chased 4-phenylbutanol by Dess-Martin oxidation.
The inhibitory effects of each aldehyde on SC and CT were
evaluated by the method of Waley,12 using Suc-Ala-Ala-Pro-
Phe-PNA as substrate. The results, which showed that each
aldehyde was a competitive inhibitor of both enzymes, are
summarized in Table 1.
Each of the aldehydes was a significantly more potent
inhibitor of CT than of SC, generally by about an order of
magnitude, but by almost two orders of magnitude with (R)-2
as the inhibitor. Furthermore, within each homologous series,
binding of an R-enantiomer is stronger than for either the
(5) (a) McPhalen, C. A.; James, M. N. G. Biochemistry 1988, 27, 6582.
(b) Bode, W.; Papamokos, E.; Musil, D. Eur. J. Biochem. 1987, 166, 673.
(c) Frigerio, F.; Coda, A.; Pugliese, L.; Lionetti, C.; Menegatti, E.; Amiconi,
G.; Schnebli, H. P.; Ascenzi, P.; Bolognesi, M. J. Mol. Biol. 1992, 225,
107. (d) Tsukada, H.; Blow, D. M. M J. Mol. Biol. 1985, 184, 703. (e)
Birktoft, J. J.; Blow, D. M. M. J. Mol. Biol. 1972, 68, 187. (f) Blevins, R.
A.; Tulinsky, A. J. Biol. Chem. 1985, 260, 4264. (g)Tulinsky, A.; Blevins,
R. A. J. Biol. Chem. 1987, 262, 7737.
(6) (a) Schechter, I.; Berger, A. Biochem. Biophys. Res. Commun. 1967,
27, 157. (B) Berger, A.; Schechter, I. Philos. Trans. R. London, Ser. B
1970, 257, 249.
(7) (a) Lienhard, G. E. Science 1973, 180, 149. (b) Wolfenden, R. V.
Acc. Chem. Res. 1972, 5, 10. (c) Koehler, K. A.; Lienhard, G. E.
Biochemistry 1971, 10, 2477. (d) Wolfenden, R. V.; Radzicka, A. Curr.
Opin. Struct. Biol. 1991, 1, 780.
(9) (a) Rawn, J. D.; Lienhard, G. E. Biochemistry 1974, 13, 3124. (b)
Tatsuta, K.; Mikami, N.; Fujimoto, K.; Umezawa, S.; Umezawa, H.; Aoyagi,
T. J. Antibiot. (Tokyo), Ser. A 1973, 26, 625. (c) Ito, A.; Tokawa, K.;
Shimizu, B. Biochem. Biophys., Res. Commun. 1972, 49, 343. (d) Westmark,
P. R.; Kelly, J. P.; Smith, B. R. J. Am. Chem. Soc. 1993, 115, 3416.
(10) Dess, D. B.; Martin, J. C. J. Org. Chem. 1983, 48, 4155.
(11) (a) Arndt, F.; Eistert, B. Ber. 1935, 68, 200. (b) Bachmann, W. E.;
Struve, W. S. In Organic Reactions; (Florkin, M., Stotz, E. H., Eds.); Wiley
and Sons Inc.: New York, 1965, Vol. 1, pp 38-62.
(8) (a) Keller, T. H.; Seufer-Wasserthal, P.; Jones, J. B. Biochem. Biophys.
Res. Commun. 1991, 176, 401. (b) Seufer-Wasserthal, P.; Martichonok, V.;
Keller, T. H.; Chin, B.; Martin, R.; Jones, J. B. Bioorg. Med. Chem. 1994,
2, 35.
(12) Waley, S. G. J. Biochem. 1982, 205, 631.

504 J. Am. Chem. Soc., Vol. 118, No. 3, 1996
Lee and Jones
corresponding S-enantiomer or its achiral parent, but with the
variations being far greater for CT inhibition than for SC. In
part, this reflects the generally weaker interactions of the
inhibitors with SC. The strongest inhibition observed was of
CT by (R)-3-phenylbutanal ((R)-2), whose KI of 8.4 µM is 61-
fold lower than that of its achiral precursor, 3-phenylpropanal
(1), and 88-fold lower than that of its S-enantiomer. This 88-
fold difference in CT-binding affinities between (R)- and (S)-
2, with its unequivocal demonstration that significant discrimi-
nation of remote stereocenters is achievable using enzymes, is
an extremely encouraging result. For example, if differences
in binding of this magnitude were realized between diastereo-
meric ES complexes during catalysis of a remote stereocenter-
containing racemic substrate, very high resolution efficiencies
would be anticipated and the transformation would be of true
asymmetric synthetic value. Even the 17.5- and 12.6-fold
differences manifest between (R)- and (S)-2 with SC and (R)-
and (S)-4 with CT, respectively, would translate into asymmetric
synthetically useful distinctions for analogous remote stereo-
center substrates if the overall rates of hydrolysis reflected these
levels of binding disparity. Indeed, if in these cases kcat
variations for the enantiomeric or diastereomeric substrate
transformations reinforced the binding differences to even a
small degree, the degree of remote stereoselectivity discrimina-
tion achieved would become very exploitable synthetically.
Figure 2. Superimposed energy-minimized EI complexes of (R)- and
(S)-3-phenylbutanal, (R)-2 (heavy solid line) and (S)-2 (heavy dashed
line), respectively, in the active site of CT. The oxyanions of the
tetrahedral complexes derived from (R)-2 and (S)-2 are located in the
oxyanion hole, with the negative charges stabilized by hydrogen
bonding (light dashed line) with the peptide NH’s of Ser195 (O--N
) 3.18 and 3.49 Å respectively for (R)-2 and (S)-2) and Gly193 (O--
N ) 2.72 and 2.79 Å respectively for (R)-2 and (S)-2). Both phenyl
groups are positioned in the hydrophobic S1 pocket. The main difference
between the two orientations is that the methyl group at the stereocenter
of (R)-2 (solid arrow) is favorably located in the S2 pocket, while that
of (S)-2 (dashed arrow) is oriented toward the outside of the active
site and does not contribute to binding.
The questions posed by the observed differences in binding
affinities of the inhibitors 1-4 are intriguing. Both SC and
CT appear to be largely indifferent to the presence of a methyl
substituent that creates a remote stereocenter of S-configuration,
with the KI’s of (S)-2 and (S)-4 with SC, and of (S)-2 with CT,
being so similar to those of their achiral analogs 1 and 3,
respectively. Indeed, for (S)-4 with CT, the presence of the
methyl group is a somewhat negative factor, with the KI of the
unsubstituted parent 3 being 5.7-fold lower. In contrast, an
R-center methyl substituent is highly advantageous to good
binding, particularly for (R)-3-phenylbutanal, which binds 61-
fold and 16-fold more strongly with CT and SC, respectively,
than does the unsubstituted inhibitor 1. The same trend is again
evident for the homologous aldehyde (R)-4, but to a lesser
degree, especially for CT where the beneficial effect of the (R)-
methyl group on binding is highly subdued.
Figure 3. Superimposed energy-minimized EI complexes of (R)- and
(S)-4-phenylbutanal, (R)-4 (heavy solid line) and (S)-4 (heavy dashed
line), respectively, in the active site of CT. The oxyanions of the
tetrahedral complexes derived from (R)-4 and (S)-4 are located in the
oxyanion hole, with the negative charges stabilized by hydrogen
bonding (light dashed line) with the peptide NH’s of Ser195 (O--N
) 3.00 and 3.15 Å respectively for (R)-4 and (S)-4) and Gly193 (O--
N ) 2.91 and 2.84 Å respectively for (R)-4 and (S)-4. The stereocenter
methyl groups of both (R)-4 (solid arrow) and (S)-4 (dashed arrow)
are obliged to locate outside of the active site and are not binding
contributors. The main differences are seen to be in the locations of
the phenyl groups, with that of (R)-4 being positioned more deeply in
the hydrophobic S1 site than that of (S)-4.
Molecular modeling was applied in order to interpret the
kinetic data more completely. The X-ray structures of SC and
CT were energy minimized by molecular mechanics and
molecular dynamics. Each aldehyde inhibitor was then docked
into the active site, with the phenyl moieties in S1, and the
aldehyde carbonyl carbon covalently connected to the active
site serine-CH2OH oxygen to form the transition state-like
tetrahedral intermediate. Each EI complex was then subjected
to energy minimization, by molecular mechanics, followed by
molecular dynamics, calculations, and the optimized conforma-
tions and the free energies of each R- and S-pair of EI complexes
were compared. The results for the chiral inhibitors are depicted
in Figures 2-5. The achiral inhibitors also occupied very
similar active site positions. In each minimized EI complex,
there were strong interactions between the oxyanion of the
tetrahedral intermediate and the oxyanion hole H-bonding
residues of the peptide backbone NH’s of Ser195 and Gly193
of CT, and the backbone NH of Ser221 and the side chain NH2
of Asn155 of SC, respectively.
corresponding methyl group of the S-enantiomer is located
outside of the active site of the enzyme and is oriented toward
the external water in a manner that does not contribute to
hydrophobic bonding. In addition, the phenyl group of the
R-enantiomer is able to locate itself deeper inside the hydro-
phobic S1 pocket than can its enantiomeric equivalent, thereby
providing a stronger phenyl-S1 contribution to binding. It is
the combination of these two factors that accounts for the
dramatic 88-fold difference in KI’s between (R)-2 and (S)-2 as
inhibitors of CT. In the cases of the homologous inhibitors,
(R)- and (S)-4-phenylpentanal ((R)- and (S)-4), the KI differences
between inhibition of CT by the enantiomers are much smaller
(12.6-fold). This also is accounted for by the molecular
modeling results (Figure 3). The superimposed structures of
(R)-4 and (S)-4 in the active site of CT reveal that the methyl
The superimposed conformations of (R)- and (S)-3-phenylbu-
tanal ((R)- and (S)-2) in the active site of CT (Figure 2) show
that the methyl group at the stereocenter of the R-enantiomer
is positioned in the S2 pocket such as to elicit favorable
hydrophobic interactions with the enzyme. In contrast, the

Discrimination of R and S with Serine Proteases
J. Am. Chem. Soc., Vol. 118, No. 3, 1996 505
The molecular modeling results on the SC inhibitors are
equally enlightening. The superimposed structures of (R)- and
(S)-3-phenylbutanal ((R)- and (S)-2) in their EI-complexes with
SC are depicted in Figure 4. In both complexes, the phenyl
residues of both (R)-2 and (S)-2 penetrate adequately, and almost
equivalently, into S1 to provide good hydrophobic binding
contributions to the corresponding KI’s. For the (R)-2 orienta-
tion, the methyl group at the stereocenter is oriented toward
the outside of the active site and does not make any binding
contribution. In contrast to this neutral binding role, the position
that the (S)-2 methyl group is obliged to occupy elicits
unfavorable steric interactions of the methyl group with the
Asn155 oxyanion hole residue. As a result, the usual oxyanion
stabilization mechanism is disturbed. For example, the distances
between the oxyanion of the S-enantiomer and the backbone N
and NH of Ser 221 have become 3.95 and 3.15 Å, respectively,
which precludes normal oxyanion hydrogen bonding. Thus the
EI complex of (R)-2 with SC is relatively favored over that of
(S)-2 by the absence of negative interactions rather than, as in
the case of CT (Figure 2), of a beneficial contribution by the
methyl substituent. For the interactions of the higher homologue
inhibitors (R)- and (S)-4-phenylpentanal ((R)- and (S)-4) with
SC, the minor (4.8-fold) KI variations and the relatively weak
binding of each enantiomer indicate that there should be little
difference in their orientations in the active site, and that there
are no strong EI interactions. This is confirmed by the
minimized complexes of Figure 5, in which the phenyl groups
of both (R)- and (S)-4 are similarly, but not deeply, located in
S1, and with the methyl groups at the stereocenter both oriented
away from the active site into locations that influence binding
neither favorably nor adversely.
Figure 4. Superimposed energy-minimized EI complexes of (R)- and
(S)-3-phenylbutanal, (R)-2 (heavy solid line) and (S)-2 (heavy dashed
line), respectively, in the active site of SC. Both phenyl residues bind
in S1. The oxyanions of the tetrahedral complexes derived from (R)-2
and (S)-2 are located in the oxyanion hole, with the negative charges
stabilized by hydrogen bonding (light dashed line) with the peptide
NH of Ser221 and the side chain -NH2 of Asn155 for (R)-2 (O--N
) 2.95 and 2.72 Å respectively), but less well for (S)-2, for which
only hydrogen bonding with Asn155 is seen (O--N ) 2.74 Å). The
stereocenter methyl group of (R)-2 (solid arrow) is oriented outside of
the active site where its binding influence is neutral. In contrast, that
of (S)-2 (dashed arrow) takes up a position close to Asn155 and is
responsible for the disruption in the oxyanion stabilization, and hence
the weak overall inhibitory properties of (S)-2.
A comparison of ∆∆Gq values for the inhibition of CT and
SC provides further support for the validity of the molecular
modeling analyses of the weaker R- and S-stereocenter dis-
crimination for structures 4 than for 2. The differences in
experimental (from KI’s) and calculated (from EI-complex
energies) ∆∆Gq’s for the R- and S-pairs of 2 and 4 are similar,
each being 50-70% lower for (R)- and (S)-4 than for (R)- and
(S)-2 for both CT and SC.
While the current results represent only a first step toward
probing remote stereocenter stereoselectivity of enzymes, the
strategy of using competitive inhibitors is clearly an effective
one. The observation that the stereocenters of 2 and 4 become
less well recognized the more distant they are from the site of
reaction is in accord with the established trend in asymmetric
catalysis generally.3 Thus in attempting to establish enzymatic
control of the configurations of remote stereocenters, the natural
proclivity of enzymes to manifest maximum stereoselectivity
close to the catalytic site will have to be conquered. Another
of our ultimate goals is to formulate guidelines for forecasting
the degrees of stereocenter discrimination to be expected in the
transformations of any new, unnatural substrate structures
catalyzed by synthetically useful enzymes. In this regard, the
molecular modeling results are encouraging in that the differ-
ences in calculated binding energies for enantiomeric pairs
studied here are proportional to the actual differences in KI’s.
If this promise holds out, it will be possible to develop protocols
for rapidly screening remote-stereocenter, synthon-precursor,
structures of interest, using calculations of their relative energies
of binding to appropriate enzymes as a basis for predicting
whether significant stereocenter discrimination will be attainable,
and hence identifying the most favorable enzyme-substrate
combination for asymmetric synthetic use. Further studies
toward these objectives, using an expanded structural range of
inhibitors as probes, are in hand.
Figure 5. Superimposed energy-minimized EI complexes of (R)- and
(S)-4-phenylbutanal, (R)-4 (heavy solid line) and (S)-4 (heavy dashed
line), respectively, in the active site of SC. The oxyanions of the
tetrahedral complexes derived from (R)-4 and (S)-4 are located in the
oxyanion hole, with the negative charges similarly stabilized by
hydrogen bonding (light dashed line) with the peptide NH of Ser221
(O--N ) 2.84 and 2.85 Å respectively for (R)-4 and (S)-4) and the
side chain -NH2 of Asn155 (O--N ) 2.80 and 2.78 Å respectively
for (R)-4 and (S)-4). The methyl groups at the stereocenters of (R)-4
(solid arrow) and (S)-4 (dashed arrow) are both located outside of the
active site and are non-contributors to binding. The only significant,
but still minor, difference between the two EI complexes is that the
phenyl group of (R)-4 penetrates more deeply into the hydrophobic S1
region than does that of (S)-4.
groups of both the R- and S-enantiomers are directed toward
the outside of the active site. The S-enantiomer is obliged to
adopt this unusual orientation in order to avoid unfavorable steric
interactions between the methyl group and the side chain of
Met 192. As a consequence, this prevents the phenyl group of
(S)-4 from penetrating as deeply into the hydrophobic S1 pocket
as it might otherwise, thereby precluding the strong phenyl-S1
hydrophobic interactions that (R)-4 enjoys. The latter enanti-
omer does not bind as strongly as (R)-2 above because it lacks
the strong hydrophobic bonding contribution that the additional
methyl-S2 interaction provides in the Figure 2 situation.

506 J. Am. Chem. Soc., Vol. 118, No. 3, 1996
Lee and Jones
To a solution of the Dess-Martin reagent (440 mg, 1.0 mmol, 1.2
mol equiv) in CH2Cl2 (5.0 mL) was added (R)-(-)-3-phenylbutan-1-ol
(120 mg, 0.87 mmol, 1.0 mol equiv) in CH2Cl2 (2.0 mL) at 20 °C. The
reaction mixture was stirred for 30 min at 20 °C, then diluted with
Et2O (7.0 mL) and poured into sodium thiosulfate (1.1 g, 7.0 mol equiv)
in saturated aqueous NaHCO3 (20 mL) and stirred for 10 min. The
layers were separated, and the organic layer was washed sequentially
with saturated aqueous NaHCO3 and deionized water. The combined
organic layers were dried (MgSO4), filtered, and concentrated under
reduced pressure. The crude product was purified by radial TLC
(Chromatotron, hexanes/EtOAc, 9:1) to give (R)-(-)-3-phenylbutanal
Experimental Section
General Methods. Anhydrous reagents were prepared according
to the literature procedure.13 Analytical TLC was performed on
precoated plates (silica gel 60F-254). Purification by radial TLC was
performed on a model 7924T Chromatotron from Harrison Research.
Plates of 4 and 2 mm thickness were coated with E. Merck Silica Gel
60 PF254 containing gypsum. Absorption measurements were per-
formed on a Perkin Elmer Lambda 2 UV/vis spectrometer. IR spectra
were determined on films on a Nicolet 5DX FTIR spectrophotometer.
NMR (1H, 13C) spectra were recorded on a Gemini 200 (200, 50 MHz,
respectively) spectrometer. 1H NMR chemical shifts are reported in
ppm relative to the TMS peak (δ ) 0.0) with CDCl3 as solvent. 13C
NMR chemical shifts are reported in ppm relative to the CDCl3 peak
(δ ) 77.0) with CDCl3 as solvent. Optical rotations were measured
on a Perkin-Elmer 141 polarimeter. Mass spectra were measured on a
Bell and Howell 21-490 (low resolution) or an AEI MS3074 instrument
(high resolution).
Reagent grade chemicals, 3-phenylpropanal (1), 4-phenylbutan-1-
ol, (()-3-phenylbutanoic acid, (R)-(+)-R-methylbenzylamine (96% ee),
(S)-(-)-R-methylbenzylamine (99% ee), Diazald (N-methyl-N-nitroso-
p-toluenesulfonamide), and diisobutylaluminum hydride (DIBAL, 1.0
M in CH2Cl2) were purchased from Aldrich. Subtilisin Carlsberg (EC
3.4.21.14), R-chymotrypsin (EC 3.4.21.1), and succinyl-L-Ala-L-Ala-
L-Pro-L-Phe-p-nitroanilide (Suc-Ala-Ala-Pro-Phe-PNA) were purchased
from Sigma Chemical Co. The concentration of SC was estimated by
the rate of the hydrolysis of Suc-Ala-Ala-Pro-Phe-PNA.14 The
concentration of CT was estimated by the rate of the hydrolysis of
p-nitrophenyl acetate.15 The Dess-Martin reagent10 and (S,S)-N,N′-
dimethyl-1,2-diphenylethylenediamine16,17 were prepared according to
the literature procedure.
((R)-2, 120 mg, 94%, >95% ee) as a colorless oil: [R]25 -38.5° (c
D
0.2 g, Et2O); lit.20 [R]25D -38.0° (c 0.2, Et2O); 1H NMR δ 1.32 (3H, d,
J ) 7.0 Hz), 2.55-2.85 (2H, m), 3.25-3.45 (1H, m), 7.18-7.37 (5H,
m), 9.71 (1H, t, J ) 2.0 Hz).
The preparation of (S)-(+)-3-phenylbutanal ((S)-2) was carried out
in the same manner, as follows:
(S)-(+)-3-Phenylbutanal ((S)-2). (S)-(+)-3-Phenylbutanoic acid
(520 mg, 3.2 mmol) and CH2N2 in Et2O gave methyl (S)-(+)-3-
phenylbutanoate (560 mg, 98%): [R]20 +35.2° (c 2.2, EtOH),
D
spectroscopically identical to the R-enantiomer.
Methyl (S)-(+)-3-phenylbutanoate (500 mg, 2.8 mmol, 1.0 mol
equiv) and DIBAL (1.0 M solution in CH2Cl2, 6.0 mmol, 2.1 mol equiv)
yielded (S)-(+)-3-phenylbutan-1-ol (410 mg, 98%), spectroscopically
identical with the R-enantiomer.
(S)-(+)-3-Phenylbutanol (410 mg, 2.7 mmol, 1.0 mol equiv) and
the Dess-Martin reagent (1.4 g, 3.3 mmol, 1.2 mol equiv) afforded
(S-(+)-3-phenylbutanal ((S)-2, 330 mg, 82%, >95% ee): [R]25D +37.1°
(c 0.2, Et2O); lit.20 [R]25 +38.0° (c 0.2, Et2O), spectroscopically
D
identical to (R)-2.
4-Phenylbutanal (3). 4-Phenylbutan-1-ol (700 mg, 4.7 mmol, 1.0
mol equiv) was oxidized with the Dess-Martin reagent (2.4 g, 5.7
mmol, 1.2 mol equiv) in CH2Cl2 (35 mL). After workup and
purification using the Chromatotron (hexanes/EtOAc, 90/10), 4-phe-
nylbutanal21 (3, 575 mg, 83%) was obtained as a colorless oil: 1H NMR
δ 1.90-2.05 (2H, m), 2.35-2.50 (2H, m), 2.62-2.75 (2H, m), 7.15-
7.35 (5H, m), 9.76 (1H, t, J ) 1.5 Hz).
(R)-(-)-4-Phenylpentanal (R-4). (R)-(-)-3-phenylbutanoic acid
(520 mg, 3.1 mmol, 1.0 mol equiv) was dissolved in thionyl chloride
(754 mg, 6.34 mmol, 2.0 mol equiv) under an N2 atmosphere and the
reaction mixture was refluxed for 2 h to remove of the excess thionyl
chloride under reduced pressure, followed by Kugelrohr distillation,
to give (R)-(-)-3-phenylbutanoyl chloride (560 mg, 97%) as a colorless
oil, which was used directly.
Resolution of (()-3-Phenylbutanoic acid. (()-3-Phenylbutanoic
acid was resolved with (R)-(+)-R-methylbenzylamine to give (R)-
(-)-3-phenylbutanoic acid (42%): [R]20 -29.4° (c 1.3, EtOH); lit.18
D
[R]25 -58.5° (c 3.0, benzene). (S)-(+)-3-Phenylbutanoic acid (30%)
D
was obtained similarly using (S)-(-)-R-methylbenzylamine: [R]20
D
+29.1° (c 1.2, EtOH); lit.19 [R]25 +54.4° (c 1.4, benzene).
D
(R)-(-)-3-Phenylbutanal ((R)-2) To a solution of (R)-(-)-3-
phenylbutanoic acid (500 mg, 3.0 mmol) in Et2O (20 mL) at 20 °C
was added an ethereal solution of CH2N2, generated from Diazald, until
the reaction mixture became yellow. The reaction mixture was stirred
for additional 2 h at 20 °C and concentrated under reduced pressure.
The residue was chromatographed on silica gel (hexanes), and methyl
(R)-(-)-3-phenylbutanoate (490 mg, 91%) was obtained as a colorless
1
oil: [R]20 -35.7° (c 2.2, EtOH); H NMR δ 1.30 (3H, d, J ) 7.0
D
(R)-(-)-3-Phenylbutanoyl chloride (560 mg, 3.1 mmol, 1.0 mol
equiv) in Et2O (5.0 mL) was added dropwise at 0 °C to a solution of
CH2N2 (generated from Diazald (2.0 g, 9.3 mmol, 3.0 mol equiv)), in
Et2O (25 mL), and the reaction mixture was then stirred at 20 °C for
3 h. The reaction mixture was then concentrated under reduced pressure
and the crude product purified by Chromatotron chromatography
(hexanes/EtOAc, 9:1) to give (R)-(-)-1-diazo-4-phenylpentan-2-one
Hz), 2.50-2.70 (2H, m), 3.20-3.40 (1H, m), 3.62 (3H, s), 7.15-7.37
(5H, m).
DIBAL (1.0 M solution in CH2Cl2, 2.6 mmol, 2.1 mol equiv) was
added dropwise under an N2 atmosphere at -78 °C to a solution of
methyl (R)-(-)-3-phenylbutanoate (215 mg, 1.2 mmol, 1.0 mol equiv)
in CH2Cl2 (10 mL). The reaction mixture was stirred for 2 h at -78
°C, then slowly warmed to 20 °C and stirred for an additional 12 h at
20 °C. The reaction was quenched by the addition of MeOH (3 mol
equiv), the aluminum oxide removed by filtration through Celite, and
the filtrate concentrated under reduced pressure. The residue was
dissolved in hexanes (10 mL), filtered through a small silica gel column,
and concentrated to give (R)-(-)-3-phenylbutan-1-ol (165 mg, 91%)
as a colorless oil, which was used without further purification: 1H NMR
δ 1.28 (3H, d, J ) 7.0 Hz), 1.35-1.45 (1H, br), 1.81-1.92 (2H, m),
2.78-2.98 (1H, m), 3.45-3.65 (2H, m), 7.15-7.37 (5H, m).
(566 mg, 97%) as a yellow oil: [R]25D -109° (c 0.5, CH2Cl2); IR (cm-1
3081, 2968, 2105, 1639; H NMR δ 1.31 (3H, d, J ) 7.0 Hz), 2.45-
2.70 (2H, m), 3.20-3.40 (1H, m), 5.10 (3H, s), 7.15-7.37 (5H, m);
13C NMR δ 21.7, 36.5, 49.2, 54.9, 126.3, 126.6, 128.4, 145.7, 193.6.
)
1
To a solution of (R)-(-)-1-diazo-4-phenylpentan-2-one (495 mg, 2.6
mmol) in methanol (19 mL) was added a catalytic amount of Ag2O at
60 °C. After N2 evolution ceased (5 min), the reaction mixture was
stirred an additional 30 min at 60 °C, Norit was added, and the reaction
mixture was filtered, concentrated under reduced pressure, and purified
by Chromatotron (hexanes) to give methyl (R)-(-)-4-phenylpentanoate
(13) 13)Perrin, D. D.; Armarego, W. L. F.; Perrin, D. R. Purification of
Laboratory Chemicals; Pergamon Press: New York, 1980.
(14) (a) Del Mar, E. G.; Largman, C.; Brodricl, J. W.; Goekas, M. C.
Anal. Biochem. 1979, 99, 316. (b) Russell, A. J.; Thomas, P. G.; Fersht, A.
R. J. Mol. Biol. 1987, 193, 803.
(15) (a) Kezdy, F. J.; Kaiser, E. T. Methods Enzymol. 1970, 19, 3. (b)
Ottensen, M.; Svendsen, I. Methods Enzym. Anal. 1984, 5, 159.
(16) Mangeney, P.; Tejero, T.; Alexakis, A.; Grosjean, F.; Normant, J.
Synthesis 1988, 255.
(17) Mangeney, P.; Grosjean, F.; Alexakis, A.; Normant, J. Tetrahedron
Lett. 1988, 29, 2675.
(18) Cram, D. J. J. Am. Chem. Soc. 1952, 74, 2137.
(19) Sorlin, G.; Bergson, G. Ark. Kemi 1968, 29, 593.
(418 mg, 83%) as a colorless oil: [R]25 -22.0° (c 1.1, CH2Cl2); IR
D
(cm-1) 3061, 2959, 1735, 1605; 1H NMR δ 1.27 (3H, d, J ) 6.9 Hz),
1.85-2.00 (2H, m), 2.15-2.27 (2H, m), 2.62-2.82 (1H, m), 3.63 (3H,
s), 7.15-7.37 (5H, m); 13C NMR δ 22.2, 32.3, 33.2, 39.4, 51.4, 126.1,
126.9, 128.4, 146.1, 173.0.
(20) Mangeney, P.; Alexakis, A.; Normant, J. F. Tetrahedron Lett. 1988,
29, 2677.
(21) 4-Phenylbutanal had been synthesized by different methods: (a)
Kumler, W. D.; Strait, L. A.; Alpen, E. L. J. Am. Chem. Soc. 1950, 72,
1463. (b) Braun, J. V. Chem. Ber. 1934, 67, 218.

Discrimination of R and S with Serine Proteases
J. Am. Chem. Soc., Vol. 118, No. 3, 1996 507
The same reduction and oxidation procedure used above to prepare
(R)-2 was then followed. Methyl (R)-(-)-4-phenylpentanoate (360 mg,
1.9 mmol, 1.0 mol equiv) and DIBAL (1.0 M solution in CH2Cl2, 4.2
mmol, 2.2 mol equiv) gave (R)-(-)-4-phenylpentan-1-ol (270 mg,
substrate, and 5% DMSO. The progress curve with inhibitor was
obtained similarly, but with 10-120 µL of the inhibitor solution (0.14-
0.27 M in DMSO) added. The progress of each hydrolysis was
recorded directly into a PC. Points for calculation12 were taken at
15,18, 21, 24, 27, 30, 33, 36, and 39% conversions.
1
88%): IR (cm-1) 3339 (br), 3060, 2932, 1603; H NMR δ 1.26 (3H,
d, J ) 6.9 Hz), 1.40-1.80 (6H, m), 2.60-2.80 (1H, m), 3.59 (2H, t,
J ) 6.4 Hz), 7.15-7.35 (5H, m); 13C NMR δ 22.4, 30.9, 34.4, 39.8,
62.9, 125.8, 126.9, 128.3, 147.2.
All kinetic measurements for CT were performed with degassed
water to minimize the inhibition of the enzyme by CO2, and the kinetic
determinations were carried out as described above for SC. KM and
(R)-(-)-4-Phenylpentanol (253 mg, 1.6 mmol, 1.0 mol equiv) and
the Dess-Martin reagent (1.0 g, 2.3 mmol, 1.5 mol equiv) yielded
k
cat values for Suc-Ala-Ala-Pro-Phe-PNA were determined (Grafit) for
the following assay conditions: substrate concentration (0.025, 0.033,
0.050, 0.100, 0.200, 0.500, and 1.000 mM), 0.1 M NaH2PO4 buffer
(pH 7.5), 0.5 M NaCl, 5% DMSO, and 2.1 nM of CT. The progress
curve without inhibitor was obtained from an assay mixture containing
0.2 M NaH2PO4 (pH 7.5) and 1.0 M NaCl (1.50 mL), 25 mM substrate
in DMSO (30 µL), DMSO (120 µL), and H2O (1.32 mL) which was
incubated in a water jacket for 5 min at 25 °C prior to initiation of the
(R)-(-)-4-phenylpentanal ((R)-4,160 mg, 63%, >95% ee): [R]25
D
1
-16.2° (c 0.9, CH2Cl2); IR (cm-1) 3064, 2967, 1724, 1603; H NMR
δ 1.28 (3H, d, J ) 7.0 Hz), 1.80-2.00 (2H, m), 2.27-2.40 (2H, m),
2.62-2.80 (1H, m), 7.15-7.37 (5H, m), 9.69 (1H, t, J ) 1.5 Hz); 13
C
NMR δ 22.2, 30.3, 39.2, 42.1, 126.2, 126.8, 128.4, 145.9, 202.1; HRMS,
calcd for M+ C11H14O m/e 162.1045, found m/e 162.1051.
The preparation of (S)-(+)-4-phenylpentanal (((S)-4) was carried out
in the same manner, as follows:
reaction by the addition of 30 µL of CT stock solution (1.6 × 10-6
M
in 1.0 mM HCl). The final volume of the assay mixture was 3 mL
with a final concentration of 0.1 M NaH2PO4, 0.5 M NaCl, 16 nM of
the enzyme, 0.25 mM of substrate, and 5% DMSO. The progress curve
with inhibitor was obtained similarly, with 100-120 µL of the inhibitor
solution (0.14-0.17 M in DMSO) added. The progress of hydrolysis
was recorded directly into a PC. The inhibition constants were
determined as described for SC. The results are recorded in Table 1.
Computational Methods. System Setup. The reference structures
used for SC5a and CT5d were from the Protein Data Bank22 at
Brookhaven National Laboratory,23 with Insight II (version 2.3.0,
Biosym Technologies, Inc., San Diego, CA, USA) as the graphics
software on a Silicon Graphics Indigo computer. To create initial
coordinates for the minimization of SC, the inhibitor (eglin C) and the
three calcium ions were removed. Hydrogens were added at the pH
(7.5) used for the kinetic measurements, which protonated all Lys and
Arg side chains, and the N-terminal NH2, and deprotonated all Glu
and Asp side chains, and the C-terminal COOH. The overall charge
of the enzyme resulting from this setup was -1. To complete the setup,
only crystallographic water molecules within 5 Å from the enzyme
surface were maintained and a 5-Å shell of water was generated to
surround the enzyme. The total number of water molecules in this
setup was 1274. In the case of CT the dimeric structure of the enzyme
was split into its individual, independent, monomers, and only one of
the monomers used to create the initial coordinates. The missing
residues Gly 12 and Leu 13 were added using Insight. The setup was
then completed as reported above for SC. The final charge of the
molecule was +2, and the total number of water molecules was 1252.
Minimizations of each enzyme were performed with the DISCOVER
program (Biosym, Version 3.1) using the Consistent Valence Force
Field (CVFF).24 The nonbonded pair list was updated every 20 cycles
and a dielectric constant of 1 was used in all calculations. A switching
function was used to scale down the electrostatic and the van der Waals
interactions from full to zero between 10 and 12 Å, with a nonbonded
cutoff of 14 Å. The energy of the system was minimized with respect
to all 3N Cartesian coordinates until the maximum derivative of 0.1
kcal/(mol‚Å) was reached. The minimized enzyme structures obtained
were then used as the bases for the molecular mechanics calculations
on enzyme-inhibitor complexes
(S)-(+)-4-Phenylpentanal ((S)-4). (S)-(+)-3-Phenylbutanoic acid
(380 mg, 2.31 mmol, 1.0 mol equiv) and thionyl chloride (540 mg,
4.62 mmol, 2.0 mol equiv) gave (S)-(+)-3-phenylbutanoyl chloride (403
mg, 95%).
(S)-(+)-3-phenylbutanoyl chloride (403 mg, 2.2 mmol, 1.0 mol
equiv) and CH2N2 (from Diazald (1.41 g, 6.6 mmol, 3.0 mol equiv))
afforded (S)-(+)-1-diazo-4-phenylpentan-2-one (356 mg, 87%): [R]25
D
+107° (c 0.5, CH2Cl2); spectroscopically identical to the R-enantiomer.
(S)-(+)-1-Diazo-4-phenylpentan-2-one (300 mg, 1.6 mmol), Ag2O
(catalytic), and MeOH (13 mL) yielded methyl (S)-(+)-4-phenylpen-
tanoate (250 mg, 82%): [R]25 +22.2° (c 1.0, CH2Cl2), spectroscopi-
D
cally identical to the R-enantiomer.
Methyl (S)-(+)-4-phenylpentanoate (167 mg, 0.9 mmol, 1.0 mol
equiv) and DIBAL (1.0 M solution in CH2Cl2, 1.9 mmol, 2.2 mol equiv)
gave (S)-(+)-4-phenylpentan-1-ol (135 mg, 95%), spectroscopically
identical to the R-enantiomer.
(S)-(+)-4-Phenylpentan-1-ol (140 mg, 0.86 mmol, 1.0 mol equiv)
and the Dess-Martin reagent (550 mg, 1.3 mmol, 1.5 mol equiv)
yielded (S)-(+)-4-phenylpentanal ((S)-4, 85 mg, 60%, >95% ee): [R]25
D
+16.7° (c 0.9, CH2Cl2), spectroscopically identical to (R)-4. HRMS,
calcd for M+ C11H14O m/e 162.1045, found m/e 162.1049.
General Procedure for the Determination of Enantiomeric
Excess. To a solution of aldehyde ((R,S)-2 and (R,S)-4, 0.15 M, 1.0
mol equiv) in Et2O was added molecular sieves followed by (S,S)-
N,N′-dimethyl-1,2-diphenyl-ethylenediamine (1.3 mol equiv) at 20 °C
under N2. The reaction mixture was stirred for 5 h at 20 °C and then
filtered through Celite. The filtrate was concentrated in Vacuo, and
the 13C NMR spectrum of the residue was taken. The integrations of
the peaks for the methyl groups at the stereocenters of each aldehyde
(22.21 ppm for (R)-2, 24.60 ppm for (S)-2, 22.52 ppm for (R)-4, and
22.90 ppm for (S)-4) were used to determine the enantiomeric excesses.
In each case, the materials were seen to be enantiomerically pure, within
the (5% confidence limits of the NMR method.
Kinetic Measurements. The enzyme kinetics for both SC and CT
were performed at 25 °C by measuring the absorbance of the PNA (ꢀ
) 8800 at 410 nm) released during the hydrolysis of Suc-Ala-Ala-
Pro-Phe-PNA. All kinetic runs were carried out in duplicate. For SC,
Energy Minimizations of Enzyme-Inhibitor Complexes. For SC,
to set up the initial structure for the energy minimization on each
inhibitor, the carbonyl carbon was covalently bound to the oxygen of
the hydroxyl group of the active site Ser221 of SC to form the oxyanion-
containing tetrahedral intermediate. The hydrogen atom of the Ser221
hydroxyl group was then transferred to the sp2 nitrogen of the imidazole
ring of the side chain of His64, rendering it positively charged. The
K
M and kcat of Suc-Ala-Ala-Pro-Phe-PNA were determined by measur-
ing initial rates of hydrolysis at different substrate concentrations (0.031,
0.042, 0.067, 0.125, 0.250, 0.500, 1.000, 2.000 mM), and fitting the
obtained data to the Michaelis-Menten equation using the Grafit
program (version 3.0, Erithacus Software Ltd., UK). Assays were run
in 0.1 M sodium phosphate buffer at pH 7.5, containing 0.5 M NaCl,
and 5% (v/v) dimethyl sulfoxide. The enzyme concentration was 0.47
nM. The KI for each inhibitor for SC was determined by Waley’s
method.12 The progress curve without inhibitor was obtained from an
assay mixture containing 0.2 M NaH2PO4 and 1.0 M NaCl (1.47 mL),
pH 7.5 buffer with 0.1 M NaH2PO4 (50 µL), 1.0 M NaCl (30 µL), 25
mM substrate in DMSO (30 µL), DMSO (120 µL), and H2O (1.29
mL). This mixture was incubated in a water jacket for 5 min at 25 °C
prior to initiation of the reaction by addition of 10 µL of SC stock
solution (1.4 × 10-6 M in 0.1 M NaH2PO4 buffer, pH 7.5). The final
volume of the assay mixture was 3 mL, with a final concentration of
0.1 M NaH2PO4, 0.5 M NaCl, 4.5 nM of the enzyme, 0.25 mM of
(22) For subtilisin Carlsberg entry 2SEC, 1.8-Å resolution; for R-chy-
motrypsin entry 4CHA, 1.68-Å resolution.
(23) (a) Bernstein, F. C.; Koetzle, T. F.; Williams, J. B.; Meyer, E. F.,
Jr.; Brice, M. D.; Rogers, J. R.; Kennard, O.; Shimanouchi, T.; Tasumi, M.
J. Mol. Biol. 1977, 112, 535. (b) Abola, E. E.; Bernstein, F. C.; Bryant, S.
H.; Koetzle, T. F.; Weng, J. In Crystallographic DatabasessInformation
Content, Software Systems, Scientific Applications; Allen, F. H., Bergerhoff,
G., Sievers, R., Eds; Data Commision of the International Union of
Crystallography: Bonn/Cambridge/Chester, 1987, 107.
(24) Hagler, A. T.; Osguthorpe, D. J.; Dauber-Osguthorpe, P.; Hempel,
J. C. Science 1985, 227, 1309.

508 J. Am. Chem. Soc., Vol. 118, No. 3, 1996
Lee and Jones
X-ray structure of the Met192Ala mutant of R-lytic protease complexed
with methoxysuccinyl-Ala-Ala-Pro-phenylalanylboronic acid25 was used
as the model for docking the tetrahedral intermediate into the active
site of SC. The oxyanion was oriented toward the oxyanion hole,
formed by the side chain NH2 of Asn155 and the backbone NH of
Ser221. The aromatic residue of each inhibitor was positioned in the
S1 pocket, defined by Ser125-Ala129, Ala152-Ser156, and Ile165-
Tyr167 in a manner that avoided all bad Van der Waals interactions.
At this time, six water molecules in the SC active site that were located
in positions that caused severe steric interactions with the docked
inhibitor were removed.
An analogous docking procedure was used for CT. The reference
structure used for docking was that of CT complexed with 2-phenyl-
ethylboronic acid.5g The oxyanion of the tetrahedral intermediate
generated by attack of the active site Ser195 was directed toward the
oxyanion hole formed by the backbone NH’s of Gly193 and Ser195,
and a positive charge was generated on His57 by transferring the
hydrogen from the Ser195 hydroxyl group. The aromatic ring of each
inhibitor was positioned in the S1 pocket, defined by Val213-Thr219,
Ser190-Asp194, Ser189, and Gly226. Again, six water molecules
located too close to the inhibitor were also removed from the CT active
site.
The partial charges on the active site serine and the enzyme-bound
aldehyde complex were generated by a semiempirical ab initio
calculation program (Mopac, Biosym). The overall negative charge
of -1 was distributed between the oxygen atoms bound to the
tetrahedral carbon formed from aldehyde. The partial charges of
residues His64 for SC and His57 for CT were assigned as reported.26
For running the energy minimization, the same strategy was applied
for both SC and CT by using a nonbonded cutoff of 14 Å, with a
switching function between 10 and 12 Å. All residues and the water
molecules located outside of the cutoff distance were fixed to reduce
the computation time while the enzyme-bound inhibitor and the water
molecules within the cutoff distance were free to move throughout the
calculation. Initially, all heavy atoms on the enzyme were tethered
using harmonic restraints, and the energy minimized with respect to
the positions of the hydrogen atoms, using steepest descents, until the
maximum derivative was <0.5 kcal/(mol‚Å). The restraints were then
applied only to the heavy atoms on the main chain, and the minimization
continued, again using steepest descents, until the maximum derivative
was <0.25 kcal/(mol‚Å). Finally, the restraints were removed from
all atoms, and the energy minimized using steepest descents until the
maximum derivative was <0.1 kcal/(mol‚Å), and then using conjugate
gradients until the maximum derivative was <0.01 kcal/(mol‚Å).
The resulting enzyme-inhibitor structure was used as the starting
point for the molecular dynamics calculations. During the molecular
dynamic simulations, the whole enzyme was kept fixed, as were water
molecules more than 14 Å away from Ser221 of SC or Ser195 of CT.
The molecular dynamic simulations were performed at 500 °K, at a
time step of 1 fs, and sampling every 10 steps over a 10 ps time period.
As an additional comparison, the lowest energy structure from the
molecular dynamics simulation was reminimized by molecular mechan-
ics as described above. In each case, the latter structures were found
to be almost identical with those of the initial, molecular mechanics-
minimized structures. For the enzyme-inhibitor complexes of (R)-
2-(S)-2 and (R)-4-(S)-4 with CT, the calculated ∆∆Gq values were
6.4 and 4.5 kcal/mol, respectively, and for SC, 5.3 and 3.2 kcal/mol,
respectively. The corresponding experimental ∆∆Gq values were 2.7
and 1.3, and 1.7 and 0.93 kcal/mol, respectively.
Acknowledgment. Support from the Natural Sciences and
Engineering Research Council of Canada and the award of
University of Toronto Open Scholarships (to T.L.) are gratefully
acknowledged.
Supporting Information Available: 1H and 13C NMR
spectra of (R)- and (S)-4 (4 pages). This material is contained
in many libraries on microfiche, immediately follows this article
in the microfilm version of the journal, can be ordered from
the ACS, and can be downloaded from the Internet; see any
current masthead page for ordering information and Internet
access instructions.
(25) Bone, R.; Silen, J. L.; Agard, D. A. Entry 1P08 in the Brookhaven
protein database, 2.25-Å resolution.
(26) Rao, S. N.; Singh, U. C.; Bash, P. A.; Kollman, P. A. Nature 1987,
328, 551.
JA952835T