Selection of enantioselective acyl transfer catalysts from a pooled peptide library through a fluorescence-based activity assay: An approach to kinetic resolution of secondary alcohols of broad structural scope

Article abstract of DOI:10.1021/ja0108584

An assay employing a fluorescently labeled split and pool peptide library has been applied to the discovery of a new class of octapeptide catalysts for the kinetic resolution of secondary alcohols. A highly diverse library of peptide-based catalysts was synthesized on solid-phase synthesis beads such that each individual bead was co-functionalized with (i) a uniform loading of a pH-sensitive fluorophore and (ii) a unique peptidebased catalyst. The library was then screened for activity in acylation reactions employing (±)-sec-phenylethanol as the substrate and acetic anhydride as the acylation agent. From the most active catalysts, a lead peptide (4) was identified that provides a selectivity-factor (k_rel) of 8.2 upon resynthesis and evaluation under homogeneous conditions. A "directed" second-generation split and pool peptide library was synthesized such that the new peptide sequences in the library were biased toward the lead structure. Random samples of the second generation library were screened in single bead assays that revealed several new peptide-based catalysts that afford improved selectivities in kinetic resolutions. Peptide catalyst 13 proves effective for the kinetic resolution of sec-phenylethanol (k_rel = 20), as well as eight other secondary alcohols of a broad substrate scope (k_rel = 4 to > 50).

Full text of DOI:10.1021/ja0108584

6496
J. Am. Chem. Soc. 2001, 123, 6496-6502
Selection of Enantioselective Acyl Transfer Catalysts from a Pooled
Peptide Library through a Fluorescence-Based Activity Assay: An
Approach to Kinetic Resolution of Secondary Alcohols of Broad
Structural Scope
Gregory T. Copeland and Scott J. Miller*
Contribution from the Department of Chemistry, Merkert Chemistry Center,
Boston College, Chestnut Hill, Massachusetts 02467-3860
ReceiVed April 2, 2001
Abstract: An assay employing a fluorescently labeled split and pool peptide library has been applied to the
discovery of a new class of octapeptide catalysts for the kinetic resolution of secondary alcohols. A highly
diverse library of peptide-based catalysts was synthesized on solid-phase synthesis beads such that each individual
bead was co-functionalized with (i) a uniform loading of a pH-sensitive fluorophore and (ii) a unique peptide-
based catalyst. The library was then screened for activity in acylation reactions employing (()-sec-phenylethanol
as the substrate and acetic anhydride as the acylation agent. From the most active catalysts, a lead peptide (4)
was identified that provides a selectivity-factor (krel) of 8.2 upon resynthesis and evaluation under homogeneous
conditions. A “directed” second-generation split and pool peptide library was synthesized such that the new
peptide sequences in the library were biased toward the lead structure. Random samples of the second generation
library were screened in single bead assays that revealed several new peptide-based catalysts that afford improved
selectivities in kinetic resolutions. Peptide catalyst 13 proves effective for the kinetic resolution of
sec-phenylethanol (krel ) 20), as well as eight other secondary alcohols of a broad substrate scope (krel ) 4 to
>50).
Introduction
the practical advantage that both the theoretical and actual
diversity of catalyst libraries is massive using the split-and-
pool synthesis strategy.5 For catalyst selection, we have been
developing fluorescence-based activity assays that allow local-
ization of a fluorescent signal on a synthesis bead that carries
upon it an active, unique acyl transfer catalyst.6,7 Selection of
the brightest beads from a pooled mixture subjected to reaction
conditions thus allows the identification of the most active
catalysts in a library.8
The discovery of small molecules that function as efficient
catalysts for synthetically important reactions stands as a major
challenge for chemists. Techniques that allow for the simulta-
neous screening of numerous catalysts hold promise for the field
since they may allow for accelerated catalyst discovery.1,2 Such
approaches also have the potential to mimic selection processes
evolved by natural systems, where evolution of complex
chemical systems proceeds by iterative screening and selection
protocols. Two problems among many to be addressed if the
full potential of diversity-oriented catalyst discovery is to be
realized are the following: (a) the synthesis of genuinely diverse
catalyst libraries and (b) the identification of the most promising
catalysts from pooled mixtures containing many less interesting
ones.
We had previously found that peptides containing the
nucleophilic N-methylimidazole (NMI) moiety are effective
catalysts for the kinetic resolution of certain secondary alcohols.9
(5) (a) Furka, A.; Sebestyen, F.; Asgedom, M.; Dibo, G. Int. J. Pept.
Protein Res. 1991, 37, 487-493. (b) Lam, K. S.; Salmon, S. E.; Hersh, E.
M.; Hruby, V. J.; Kazmierski, W. M.; Knapp, R. J. Nature 1991, 354, 82-
84. (c) For a review of the “one-bead-one-compound” combinatorial library
method see: Lam, K. S.; Lebl, M.; Krchna´k, V. Chem. ReV. 1997, 97, 411-
448.
(6) (a) Copeland, G. T.; Miller, S. J. J. Am. Chem. Soc. 1999, 121, 4306-
4307. (b) Harris, R. F.; Nation, A. J.; Copeland, G. T.; Miller, S. J. J. Am.
Chem. Soc. 2000, 122, 11270-11271.
(7) For other fluorescence-based assays of catalyst libraries, see: (a)
Reddington, E.; Sapienza, A.; Gurau, B.; Viswanathan, R.; Sarangapani,
S.; Smotkin, E. S.; Mallouk, T. E. Science 1998, 280, 1735-1737. (b)
Shaughnessy, K. H.; Kim, P.; Hartwig, J. F. J. Am. Chem. Soc. 1999, 121,
2123-2132. (c) Yeung, E. S.; Su, H. J. Am. Chem. Soc. 2000, 122, 7422-
7423.
(8) For several examples of other studies where catalysts of interest have
been identified from pooled mixtures of catalysts, see: (a) Francis, M. B.;
Jacobsen, E. N. Angew. Chem., Int. Ed. 1999, 38, 937-941. (b) Berkessel,
A.; Herault, D. A. Angew. Chem., Int. Ed. 1999, 38, 102-105. (c) Taylor,
S. J.; Morken, J. M. Science 1998, 280, 267-270.
(9) For reviews of catalytic kinetic resolution, see: (a) Keith, J. M.;
Larrow, J. F.; Jacobsen, E. N. AdV. Synth. Catal. 2001, 343. 5-26. (b)
Hoveyda, A. H.; Didiuk, M. T. Curr. Org. Chem. 1998, 2, 537-574.
We have been focusing on peptide-based catalyst libraries
for several reasons, including a general interest in the minimal
structural basis of a polypeptide that is required to achieve
efficient enantioselective catalysis.3,4 In addition, peptides offer
(1) Jandeleit, B.; Schaefer, D. J.; Powers, T. S.; Turner, H. W.; Weinberg,
W. H. Angew. Chem., Int. Ed. 1999, 38, 2494-2532.
(2) For several recent reviews of combinatorial catalysis, see: (a) Reetz,
M. T. Angew. Chem., Int. Ed. 2001, 40, 284-310. (b) Crabtree, R. H. Chem.
Commun. 1999, 17, 1611-1616. (c) Kuntz, K. W.; Snapper, M. L.;
Hoveyda, A. H. Curr. Opin. Chem. Biol. 1999, 3, 313-319. (d) Francis,
M. B.; Jamison, T. F.; Jacobsen, E. N. Curr. Opin. Chem. Biol. 1998, 2,
422-428.
(3) For an overview of enzymatic catalysts and mechanisms, see: (a)
Wong, C.-H.; Whitesides, G. M. Enzymes in Synthetic Organic Chemistry;
Elsevier Science Ltd.: Oxford, 1994; Chapter 2. (b) Lamzin, V. S.; Dauter,
Z.; Wilson, K. S. Curr. Opin. Struct. Biol. 1995, 5, 830-836.
(4) Jarvo, E. R.; Copeland, G. T.; Papaioannou, N.; Bonitatebus, P. J.;
Miller, S. J. J. Am. Chem. Soc. 1999, 121, 11638-11643.
10.1021/ja0108584 CCC: $20.00 © 2001 American Chemical Society
Published on Web 06/19/2001

Kinetic Resolution of Secondary Alcohols of Broad Structural Scope
J. Am. Chem. Soc., Vol. 123, No. 27, 2001 6497
Scheme 1
Scheme 2
and acetic anhydride proceeds, generation of product (ester +
acetic acid) is signaled by an increase in fluorescence intensity.
A key feature of this assay, within the context of enantioselective
acylation reactions, is that the fluorescence signal is coupled to
catalyst actiVity, but not the degree of enantioselectiVity that a
particular catalyst may afford. Nevertheless, we formulated a
hypothesis that selection of catalysts based on activity alone
might lead to catalysts that also exhibited good degrees of
selectivity.
Kinetic resolutions of functionalized substrates such as (()-1
(Scheme 1) can be achieved with krel > 50 (recovered starting
material, 96% ee at 52% conversion).10-13 In particular, amide-
functionalized alcohols that have the potential to form amide-
based hydrogen bonds to the catalyst have been among the best
substrates we have studied. To expand the scope of the catalysts
to include resolutions of unfunctionalized substrates (i.e.,
substrates lacking additional hydrogen bond donors and accep-
tors), we set out to screen a highly diverse library of potential
catalysts. Notably, at the outset of the study, we were unclear
as to what peptide sequences (if any) would be optimal for the
kinetic resolution of this class of substratessnone that we had
prepared previously exhibited appreciable selectivity in at-
tempted resolutions of alcohols such as 2 (Scheme 1). This lack
of selectivity was a particular concern since these new substrates
would lack functional groups that could lead to obvious multi-
point contacts with the catalyst during the stereodifferentiating
step of the reaction. Herein, we disclose the results of this study,
which include the discovery of a new set of octapeptide catalysts
that are effective for the kinetic resolution of a number of
unfunctionalized substrates.
Our hypothesis was predicated on the fact that each of the
active catalysts will exhibit a level of activity associated with
the core alkylimidazole resident in each peptide. (Alkylimida-
zoles are fully competent catalysts in the absence of the
peptide.15) Then, the most active of the catalysts would exhibit
an additional increment of rate acceleration that could arise from
a secondary effect in the rate-determining transition state (TS1
versus TS2, Scheme 3). The increment could come from a well-
placed hydrogen bond, a π-stacking interaction, favorable ion-
pairing geometries, etc. The essential feature of the fastest
catalysts therefore could be a multi-point contact between
catalyst and substrate in the rate-determining TS. Therefore, we
speculated that since the peptide-based catalysts are chiral and
optically pure, and the substrates are racemic, on average the
rate accelerations would be accompanied by enantioselectivity.
Screening catalysts en masse could allow a means of evaluating
the hypothesis: that selection of catalysts that exhibit substantial
rate acceleration would lead to catalysts that also exhibited
appreciable enantioselectivities. Of course, this approach could
select for catalysts that accelerate both enantiomers of a racemate
equally, leading to low krel values for some catalysts that also
turn out to be highly active. However, such a screening approach
substantially narrows the field of catalyst candidates, focusing
our investigation exclusively on those that are highly active.
Results and Discussion
Assay, Library Design, and Screening. The assay that we
have described previously relies upon proton-activated fluores-
cence within a bead such as that illustrated by structure 3
(Scheme 2).6,14 Thus, as an acylation reaction between an alcohol
(10) (a) Copeland, G. T.; Jarvo, E. R.; Miller, S. J. J. Org. Chem. 1998,
63, 6784-6785. (b) Miller, S. J.; Copeland, G. T.; Papaioannou, N.;
Horstmann, T. E.; Ruel, E. M. J. Am. Chem. Soc. 1998, 120, 1629-1630.
(11) For reports of techniques that allow high throughput assays for
enantioselectivity of each catalyst within a library, see: (a) Korbel, G. A.;
Lalic, G.; Shair, M. D. J. Am. Chem. Soc. 2001, 123, 361-362. (b) Guo,
J.; Wu, J.; Siuzdak, G.; Finn, M. G. Angew. Chem., Int. Ed. 1999, 38, 1755-
1758. (c) Reetz, M. T.; Becker, M. H.; Klein, H.-W.; Stockigt, D. Angew.
Chem., Int. Ed. 1999, 38, 1758-1761.
Scheme 3
(12) For alternative representative nonenzymatic catalysts that effect
kinetic resolution of racemic alcohols, see: (a) Fu, G. C. Acc. Chem. Res.
2000, 33, 412-420. (b) Vedejs, E.; Daugulis, O. J. Am. Chem. Soc. 1999,
121, 5813-5814. (c) Spivey, A. C.; Fekner, T.; Spey, S. E. J. Org. Chem.
2000, 65, 3154-3159. (d) Sano, T.; Miyata, H.; Oriyama, T. Enantiomer
2000, 5, 119-123. (e) Kawabata, T.; Nagato, M.; Takasu, K.; Fuji, K. J.
Am. Chem. Soc. 1997, 119, 3169-3170.
(13) S-values ()krel) were calculated according to the method of Kagan.
See: Kagan, H. B.; Fiaud, J. C. Top. Stereochem. 1988, 18, 249-330.
(14) For pioneering studies on the use of aminomethylanthracenes as
pH and metal-ion sensors, see: (a) Fluorescent Chemosensors for Ion And
Molecule Recognition; Czarnik, A. W., Ed.; American Chemical Society:
Washington, DC, 1993. (b) de Silva, A. P.; Gunaratne, H. Q. N.;
Gunnlaugsson, T.; Huxley, A. J. M.; McCoy, C. P.; Rademacher, J. T.;
Rice, T. E. Chem. ReV. 1997, 97, 1515-1566.
Our library design is illustrated below (Figure 1). Specifically,
an octapeptide format was selected wherein the first and last
amino acids were held constant with alanine and π-(Me)-
histidine (Pmh), respectively. Fourteen unique amino acid
(15) Nucleophilic versus general base catalysis with alkylimidazoles has
been a subject of some debate. We have adopted the nucleophilic paradigm
for this analysis. (a) Guibe-Jampel, E.; Bram, G.; Vilkas, M. Bull. Soc.
Chim. Fr. 1973, 1021-1027. (b) Ho¨fle, G.; Steglich, W.; Vorbru¨ggen, H.
Angew. Chem., Int. Ed. Engl. 1978, 17, 569-583. (c) See: Pandit, N. K.;
Connors, K. A. J. Pharm. Sci. 1982, 71, 485-491.

6498 J. Am. Chem. Soc., Vol. 123, No. 27, 2001
Copeland and Miller
mobilized,18 and in particular in single bead assays, we sought
to determine the structures of the most interesting catalysts, and
then to screen them under more favorable homogeneous
conditions.
Catalyst Identification and Validation. The structures of
the most selective catalysts were determined by submitting the
beads to single bead Edman degradation analysis.19 The
sequences that were obtained for two of the peptides are shown
in Figure 3a. Resynthesis and rescreening of these peptide
catalysts (2.5 mol %) were then carried out under homogeneous
conditions. Kinetic resolutions were performed that verified that
selective catalysts had indeed been obtained. In particular,
peptide 4 exhibits a krel ) 8.2 for the resolution of sec-
phenylethanol 2 (-65 °C, 40% conversion). Peptide 5 affords
a krel ) 5.9 in the same resolution, but preferentially acylates
the opposite enantiomer of the racemic substrate, despite
possessing the same absolute configuation of the key nucleo-
philic amino acid. Furthermore, these results stand in contrast
to the best previously disclosed peptide-based asymmetric
acylation catalysts from our laboratory; none was found to be
selective (krel < 1.8) with substrate 2 under optimized conditions.
In addition to the appreciable enantioselection exhibited by
catalysts 4 and 5, their exceptional catalytic actiVity is of note.
In particular, the reaction rate exhibited by peptide 4 (toward
the faster reacting enantiomer) is substantially greater than that
of N-methylimidazole (NMI), highlighting the fact that the
enantioselectivity observed is due to accelerative effects. Indeed,
for the fast reacting enantiomer, peptide 4 exhibits activity
greater than N,N-(dimethylamino)pyridine (DMAP), which is
known to be substantially more active than NMI (Figure 3b).15
Figure 1. Library design and composition.
monomers were then incorporated in the remaining six positions
using the split-and-pool method. The theoretical diversity of
such a library (146) is 7.5 million unique catalysts.16 We
employed 5 g of 400-500 µm beads in our synthesis; therefore,
we estimate that our library contained approximately 100,000
unique peptide catalysts on beads containing a uniform loading
of fluorophore.17
A sample of the library was selected at random and screened
for the reaction of racemic sec-phenylethanol (2). Specifically,
about 1000 catalysts were screened conveniently over the course
of a 60 min period. From the assay (Figure 2a), 27 of the
brightest beads were manually selected, and 16 of the least bright
beads were selected for secondary screening. The beads were
then washed and subjected to single bead kinetic resolutions,
with encouraging results. In the single-bead, on-resin assay, the
k
rel values of the brightest beads were, on average, higher than
those observed when the darker beads were employed as
catalysts (Figure 2b). Since, in our experience, the enantio-
selectivities exhibited by the peptide-based catalysts we have
been studying tend to be lower when the catalysts are im-
Figure 2. (a) Fluorescence micrograph showing distribution of bead intensities during catalytic reaction. (b) On-resin krel data for single bead
assays. Data for the “bright” beads are to the left; data for the “dark” beads are to the right.

Kinetic Resolution of Secondary Alcohols of Broad Structural Scope
J. Am. Chem. Soc., Vol. 123, No. 27, 2001 6499
Figure 4. (a) Theoretical composition of the second-generation split/
pool library. (b) Structure of the noncleavable library to be screened
“on-resin.” (c) Format of the co-functionalized library. The strand with
the safety catch linker is cleaved into the multi-well plate for screening.
The other strand is retained on the bead for single bead Edman
degradation.
Figure 3. (a) New peptide catalysts that exhibit appreciable krel values
for unfunctionalized substrate 2. (b) Plot verifying the high activity of
peptide 4 relative to DMAP.
library was obtained that possessed the theoretical composition
shown in Figure 4a. Of the approximately 6,000 unique
structures that would be represented,22 5% were likely to be
identical to the parent structure, providing an important internal
control. Eighteen percent of the library would differ from the
parent by only one amino acid, but that residue might be inserted
anywhere in the sequence (except the final residue which was
always BOC-Pmh). Thirty percent would differ from the parent
by two residues, and so on, as indicated in Figure 4a.
The second generation library was independently synthesized
twice so that single bead screening experiments could be
conducted with two different formats. In both cases, a random
sample of the biased library was screened in a multi-well format.
In the first synthesis, the Wang linker was used to prepare the
second generation library. Single bead assays were then
conducted in each well of a multi-well plate with the catalyst
immobilized on the bead (Figure 4b). In the second case, the
library was synthesized with beads that were co-functionalized
with both a permanent, noncleavable linker and also the safety
catch linker,23 such that catalyst samples could be cleaved
directly into a multi-well plate (Figure 4c). This latter approach
would allow for “single bead” screening of peptides, but under
homogeneous conditions. It was hoped that this latter approach
might provide a more realistic prediction of how the catalysts
derived from a single bead might perform once they were
resynthesized, purified, and rescreened under homogeneous
conditions. For this latter format of the second generation library,
cleavage was accomplished such that the C-terminal N-propyl
amide was obtained.
These observations underscore the fact that the fluorescence-
based activity screen is effective in identifying highly active
catalysts.
Directed Library Synthesis and Screening. With catalyst
4 identified from the fluorescence-based activity screen, we
sought to identify catalysts that would provide enhanced degrees
of selectivity for the kinetic resolution of the unfunctionalized
substrates such as sec-phenylethanol. For this goal, we synthe-
sized a second generation split and pool peptide library, such
that the members were biased toward the structure of the best
“hit” structure (4). The new library was sufficiently different,
however, to allow for the identification of improved catalysts.20
Experimentally, a split and pool library was synthesized wherein
the beads were partitioned unevenly at each split step: 65% of
the beads were functionalized with the residue corresponding
to that of parent peptide 4 at each position. The remaining 35%
of the beads were divided into 14 equal portions, and coupled
with 14 unique amino acid monomers.21 Accordingly, a new
(16) This value is achieved by the randomization of 14 monomers in 6
unique positions ()146).
(17) The 400-500 µm beads contain ∼20-25 beads per mg. 5 g of
beads therefore contains 100 000 to 125 000 beads.
(18) For representative examples of resin-bound catalysts that afford
similar selectivities on-resin and in solution, see: (a) Clapham, B.; Cho,
C. W.; Janda, K. D. J. Org. Chem. 2001, 66, 868-873. (b) Shimizu, K. D.;
Cole, B. M.; Krueger, C. A.; Kuntz, K. W.; Snapper, M. L.; Hoveyda, A.
H. Angew. Chem., Int. Ed. Engl. 1997, 36, 1703-1707.
(19) Single bead Edman degradations were carried out by Midwest
Analytical (St. Louis, MO).
(20) (a) This technique is related to the “Directed Evolution” approach
to enzyme optimization. See: Arnold, F. H. Acc. Chem. Res. 1998, 31,
125-131 and references therein. (b) Liebeton, K.; Zonta, A.; Schimossek,
K.; Nardini, M.; Lang, D.; Dijkstra, B. W.; Reetz, M. T.; Jaeger, K.-E.
Chem. Biol. 2000, 7, 709-718.
(21) Figure 4a specifically relates to the version of the library that was
synthesized with the safety catch linker. The version of the library employing
the Wang linker involved a C-terminus that was fixed as alanine, and was
consequently variable at only six positions. See Supporting Information
for details.
The single bead assays were carried out under kinetic
resolution conditions where sec-phenylethanol (2) was used as
the substrate. Reactions were quenched after a uniform period
of time to allow for a comparison of the relative activity and
selectivity afforded by each catalyst. A conventional chiral GC
(22) The library size is limited by the number of beads used in the split
and pool synthesis. See Supporting Information for details.
(23) Backes, B. J.; Virgilio, A. A.; Ellman, J. A. J. Am. Chem. Soc.
1996, 118, 3055-3056.

6500 J. Am. Chem. Soc., Vol. 123, No. 27, 2001
Copeland and Miller
assay was employed for this purpose. The results obtained with
the second generation library are displayed in Figure 5. Figure
5a depicts the activity and selectivity afforded by the catalysts
that were screened while still immobilized on the beads (i.e.,
the library shown in Figure 4b). Figure 5b shows the data for
the library that was subjected to cleavage such that the
C-terminal N-propyl amides were deposited into wells, and
screened in solution (i.e., the library depicted in Figure 4c). A
number of observations are worthy of note: (i) These data
unambiguously reveal new catalysts of both improved absolute
activity and improved enantioselectivity relative to the parent
peptide (4). (ii) The majority of the library consists of catalysts
which afford krel values of near unity for the kinetic resolution
of sec-phenylethanol. This is perhaps expected since the majority
of peptide-based catalysts in this library differ substantially from
the parent structure upon which the library was based (cf., Figure
4a). (iii) In both formats (Figure 5a,b), the second generation
library reveals a modest trend among those catalysts that are
approximately as selective, or more so, than the parent: there
appears a loose correlation between the more active and the
more selective of the catalysts. (iv) As a control, shown in red
in Figure 5a is a cluster of catalysts that are identical in structure
to the parent peptide sequence (4). Notably, these exhibit on-
resin krel values of ∼2.0 for the kinetic resolution, providing an
important point of calibration. (v) The second generation
directed library also contained a number of catalysts that
preferentially acylate the opposite enantiomer of the substrate,
but with lower selectivity. (Catalysts selective for the (S)-
enantiomer are shown with blue dots; catalysts selective for the
(R)-enantiomer are denoted with green dots.)
Directed Library Results. Because the single bead assays
afforded catalysts exhibiting enhanced enantioselectivity relative
to the parent (4), it became critical to validate the data shown
in Figure 5 through catalyst identification and rescreening. Four
of the most selective beads from each format of the directed
library were submitted to single bead Edman degradation to
determine the structures of the catalysts. The sequences of the
eight new peptides are shown in Table 1. Each sequence was
then independently synthesized as the C-terminal methyl ester
Figure 5. Plots of “activity vs selectivity” for kinetic resolutions of
sec-phenylethanol employing the second generation split and pool
peptide library. The two plots were obtained from two directed libraries
independently synthesized and independently screened.
Table 1. Sequences for Eight New Peptides Selected from the Second Generation Libraries and Kinetic Resolution Data for sec-Phenylethanol
a Single bead screens were conducted either on resin (catalysts 6-9) or following single-bead cleavage to the N-propyl amide (catalysts 10-13).
Kinetic resolutions were conducted at 25 °C. b Solution krel values were obtained with HPLC purified peptides under homogeneous conditions at
-65 °C. See Experimental Section for details.

Kinetic Resolution of Secondary Alcohols of Broad Structural Scope
J. Am. Chem. Soc., Vol. 123, No. 27, 2001 6501
Table 2. Kinetic Resolution Results Employing Peptide 13 (2.5
and purified by reverse-phase HPLC. Also shown in Table 1
are the krel values obtained with each catalyst in (i) the single
bead assay and (ii) the kinetic resolution following resynthesis
and purification. Peptides 6-9 were identified from the “on-
resin” single bead assays, and upon rescreening in solution
produced krel values that were comparable to peptide 4. Peptide
9 actually affords a krel value of 9.6, which is perhaps an
experimentally significant improvement over peptide 4 (cf., krel
) 8.2). Peptides 10-13 were identified from the screens of
peptides cleaved to the C-terminal N-propyl amide. Each of these
catalysts also performs similarly, or better than parent peptide
catalyst 4. Among these peptides, catalysts 12 and 13 afford
krel values that are 12 and 20, respectively. These results reflect
unambiguous improvements for enantioselectivity with respect
to catalyst 4. These results stand in stark contrast to our
originally reported peptide-based catalysts that afforded es-
sentially no enantioselectivity for unfunctionalized substrates.
mol %) as a Catalysta
It is important to note that the single bead assays (either with
resin-bound peptides or those that had been cleaved from single
beads to the N-propyl amide) do not provide an exact prediction
of which will be the most enantioselective upon rescreening
under homogeneous conditions as the methyl ester. For example,
catalyst 10 affords a krel value of 3.6 in the single bead assay,
and a krel value of 8.1 under optimized homogeneous conditions;
catalyst 13 affords a krel value of 3.2 in the single bead assay,
but a krel value of 20 under homogeneous conditions. While
the reasons for this are currently under study, from a pragmatic
point of view, the selection process meets its principle objec-
tive: identification of individual catalysts that are most interest-
ing for further study. It is the case that the rescreening
experiments (homogeneous conditions with purified peptides)
are conducted under different conditions than the selection
experiments (single bead assays), which may account for
differences that we observe. Nevertheless, while screens of the
type we have employed may miss certain selective catalysts,
the approach effectively narrows the field of catalyst candidates
to a manageable number for further study by conventional
means.
a Reactions were carried out at -65 °C in PhMe (2.5 mol % 13).
b
Reactions were run to 40-50% conversion. krel values were deter-
The structures of the new catalysts are of interest with respect
to the selectivities that each affords; in particular, the improve-
ments afforded by 12 and 13 are noteworthy. All eight catalysts
share sequence homology with respect to the parent (4) in the
i+2 position (D-Val), the i+3 position (L-(trt)His), and the i+5
position (D-Val). But, the two most selective catalysts lack the
L-Ala of the parent in the C-terminal position and have L-Leu
(12) or L-Ile (13) in its place. Perhaps this change in a remote
position relative to the Pmh residue reflects conformational
proximity of the N-terminal and C-terminal residues of the
octapeptide during the stereochemistry-determining transition
state. Conformational studies of peptides 4, 12, and 13 will
investigate this possibility. Preliminary studies reveal that
peptide 13 exhibits a very sharp 1H NMR spectrum (CDCl3, 25
°C) that may reflect a unique, monomeric conformation in
solution.
mined according to the method of Kagan (ref 13) employing chiral
GC or HPLC analyses. See Supporting Information for details.
evaluating the scope of a given catalyst without respect to
optimization on a per substrate basis. Notably, catalyst 13
provides a krel > 50 for R-naphthyl-substituted alcohol 14 (entry
2; cf. krel ) 20 for compound 2, entry 1). Para substitution results
in a modest drop in selectivity with p-methoxy-bearing substrate
15 affording krel ) 16 (entry 3; p-fluoro-substituted substrate
16 affords krel ) 11, entry 4). trans-2-Phenylcyclohexanol (17)
is an excellent substrate for kinetic resolution, undergoing
resolution with a krel > 50 under these conditions (entry 5).
Higher degrees of substitution on aryl-alkyl secondary alcohols
also modulate the observed selectivity. For example, sec-
phenylpropanol (18) exhibits krel ) 8.2 with catalyst 13 (entry
6) and tert-butyl-substituted alcohol 19 is resolved with krel
)
30 (entry 7). Finally, we examined whether secondary alcohols
that do not possess an aromatic substituent might be substrates
for kinetic resolution with catalyst 13. Indeed, cyclohexyl
Generalization to Other Substrates. The finding that
catalyst 13 affords appreciable enantioselectivity in the kinetic
resolution of an unfunctionalized substrate is one that could have
implications for the application of peptide-based enantioselective
catalysts across a broader substrate scope than had been
previously possible. To examine this notion, we investigated
octapeptide 13 as a catalyst for kinetic resolution of a range of
substrates (Table 2). Reactions were conducted with a uniform
set of conditions (2.5 mol % catalyst 13, -65 °C, 40-50%
conversion). This decision was made with the goal of rapidly
ethanol 20 is resolved with a moderate, but significant krel
)
9.0 (entry 8). Even sec-butanol 21 is resolved with a krel ) 4.0
under these conditions (entry 9), indicating that the catalyst can
discriminate to some degree between a methyl group and an
ethyl group. These latter two substrates are rarely reported in
studies on nonenzymatic kinetic resolution. The appreciable
kinetic resolution of these substrates suggests that catalyst 13

6502 J. Am. Chem. Soc., Vol. 123, No. 27, 2001
Copeland and Miller
may well be a reasonable lead catalyst for further optimization
with respect to these types of substrates.24
by rational design methods. Nevertheless, from diversity-
oriented screens we have found catalysts that are effective for
kinetic resolutions of unfunctionalized substrates. These same
octapeptides represent reasonable lead structures for the devel-
opment of new catalysts that could benefit from further diversity-
based screening techniques, and also mechanistic studies. The
application of screens for rate acceleration with catalyst libraries
thus offers a potentially powerful method for the rapid identi-
fication of new chiral catalyst motifs that are both mechanisti-
cally interesting and of potential utility to the synthetic
community.
Conclusions
Collectively, the results in Table 2 reveal catalyst 13 to be
somewhat general in scope. In addition, these data indicate that
peptide-based catalysts can be found for asymmetric acylation
of a wide variety of alcohol types. Such catalysts are not limited
to substrates that bear acetamide functional groups. In addition,
it is worth noting that catalyst 13 was identified from combi-
natorial screening techniques that were designed for evaluation
of sec-phenylethanol. Yet, catalyst 13 provides enantioselection
with a range of different alcohols. Presumably, screening similar
catalyst libraries against a particular substrate would result in
the identification of different catalysts that would afford different
and perhaps improved selectivities for these and other alcohols.
That one such set of screens resulted in the identification of an
effective catalyst for the targeted substrate (as well as others)
bodes well for the application of this approach to a wide variety
of structural types.
Acknowledgment. This research was supported by the U.S.
National Science Foundation (CHE-9874963) and the National
Institutes of Health (GM-57595). G.T.C. is grateful to Smith-
Kline Beecham and the Organic Division of the ACS for a
Graduate Fellowship. In addition, we are grateful to DuPont,
Eli Lilly, Glaxo-Wellcome, and the Merck Chemistry Council
for research support. S.J.M. is a Fellow of the Alfred P. Sloan
Foundation, a Cottrell Scholar of Research Corporation, and a
Camille Dreyfus Teacher-Scholar.
The mechanistic basis for the enantioselectivities achieved
with the peptides reported in this article is unclear at this time.
It is our contention that the sequences of peptides 4 and 13 are
sufficiently nonobvious that they would not have been identified
Supporting Information Available: Experimental details
for all experimental aspects of the study, including compound
characterization (PDF). This material is available free of charge
via the Internet at http://pubs.acs.org.
(24) Reactions may be run with reduced catalyst loading (0.5 mol %)
without erosion of krel or conversion. In these experiments, Hu¨nig’s base
(0.4 equiv) is employed as an additive. See Supporting Information for
details.
JA0108584