Enzymatic, Polymer-Supported Formation of an Analog of the Trypsin Inhibitor A90720A: A Screening Strategy for Macrocyclic Peptidase Inhibitors

Full text of DOI:10.1021/ja971216c

J. Am. Chem. Soc. 1997, 119, 12697-12698
Enzymatic, Polymer-Supported Formation of an
12697
Analog of the Trypsin Inhibitor A90720A: A
Screening Strategy for Macrocyclic Peptidase
Inhibitors
Matthew T. Burger and Paul A. Bartlett*
Department of Chemistry, UniVersity of California
Berkeley, California 94720-1460
ReceiVed April 17, 1997
Macrocyclic constraint is effective in reducing conformational
flexibility and thereby enhancing binding affinity and metabolic
stability in enzyme inhibitors and other protein ligands.1 Such
derivatives present challenges both in design and synthesis: ring
systems must be identified that induce the correct conformation
in the binding region while avoiding unfavorable contacts with
the receptor, and a significant investment in effort is required
to prepare the analogs. A rapid and convenient method for
identifying favorable ring systems prior to synthesis would be
useful, especially if the method could be applied as a combi-
natorial approach. In this paper, we propose such a strategy
for the discovery of macrocyclic peptidase inhibitors, demon-
strating the key features of the method with the synthesis of an
analog (2) of A90720A (1), a naturally-occurring inhibitor of
trypsin.2
Figure 1. Strategy for distinguishing, from a library of analogs,
molecules that cyclize from those that do not.
linkage were replaced with a transition state analog). Since both
the forward (hydrolysis) and reverse (amide synthesis) reactions
catalyzed by a peptidase involve the same intermediate,
discovering which linear derivatives are readily cyclized by the
enzyme should also point to structures that would make good
macrocyclic inhibitors.3 Linear molecules are inherently easier
to synthesize than macrocycles, and a variety of strategies are
available for inducing peptide bond formation by a peptidase
(activated precursors, lower pH, low-water content, etc.4).
Key for the success of this strategy is a simple assay for
cyclization that can identify individual beads from an encoded
library and thus take full advantage of the combinatorial
approach.5 The scheme depicted in Figure 1 relies on the
cyclization reaction to establish a connection between the
polymeric synthesis support and a dye that is stable to conditions
which cleave the molecule at a different point. Cleavage
removes the dye from beads carrying derivatives that did not
cyclize, while dye is retained on those that have the desired
analogs. We have reduced to practice the major elements in
this strategy by synthesizing macrocycle 2 on a solid phase by
enzymatic cyclization and demonstrating that beads with cy-
clized and uncyclized molecules can be differentiated by
hydrolysis of the ester linkage.
A90720A (1) furnished an ideal test system for this approach;
it binds to trypsin with the Thr-Arg-Ahp [Ahp ) 3-amino-6-
hydroxy-2-piperidone] tripeptide in the S2-S1-S1′ subsites, with
the tyrosine N-methyl projecting away from the enzyme, and
the sulfoglyceryl-Leu residues in an open region on the protein
surface.2 These positions on A90720A were thus logical places
to locate the dye molecule (a derivative of Disperse Red 1) and
the tether point, respectively, since they are on opposite sides
of the cleavable ester linkage. Design of the synthetic target 2
was completed by substituting Ser for the Ahp subunit for ease
of synthesis, and replacing the sulfoglyceryl-Leu moiety with
a simple acid- and base-stable attachment to a photocleavable
tether to the resin. The acyclic precursor was assembled in
solution in protected form (see the Supporting Information),
We reasoned that a ring system that is favorably bound and
turned over as a substrate by the enzyme should be similarly
effective in the form of an inhibitor (for example, if the scissile
(1) Inter alia: Ksander, G. M.; de Jesus, R.; Yuan, A.; Ghai, R. D.;
Trapani, A.; McMartin, C.; Bohacek, R. J. Med. Chem. 1997, 40, 495.
Ksander, G. M.; de Jesus, R.; Yuan, A.; Ghai, R. D.; McMartin, C.; Bohacek,
R. J. Med. Chem. 1997, 40, 506. Ettmayer, P.; Billich, A.; Hecht, P.;
Rosenwirth, B.; Gstach, H. J. Med. Chem. 1996, 39, 3291. Kim, C. U.;
McGee, L. R.; Krawczyk, S. H.; Harwood, E.; Harada, Y.; Swaminathan,
S.; Bischofberger, N.; Chen, M. S.; Cherrington, J. M.; Xiong, S. F.; Griffin,
L.; Cundy, K. C.; Lee, A.; Yu, B.; Gulnick, S.; Erickson, J. W. J. Med.
Chem. 1996, 39, 3431. March, D. R.; Abbenante, G.; Bergman, D. A.;
Brinkworth, R. I.; Wickramasinghe, W.; Begun, J.; Martin, J. L.; Fairlie,
D. P. J. Am. Chem. Soc. 1996, 118, 3375. Morgan, B. P.; Bartlett, P. A.;
Holland, D. R.; Matthews, B. W. J. Am. Chem. Soc. 1994, 116, 3251.
Podlogar, B. L.; Farr, R. A.; Friedrich, D.; Tarnus, C.; Huber, E. W.; Cregge,
R. J.; Schirlin, D. J. Med. Chem. 1994, 37, 3684. Hruby, V. J.; Bonner, G.
G. Methods Mol. Biol. 1994, 35, 201. Davies, J. S. Amino Acids, Pept.,
Proteins 1994, 25, 246. MacPherson, L. J.; Bayburt, E. K.; Capparelli, M.
P.; Bohacek, R. S.; Clarke, F. H.; Chai, R. D.; Sakane, Y.; Berry, C. J.;
Peppard, J. V.; Trapani, A. J. J. Med. Chem. 1993, 36, 3821. Yang, L.;
Weber, A. E.; Greenlee, W. J.; Patchett, A. A. Tetrahedron Lett. 1993, 34,
7035.
(3) Because some ring systems are inherently easier to form, no direct
relationship between the rate of enzymatic cyclization and transition state
binding affinity is expected (Radzicka, A.; Wolfenden, R. Methods Enzymol.
1995, 249, 284); however, the fact that ring systems which are easily formed
would be favored in this screen may have practical benefits in identifying
macrocyclic inhibitors that are easier to synthesize.
(4) Heiduschka, P.; Dittrich, J.; Barth, A. Pharmazie 1990, 45, 164.
(5) Joyce, G. F.; Still, W. C.; Chapman, K. T. Curr. Opin. Chem. Biol.
1997, 1, 3 and companion articles in that issue.
(2) Bonjouklian, R.; Smitka, T. A.; Hunt, A. H.; Occolowitz, J. L.; Perun,
T. J., Jr.; Doolin, L.; Stevenson, S.; Knauss, L.; Wijayaratne, R.; Szewczyk,
S.; Patterson, G. M. L. Tetrahedron 1996, 52, 395. Lee, A. Y.; Smitka, T.
A.; Bonjouklian, R.; Clardy, J. Chem. Biol. 1994, 1, 113.
S0002-7863(97)01216-X CCC: $14.00 © 1997 American Chemical Society

12698 J. Am. Chem. Soc., Vol. 119, No. 51, 1997
Communications to the Editor
attached to PEGA (polyethyleneglycol-polyacrylamide)6 beads
bearing the photolinker,7 and deprotected to the methyl ester
3b.8 Photolytic detachment of this material in methanol at 366
nM yielded 3a in solution as a single compound, as characterized
1
by H NMR, HRMS, and RPLC.
Figure 2. A mixture of beads gathered before and after treatment with
trypsin, combined, and saponified.
treated with trypsin prior to ester cleavage, the red color is not
removed by the saponification step, and the hydrolyzed product
7 can be identified in the eluant after photolysis. The difference
in coloration is illustrated in Figure 2, which shows a mixture
of beads that were taken before (3b) and after (2b and 4b)
trypsin treatment, combined, and saponified together.
Not surprisingly, analog 2a is a substrate for trypsin in
aqueous solution (kcat ≈ 0.1 s-1 in 5% DMSO, pH 8.0), but it
persists long enough at low enzyme concentrations (22 nM)
that an inhibition constant of 230 ( 30 nM can be determined.11
The binding affinity of 2a is only 25-fold less than that of
A90720A itself (Ki ) 9 nM),2 reflecting loss of the active site
contacts of the sulfoglyceryl-Leu moiety of the natural product
and the rigidification induced by the Ahp residue. The acyclic,
unconstrained hydrolysis product 4a is a weak inhibitor: Ki )
15 µM. Although modest alterations in substrate selectivity may
be expected for enzymatic reactions in organic versus aqueous
solutions,12 it is clear that the ability of 3a to cyclize in the
mixed solvent translates into the ability of 2a to inhibit the
enzyme in water.
An eventual goal of this project is to determine whether the
information needed for discovery of potent, macrocyclic pep-
tidase inhibitors can be obtained more easily than through
iterative design and synthesis or by the preparation and assay
of libraries of macrocyclic structures. The results presented here
demonstrate the essential elements that make screening a library
of linear analogs for enzymatic cyclization a viable approach.
This method should be general for a variety of proteases,
macrocyclic motifs, and cleavage sites, requiring only that the
enzyme be able to function in the direction of amide synthesis.
Cyclization of soluble amino ester 3a by trypsin was readily
achieved in 30:30:40 DMF/ethanol/pH 6.5 Tris buffer, affording
a 2:1 ratio of cyclic/hydrolyzed products 2a and 4a. Conditions
for on-bead cyclization of 3b were explored by incubation with
trypsin, photolysis, and analysis for precursor 3a, macrocycle
2a, and the hydrolysis product 4a by HPLC. In a preparative
experiment, treatment of 2 mg of 3b (ca. 2000 beads, 0.6 µmol)
with 40 mg (ca. 3 mol equiv) of trypsin in the mixed solvent
system for 26 h, followed by washing and photolysis, gave a
1:1 mixture of 2a and 4a.9 The formation of both products
presumably proceeds via the acyl-enzyme as a common
intermediate, which partitions to the hydrolyzed product 4b by
attack of water and to the macrocycle 2b by intramolecular
attack of the terminal amine.10 Only a small amount of
precursor 3a was present in the eluant, indicating that most of
the material on the resin is accessible to the enzyme; a control
experiment without the enzyme afforded only the unaltered
amino ester 3a after photolysis.
Treatment of the resin-bound precursor 3b with 1:1 metha-
nol/1 M NaOH for 7 h cleaves the Val-Thr ester linkage, and
the beads turn colorless after filtration and washing. MS
analysis of the filtrate confirms the presence of 5, and after
photolysis of the remaining resin, the expected dipeptide
derivative 6 is released. In contrast, from beads that have been
(6) Meldal, M.; Svendsen, I.; Breddam, K.; Auzanneau, F.-I. Proc. Natl.
Acad. Sci. U.S.A. 1994, 91, 3314. Meldal, M.; Auzanneau, F.-I.; Hindsgaul,
O.; Palcic, M. M. J. Chem. Soc., Chem. Commun. 1994, 1849.
(7) Holmes, C. P.; Jones, D. G. J. Org. Chem. 1995, 60, 2318.
(8) We first explored Tentagel (PEG-polystyrene) as the support;
however, no products from enzymatic transformation, either cyclization or
hydrolysis, were detected under a variety of conditions (Lowe, G.; Quarrell,
R. METHODS: A Companion to Methods in Enzymology 1994, 6, 411.
Va´gner, J.; Barany, G.; Lam, K. S.; Krchna´k, V.; Sepetov, N. F.; Ostrem,
J. A.; Strop, P.; Lebl, M. Proc. Nat. Acad. Sci. U.S.A. 1996, 93, 8194).
(9) It should be noted that 2 mg of resin beads may represent more than
1000 compounds in the context of a combinatorial library; hence, even
without optimization, the conditions of the preparative experiment could
be applied efficiently as a screen.
Acknowledgment. This work was supported by a grant from the
National Institutes of Health (grant no. GM-30759) and an NIH
postdoctoral fellowship to M.T.B.
Supporting Information Available: Experimental procedures for
preparation of intermediates, solution synthesis, characterization, and
enzymatic evaluation of macrocycle 2a (17 pages). See any current
masthead page for ordering and Internet access instructions.
JA971216C
(11) Erlanger, B. F.; Kokowsky, N.; Cohen, W. Arch. Biochem. Biophys.
1961, 95, 271.
(12) Jakubke, H.-D.; Kuhl, P.; Ko¨nnecke, A. Angew. Chem., Intl. Ed.
Engl. 1985, 24, 85.
(10) The ratio of 4a:2a did not appear to increase during the course of
the reaction, indicating that 4b did not arise from hydrolysis of 2a.