From peptides to non-peptide peptidomimetics: design and synthesis of new piperidine inhibitors of aspartic peptidases.

Article abstract of DOI:10.1021/ol016092u

[reaction: see text] The 3-alkoxy-4-arylpiperidine inhibitors of aspartic peptidases are shown to be a new type of non-peptide peptidomimetic inhibitor. These piperidines can be designed from peptide-derived inhibitors by use of a structure-generating pro

Full text of DOI:10.1021/ol016092u

ORGANIC
LETTERS
2001
Vol. 3, No. 15
2317-2320
From Peptides to Non-Peptide
Peptidomimetics: Design and Synthesis
of New Piperidine Inhibitors of Aspartic
Peptidases
,†,‡
Matthew G. Bursavich,† Chris W. West,† and Daniel H. Rich*
Department of Chemistry, UniVersity of WisconsinsMadison, 1101 UniVersity AVenue,
Madison, Wisconsin 53706, and School of Pharmacy, UniVersity of Wisconsins
Madison, 777 Highland AVenue, Madison, Wisconsin 53705
dhrich@facstaff.wisc.edu
Received May 9, 2001
ABSTRACT
The 3-alkoxy-4-arylpiperidine inhibitors of aspartic peptidases are shown to be a new type of non-peptide peptidomimetic inhibitor. These
piperidines can be designed from peptide-derived inhibitors by use of a structure-generating program but only after the enzyme active site
conformation has been modified in a mechanistically related fashion. New enantioselective syntheses of 3-alkoxy-4-arylpiperidine analogues
are described.
Structure-based design of peptidomimetic peptidase inhibitors
is important for developing new drugs.1 This approach has
proven highly successful with the aspartic peptidase inhibi-
tors,2,3 as illustrated by the development of the HIV protease
inhibitors for treating AIDS.4 Most of these inhibitors have
been designed to emulate the ligand-bound extended â-strand
conformation5 observed with most substrate-based peptidase
inhibitors.6 In this Letter we define these inhibitors as
peptide-derived peptidomimetics,7-9 due to their close rela-
tionship with the enzyme-bound peptide-substrate confor-
mation. A few structurally distinct aspartic protease inhibi-
tors10,11 have been discovered by high-throughput screening
methods and developed into useful HIV protease inhibitors,
e.g., 1 and 2 (Figure 1). We define these inhibitors as non-
peptide peptidomimetics12 because of their remote structural
relationship to peptide substrates. However, X-ray crystal
structures of inhibitors 1 and 2, as well as the rationally
(6) Fairlie, D. P.; Tyndall, J. D. A.; Reid, R. C.; Wong, A. K.; Abbenante,
G.; Scanlon, M. J.; March, D. R.; Bergman, D. A.; Chai, C. L. L.; Burkett,
B. A. J. Med. Chem. 2000, 43, 1271.
† Department of Chemistry, University of WisconsinsMadison.
‡ School of Pharmacy, University of WisconsinsMadison.
(1) Leung, D.; Abbenante, G.; Fairlie, D. P. J. Med. Chem. 2000, 43,
305.
(2) Rich, D. H. Peptidase Inhibitors. In ComprehensiVe Medicinal
Chemistry. The Rational Design, Mechanistic Study and Therapeutic
Application of Chemical Compounds; Hansch, C., Sammes, P. G., Taylor,
J. B., Eds.; Pergamon Press: 1990; pp 391-441.
(3) Babine, R. E.; Bender, S. L. Chem. ReV. 1997, 97, 1359.
(4) Lebon, F.; Ledecq, M. Curr. Med. Chem. 2000, 7, 455.
(5) Smith, A. B., III.; Hirschmann, R.; Pasternak, A.; Guzman, M. C.;
Yokoyama, A.; Sprengeler, P. A.; Darke, P. L.; Emini, E. A.; Schleif, W.
A. J. Am. Chem. Soc. 1995, 117, 11113.
(7) For definitions of different types of peptidomimetics, see: Ripka,
A. S.; Rich, D. H. Curr. Opin. Chem. Biol. 1998, 2, 439.
(8) Wiley: R. A.; Rich, D. H. Med. Res. ReV. 1993, 13, 327.
(9) Freidinger, R. M. Curr. Opin. Chem. Biol. 1999, 3, 395.
(10) Turner, S. R.; Strohbach,; J. W.; Tommasi, R. A.; Aristoff, P. A.;
Johnson, P. D.; Skulnick, H. I.; Dolak, L. A.; Seest, R. P.; Tomich, P. K.;
Bohanon, M. J.; Horng, M.-M.; Lynn, J. C.; Chong, K.-T.; Hinshaw, R.
R.; Watenpaugh, K. D.; Janakiraman, M. N.; Thaisrivongs, S. J. Med. Chem.
1998, 41, 3467.
(11) Hagen, S.; Vara Prasad, J. V. N.; Tait, B. D. AdV. Med. Chem. 2000,
5, 159.
10.1021/ol016092u CCC: $20.00 © 2001 American Chemical Society
Published on Web 07/04/2001

enzyme subsite is close to where the P1 side chain of peptide-
derived inhibitors bind. However, the topography of the
inhibited active site is fundamentally different from the
extended â-strand conformational topography. Piperidine 5
constitutes a new type of non-peptide peptidomimetic inhibi-
tor.
In the preceding two Letters we described our use of a
structure-generating computer program to generate novel
aspartic peptidase inhibitors and to “rediscover” known
peptide-derived inhibitors related to pepstatin.18 Therefore,
we attempted to generate the Roche-type structures in the
active sites of pepsin and Rhizopus chinensis pepsin, two
enzymes not inhibited by the Roche compounds. We began
with the X-ray structure of bis-S-benzyl peptide 619 bound
to porcine pepsin (Figure 2, Supporting Information) and
attempted to grow the piperidine unit from the P1 S-benzyl
side chain.20 Growth from that point on CySta 7 toward the
catalytic carboxyls (Figure 3) generated only acyclic amines,
Figure 1. Non-peptide peptidomimetic peptidase inhibitors.
designed non-peptide peptidomimetic 3,13 complexed with
HIV protease revealed that the inhibited active site topog-
raphy is very similar to the inhibited active site topography
for the peptide-derived peptidomimetics.
In this Letter we show that GrowMol, a computer program
for generating libraries of structures complementary to an
enzyme active site,14 can be used to identify known “non-
peptide peptidomimetics” that bind to a structurally distinct
(nonextended â-strand conformation) aspartic peptidase
active site conformation. Piperidine 4 inhibits renin at about
26 µΜ while the optimized piperidine inhibitor 5 inhibits
renin at low nM concentrations.15 The discovery of these
piperidines as a new class of aspartic peptidase inhibitor
represents a major advance in the design of inhibitors. These
compounds are simple and contain no amide bonds, and some
are orally active. Moreover, portions of both lead compound
4 and optimized inhibitor 5 bind in the active site of human
renin as mechanism-based inhibitors. The binding of the
piperidine nitrogen to the enzyme catalytic carboxyl groups
is similar to the binding of the statine hydroxyl16 and
aminostatine nitrogen17 in peptide-derived inhibitors. In
addition, the binding of the 3-alkoxy group in the S1-S3
Figure 3. GrowMol growth point defined.
e.g., 8 (Figure 4), and it became evident that we could not
generate piperidines such as 4 without altering the active
(12) For the original definition of non-peptide peptidomimetic, see:
Farmer, P. S. In Drug Design; Ariens, E. J., Ed.; Academic Press: New
York, 1980; Vol. 10, p 119. However, those HIV protease inhibitors in
refs 3-5 that do not contain a dipeptide unit are a different type of non-
peptide peptidomimetic.
(13) Lam, P. Y.-S.; Jadhav, P. K.; Eyermann, C. J.; Hodge, C. N.; Run,
Y.; Bacheler, L. T.; Meek, J. L.; Otto, M. J.; Rayner, M. M.; Wong, Y. N.;
Chang, C.-H.; Weber, P. C.; Jackson, D. A.; Sharpe, T. R.; Erickson-Vitanen,
S. Science 1994, 263, 380.
(14) Bohacek, R. S.; McMartin, C.; Guida, W. C. Med. Res. ReV. 1996,
16, 3.
(15) (a) Oefner, C.; Binggeli, A.; Breu, V.; Bur, D.; Clozel, J.P.; D’arcy,
A.; Dorn, A.; Fischli, W.; Gruninger, F.; Guller, R.; Hirth, G.; Marki, H.;
Mathews, S.; Miller, M.; Ridley, R. G.; Stadler, H.; Viera, E.; Wilhelm,
M.; Winkler, F.; Wostl, W. Chem. Biol. 1999, 6, 127. (b) Viera, E.; Binggeli,
A.; Breu, V.; Bur, D.; Fischli, W.; Gu¨ller, R.; Hirth, G.; Ma¨rki, H. P.; Mu¨ller,
M.; Oefner, C.; Scalone, M.; Stadler, H.; Wilhelm, M.; Wostl, W. Bioorg.
Med. Chem. Let. 1999, 9, 1397. (c) Gu¨ller, R.; Binggeli, A.; Breu, V.; Bur,
D.; Fischli, W.; Hirth, G.; Jenny, C.; Kansy, M.; Montavon, F.; Mu¨ller,
M.; Oefner, C.; Stadler, H.; Vieira, E.; Wilhelm, M.; Wostl, W.; Ma¨rki, P.
Bioorg. Med. Chem. Let. 1999, 9, 1403.
Figure 4. GrowMol-generated structures.
site conformation. Aspartic peptidases contain a hairpin turn
structure, or “flap” region, in the enzyme active site that
(16) Rich, D. H. J. Med. Chem. 1985, 28, 263.
(17) (a) Jones, D. M.; Sueiras-Diaz, J.; Szelke, M.; Leckie, B. J.; Beattie,
S. R.; Morton, J.; Neidle, S.; Kuroda, R. J. Pept. Res. 1997, 50, 109. (b)
Sueiras-Diaz, J.; Jones, D. M.; Szelke, M.; Leckie, B. J.; Beattie, C.; Morton,
J. J. Pept. Res. 1997, 50, 239.
(18) (a) Ripka, A. S.; Satyshur, K. A.; Bohacek, R. S.; Rich, D. H. Org.
Lett. 2001, 3, 2309. (b) Dales, N. A.; Satyshur, K. A.; Bohacek, R. S.;
Rich, D. H. Org. Lett. 2001, 3, 2313.
2318
Org. Lett., Vol. 3, No. 15, 2001

moves in a “hinge motion” during peptide inhibitor binding,
with the end of the flap moving up to 4 Å.21 Since flap
opening is a low-energy process that can occur up to 100
times per second for good substrates,22 we decided to open
this enzyme flap about 1 Å (Figure 5A); GrowMol now
Scheme 1
Figure 5. (A, left) Growth point in porcine pepsin active site after
opening of flap region: flap, light brown; S-benzyl, pink; Asp32
and Asp215, red; Try75, yellow; Trp39, orange. (B, right) Active
site of porcine pepsin after flap opening and Tyr75 side-chain
rotation. Note the opening between Tyr75 and Trp39 where C4
phenyl growth occurs.
(>95% ee). The remaining secondary hydroxyl of 15 was
alkyated with NaH and p-bromobenzyl bromide in DMF to
form the Boc-protected piperidine 16. Subsequent removal
of the Boc protecting group with HCl-dioxane provided
piperidine 12, which was used directly in the enzyme assays.
The enantioselective synthesis of piperidine 11 (Scheme
2) required slight modifications in this route. tert-Butyl
generated a series of piperidines, e.g., 9, but not the
4-arylpiperidines. Further examination of the active site by
molecular modeling revealed that a simple -120° rotation
of ø1 in Tyr75 would provide the space needed for growth
at C4 of the piperidine system (Figure 5B). Running
GrowMol now created the 3,4-disubstituted piperidine 10, a
direct analogue of the Roche HTS lead 4. In addition, Roche
scientists had discovered that the C4′-position of the 4-phen-
ylpiperidines could be substituted to produce tight-binding
renin inhibitors. This was possible because a tryptophan
indole in human renin rotated out of the way to provide an
additional binding site for the C4′ substituents. Therefore,
we moved Trp39 in pepsin the same way (Figure 6,
Supporting Information) and GrowMol now created aceto-
naphthone analogue 11, which is closely related to the
optimized Roche renin inhibitors.
Scheme 2
These results show that piperidines 9 and 10 bind to an
active site conformation that is fundamentally different from
the extended â-strand topography, yet the new active site
conformation can be reached via mechanistically related local
conformational changes.
We developed a new and expedient enantioselective
synthesis to prepare piperidine 12 (Scheme 1). N-Protection
of 4-phenyl-1,2,3,6-tetrahydropyridine with Boc2O gave 13.
Sharpless asymmetric dihydroxylation (AD)23 was employed
to generate diol 14. Stereoselective reduction24 of benzylic
alcohol 14 with Raney nickel in refluxing EtOH gave 15
(19) Szewczuk, Z.; Satyshur, K,; Rich, D. H. Unpublished results. Ki )
320 nM against porcine pepsin.
4-oxypiperidine-1-carboxylate was converted to the corre-
sponding tetrahydropyridinyl triflate 17 by use of LDA and
N-phenyltrifluoromethanesulfonamide.25 Palladium-mediated
coupling of triflate 17 with the readily prepared arylboronic
(20) The same process was carried out with R. chinensis pepsin.
(21) (a) Hsu, I.-N.; Delbaere, L. T.J.; James, M. N. G.; Hofmann, T.
Nature 1977, 40, 233. (b) James, M. N. G.; Sielecki, A.; Salituro, F.; Rich,
D. H.; Hofmann, T. Proc. Natl. Acad. Sci. U.S.A. 1982, 79, 6137.
(22) Kcat for aspartic peptidases typically varies from 1 to 100 s-1
.
(23) Sharpless, K. B.; Amberg, W.; Bennani, Y. L.; Crispino, G. A.;
Hartung, J.; Jeong, K. S.; Kwong, H. L.; Morikawa, K.; Wang, Z. M.; Xu,
D.; Zhang, X. L. J. Org. Chem. 1992, 57, 2768.
(24) Synthesis of the related trans-2-phenylcyclohexanol: King, S. B.;
Sharpless, K. B. Tetrahedron Lett. 1994, 35, 5611.
(25) Wustriw, D. J.; Wise, L.D. Synthesis 1991, 993.
Org. Lett., Vol. 3, No. 15, 2001
2319

acid 18 afforded the desired aryltetrahydropyridine 19 in
moderate yield. Utilizing Sharpless AD methods, the enan-
tiomerically pure diol 20 was synthesized. However, Raney
nickel reduction of the benzylic alcohol in phenol-protected
analogues of 20 was not successful.26 After much experi-
mentation, we determined that the desired Raney nickel
reduction of 20 to 21 could be achieved after first removing
the silyl-protecting group. Selective TIPS-protection of the
phenol followed by alkylation of the secondary alcohol with
NaH and p-bromobenzyl bromide in THF provided 22.
Deprotection of the TIPS groups with TBAF in THF gave
phenol 23 whose stereochemical assignments were confirmed
by X-ray structure.27 Alkylation of 23 with potassium
carbonate and 2-bromo-2′-acetonaphthone in acetone af-
forded the desired Boc-protected piperidine 24. Removal of
the Boc protecting group with HCl-dioxane provided
piperidine 11, which was used directly in the enzyme assays.
Inhibition of substrate hydrolysis by porcine pepsin and
R. chinensis pepsin were determined using reported assay
conditions.28 Piperidine 11 inhibited porcine pepsin with an
IC50 ) 0.2 µM and piperidine 12 inhibited R. chinensis
pepsin with an IC50 ) 2 µM. Several of these piperidines
are poorly water-soluble, a property first noted by Oefner et
al.11 Some compounds saturated the buffer before full
inhibition was obtained, which was detected by titrating
enzyme activity from low to high inhibitor concentration.
Our results show that judicious application of a structure-
generating program by medicinal chemists can be used to
convert peptide-derived inhibitors into non-peptide peptido-
mimetics related to those found by HTS methods. To achieve
these results, it was necessary to alter the conformation of
portions of the enzyme active site. These conformational
changes were implemented after careful consideration of
plausible enzyme intermediates that could be formed during
catalysis. All changes were predictable, low-barrier confor-
mational changes. However, it should be emphasized that
our success was made possible only because we knew this
type of piperidine inhibited human renin. In our earlier efforts
to identify novel structures,29 various potential inhibitors were
generated that contained amines but these were not studied
further as we focused on hydroxyl inhibitors. Even when
we decided to evaluate amines as a result of the Roche
publications, the generation of piperidine structures by
GrowMol was rare; only about 1% of the grown structures
contained the piperidine nucleus. Furthermore, these pip-
eridines were obtained only when the growth point (Figure
3) was moved along the hydrophobic surface comprising the
S1-S3 subsites. Finally, successful growth was achieved only
when we began with X-ray structures of moderate inhibitors
(e.g., 6) bound to the enzyme.
GrowMol and the crystal structure of a peptide-derived
inhibitor bound to the enzyme. The process we have
described represents a simple protocol for altering enzyme
active sites to permit design of non-peptide peptidomimetic
inhibitors that bind to novel enzyme active site conformers.
Our results are consistent with recent calculations that show
ligands bind to the dynamic ensemble of preexisting enzyme
conformations30 such that “binding of an inhibitor selectively
stabilizes those conformational states in which the binding
site is formed.” 31 Thus, we propose that the surprising renin
conformations discovered by Oefner et al.11 are not “induced”
but simply preexisting enzyme conformations selectively
stabilized by the inhibitor to afford the crystallographically
observed complex.
Our work demonstrates that a third approach to peptido-
mimetic design can include inhibitors designed to stabilize
potential preexisting enzyme active site conformations in
addition to those observable in both native and inhibitor-
enzyme crystal structures. Rather than design peptidomi-
metics to emulate only the enzyme-bound conformation of
the peptide-derived inhibitor (the extended â-strand topog-
raphy), as is currently done,5 design should be targeted to
the complete ensemble of potential preexisting active site
conformations. Structure-generating programs that allow
systematic variation of the position of the starting growth
point, that permit systematic evaluation of the dynamic
ensembles of preexisting enzyme active site conformations,
and that provide systematic evaluation (scoring) of the grown
structures will greatly accelerate the discovery of non-peptide
peptidomimetic inhibitors.32 Finally, inhibition of two ad-
ditional aspartic peptidases by these piperidines supports the
proposal of Oefner et al.11 that the piperidines may become
general scaffolds for inhibiting this enzyme class. Since
piperidines 11 and 12 are structurally related to paroxetine,33
a known CNS active drug, this scaffold may become
especially effective at inhibiting aspartic peptidases located
in the CNS.
Acknowledgment. We thank the NIH (GM50113) for
financial support, Dr. George Flentke for providing R.
chinensis pepsin and initial assistance with the biological
assays, Drs. Ken Satyshur and Regine Bohacek for assistance
with molecular modeling, Dr. Colin McMartin for the use
of his software Flo-01, Dr. Ilia Guzei for the X-ray crystal
structure of 23, and Dr. Martha Vestling for MS data.
Supporting Information Available: Figures 2 and 6-9
as well as experimental procedures for the synthesis and
characterization of compounds 11-24. This material is
available free of charge via the Internet at http://pubs.acs.org.
OL016092U
Despite these caveats, our work demonstrates that it is
possible to identify non-peptide peptidomimetics utilizing
(29) Rich, D. H.; Bohacek, R. S.; Dales, N. A.; Glunz, P.; Ripka, A. S.
Chimia 1997, 51, 45.
(30) Kumar, S.; Ma, B.; Tsai, C.-J.; Sinha, N.; Nussinov, R. Protein
Sci. 2000, 9, 10.
(31) Todd, M. J.; Freire, E. Proteins: Struct., Funct., Gene. 1999, 36,
147.
(26) The silyl-ether analogue did not react presumably for steric reasons;
other ether analogues underwent reductive cleavage of the C4′ C-O bond.
(27) See Figure 7 in Supporting Information.
(28) (a) Flentke, G. R.; Glinski, J.; Satyshur, K.; Rich, D. H. Protein
Expression Purif. 1999, 16, 213. (b) Peranteau, A. G.; Kuzmic, P.; Angell,
Y.; Rich, D. H. Anal. Biochem. 1995, 227, 242.
(32) Programs to perform these calculations have not been reported to
date, but seem possible.
(33) For structural comparison, see Figure 8 in Supporting Information.
2320
Org. Lett., Vol. 3, No. 15, 2001