Nonpeptide inhibitors of cathepsin G: Optimization of a novel β-ketophosphonic acid lead by structure-based drug design

Article abstract of DOI:10.1021/ja017506h

The serine protease cathepsin G (EC 3.4.21.20; Cat G), which is stored in the azurophilic granules of neutrophils (polymorphonuclear leukocytes) and released on degranulation, has been implicated in various pathological conditions associated with inflammation. By employing high-throughput screening, we identified β-ketophosphonic acid 1 as a moderate inhibitor of Cat G (IC₅₀ = 4.1 μM). We were fortunate to obtain a cocrystal of 1 with Cat G and solve its structure by X-ray crystallography (3.5 A). Structural details from the X-ray analysis of 1·Cat G served as a platform for optimization of this lead compound by structure-based drug design. With the aid of molecular modeling, substituents were attached to the 3-position of the 2-naphthyl ring of 1, which occupies the S1 pocket of Cat G, to provide an extension into the hydrophobic S3 region. Thus, we arrived at analogue 7 with an 80-fold potency improvement over 1 (IC₅₀ = 53 nM). From these results, it is evident that the β-ketophosphonic acid unit can form the basis for a novel class of serine protease inhibitors. Copyright

Full text of DOI:10.1021/ja017506h

Published on Web 03/20/2002
Nonpeptide Inhibitors of Cathepsin G: Optimization of a Novel
â-Ketophosphonic Acid Lead by Structure-Based Drug Design
Michael N. Greco,† Michael J. Hawkins,† Eugene T. Powell,† Harold R. Almond, Jr.,†
Thomas W. Corcoran,† Lawrence de Garavilla,† Jack A. Kauffman,† Rosario Recacha,‡
Debashish Chattopadhyay,‡ Patricia Andrade-Gordon,† and Bruce E. Maryanoff*,†
Johnson & Johnson Pharmaceutical Research & DeVelopment, Spring House, PennsylVania 19477-0776, Center for
Biophysical Sciences and Engineering, UniVersity of Alabama at Birmingham, Birmingham, Alabama 35294
Received November 9, 2001
Table 1. Effect of Carboxamide Substituents on Bioactivity
The serine protease cathepsin G (EC 3.4.21.20; Cat G) is stored
in the azurophilic granules of neutrophils (polymorphonuclear
leukocytes) and released on degranulation.1 This chymotrypsin-
like enzyme has been implicated in a variety of pathological con-
ditions associated mainly with inflammation.2 For example, Cat G
is involved in tissue remodeling at sites of injury via the cleavage
of matrix components, including proteoglycans, collagen, fibronec-
tin, and elastin.3 Thus, specific inhibitors of Cat G could be useful
for treating emphysema, asthma, reperfusion injury, psoriasis, and
rheumatoid arthritis. We now report the discovery of nonpeptide
inhibitors of Cat G possessing a novel â-ketophosphonic acid sub-
unit, by use of a structure-based drug design approach.
There are several reports on Cat G inhibitors,2c-e,4 some of which
are reasonably potent; however, the inhibitors have generally been
peptide-like structures or irreversible inactivators.2c-e,4a-j We were
able to identify bis-naphthyl â-ketophosphonic acid 15 as a moder-
ately potent, reversible Cat G inhibitor (Table 1) by high-throughput
screening of a diverse chemical library with a chromogenic assay.6
This compound is very intriguing because of its novel nonpeptide
structure and competitive, reversible kinetics. To enhance the
potency, and to develop an initial structure-activity profile, we
synthesized numerous analogues with alterations of the aromatic
rings and ketone unit. Unfortunately, this exercise did not improve
the IC50 value into the nM range. For instance, replacement of the
2-naphthyl ring of 1 with phenyl or 1-naphthyl, or replacement of
the 1-naphthyl ring with phenyl or 2-naphthyl, abolished Cat G
activity (IC50 > 100 µM). Unfavorable results were also realized
on converting the CdO group to a CHOH, CH2, or SO2 group.
Since 1 appeared to represent a minimum-acceptable pharmaco-
phore, we resorted to structure-based drug design to enhance
potency.
complexes in the asymmetric unit were established. The two
complexes are essentially the same when superimposed except for
the Ile-35/Arg-41 loop in the S1′ region, which has two distinctly
different conformations. The two ligands had a displacement of
RMSD ) 1.79 Å, while the R-carbons of the proteins (minus the
35-41 loop) had RMSD ) 0.08 Å. One of the complexes is shown
in Figure 1.
In the complex, the active site of Cat G is occupied by the
enantiomer of 1 possessing the R absolute configuration. The
2-naphthyl group resides in the hydrophobic S1 specificity pocket
and the 1-naphthyl group resides in the S2 region (Figure 2).
Significantly, the distal benzene ring of the 1-naphthyl appears to
be involved in a π-stacking interaction with the imidazole ring of
the catalytic His-57, with the nearly parallel planes having a closest
approach of 3.6 Å. The phosphonic acid is strategically deployed,
with one oxygen atom H-bonded to Nꢀ of His-57,8 another partially
inserted into the “oxyanion hole”, H-bonded to NR of Gly-193,
and a third H-bonded to the side chain Nꢀ of Lys-192 (Figure 3).
The unenolized ketone of 1 is also H-bonded to Nꢀ of Lys-192.
Notably, this structure reveals a vacant S3 region in the active-site
cleft, which offers an opportunity for occupation by a suitable
appendage to enhance the potency (Figures 2 and 3). The
hydrophobic surfaces provided by the side chains of Ile-99, Phe-
172, and Tyr-215 are attractive for added interactions. By employing
computer-assisted molecular modeling, we tested different ideas
for substituting 1 with probes of the S3 subsite of Cat G. Sub-
stituents on the 1-naphthyl ring were perceived to have difficulty
in reaching the S3 pocket because of impeded access by Ile-99.
Our analysis indicated that attachment would be preferred on the
2-naphthyl ring at the 3-position. Thus, we designed versions of 1
with a carboxamide group located on the 2-naphthyl ring to anchor
a substituent that could nestle comfortably in the S3 region with 2
and 2,3-naphthalic anhydride serving as building blocks.9 Since our
modeling gave priority to structures containing an aromatic group
tethered by 6-8 Å to the 3-carboxyl, we incorporated arene-bearing
amines of varying chain lengths to obtain ligands of general type
3,9 namely 4-8, which were assayed for inhibition of Cat G (Table
1; N ) number of experiments).6a
With the work of Hof et al.4a as a backdrop, we prepared an
X-ray diffractable crystal of 1‚Cat G by the hanging-drop method
and determined the molecular structure.7 The tetragonal crystals,
space group P41 (a ) b ) 59.44 Å, c ) 130.62 Å), diffracted
X-rays to 3.0 Å resolution. Refinement of 7065 reflections (of 9299
independent reflections; R-merge ) 9.8%) gave an R-factor of 0.24
using 15 to 3.5 Å data. The positions of the two ligand/protein
† Johnson & Johnson Pharmaceutical Research & Development (formerly, The
R. W. Johnson Pharmaceutical Research Institute).
‡ University of Alabama at Birmingham.
9
3810
J. AM. CHEM. SOC. 2002, 124, 3810-3811
10.1021/ja017506h CCC: $22.00 © 2002 American Chemical Society

C O M M U N I C A T I O N S
hibitors also usually possess this structural element. Consequently,
we have identified a new inhibitor motif involving a ketone on the
carbon â to phosphorus. Such a â-ketophosphonic acid core struc-
ture has the potential for wider applicability. One can imagine
designing potent inhibitors for other types of serine proteases by
incorporating structural features that achieve favorable interactions
within the S1-S3 subsites.13
Acknowledgment. We thank Stephen Yabut for technical
contributions and Dr. Michael Costanzo for discussions and advice.
Supporting Information Available: Experimental details and
Figure 1. Structure of 1 (magenta) and Cat G (green ribbon), showing
key side chains (blue).
compound characterization, Figure 4, schematic of the interactions of
1 with Cat G, stereoview of Figure 1, diagrams of the electron density
of 1 in Cat G (PDF). This material is available free of charge via the
Internet at http://pubs.acs.org.
References
(1) Bainton, D. F.; Ullyot, J. L.; Farquhar, M. G. J. Exp. Med. 1971, 134,
907. Baggiolini, M.; Bretz, U.; Dewald, B.; Feigenson, M. E. Agents
Actions 1978, 8, 3. Hanson, R. K.; Connolly, N. L.; Burnett, D.; Campbell,
E. J.; Senior, R. M.; Ley, T. J. J. Biol. Chem. 1990, 265, 1524. Owen, C.
A.; Campbell, E. J. J. Leukocyte Biol. 1999, 65, 137.
Figure 2. View of 1 (magenta) within the active site of Cat G (green
electron-density surface); the vacant S3 pocket is on the left side.
(2) (a) Groutas, W. C. Med. Res. ReV. 1987, 7, 227. (b) Lomas, D. A.; Stone,
S. R.; Llewellyn-Jones, C.; Keogan, M.-T.; Wang, Z.; Rubin, H.; Carrell,
R. W.; Stockley, R. A. J. Biol. Chem. 1995, 270, 23437. (c) Groutas, W.
C.; Kuang, R.; Venkataraman, R.; Epp, J. B.; Ruan, S.; Prakash, O.
Biochemistry 1997, 36, 4739. (d) Gu¨tschow, M.; Neumann, U. Bioorg.
Med. Chem. 1997, 5, 1935. (e) Kuang, R.; Epp, J. B.; Ruan, S.; Yu, H.;
Huang, P.; He, S.; Tu, J.; Schechter, N. M.; Turbov, J.; Froelich, C. J.;
Groutas, W. C. J. Am. Chem. Soc. 1999, 121, 8128. (f) Maclov, D. M.;
Shapiro, S. D.; Pham, C. T. N.; Belaaouaj, A.; Abraham, S. N.; Ley, T.
J. Blood 1999, 94, 4282. (g) Owen, C. A.; Campbell, E. J. Semin. Cell
Biol. 1995, 6, 367.
Figure 3. Stereoview of the interactions of with Cat G. Hydrogen bonds
(3) Roughley, P. J.; Barrett, A. J. Biochem. J. 1977, 167, 629. Capodici, C.;
Berg, R. A. Inflammation 1989, 13, 137. Vartio, T.; Seppa, H.; Vaheri,
A. J. Biol. Chem. 1981, 256, 471. Reilly, C. F.; Travis, J. Biochim. Biophys.
Acta 1980, 621, 147. Bieth, J. G. In Biology of the Extracellular Matrix;
Mecham, R. P., Ed.; Academic Press: New York; 1986; Vol. 1, pp 217-
320.
are indicated by the broken red lines.
Compounds 4 and 5 exhibit a 3-fold potency increase over 1,
while 6 exhibits an 8-fold increase, indicating that added hydro-
phobicity can enhance affinity. There was a notable 80-fold
improvement over 1 with the aromatic moiety attached via a cyclic
tether, as in 7. Presumably, the conformational constraint imparted
by the piperidine ring favorably orients the phenyl portion of the
ligand within Cat G’s cleft. The 15-fold weaker potency of acyclic
variant 8, compared with 7, supports this view. For additional
insight, we carried out a simulated annealing experiment in which
7 was docked into the active site of Cat G (Figure 4).10 The inter-
actions in the S1, S2, and catalytic regions of the 7‚Cat G model
are analogous to those observed for the X-ray crystal structure of
1‚Cat G (Figure 1). However, the piperidine ring (chair form) serves
as a scaffold to position the phenyl ring of 7 within the S3 pocket,
such that it makes hydrophobic contacts with the side chains of
Phe-172, Tyr-215, and Ile-99. A more detailed picture of the
interaction between 7 and Cat G awaits an X-ray structure determi-
nation of the complex.
Compound 7 shows reversible, competitive inhibition with IC50
and Ki values of 53 ( 12 (N ) 10) and 63 ( 14 nM (N ) 5),
respectively. Another attribute of 7 relates to its selectivity vs other
serine proteases. It weakly inhibits chymotrypsin (Ki ) 1.5 ( 0.2
µM), and poorly inhibits (<50% inhibition at 100 µM) thrombin,
factor Xa, factor IXa, plasmin, trypsin, tryptase, proteinase 3, and
human leukocyte elastase.
Serine protease inhibitors with phosphonate or phosphinate
groups that can occupy the active site in the vicinity of the key
catalytic machinery, that is, Ser-195 and His-57, are known.5a,11
These have generally been P-ester forms, especially diphenyl
phosphonates, which operate by slow-tight or irreversible binding.5a,11
The formation of a covalent bond with Oγ of Ser-195 to yield a
phosphonylated enzyme species has been confirmed by three dif-
ferent X-ray studies.5a,11e,h In the reports that describe serine protease
inhibitors with a free phosphonic or phosphinic acid end-group,
all of the compounds have an acylamino group on the carbon R to
phosphorus.11a-c,12 The known phosphonate/phosphinate ester in-
(4) (a) Hof, P.; Mayr, I.; Huber, R.; Korzus, E.; Potempa, J.; Travis, J.; Powers,
J. C.; Bode, W. EMBO J. 1996, 15, 5481. (b) Groutas, W. C.; Kuang, R.;
Ruan, S.; Epp, J. B.; Venkataraman, R.; Truong, T. M. Bioorg. Med. Chem.
1998, 6, 661. (c) Iijima, K.; Katada, J.; Yasadu, E.; Uno, I.; Hayashi, Y.
J. Med. Chem. 1999, 42, 312. (d) Powers, J. C.; Tanaka, T.; Harper, J.
W.; Minematsu, Y.; Baker, L.; Lincoln, D.; Crumley, K. V. Biochemistry
1985, 24, 2040. (e) Harper, J. W.; Powers, J. C. Biochemistry 1985, 24,
7200 and references cited. (f) Peet, N. P.; Burkhart, J. P.; Angelastro, M.
R.; Giroux, E. L.; Mehdi, S.; Bey, P.; Kolb, M.; Neises, B.; Schirlin, D.
J. Med. Chem. 1990, 33, 394. (g) Groutas, W. C.; Brubaker, M. J.;
Venkataraman, R.; Epp, J. B.; Stanga, M. A.; McClenahan, J. J. Arch.
Biochim. Biophys. 1992, 297, 144. (h) Knight, W. B.; Chabin, R.; Green,
B. Arch. Biochim. Biophys. 1992, 296, 704. (i) Angelastro, M. R.; Bey,
P. Mehdi, S.; Janusz, M. J.; Peet, N. P. Bioorg. Med. Chem. Lett. 1993,
3, 525. (j) Zembower, D. E.; Kam, C.-M.; Powers, J. C.; Zalkow, L. H.
J. Med. Chem. 1992, 35, 1597.
(5) Schwender, C. F.; Beers, S. A.; Malloy, E.; Demarest, K.; Minor, L.;
Lau, K. H. W. Bioorg. Med. Chem. Lett. 1995, 5, 1801.
(6) (a) Details of the Cat G assay and IC50 determinations are presented in
the Supporting Information. (b) We screened our proprietary chemical
library of ca. 250 000 compounds.
(7) Details of the X-ray work will be published separately.
(8) The Oγ of Ser-195 is proximal to this oxygen atom (3.3 Å) and may be
involved in a hydrogen bond.
(9) The adduct from acylation of the lithio anion of 2 with anhydride was
reacted with dicyclohexylcarbodiimide and RMeNH, followed by TMS-
Br/pyridine, then 1 N HCl. Synthetic details are contained in the
Supporting Information. Use of a primary amine in the DCC coupling
gave an undesired phthalimide product and little of the desired NH amide.
(10) See Supporting Information for Figure 4 and for details of the simulated
annealing studies, performed with AMBER (version 5.0).
(11) (a) Cheng, L.; Goodwin, C. A.; Scully, M. F.; Kakkar, V. V.; Claeson,
G. Tetrahedron Lett. 1991, 32, 7333. (b) Wang, C.-L.; Taylor, T. L.; Mical,
A. J.; Spitz, S.; Reilly, T. M. Tetrahedron Lett. 1992, 33, 7667. (c) Fastrez,
J.; Jespers, L.; Lison, D.; Renard, M.; Sonveaux, E. Tetrahedron Lett.
1989, 30, 6861. (d) Bartlett, P. A.; Lamden, L. A. Bioorg. Chem. 1986,
14, 356. (e) Bertrand, J. A.; Oleksyszyn, J.; Kam, C.-M.; Boduszek, B.;
Presnell, S.; Plaskon, R. R.; Suddath, F. L.; Powers, J. C.; Williams, L.
D. Biochemistry 1996, 35, 3147. (f) Oleksyszyn, J.; Powers, J. C. Methods
Enzymol. 1994, 244, 423. (g) Oleksyszyn, J.; Powers, J. C. Biochemistry
1991, 30, 485. (h) Bone, R.; Sampson, N. S.; Bartlett, P. A.; Agard, D.
A. Biochemistry 1991, 30, 2263. (i) Sampson, N. S.; Bartlett, P. A.
Biochemistry 1991, 30, 2255. (j) Collinsova, M.; Jiracek, J. Curr. Med.
Chem. 2000, 7, 629.
(12) Li, M.; Lin, Z.; Johnson, M. E. Bioorg. Med. Chem. Lett. 1999, 9, 1957.
(13) Note Added in Proof: The X-ray coordinates for 1 Cat G were deposited
in the PDB under the code number 1KYN.
JA017506H
9
J. AM. CHEM. SOC. VOL. 124, NO. 15, 2002 3811