Conformationally constrained 1,3-diamino ketones: A series of potent inhibitors of the cysteine protease cathepsin K [3]

Full text of DOI:10.1021/jm980295f

J . Med. Chem. 1998, 41, 3563-3567
3563
cathepsin K revealed that the enzyme could potentially
accommodate cyclization of the 1,3-diamino ketone
template in two alternative fashions with no deleterious
steric interactions with the peptide backbone residues.8
As seen in Figure 1, cyclization of 1 along path A would
produce the 2,5-diaminocyclopentanone and 2,6-diami-
nocyclohexanone inhibitors 2 and 3. No overall reor-
ganization of the original 1,3-diamino ketone template
is observed in these analogues. Alternatively, cycliza-
tion along path B would provide the 3-aminopyrrolidi-
none and 4-aminopiperidinone inhibitors 4 and 5,
respectively. Molecular modeling revealed that this
cyclization produces a change in the conformation of the
original 1,3-diamino ketone substructure. All four of
the proposed conformationally constrained scaffolds 2-5
have been modeled as tetrahedral adducts with the
active site cysteine 25 of cathepsin K. The interactions
between the inhibitor and enzyme have been optimized
by overlapping the tetrahedral adducts with that from
the inhibitor 1/cathepsin K crystal structure, adjusting
torsion angles to mimic the conformation of 1 bound to
the enzyme. No apparent deleterious enzyme/inhibitor
interactions were seen. Prediction of the binding of
inhibitors using path B is much less obvious than for
path A. Indeed, the exceptional differences in potency
achieved by the analogues of path B compared to those
of path A were not predicted and remain a matter of
conjecture despite the extensive structure data avail-
able.
Con for m a tion a lly Con str a in ed
1,3-Dia m in o Keton es: A Ser ies of P oten t
In h ibitor s of th e Cystein e P r otea se
Ca th ep sin K
Robert W. Marquis,† Dennis S. Yamashita,† Yu Ru,†
Stephen M. LoCastro,† Hye-J a Oh,† Karl F. Erhard,†
Renee L. DesJ arlais,‡ Martha S. Head,‡
Ward W. Smith,§ Baoguang Zhao,§ Cheryl A. J anson,§
Sherin S. Abdel-Meguid,§ Thaddeus A. Tomaszek,|
Mark A. Levy,| and Daniel F. Veber*,†
Departments of Medicinal Chemistry, Structural and
Physical Chemistry, Structural Biology, and Molecular
Recognition, SmithKline Beecham Pharmaceuticals, 709
Swedeland Road, King of Prussia, Pennsylvania 19406
Received May 11, 1998
The maintenance of bone homeostasis is a dynamic
equilibrium of bone formation by osteoblasts and re-
sorption of the bone matrix by osteoclasts.1 Shifts in
this equilibrium toward the excessive resorption of bone
matrix produce clinical manifestations such as os-
teoporosis, Pagets disease, and certain types of arthritis.
Over the past several years cysteine proteases have
been suggested to play a role in the osteoclast-mediated
resorption of the bone matrix.2 Recently it has been
shown that cathepsin K (EC 3.4.22.38), a cysteine
protease of the papain superfamily, is both selectively
and highly expressed in osteoclasts.3 The implication
of this abundant and selective expression suggests that
this enzyme may be playing an important and specific
role in the resorption phase of bone remodeling. Inhibi-
tion of this enzyme may therefore be a potential strategy
for therapeutic intervention in diseases of excess bone
resorption. Further support for this hypothesis is that
mutations in the gene which encodes for this enzyme
are associated with a rare autosomal disorder of bone
remodeling known as pycnodysostosis.4 Also, Yama-
mura and co-workers have shown that a cathepsin K
antisense construct inhibits osteoclastic bone resorption
in pit formation.5
A series of potent, reversible inhibitors of cathepsin
K based on a 1,3-bis(Cbz-Leu-amino)-2-propanone tem-
plate has recently been disclosed from these laborato-
ries.6 The design of these inhibitors was based on the
observation that two aldehyde inhibitors, leupeptin (Ac-
Leu-Leu-Arg-H) and Cbz-Leu-Leu-Leu-H, bound in op-
posite directions in the active site of papain. In an effort
to design more potent inhibitors of cathepsin K, we have
attempted to further exploit this design element by
incorporating the 1,3-bis(Cbz-Leu-amino)-2-propanone
inhibitor template 1 into conformationally constrained
ring systems (see Figure 1). Historically, the introduc-
tion of a conformational constraint has been used to
capture a bioactive orientation of a molecule.7
The designed inhibitors were synthesized as outlined
in Scheme 1. Compounds 3 and 2 were synthesized via
bis-acylation of the known racemic trans-1,3 diamino-
2-cyclohexanol (7)12 and by analogy the trans-1,3-
diamino-2-cyclopentanol (6) with Cbz-leucine and HBTU
to give the alcohols 9 and 8. J ones oxidation gave the
desired ketones 3 and 2. The diastereomers of 2 were
not separable and were therefore tested as a mixture.
The diastereomeric mixture 3 was separable by silica
gel chromatography, and each diastereomer was char-
acterized as 1,3-trans by measuring the proton coupling
constants (J 1ax,2ax ) 6.8 Hz and J 1ax,2eq ) 3.3 Hz) of the
corresponding ketone-derived alcohols (NaBH4/CH3OH).
On the basis of circular dichroism studies (n to π*
transitions of the cyclohexanones13), the absolute ste-
reochemistries were assigned as R,R for the first eluted
diastereomer and S,S for the second eluted diastere-
omer. Compounds 4 and 5 were synthesized as dia-
stereomeric mixtures by EDC-mediated acylation of the
amino alcohols 1014 and 1115 with Cbz-leucine. Depro-
tection of the Boc protecting group followed by a second
EDC-mediated acylation gave the diastereomeric alco-
hols 12 and 13. Oxidation of the alcohols was ac-
complished by employing either J ones reagent or Swern
oxidation to give ketones 4 and 5. The bis-Cbz-leucine-
4-aminopyrrolidinone diastereomers 4 were separated
by reverse-phase HPLC under neutral conditions (CH3-
CN/pH ) 7 phosphate buffer). These diastereomers
were not configurationally stable and equilibrated to the
original mixture upon concentration of the eluent.
HPLC separation of the diastereomers of 5 revealed the
Modeling experiments based on the crystal structure
of inhibitor 1 bound to the active site cysteine of
† Department of Medicinal Chemistry.
‡ Department of Structural and Physical Chemistry.
§ Department of Structural Biology.
| Department of Molecular Recognition.
S0022-2623(98)00295-7 CCC: $15.00 © 1998 American Chemical Society
Published on Web 08/19/1998

3564 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 19
Communications to the Editor
F igu r e 1. Conformationally constrained diamino ketone inhibitors of cathepsin K.
Sch em e 1. Synthesis of Cyclized Cyclic 1,3-Diamino Ketone Inhibitors 2-5a
a
Reagents and conditions: (1) HBTU, Cbz-leucine, DMF; (2) J ones reagent, acetone; (3) Cbz-leucine, EDC, HOBt, CH2Cl2; (4) HCl/
EtOAc; (5) DMSO, (COCl)2, TEA, CH2Cl2, -78 °C to rt.
presence of two diastereomers plus a third slower
eluting compound. Attempted isolation of this slower
eluting compound was not successful as it rapidly
reverted to a mixture of three compounds. Mass
spectral analysis revealed an M + 18 molecular ion
indicative of the hydrate of ketone 5. Hydration of the
ketone was not observed in the attempted isolation of
4. The diastereomers of 5 were isolated and were also
found to equilibrate to a mixture upon concentration of
the eluent.
Cathepsin K inhibition and selectivity data are pre-
sented in Table 1.16 Compound 2, which was tested as
a mixture of diastereomers, possessed weak time-
dependent inhibitory activity with a kobs/[I] of 56 M-1
s-1. Each diastereomer of 3 was found to have weak
inhibitory activity against cathepsin K with Ki,app of 16
µM (R,R) and 15 µM (S,S), respectively. Compounds 4
and 5 were potent inhibitors of cathepsin K with Ki,app
of 2.3 and 2.6 nM, respectively. Diastereomers 4 were
found to selectively inhibit cathepsin K over cathepsins
B and L. As a diastereomeric mixture these inhibitors
were greater than 400-fold selective for cathepsin K
over cathepsin B and approximately 17-fold selective
over cathepsin L. The six-membered ring cyclic diamino
ketones 5 were somewhat less selective than 4. The
alcohols 12 and 13 were approximately 1/3000 as potent
as the ketone analogues. This is in agreement with a
transition-state analogue mechanism of action for the
ketone analogues.17 The X-ray crystal structure of
inhibitors 4 with cathepsin K confirmed this hypothesis

Communications to the Editor
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 19 3565
Ta ble 1. Cathepsin K, B, and L Inhibition Data
as it is consistent with a covalent interaction of the
active site cysteine 25 with the carbonyl group of 4 to
form a tetrahedral adduct (Figure 2). The hemith-
ioketal was flanked by the N1 Cbz-leucine which was
observed to bind on the unprimed side of the active
site.18 The isobutyl group of the leucine was bound in

3566 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 19
Communications to the Editor
F igu r e 2. View of the X-ray cocrystal structure of inhibitor 4
bound in the active site of cathepsin K. The isobutyl group of
the N1 Cbz-leucine lies buried in the hydrophobic S2 pocket
formed by the side chains of Leu 160, Ala 134, and Met 68. A
hydrogen bond is formed between the N1 amide carbonyl and
Gly 66 of the enzyme.
F igu r e 3. View of the X-ray cocrystal of inhibitor 14 bound
in the active site of cathepsin K. The hydrogen bond between
the N1 amide carbonyl of the inhibitor and Gly 66 of the
enzyme that was present for inhibitor 4 is lacking in this
inhibitor/enzyme complex.
has been replaced with the same p-phenoxybenzamide,
is 30-fold less active relative to analogue 15. Incorpora-
tion of the sulfonamide into the peptidomimetic gives
the potent 4-phenoxybenzenesulfonamide derivative 17.
This modification has removed most of the structural
liabilities commonly associated with peptide amide
linkages. Complete removal of the peptide elements of
the Cbz-leucine via the incorporation of a reduced amide
while retaining the recognition element of the isohexyl
group gave rise to analogue 18 which incorporates the
N-isohexyl peptidomimetic.
In this communication we have disclosed a potent
class of conformationally constrained ketone inhibitors
of the cysteine protease cathepsin K. Peptidomimetic
elements have been incorporated onto this scaffold
which remove hydrogen bonding, a feature postulated
to reduce rates of intestinal transfer.19 More detailed
in vivo evaluation of such inhibitors will be the subject
of future studies. The inhibition of other proteases may
be possible with this cyclic diamino ketone template by
the incorporation of the appropriate specificity elements
required by these enzymes.
the hydrophobic S2 pocket which was formed from the
side-chain atoms of Leu 160, Ala 134, and Met 68. A
hydrogen bond between the amide carbonyl of the N1
Cbz-leucine and Gly 66 of the protein backbone is also
seen in the X-ray structure of 4 complexed with the
enzyme. The N3 Cbz-leucine was seen bound on the
primed side of the active site. The stereochemistry of
the diastereomeric C3 center was consistent with the
R configuration in the X-ray crystal structure and may
therefore represent the more potent of the two diaster-
eomeric forms. Alternatively, this stereochemical pref-
erence may be a function of crystal packing forces.
A possible rationalization for the dramatic difference
in activities for the compounds resulting from the two
different modes of cyclization is that, despite favorable
modeling predictions and the presence of the correct
specificity elements required for initial Michaelis bind-
ing in both sets of compounds, inhibitors 4 and 5 have
the appended N1 amine group oriented in such a fashion
as to allow for the facile attack by the active site cysteine
nucleophile on the electrophilic carbonyl of the designed
inhibitors. This may not be true for inhibitors 2 and 3
in which the diamino ketone template of 1 has es-
sentially been substituted in both the R and R′ positions.
This substitution may cause a steric interaction which
was unforeseen in the original modeling experiments
on the thiol adducts.
As seen in Table 1, analogue 14 in which a methylene
group has replaced the tertiary amide carbonyl moiety
retains the overall potency of the parent cyclic diamino
ketone 4. The X-ray cocrystal structure of inhibitor 14
with cathepsin K shows no hydrogen bond in this region
of the inhibitor/protein complex (see Figure 3). This
change to a reduced amide has served to remove the
amide linkage and simultaneously increase overall
aqueous solubility. The leucine binding on the unprimed
side of the active site of the enzyme could be replaced
with several peptidomimetics and still retain the overall
potency of this class. Analogue 15, which incorporates
a p-phenoxybenzamide peptidomimetic, results in a 13-
fold loss of activity relative to 4. Compound 16, in which
the leucine binding on the primed side of the active site
Ack n ow led gm en t. Dr. J ohn Gleason and Dr. Brian
Metcalf are thanked for their support of this work.
Su p p or tin g In for m a tion Ava ila ble: Coordinates for
compounds 4 and 14 bound to cathepsin K have been deposited
in the Brookhaven Protein Databank, accession numbers 1AU3
and 1AU4, respectively. Crystallization data of cathepsin K
complexed with 4 and 14 (2 pages). Ordering information is
given on any current masthead page.
Refer en ces
(1) Baron, R. Anatomy and Ultrastructure of Bone. In Primer on
the Metabolic Bone Diseases and Disorders of Mineral Metabo-
lism, 1st ed.; Favus, M. J ., Ed.; American Society for Bone and
Mineral Research: Kelseyville, CA, 1990; pp 3-9.
(2) (a) Votta, B. J .; Levy, M. A.; Badger, A.; Bradbeer, J .; Dodds, R.
A.; J ames, I. A.; Thompson, S. T.; Bossard, M. J .; Carr, T.;
Connor, J . R.; Tomaszek, T. A.; Szewczuk, L.; Drake, F. H.;
Veber, D. F.; Gowen, M. Peptide Aldehyde Inhibitors of Cathe-
psin K Inhibit Bone Resorption Both In Vitro and In Vivo. J .
Bone Miner. Res. 1997, 12, 1396-1406. (b) Delaisse, J .-M.;
Eeckhout, Y.; Vaes, G. In Vivo and In Vitro Evidence for the
Involvement of Cysteine Proteinases in Bone Resorption. Bio-
chem. Biophys. Res. Commun. 1984, 125, 441-447. (c) Delaisse,

Communications to the Editor
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 19 3567
J . M.; Boyde, A.; Macconachie, E.; Ali, N. N.; Sear, C. H. J .;
Eeckhout, Y.; Vaes, G.; J ones, S. J . The Effects of Inhibitors of
Cysteine-Proteinases and Collagenase on the Resorptive Activity
of Isolated Osteoclasts. Bone 1987, 8, 305-313.
(14) Hall, S. E.; Ballas, L. M.; Kulanthaivel, P.; Boros, C.; J iang, J .
B.; J agdmann, G. E.; Lai, Y.-S.; Biggers, C. K.; Hu, H.; Hallock,
Y.; Hughs, P. F.; DeFauw, J . M.; Lynch, M. P.; Lampe, J . W. C.;
Menaldino, D. S.; Heerding, J . M.; J anzen, W. P.; Hollinshead,
S. P. Balanoids. PCT Application WO 94/20062, 1994.
(3) (a) Drake, F. H.; Dodds, R. A.; J ames, I. A.; Connor, J . R.;
Debouck, C.; Richardson, S.; Lee-Rykaczewski, L.; Coleman, L.;
Riemann, D.; Barthlow, R.; Hastings, G.; Gowen, M. Cathepsin
K, but Not Cathepsins B, L or S is Abundantly Expressed in
Human Osteoclasts. J . Biol. Chem. 1996, 271, 12511-12516. (b)
Bromme, D.; Okamoto, K. Human Cathepsin O2, a Novel
Cysteine Protease Highly Expressed in Osteoclastomas and
Ovary. Molecular Cloning, Sequencing and Tissue Distribution.
Biol. Chem. Hoppe-Seyler 1995, 376, 379-384. (c) Shi, G.-P.;
Chapman, H. A.; Bhairi, S. M.; DeLeeuw, C.; Reddy, V. Y.; Weiss,
(15) The Boc-protected amino alcohol 11 was prepared in an analo-
gous fashion as that described for the Cbz derivative 27.
Treatment of epoxide 19 with sodium azide gave an inseparable
mixture of azides 21 and 23 in a ratio of approximately 1:4-5
as determined by 1H NMR analysis. The mixture of azido
alcohols 21 and 23 was reduced by catalytic hydrogenation over
10% Pd/C in methanol to give the amino alcohols 25 and 11.
The undesired regioisomeric amino alcohol 25 was removed by
column chromatography at a later stage in the synthesis. See:
Lai, Y.-S.; Mendalino, D. S.; Nichols, J . B.; J agdmann, G. E.,
J r.; Mylott, F.; Gillespie, J .; Hall, S. E. Ring Size Effect in the
PKC Inhibitory Activities of Perhydroazepine Analogues of
Balanol. Bioorg. Med. Chem. Lett. 1995, 5, 2151-2154. In this
paper the major product of sodium azide treatment of epoxide
20 is reported to be regioisomer 22 (4.8:1 22:24). In contrast to
these results we have found that the desired regioisomer 24 is
indeed the major product (4-5:1 24:22) of epoxide opening by
sodium azide. These results have been confirmed by a series of
1H NMR COSY studies on both 22 and 24. This result was
consistent with epoxide ring openings that have been reported
for the analogous 3,4-epoxytetrahydropyrans. See: Berti, G.;
Catelani, G.; Ferretti, M.; Monti, L. Regio- and Stereoselectivity
of the three-membered Ring Opening of 3,4-Epoxytetrahydro-
pyrans and of the Corresponding Epibromonium Ions. Tetrahe-
dron 1974, 30, 4013-4020.
S. J . Molecular Cloning of Human Cathepsin O,
a Novel
Endoproteinase and Homologue of Rabbit OC2. FEBS Lett. 1995,
357, 129-134. (d) Littlewood-Evans, A.; Kokubo, T.; Ishibashi,
O.; Inaoka, T.; Wlodarski, B.; Gallagher, J . A.; Bilbe, G.
Localization of Cathepsin K in Human Osteoclasts by In Situ
Hybridization and Immunohistochemistry. Bone 1997, 20, 81-
86. (e) Bossard, M. J .; Tomaszek, T. A.; Thompson, S. K.;
Amegadzie, B. Y.; Hanning, C. R.; J ones, C.; Kurdyla, J . T.;
McNulty, D. E.; Drake, F. H.; Gowen, M.; Levy, M. A. Proteolytic
Activity of Human Osteoclast Cathepsin K. Expression, Activa-
tion, and Substrate Identification. J . Biol. Chem. 1996, 271,
12517-12524.
(4) (a) Gelb, B. D.; Moissoglu, K.; Zhang, J .; Martignetti, J . A.;
Bromme, D.; Desnick, R. J . Cathepsin K; Isolation and Char-
acterization of the Murine cDNA and Genomic Sequence, the
Homologue of the Human Pycnodysostosis Gene. Biochem. Mol.
Med. 1996, 59, 200-206. (b) J ohnson, M. R.; Polymeropoulos,
M. H.; Vos, H. L.; Oritz de Luna, R. I.; Francomano, C. A. A
Nonsense Mutation in the Cathepsin K Gene Observed in a
Family with Pycnodysostosis. Genome Res. 1996, 6, 1050-1055.
(c) Gelb, B. D.; Shi, G.-P.; Chapman, H. A.; Desnick, R. J .
Pycnodysostosis, A Lysosomal Disease Caused by Cathepsin K
Deficiency. Science 1996, 273, 1236-1238.
(5) Inui, T.; Ishibashi, O.; Inaoka, T.; Origane, Y.; Kumegawa, M.;
Kokubo, T.; Yamamura, T. Cathepsin K Antisense Oligodeoxy-
nucleotide Inhibits Osteoclastic Bone Resoption. J . Biol. Chem.
1997, 272, 8109-8112.
(6) Yamashita, D. S.; Smith, W. W.; Zhao, B.; J anson, C. A.;
Tomaszek, T. A.; Bossard, M. A.; Levy, M. A.; Oh, H.-J .; Carr,
T. J .; Thompson, S. T.; Ijames, C. F.; Carr, S. A.; McQueney,
M.; D′Alessio, K. J .; Amegadzie, B. Y.; Hanning, C. R.; Abdel-
Meguid, S.; DesJ arlais, R. L.; Gleason, J . G.; Veber, D. F.
Structure and Design of Potent and Selective Cathepsin
K
Inhibitors. J . Am. Chem. Soc. 1997, 119, 11351-11352.
(7) (a) Liskamp, R. M. J . Conformationally Restricted Amino Acids
and Dipeptides, (Non)peptidomimetics and Secondary Structure
Mimetics. Recl. Trav. Chim. Pays-Bas 1994, 113, 1-19. (b)
Marshall, G. R.; Fedric, G. A.; Moore, M. L. Peptide Conforma-
tion and Biological Activity. Annu. Rep. Med. Chem. 1978, 13,
227-238.
(8) The analogues were modeled with MacroModel version 5.5.9 The
structures were then minimized using the MacroModel imple-
mentation of the AMBER10 force field and fit using UCSF
MidasPlus.11
(9) Mohamadi, F.; Richards, N. G. J .; Guida, W. C.; Liskamp, R.;
Lipton, M.; Caufield, C.; Chang, G.; Hendrickson, T.; Still, W.
C. MacroModel-An Integrated Software System for Modeling
Organic and Bioorganic Molecules Using Molecular Mechanics.
J . Comput. Chem. 1990, 11, 440.
(10) (a) Weiner, S. J .; Kollman, P. A.; Case, D. A.; Singh, U. C.; Ghio,
C.; Alagona, G.; Profeta, S., J r.; Weiner, P. A New Force Field
for Molecular Mechanical Simulation of Nucleic Acids and
Proteins. J . Am. Chem. Soc. 1984, 106, 765-784. (b) Weiner, S.
J .; Kollman, P. A.; Nguyen, D. T.; Case, D. A. An All Atom Force
Field for Simulations of Proteins and Nucleic Acids. J . Comput.
Chem. 1986, 7, 230-252. (c) McDonald, D. Q.; Still, W. C.
AMBER Torsional Parameters for the Peptide Backbone. Tet-
rahedron Lett. 1992, 33, 7743-7746.
(11) Ferrin, T. E.; Huang, C. C.; J arvis, L. E.; Langridge, R. The
MIDAS Display System. J . Mol. Graph. 1988, 6, 13-27.
(12) Sammes, P. G.; Thetford, D. Stereocontrolled Preparation of
Cyclohexane Amino Alcohols Utilising a Modified Mitsunobu
Reaction. J . Chem. Soc. PTI 1989, 655-661.
(16) Testing of the inhbitors was conducted with purified recombinant
cathepsin K as described in ref 3e.
(17) (a) Abeles, R. H. Enzyme Inhibitors: Ground State/Transition
State Analogues. Drug Dev. Res. 1987, 10, 221-234. (b) Douglas,
D. T. Transition-state Analogues in Drug Design. Chem. Ind.
(London) 1983, 8, 311-315. (c) Brodbeck, U. Recent Develop-
ments in the Field of Enzyme Inhibitors. Chima 1980, 34, 415-
421. (d) Lienhard, G. E. Transition State Analogues as Enzyme
Inhibitors. Annu. Rep. Med. Chem. 1972, 7, 249-258. (e)
Radzika, A.; Wolfenden, R. Transition State and Multisubstrate
Analogue Inhibitors. Methods Enzymol. 1995, 249, 284-312.
(18) Schecter, I.; Berger, A. On the Size of the Active Site in
Proteases. I. Papain. Biochem. Biophys. Res. Commun. 1967, 27,
157-162.
(19) Spatola, A. Peptide Backbone Modifications: A Structure-
Activity Analysis of Peptides Containing Amide Bond Sur-
rogates, Conformational Constraints, and Related Backbone
Replacements. Chem. Biochem. Amino Acids Pept. Proteins 1983,
7, 267-357.
(13) Lightner, D. A.; Bouman, T. D.; Crist, B. V.; Rodgers, S. L.;
Knobeloch, M. A.; J ones, A. M. The Octant Rule. 21. Circular
Dichroism Dependence on R-Methyl Configuration in Cyclohex-
anones. Experiments and RPA Calculations. J . Am. Chem. Soc.
1987, 109, 6248-6259 and references therein.
J M980295F