Structure-based design of potent and selective inhibitors of collagenase-3 (MMP-13)

Article abstract of DOI:10.1016/j.bmcl.2004.12.016

Computer aided drug design led to a new class of spiro-barbiturates (e.g., 4a, MMP-13 K_i = 4.7 nM) that are potent inhibitors of MMP-13.

Full text of DOI:10.1016/j.bmcl.2004.12.016

Bioorganic & Medicinal Chemistry Letters 15 (2005) 1101–1106
Structure-based design of potent and selective inhibitors
of collagenase-3 (MMP-13)
Soong-Hoon Kim,* Andrew T. Pudzianowski, Kenneth J. Leavitt, Joseph Barbosa,
Patricia A. McDonnell, William J. Metzler, Bruce M. Rankin, Richard Liu,
Wayne Vaccaro and William Pitts
Bristol Myers Squibb Co., Pharmaceutical Research Institute, PO Box 4000, Princeton, NJ 08560, USA
Received 17 September 2004; revised 1 December 2004; accepted 8 December 2004
Available online 7 January 2005
Abstract—Computer aided drug design led to a new class of spiro-barbiturates (e.g., 4a, MMP-13 Ki = 4.7 nM) that are potent
inhibitors of MMP-13.
Ó 2004 Elsevier Ltd. All rights reserved.
The matrix metalloproteinases (MMPs) belong to a
family of zinc proteases that are important in homeosta-
sis of extra-cellular matrix proteins.1 They have been
implicated in several pathologies including cancer,
osteoarthritis, and rheumatoid arthritis.2
to discover selective small molecule inhibitors of
MMP-13.5
Historically, most MMP inhibitors have been based on
hydroxamic acids.6 In general, hydroxamic acid based
inhibitors are remarkably potent (e.g., RS-130830 1,
MMP-13 Ki = 0.52 nM, Fig. 1).7 Nevertheless, a grow-
ing body of evidence suggests that hydroxamic acids
have poor pharmacokinetic profiles, exemplified by poor
absorption and rapid metabolism.8 Despite these poten-
tial shortcomings several hydroxamic acid based MMP
inhibitors have been in human clinical trials.9 Because
of these potential liabilities, our goal was to design a
Osteoarthritis is a degenerative disease characterized by
the loss of joint function and articular cartilage. Current
therapies, such as COX-2 inhibitors and non-steroidal
anti-inflammatory agents, target signs, and symptoms.
Disease modifying osteoarthritis drugs that may delay
disease progression are not presently available.
Although the inhibition of MMP-13 as a treatment for
osteoarthritis has not been clinically validated, pre-clin-
ical data suggest MMP-13 as a compelling target for the
modulation of osteoarthritis. Articular cartilage consists
of type II collagen and proteoglycan, and studies have
shown that MMP-13 (collagenase-3) cleaves type II col-
lagen.3 MMP-13 over-expression induces osteoarthritis
in transgenic mice, and MMP-13 mRNA expression is
co-distributed with the type II collagenase activity in
osteoarthritis cartilage.4 Thus, the data strongly suggest
that inhibition of MMP-13 may lead to inhibition of
cartilage degradation associated with osteoarthritis. A
major effort is in progress by several research groups
O
O
1 (RS-130830)
(MMP13 Ki = 0.52nM)
HOHN
S
O O
Cl
O
Zinc Binder--Linker--Hydrophobic Group
O
HN
2
O
CO2Et
HN
Keywords: MMP-13 inhibition; Barbituric acid; Structure-based
design.
O
*
Corresponding author. Tel.: +1 609 252 4079; fax: +1 609 252
3993; e-mail: soonghoon.kim@bms.com
Figure 1. Functional mapping of MMP inhibitors.
0960-894X/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.12.016

1102
S.-H. Kim et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1101–1106
series of potent MMP-13 inhibitors which did not con-
tain a hydroxamic acid. An additional goal was to iden-
tify an inhibitor of MMP-13 with a good degree of
selectivity among MMP and ADAM family members
(e.g., MMP-1, MMP-2, MMP-12, and TACE).
butyl hydroxamate (HONHCO-tert-butyl) and tris-
imidazole Zn(II), reproduced the bidentate mode of
˚
interaction with slightly longer (calculated: 2.13 A and
˚
2.34 A; the complexation enthalpy, relative to separated
tert-butyl hydroxamate and tris-imidazole Zn(II) is
DHcx = ꢀ108.3 kcal/mol) zinc–oxygen bond distances
when compared to those found in RS-130830/MMP-13
MMP inhibitors have been widely studied and exten-
sively reviewed.5,6 Most potent MMP inhibitors share
two common structural features: a Lewis base that inter-
acts with the catalytic Zn(II); and a hydrophobic group
that is in contact with the binding pocket (Fig. 1).
Among the zinc chelators, hydroxamic acids have been
used extensively to produce potent enzyme inhibition.
Barbituric acid based inhibitors, as exemplified by inhib-
itor 2, appear to provide an interesting alternative to
hydroxamic acid based inhibitors. Though its potency
against MMP-13 has not been reported, compound 2
is reported to be an inhibitor of MMP-8
(IC50 = 107 nM), and MMP-9 (IC50 = 20 nM).5,10 Since
the appearance of 2 in the literature, there have been re-
ports from other laboratories reporting successful incor-
poration of barbituric acids into inhibitors of MMPs,
including MMP-13.10c Thus the use of a barbiturate as
a zinc ligand seemed to be a reasonable starting point
for the design of a potent and selective inhibitor of
MMP-13.
˚
crystal structure (X-ray structure bond lengths: 1.9 A
14
˚
and 2.1 A). The longer bond length is characteristic
of AM1 calculations. AM1 optimization with a barbitu-
ric acid 3 showed that a monodentate nitrogen–Zn(II)
interaction was the most enthalpically favored, with a
˚
Zn(II)–nitrogen distance of 2.09 A (enthalpy of com-
plexation: DHcx = ꢀ91.8 kcal/mol). An overlay of tert-
butyl hydroxamate and barbituric acid 3 is shown in
Figure 2.15
Based on our understanding of the barbiturate–zinc
interaction and the reported X-ray structure of RS-
130830 1 bound to MMP-13 we sought to apply struc-
ture-based design to produce potent barbiturate
containing inhibitors of this enzyme (see Fig. 3). The X-
ray structure of RS-130830 provides the distance from
hydroxamate to zinc ion and the relative distance from
the hydroxamate to the biaryl ether group. The barbitu-
rate zinc model system described above was utilized to
suggest a reasonable starting conformation for projec-
tion into the binding pocket from the barbiturate ligand.
Finally, a linker structure that retained the bound con-
formation of RS-130830 was designed to provide both
constraint and a novel structural motif. The linker re-
gion of RS-130830 that joins the hydroxamic acid to
the biaryl ether group contains at least two bonds that
are to some extent freely rotating and it is unknown if
Barbituric acid could in principle bind to a zinc atom via
interactions with either a carbonyl oxygen, a ring nitro-
gen or a combination of the two atoms.11 To aid inhib-
itor design, we first sought to gain
a
better
understanding of how a barbituric acid might be bound
to zinc in the active site of an MMP by building a model
system.12 Semiempirical AM113 calculations with tert-
Figure 2. Structure of Zn(II) tris-imidazole with barbituric acid 3 (gray) and its overlay with tert-butyl hydroxamate (magenta). Note that complex is
electrically neutral.
Linker that mimics the
bound conformation
O
O
HOHN
O
O
N
O
HN
S
Cl
O
N
H
X
O O
Rapid SAR of with respect
to potency and selectivity
Babaturic acid
as zinc ligand
RS-130830 1
Figure 3. Design features for a selective, non-hydroxamate acid inhibitor of MMP-13.

S.-H. Kim et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1101–1106
1103
the compound shows significant conformational prefer-
ence in solution. In principle, a linker structure that
mimics the bound conformation of acyclic 1 should be
entropically favored (pre-organization).16
culated constraint of the barbituric acid–zinc interac-
tion. Specifically, modeling predicts that the ground
state conformation of spiro-barbiturate 4a to be nearly
identical with the bound conformation of inhibitor 1
in the MMP-13 active site.17 An overlay of spiro-barbi-
turate 4a with RS-130830 1 is shown in Figure 4. It
is interesting to note that the carbonyl oxygen in the
Several constrained linkers fit both the conformational
profile of the docked inhibitor 1 to MMP-13 and the cal-
Figure 4. Modeled structure of barbituric acid 4a in MMP-13 and its overlap with the bound structure of RS-130830 1 (magenta). Also shown is the
MOL2MAP methyl probe interaction energy contour at ꢀ1.8 kcal/mol (yellow) for the RS-130830 binding site.
R
R
O
O
O
O
N
O
b
H2N
N
O
HN
a
EtO
EtO
O
N
H
O
O
O
5
6a, R = H
6b-c, R = Et
4a, R = H
4b-c, R = Et
Scheme 1. Reagents and conditions: (a) diethylbromomalonate, toluene, 100 °C, 36 h (45%); ethylacrylate (6a) or methyl trans-2-pentenoate (6b and
c), NaOEt/EtOH, 80 °C, 3 h (69–71%); (b) urea, NaOEt, EtOH, 80 °C, 5 h (36–44%).
O
O
O
O
O
Br
O
b
N
O
a
N
EtO
EtO
EtO
EtO
EtO
EtO
H2N
OBn
H2N
O
Br
11
8
10
Br
O
OBn
9
7
d
c
O
O
O
O
B(OH)2
O
OH
Ar
Ar
N
N
O
N
O
EtO
EtO
EtO
EtO
EtO
EtO
e
f
O
OH
B(OH)2
X
12
10, 11, or 14
13
h
g
O
O
O
O
N
O
EtO
EtO
N
HN
Br
O
N
H
O
X
10
4d-p
Scheme 2. Reagents and conditions: (a) 7, DMF, 100 °C, ethyl acrylate, NaOEt, EtOH, 3 h (41%); (b) 9, DMF, 100 °C, ethyl acrylate, NaOEt,
EtOH, 3 h (45%); (c) Pd(dppf), bis(pinacolato)diboron, KOAc, DMSO, 65 °C, 3 h; NaIO4, NH4OAc, acetone/H2O (1:1), rt, 3 h (82%); (d) H2
(balloon), 10% Pd–C, 1 h, DMF (92%); (e) ArOH, Cu(OAc)2, TEA, CH2Cl2, rt, 12 h (57–62%); (f) ArB(OH)2, Cu(OAc)2, TEA, CH2Cl2, rt, 12 h (79–
85%); (g) for X = H: H2 (balloon), 10% Pd–C, EtOH, rt, 5 h (86%); for X = Ph: PhB(OH)2, Pd(dppf)Cl2, K3PO4, 65 °C, 3.5 h (93%); (h) urea, NaOEt,
EtOH, 80 °C (35–55%).

1104
S.-H. Kim et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1101–1106
Table 1. SAR of aromatic substitution
Spiro-barbiturate 4h is a potent inhibitor of MMP-13
(Ki = 0.95 nm). Its potency suggests that it interacts with
MMP-13 in a manner similar to RS-130830 and vali-
dates the structure based de novo design of an MMP-
13 inhibitor. Examination of the SAR (Table 1) shows
that MMP-13 potency is maintained with a variety of
aryloxy systems. Replacement of the aryloxy group by
hydrogen (4d) results in nearly a 500-fold loss in po-
tency. It is interesting to note that the biphenyl analogue
4f (Ki = 934 nm) is significantly less potent against
MMP-13 than compound 4a. In general para-substitu-
tion on the aryl ring provided the most potent analogs
(4h,m,n,o,p).
R
O
O
N
HN
O
N
O
X
H
R = H
Compd
X
MMP-13 Ki (nM)a
4d
4e
4a
4f
H
Br
2200
130
4.7
OPh
Ph
934
1100
0.95
54
4g
4h
4i
OBn
OPh(4-Cl)
OPh(3-Cl)
OPh(3-OMe)
MMPs are structurally similar but show differences in
the S10 binding pocket, thus the region of inhibitor
(P10) that interacts with S10 is likely to be important
to selectivity. Deep S10 binding pockets are found in
MMP-13, MMP-12, MMP-9, MMP-3, MMP-2, and
TACE; shallow S10 pockets are found in MMP-1 and
MMP-7.21 In general the compounds in Table 2 have
a improved selectivity for MMP-13 compared to that re-
ported for RS-130830 1;7 however, there was no signifi-
cant difference among the compounds with respect to
selectivity against MMP-2 and MMP-9. The data for
enantiomers 4b and 4c suggest that chirality in the linker
region had an effect on MMP-13 potency but no major
effect on the selectivity profile.
4j
1.9
4k
4l
OPh(2-OMe)
110
5.8
OPh-3,4-(OCH2O)
O(4-OMe-3-pyridyl)
OPh(4-CO2Me)
OPh(4-CO2H)
OPh(4-OPh)
4m
4n
4o
4p
25
4.2
2.7
0.33
a With the exception of compounds 4d, 4e, and 4g (single runs, MMP-
13 IC50), all compounds in Tables 1 and 2 are reported as Kis and
based on triplicate runs with standard deviation of less than 15%.20
lactam ring of 4a overlaps well with one of the sulfone
oxygen atoms in inhibitor 1.
The synthesis of compounds 4a–c is shown in Scheme 1.
Alkylation of 4-phenoxyaniline 5 with bromodiethyl-
malonate followed by condensation with either ethyl
acrylate or methyl trans-2-pentenoate provided 6a–c.
Condensation of 6a–c with urea provided the desired
spiro-barbiturates 4a–c.18
Pre-clinical evidence suggests that inhibition of MMP-
13 could be useful in osteoarthritis and rheumatoid
arthritis. The spiro-barbiturates 4b and 4p have a similar
potency and improved MMP-3 selectivity compared to
the hydroxamic acid based inhibitor 1, (RS-130830).
MMP-13 structural information, together with AM1
calculation of the Zn(II)–barbiturate interaction, led to
a predictive binding model which played a crucial role
in the selection process of the initial constrained lead
4a.15,22 Although significant selectivity against the
MMP-2 and MMP-9 was not achieved for the com-
pounds reported, this work demonstrates that appropri-
ately substituted barbituric acids can achieve similar
potency to hydroxamic acid based systems.
Once it was determined that 4a was a potent inhibitor of
MMP-13, a flexible synthetic route which permitted late
stage diversification of the terminal aryl group was de-
signed as shown in Scheme 2. Boronic acid and
Cu(OAc)2 mediated oxidative couplings were chosen
as a means of accessing a wide range of substitution.19
Condensation of malonates 10, 11, or 14 with urea pro-
vided inhibitors 4d–p.
Table 2. Selectivity SAR of compounds 4a–p
R
O
O
N
X
HN
O
N
O
O
H
Compd
R
X
MMP-1 Ki (nM) MMP-2 Ki (nM) MMP-3 Ki (nM) MMP-9 Ki (nM) MMP-13 Ki (nM) TACE Ki (nM)
1a
—
H
Et
Et
H
H
H
H
—
H
H
H
Cl
590
>5000
>5000
>5000
>5000
0.22
1.6
9.3
0.55
4.7
0.52
4.7
—
4a
4b
4c
4h
4n
4o
4p
4500
92
>1000
>1000
>1000
>1000
>1000
>1000
>1000
0.40
1.7
0.87
7.2
0.54
4.1
1200
430
0.43
4.7
0.72
1.2
0.95
4.2
CO2Me >5000
CO2H
OPh
1300
3200
110
>5000
>5000
0.23
1.8
0.72
1.9
2.7
0.33
a Data for compound 1 reported Ref. 7.

S.-H. Kim et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1101–1106
1105
Staples, D. J.; Baldwin, E. T.; Finzel, B. C. J. Med. Chem.
1999, 42, 1525.
References and notes
11. Wang, B. C.; Craven, B. M. J. Chem. Soc., Chem.
Commun. 1971, 7, 290. An X-ray structure of a barbituric
acid–Zn(II) inorganic complex.
12. The calculations of zinc–barbituric (BA) interaction,
target synthesis, and in vitro evaluation were carried
out prior to the public disclosure of X-ray crystal
structures of barbituric acid bound to MMPs (Ref. 10d
and e).
13. AM1: Dewar, M. J. S.; Zoebisch, E. G.; Healy, E. F.;
Stewart, J. P. J. Am. Chem. Soc. 1985, 107, 3902.
14. Lovejoy, B.; Welch, A. R.; Carr, S.; Luong, C.; Broka, C.;
Hendricks, T.; Campbell, J. A.; Walker, K. A. M.; Martin,
R.; Wart, H. V.; Browner, M. F. Nat. Struct. Biol. 1999, 6,
217.
1. (a) Johnson, L. L.; Dyer, R.; Hupe, D. J. Curr. Opin.
Chem. Biol. 1998, 2, 466; (b) Shapiro, S. D. Curr. Opin.
Chem. Biol. 1998, 10, 602; (c) Whittaker, M.; Floyd, C. D.;
Brown, P.; Gearing, J. H. Chem. Rev. 1999, 9, 2735; (d)
Nagase, H.; Woessner, J. F., Jr. J. Biol. Chem. 1999, 274,
21491.
2. (a) Firestein, G. S.; Paine, M. M.; Littman, B. H.
Arthritis Rheum. 1992, 34, 1094; (b) Walakovits, L. A.;
Bhardwaj, N.; Gallick, G. S.; Lark, M. W.; Clark, I. M.
Arthritis Rheum. 1992, 35, 35; (c) Clark, I. M.; Rowan, A.
D.; Cawston, T. E. Curr. Opin. Anti-Inflam. Immunomod.
Invest. Drugs 2000, 2, 16; (d) Bottomley, K. M.;
Johnson, W. H.; Walter, D. S. J. Enzyme Inhib. 1998,
13, 79; (e) Yip, D.; Ahmad, A.; Karapetis, C. S.;
Hawkins, C. A.; Harper, P. G. Invest. New Drugs 1999,
17, 387.
15. The overlap of calculated structure in Figure 2 versus
reported barbiturate MMP inhibitor structures is very
˚
˚
good. The calculated zinc–nitrogen distance is 2.09 A,
˚
3. Freemont, A. J.; Byers, R. J.; Taiwo, Y. O.; Hoyland, J. A.
Ann. Rheum. Dis. 1999, 58, 357.
which is in line with the distances of 2.09 A and 2.17 A
found for zinc–nitrogen bond (Ref. 10d and e, respec-
4. (a) Neuhold, L. A.; Killar, L.; Zhao, W.; Sung, M.;
Warner, L.; Kulik, J.; Turner, J.; Wu, W.; Billinghurst, C.;
Meijers, T.; Poole, A. R.; Babij, P.; DeGennaro, L. J.
J. Clin. Invest. 2001, 1, 35; (b) Konttinen, Y. T.; Ainola,
M.; Valleala, H.; Ma, J.; Ida, H.; Mandelin, J.; Kinne, R.
W.; Santavirta, S.; Sorsa, T.; Lopez-Otin, C.; Takagi, M.
Ann. Rheum. Dis. 1999, 58, 691.
5. For an excellent review on the design and application of
matrix metalloproteinase inhibitors see Skiles, J. W.;
Gonnella, N. C.; Jeng, A. Y. Curr. Med. Chem. 2001, 8,
425.
6. (a) Skotnicki, J. S.; Levin, J. I.; Zask, A.; Killar, L. M. In
Metalloproteinases as Targets for Anti-Inflamma-
tory Drugs; Bottomley, K. M. K., Bradshaw, D.,
Nixon, J. S., Eds.; Birkhauser: Basel, 1999; pp 17–57; (b)
Clark, I.; Parker, A. E. Exp. Opin. Ther. Targets 2003, 7,
19.
7. Campbell, J. A. Book of Abstracts, 216th ACS National
Meeting of the American Chemical Society, 1998.
8. (a) Bender, S. Lecture at the Second Winter Conference on
Medicinal Chemistry, Steamboat Springs, CO, Jan 26–31,
1997; (b) Tamura, Y.; Watanabe, F.; Nakatani, T.; Yasui,
K.; Fuji, M.; Komurasaki, T.; Tsuzuki, H.; Maekawa, R.;
Yoshioka, T.; Kawada, K.; Sugita, K.; Ohtani, M. J. Med.
Chem. 1998, 41, 1749; (c) Whittaker, M.; Floyd, C. D.;
Brown, P.; Gearing, A. J. Chem. Rev. 1999, 99,
2735.
˚
tively). The reported zinc–oxygen distances of 2.9 A
˚
(MMP-8, Ref. 10d) and 3.0 A (Stromelysin, Ref. 10e)
are significantly longer than the zinc–oxygen distance of
1.9 A found in a hydroxamate/MMP structure.14 Interest-
˚
ingly an inorganic-complex containing of Zn(II) and
barbituric acid shows little evidence that it forms a
bidentate chelate.11 Thus it may not be necessary to
invoke a bidentate binding mode for the barbituric acid–
zinc interaction (penta-coordination around zinc). Impor-
tantly, there is a good agreement between the X-ray and
calculated structures which suggests that calculations may
play an important role in the design of new non-
hydroxamic acid based inhibitors of MMPs (see for
instance, Ref. 10e). Another consideration is the ioniza-
tion state of the barbituric acid. The calculation was
carried out using deprotonated barbituric acid, with the
ligand bearing a formal ꢀ1 charge. Calculations could
not produce a stable complex when the neutral barbitu-
rate forms were substituted (data not shown.) The pKa for
4a was experimentally determined to be 6.4( 0.2)
(spectrophotometric titration in water). Thus, the possi-
bility that the ionized specie is the catalytically active
form cannot be ruled out, however both Ref. 10d and e
suggest the barbituric acid is bound to MMPs in a neutral
form.
16. Cram, D. J.; Trueblood, K. N. In Host Guest Complex
Chemistry; Vogtle, E., Weber, E., Eds.; Springer: Berlin,
1985; pp 1–42.
9. Shaw, T.; Nixon, J. S.; Bottomley, K. M. Exp. Opin.
Invest. Drugs 2000, 9, 1469–1478.
17. (a) The available X-ray coordinates of 1 with MMP-1314
were used in the SYBYL program to arrive at binding site
descriptors, including MOL2MAP surface grids that provide
chemical probe interaction energy contours of the binding
site. As shown in Figure 3, a tris-imidazole complex was
used with the AM1 semiempirical SCF-MO methodology
to examine zinc–barbiturate interaction. The calculated
10. Barbituric acids: (a) Oliva, A.; De Cellis, G.; Grams, F.;
Livi, V.; Zimmermann, G.; Menta, E.; Krell, H. -W., PCT
Int. Appl. WO 9858925, Dec 30, 1998; (b) Grams, F.;
Mermann, G. PCT Int. Appl. WO 9858915, Dec 30, 1998;
(c) Foley, L. H.; Palermo, R.; Dunten, P.; Wang, P.
Bioorg. Med. Chem. Lett. 2001, 11, 969; (d) Brandstetter,
H.; Grams, F.; Glitz, D.; Lang, A.; Huber, R.; Bode, W.;
Krell, H.; Engh, R. A. J. Biol. Chem. 2001, 20, 17405; (e)
Dunten, P.; Kammloft, U.; Crother, W. L.; Foley, L. H.;
Wang, P.; Palermo, R. Protein Sci. 2001, 10, 923; (f) Other
heterocycles: Puerta, T. D.; Lewis, J. A.; Cohen, S. M.
J. Am. Chem. Soc. 2004, 126, 8388; (g) Jacobsen, E. J.;
Mitchell, M. A.; Hendges, S. K.; Belonga, K. L.;
Skaletzky, L. L.; Stelzer, L. S.; Lindberg, T. J.; Fritzen,
E. L.; Schostarez, H. J.; OÕSullivan, T. J.; Maggiora, L. L.;
Stuchly, C. W.; Laborde, A. L.; Kubicek, M. F.; Poor-
man, R. A.; Beck, J. M.; Miller, H. R.; Petzold, G. L.;
Scott, P. S.; Truesdell, S. E.; Wallace, T. L.; Wilks, J. W.;
Fisher, C.; Goodman, L. V.; Kaytes, P. S.; Ledbtter, S. R.;
Powers, E. A.; Vogeli, G.; Mott, J. e.; Trepod, C. M.;
˚
Zn–N barbiturate distance (2.09 A) and geometric con-
straints (monodentate nitrogen chelation, tetrahedral L–
Zn–L of 109.5°) were built into the Tripos molecular
mechanics force field (Tripos FF). Inhibitor modeling was
initiated with a SAM1 optimized structure of the inhibitor,
for example, 4a, inserted into the MMP-13 binding site.
The entire protein/inhibitor aggregate was then allowed to
relax for 1000 iterations of geometry optimization with the
Tripos force field, resulting in the structure shown in
Figure 4; (b )SYBYL: Tripos, Inc., 1699 S. Hanley Road, St.
Louis MO 63144-2913(c) MOL2MAP: MOL2MAP is a FOR-
TRAN program with SYBYL SPL interface (A. T. Pudzia-
nowski) based on the original GRID methodology:

1106
S.-H. Kim et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1101–1106
Goodford, P. J. J. Med. Chem. 1985, 28, 849; (d) Tripos
19. Lam, P. Y. S.; Clark, C. G.; Saubern, S.; Adams, J.;
Averill, K. M.; Chan, D. M. T.; Combs, A. Synlett 2000,
674.
20. Copeland, R. A.; Lombardo, D.; Giannaras, J.; Decicco,
C. P. Bioorg. Med. Chem. Lett. 1995, 5, 1947.
21. Borkakoti, N. In Metalloproteinases as Targets for Anti-
Inflammatory Drugs; Bottomley, K. M. K., Bradshaw, D.,
Nixon, J. S., Eds.; Birkhauser: Basel, 1999; pp 17–57.
22. Structural characterization of 4a bound to MMP-13 will
be reported in due course.
FF: Clark, M.; Cramer, R. D., III; Van Opdenbosch, N. J.
Comp. Chem. 1989, 10, 982; (e) SAM1: Dewar, M. J. S.;
Jie, C.; Yu, G. Tetrahedron 1993, 23, 5003, SAM1 is
implemented in AMPAC 6.0, Semichem, Inc., 7204 S.
Mullen, Shawnee KS 66216 (1997).
18. Enantiomers 4b and 4c were separated by preparatory
chiral HPLC using a Chiracel AS column, isocratic
gradient 40% (1:1 MeOH/EtOH) 60% hexane, 4b
(Rt = 15 min), 4c (Rt = 18.5 min).