Encounter with unexpected collagenase-1 selective inhibitor: Switchover of inhibitor binding pocket induced by fluorine atom

Article abstract of DOI:10.1016/S0960-894X(01)00796-X

Phosphonamide-based inhibitors having trifluoromethyl moiety showed highly selective inhibition against MMP-1. A possible mechanism of the selectivity of MMP-1 inhibitors through the switchover of the binding pocket was speculated by computational calculations. As a consequence of the unexpected selectivity, the specific interaction of CF₃ group of the inhibitor and Arg214 in the S1′ pocket of MMP-1 conducted a low binding energy.

Full text of DOI:10.1016/S0960-894X(01)00796-X

Bioorganic & Medicinal Chemistry Letters 12 (2002) 581–584
Encounter with Unexpected Collagenase-1 Selective Inhibitor:
Switchover of Inhibitor Binding Pocket Induced byFluorine Atom
Masaaki Sawa,a,* Hirosato Kondoa,y and Shin-ichiro Nishimurab
aDepartment of Chemistry, R&D Laboratories, Nippon Organon K.K., 1-5-90, Tomobuchi-cho,
Miyakojima-ku, Osaka 534-0016, Japan
bDivision of Biological Sciences, Graduate School of Science, Hokkaido University, Sapporo 060-0810, Japan
Received 9 October 2001;accepted 26 November 2001
Abstract—Phosphonamide-based inhibitors having trifluoromethyl moiety showed highly selective inhibition against MMP-1. A
possible mechanism of the selectivity of MMP-1 inhibitors through the switchover of the binding pocket was speculated by com-
putational calculations. As a consequence of the unexpected selectivity, the specific interaction of CF3 group of the inhibitor and
Arg214 in the S10 pocket of MMP-1 conducted a low binding energy. # 2002 Elsevier Science Ltd. All rights reserved.
Matrix metalloproteinases (MMPs) are a family of zinc-
dependent metalloproteinases that have been shown to
play a significant physiological role in extracellular
matrix remodeling.1 The implication of MMPs in a
number of pathological processes has been reported,
and, thus, they are considered to be important ther-
apeutic targets for the treatment of a wide array of dis-
ease processes such as rheumatoid arthritis, tumor
metastasis, multiple sclerosis, and congestive heart fail-
ure.2 It has been reported that side effects were observed
in the clinical studies of MMP inhibitors, because they
showed broad-spectrum inhibition.3 Therefore, specific
inhibitors of MMPs are considered to be attractive tar-
gets in drug discovery research. In general, the studies of
selective inhibitors for MMPs have focused on the
optimization of the S10 binding groups of the inhibitors,
because X-ray analyses of the enzyme–inhibitor com-
plex suggested that the S10 pocket is a selectivity pocket
for MMP inhibitors.4ꢀ6 However, such an optimization
strategy seems to be limited in elucidating the selectivity,
because of their structural homology. We describe
herein the first example of fluorine atom-induced highly
selective inhibitors for collagenase-1, a novel concept
for the design of selective inhibitors for MMPs.
Collagenase-1 (MMP-1) is known to cleave fibrillar
collagens, and has been implicated in the process of
arthritis.7 On the other hand, it has been thought that
the inhibition of MMP-1 may cause a side effect, such as
musculoskeletal syndrome.8 Therefore, a specific inhib-
itor for MMP-1 would be a very useful tool for investi-
gating the physiological role of MMP-1 in vivo.
Recently, we demonstrated that the phosphonamide
derivatives exhibited potent inhibition against MMPs.9
Only the (R)-isomer 1a at the phosphorus atom was
found to be a potent inhibitor of MMP-1, -3 and -9,
while the (S)-isomer 1b was inactive for all MMPs
(Table 1). This observation can be explained by the
binding mode of the inhibitor in the active site of the
MMPs. As shown in Figure 1a, the (R)-isomer could
*Corresponding author at current address: Chemistry Research
Laboratories, Drug Research Division, Dainippon Pharmaceutical
Co., Ltd. 33-94 Enoki-cho, Suita, Osaka 564-0053, Japan. Tel.:
+81-6-6337-5902;fax: +81-6-6338-7656;e-mail: masaaki-sawa@
dainippon-pharm.co.jp
yCurrent address: Manufacturing Technology R&D Laboratories,
Shionogi & Co., Ltd., 1-3 Kuise Terajima 2-chome, Amagasaki,
Hyogo 660-0813, Japan.
Figure 1. Expected binding mode of the phosphonamide inhibitors in
MMPs.
0960-894X/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00796-X

582
M. Sawa et al. / Bioorg. Med. Chem. Lett. 12 (2002) 581–584
Table 1. In vitro profile of the phosphonamide derivatives
Compd
R
Ki (nM)a
MMP-1
MMP-3
MMP-9
1a (R)
1b (S)
2a (R)
2b (S)
3a (R)
3b (S)
4a (R)
4b (S)
5a (R)
5b (S)
6a (R)
6b (S)
CH2CH3
CH2CH3
CH2CH2F
CH2CH2F
CH2CHF2
CH2CHF2
CH2CF3
CH2CF3
4.59
>850
10.5
686
6.00
125
5.20
>650
9.07
>650
6.67
>650
6.75
617
5.05
>800
10.5
>800
4.97
483
3.68
358
4.46
411
Scheme 1. (a) ROH, NaH, THF;(b) KOH, then SOCI 2, cat DMF;(c)
(3R)-N-benzyloxy-1,2,3,4-tetrahydroisoquinoline-3-hydroxamide (10),
diisopropylethamine, THF;(d) H 2, Pd/C.
6.57
35.6
2.21
6.23
10.5
13.2
CH2CH2CF3
CH2CH2CF3
CH2CH2CH2CF3
CH2CH2CH2CF3
3.33
>650
29.4
>650
5.53
>800
aSee ref 9(a) for assay conditions.
bind to MMPs in the following fashion: the p-methoxy-
phenyl group attached to the phosphonamide would
bind to the S10 pocket of the MMPs, and one of the
oxygen atoms of the phosphonamide would H-bond
with the N–H of the mainchain of the MMPs. On the
other hand, the (S)-isomer could not bind to the MMPs,
because of the steric hindrance of the ester group
attached to the phosphonamide (Fig. 1b).
Figure 2. Proposed binding mode of (S)-inhibitors in MMP-1.
increase in the inhibitory activity for MMP-1, while lit-
tle or no inhibition of MMP-3 and -9 was observed. The
trifluoroethyl ester 4b showed excellent inhibitory activ-
ity and selectivity against MMP-1 (Ki: 35.6 nM, selec-
tivities for MMP-3 and -9: 17-fold and 10-fold,
respectively). The extension of the methylene chain
resulted in a further increase in the inhibitory activity
for MMP-1, and it was found that the trifluoropropyl
ester in 5b was the optimum substituent. The Ki value of
compound 5b against MMP-1 was 6.23 nM, and the
selectivities for MMP-3 and -9 were significantly
increased to >104-fold and 66-fold, respectively.
In the course of the study, we have found that the
introduction of fluorine atoms into the ester moiety of
the (S)-isomers led to highly potent and selective inhi-
bition against MMP-1. Therefore, we have investigated
the effects of fluorine atoms on the activity and selec-
tivity profile in more detail.
The synthesis of compounds 2–6 is shown in Scheme 1.
Commercially available p-methoxyphenylphosphonic
dichloride 7 was treated with the appropriate sodium
alkoxide to provide the diester 8, and then was selec-
tively hydrolyzed by potassium hydroxide to give the
corresponding monoester, which was converted to the
corresponding phosphonyl chloride 9 by refluxing with
thionyl chloride in the presence of catalytic DMF. The
phosphonyl chloride 9 was then coupled with the (3R)-
1,2,3,4-tetrahydroisoquinoline derivative 10 in THF in
the presence of diisopropylethylamine (DIEA), to yield
the (3R)-phosphonamide 11 as a mixture of diastereo-
mers (1:1). Deprotection of the benzyl group in 11 with
10% Pd/C gave the diastereomerically pure two hydrox-
amic acids, which were successfully separated by HPLC
purification.10 The compound 1b was crystallized and
analyzed by X-ray diffraction analysis,9b and the ste-
reochemistry at phosphorus atom of compound 1b was
determined to be S configuration. The stereochemistries
of the other compounds were assigned by a character-
This unexpected appearance of the inhibitory activity of
the (S)-isomer could be explained by the switching of
the binding mode in MMP-1, as shown in Figure 2. This
new binding mode would bring the ester moiety in the
(S)-isomer into the S10 pocket, and therefore, the con-
flict of the ester moiety with the enzyme would be alle-
viated. To confirm this novel binding mode, we have
synthesized compounds 12–13. The introduction of long
alkoxy chains to the para-position of the phenyl ring,
such as the ethoxyethoxyphenyl group, is known to
eliminate the inhibitory activity against MMP-1,9
because of the small, closed shape of the S10 pocket of
MMP-1.4 If the trifluoroalkylated (S)-isomer binds as in
Figure 1b, then the (S)-isomers of the ethoxyethoxy-
phenyl derivatives could not inhibit MMP-1. Com-
pounds 12–13 were synthesized from diethyl 4-(2-
ethoxyethoxy)phenylphosphonate using a procedure to
that described above.12
1
istic signal pattern in H NMR, especially in an aro-
matic region.
As expected, compound 13b showed potent inhibitory
activity against MMP-1, while the activity for MMP-1
of the (R)-isomer 13a was completely lost (Table 2).
Compound 13b also showed a selective inhibition profile
(Ki:>650 and>800 nM for MMP-3 and -9, respectively).
As seen from Figure 2, the ethoxyethoxyphenyl group
of (S)-isomers would be directed toward the S20/S30 site
and have the potentiality to interact with these sites.
Therefore, the weak activity for MMP-1 observed with
Compounds 2–6 were evaluated for the inhibition of
MMP-1, -3 and -9 (Table 1).11 Whereas the (S)-isomer
of the ethyl ester derivative (1b) showed no inhibition in
the nanomolar range against MMPs, the (S)-isomers of
the fluorine-containing ester derivatives exhibited inhi-
bitory activity against MMP-1. An increase in the
number of fluorine atoms resulted in a dramatic

M. Sawa et al. / Bioorg. Med. Chem. Lett. 12 (2002) 581–584
Table 2. In vitro profile of the phosphonamide derivatives
583
In summary, we have described the first example of a
highly selective MMP inhibitor by the fluorine atom-
induced switching of the binding mode. The (S)-form of
the 3,3,3-trifluoropropyl ester derivative (5b) showed
potent inhibitory activity against MMP-1 with a highly
selective profile. The different binding mode of this type
of compound is likely to enable the inhibition of MMP-
1. This study reveals the potential of the phosphon-
amide derivatives as a new type of MMP inhibitor, and
provides a novel concept for the design of selective
inhibitors.
Compd
R
Ki (nM)a
MMP-1
MMP-3
MMP-9
12a (R)
12b (S)
13a (R)
13b (S)
CH2CH3
CH2CH3
CH2CH2CF3
CH2CH2CF3
>850
573
>850
31.4
30.0
>650
47.5
14.2
>800
24.1
>650
>800
aSee ref 9(a) for assay conditions.
Acknowledgements
Special thanks are addressed to Ms. Kiriko Kurokawa
and Ms. Etsuko Ishibushi for expert technical assistance.
References and Notes
1. Woessner, J. F., Jr. FASEB J. 1991, 5, 2145.
2. Whittaker, M.;Floyd, C. D.;Brown, P.;Gearing, A. J. H.
Chem. Rev. 1999, 99, 2735 a review and references therein.
3. (a) Nemunaitis, J.;Poole, C.;Primrose, J.;Rosemurgy, A.;
Malfetano, J.;Brown, P.;Berrington, A.;Cornish, A.;Lynch,
K.;Rasmussen, H.;Kerr, D.;Cox, D.;Millar, A. Clin. Cancer
Res. 1998, 4, 1101. (b) Heath, E. I.;Grochow, L. B. Drugs
2000, 59, 1043.
Figure 3. The interactions of compound 5b bound to MMP-1.
4. Lovejoy, B.;Welch, A. R.;Carr, S.;Luong, C.;Broka, C.;
Hendricks, R. T.;Campbell, J. A.;Walker, K. A. M.;Martin,
R.;Wart, H. V.;Browner, M. F. Nat. Struct. Biol. 1999, 6,
217.
compound 12b would be induced by such a new inter-
action.
5. Pavlovsky, A. G.;Williams, M. G.;Ye, Q.-Z.;Ortwine,
D. F.;Purchase, C. F., II;White, A. D.;Dhanaraj, V.;Roth,
B. D.;Johnson, L. L.;Hupe, D.;Humblet, C.;Blundell, T. L.
Protein Sci. 1999, 8, 1455.
6. Babine, R. E.;Bender, S. L. Chem. Rev. 1997, 97, 1359.
7. (a) Mitchell, P. G.;Magna, H. A.;Reeves, L. M.;Lopresti-
Morrow, L. L.;Yocum, S. A.;Rosner, P. J.;Geoghegan,
K. F.;Hambor, J. E. J. Clin. Invest. 1996, 97, 761. (b) Bill-
inghurst, R. C.;Dahlberg, L.;Ionescu, M.;Reiner, A.;
Bourne, R.;Rorabeck, C.;Mitchell, P.;Hambor, J.;Diek-
mann, O.;Tschesche, H.;Chen, J.;Wart, H. V.;Poole, A. R.
J. Clin. Invest. 1997, 99, 1534.
8. Leff, R. L. Ann. N.Y. Acad. Sci. 1999, 878, 201.
9. (a) Sawa, M.;Kiyoi, T.;Kurokawa, K.;Kumihara, H.;
Yamamoto, M.;Miyasaka, T.;Ito, Y.;Hirayama, R.;Inoue,
T.;Kirii, Y.;Nishiwaki, E.;Ohmoto, H.;Maeda, Y.;Ishi-
bushi, E.;Inoue, Y.;Yoshino, K.;Kondo, H. Submitted for
publication. (b) Kurokawa, K.;Kanehisa, N.;Sawa, M.;
Kondo, H.;Kai, Y. Acta Cryst. 2001, E57, 906.
10. Satisfactory characteristics data were obtained for all new
compounds. Characteristics are given for a representative
compound: 5b;colorless solids; 1H NMR (DMSO-d6) d 2.70–
2.85 (m, 2H), 2.95–3.15 (m, 2H), 3.75 (s, 3H), 3.92 (dd, J=7.5
and 16.1 Hz, 1H), 4.20–4.50 (m, 3H), 4.55–4.60 (m, 1H), 6.90–
7.00 (m, 3H), 7.00–7.15 (m, 3H), 7.53 (dd, J=8.8 and 12.5 Hz,
2H), 8.78 (s, 1H), 10.59 (br s, 1H);MALDI-TOF MS 497
[M+K]+, 481 [M+Na]+, 459 [M+H]+. Anal. calcd for
C20H22F3N2O5P: C, 52.41;H, 4.84;N, 6.11. Found: C, 52.15;
H, 4.94;N, 6.02.
To explain the observed selectivity, the complex model
of MMP-1 and compound 5b was constructed (Fig. 3).13
This model strongly suggested the possibility of an H-
bond interaction between the fluorine atom and Arg 214
at the bottom of the S10 pocket. One of the fluorine
atoms in the trifluoromethyl group was placed 2.8 A
away from the terminal nitrogen atom of Arg 214. As
this Arg 214 is known to undergo a conformational
change to accommodate inhibitors with large sub-
stituents,4 MMP-1 could tolerate the chain length of the
ester group (4b, 5b, and 6b). It has been reported that
the specific interaction of a CF3 group and an Arg resi-
due could induce a conformational change within an
enzyme.14 On the other hand, the shape of the S10
pockets of MMP-315 and -916 is large and deep, which is
much different from that of MMP-1. Therefore, the
fluorine atom of the (S)-form inhibitors could not
interact as in MMP-1, and thus the change of the bind-
ing mode would be difficult to be induced. However, it
seems not only the H-bond interaction between the
fluorine atoms and Arg 214 but the hydrophobic inter-
action between the fluorinated ester and the hydro-
phobic S10 pocket slightly contributed to the binding
affinity. So the fluorinated compounds 3b, 4b and 5b
would show weak activity against MMP-9 without Arg
214 equivalent in the S10 pocket. However, a more
detailed study, such as an X-ray analysis of the enzyme–
inhibitor complexes, will be required to understand the
observed results. These studies are now in progress.
11. Recombinant human collagenase-1 (MMP-1), stromelysin
1 (MMP-3), and gelatinase B (MMP-9) were used in our
studies [see ref 9(a) for assay conditions].

584
M. Sawa et al. / Bioorg. Med. Chem. Lett. 12 (2002) 581–584
12. Satisfactory characteristics data were obtained for all new
compounds. Characteristics are given for a representative
compound: 13b;colorless solids; 1H NMR (DMSO-d6) d 1.08
(t, J=7.0 Hz, 3H), 2.65–2.85 (m, 2H), 2.95–3.15 (m, 2H), 3.45
(q, J=7.0 Hz, 2H), 3.60–3.70 (m, 2H), 3.92 (dd, J=7.5 and
16.5 Hz, 1H), 4.05–4.10 (m, 2H), 4.20–4.45 (m, 3H), 4.50–4.60
(m, 1H), 6.90–7.00 (m, 3H), 7.00–7.15 (m, 3H), 7.52 (dd,
J=8.7 and 12.5 Hz, 2H), 8.77 (d, J=1.5 Hz, 1H), 10.59 (d,
J=1.5 Hz, 1H);MALDI-TOF MS 555 [M+K] +, 539
[M+Na]+, 517 [M+H]+. Anal. calcd for C23H28F3N2O6P: C,
53.49;H, 5.46;N, 5.42. Found: C, 53.29;H, 5.53;N, 5.37.
13. Models of MMP-1 complexed with the phosphonamide
inhibitors were constructed based on the crystal structure of a
MMP-1/peptide-based inhibitor complex (PDB code, 2TCL5).
Firstly, the phosphonamide inhibitor was placed in the active
site of the MMP-1 to occupy the S10 pocket with the R group.
Then, the hydroxamate unit of the phosphonamide inhibitor
was superimposed on that of the peptide-based inhibitor. The
peptide-based inhibitor was replaced by the phosphonamide
inhibitor, and the obtained protein–ligand complex was sub-
sequently energy minimized treating all ligand atoms and
retaining all protein atoms. Docking and energy minimiza-
tions were performed with the TRIPOS force field within the
SYBYL program. All computations were performed on an
Octane Silicon Graphics workstation. The two complexes (1b/
MMP-1 and 5b/MMP-1) were energy minimized and the dif-
ferences of the binding energies between 1b/MMP-1 and 5b/
MMP-1 were obtained. As a result, the total binding energy of
5b/MMP-1 was lower than that of 1b/MMP-1, and the energy
difference is about 7 kcal/mol. This difference is equal to
2.5ꢁ105 times difference of Ki value. Moreover, it was found
that the interactions of the CF3 and Arg214 mainly con-
tributed to this stabilization.
14. Mattos, C.;Giammona, D. A.;Petsko, G. A.;Ringe, D.
Biochemistry 1995, 34, 3193.
15. (a) Becker, J. W.;Marcy, A. I.;Rokosz, L. L.;Axel,
M. G.;Burbaum, J. J.;Fitzgerald, P. M. D.;Cameron, P. M.;
Esser, C. K.;Hagmann, W. K.;Hermes, J. D.;Springer, J. P.
Protein Sci. 1995, 4, 1966. (b) Esser, C. K.;Bugianesi, R. L.;
Caldwell, C. G.;Chapman, K. T.;Durette, P. L.;Girotra,
N. N.;Kopka, I. E.;Lanza, T. J.;Levorse, D. A.;MacCoss,
M.;Owens, K. A.;Ponpipom, M. M.;Simeone, J. P.;Harri-
son, R. K.;Niedzwiecki, L.;Becker, J. W.;Marcy, A. I.;Axel,
M. G.;Christen, A. J.;McDonnell, J.;Moore, V. L.;Ols-
zewski, J. M.;Saphos, C.;Visco, D. M.;Shen, F.;Colletti, A.;
Krieter, P. A.;Hagmann, W. K. J. Med. Chem. 1997, 40, 1026.
16. Olson, M. W.;Bernardo, M. M.;Pietila, M.;Gervasi,
D. C.;Toth, M.;Kotra, L. P.;Massova, I.;Mobashery, S.;
Fridman, R. J. Biol. Chem. 2000, 75, 2661.