Inhibition of matrix metalloproteinase-9 by "multi-prong" surface binding groups

Article abstract of DOI:10.1039/b501780g

A novel strategy of blocking the active site accessibility of MMP-9 by "multi-prong" surface binding groups is described. The Royal Society of Chemistry 2005.

Full text of DOI:10.1039/b501780g

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Inhibition of matrix metalloproteinase-9 by ‘‘multi-prong’’ surface
binding groups{
Abir L. Banerjee, Shakila Tobwala, Manas K. Haldar, Michael Swanson, Bidhan C. Roy, Sanku Mallik* and
D. K. Srivastava*
Received (in Cambridge, MA, USA) 3rd February 2005, Accepted 15th February 2005
First published as an Advance Article on the web 7th March 2005
DOI: 10.1039/b501780g
group’’ inhibitors. Whereas CAs are preferentially inhibited by the
sulfonamide derivatives, MMPs are inhibited by hydroxamates.2
The question arose can benzenesulfonamide (being a weak
inhibitor for MMPs) be converted into a strong inhibitor for
MMP-9 by attaching the IDA-Cu2+ moieties as tether groups, such
that the latter could interact with the surface exposed histidine
residues of the enzyme. To probe this, we synthesized differently
branched benzenesulfonamide derivatives (see ESI{), containing
short chain spacers, which terminated into the IDA-Cu2+ moieties
(Fig. 1). Note that the number of IDA-Cu2+ moieties corresponds
to the conjugate number except for conjugate-5, which contains
4 IDA-Cu2+ moieties but does not contain the benzenesulfonamide
group.
A novel strategy of blocking the active site accessibility of
MMP-9 by ‘‘multi-prong’’ surface binding groups is described.
One of the ultimate goals of biomedical research is to design/
discover enzyme inhibitors (against pathogenic enzymes) as
potential drugs for the treatment of various human diseases.
Despite recent advances in the areas of genomics and proteomics,
the basic approaches in drug discovery have remained more or less
the same.1 The prevailing idea has been to synthesize small
molecular weight compounds (either via ‘‘rationale’’ or ‘‘combi-
natorial’’ methods), which could snugly ‘‘fit’’ into the active site
pockets of putative pathogenic enzymes, precluding the accessi-
bility of their cognate substrates.1,2 However, due to marked
similarity in the active site pockets of functionally similar enzymes
(particularly ‘‘isozymes’’), there have been major challenges in
designing the isozyme selective inhibitors, which could only inhibit
the pathogenic isozymes, but not their physiologically desirable
counterparts.3
Using the recombinant form of human MMP-9, we performed
steady-state kinetic experiments7 for the inhibition of the enzyme
in the presence of the different conjugates shown in Fig. 1
(see ESI{). By measuring the initial rates of the enzyme catalyzed
reaction as a function of increasing concentrations of the
conjugates in Fig. 1, we determined their IC50 values (the
Instead of exclusively focusing on the confines of the enzyme’s
active site pockets, there have been efforts in designing enzyme
inhibitors which interact at the enzyme’s surface exposed residues
as well.4 We recently demonstrated that a weak inhibitor of
carbonic anhydrase (viz., benzenesulfonamide) can be converted
into a strong inhibitor by attaching an iminodiacetate IDA-Cu2+
group via a suitable spacer.5 In such a ‘‘two-prong’’ inhibitor,
whereas the benzenesulfonamide group binds at the active site
pocket of the enzyme, the IDA-Cu2+ moiety loops around and
interacts with one of the surface exposed histidine residues. Since
the latter are unlikely to be conserved during the course of
evolution, our two-prong approach exhibits potential for designing
highly potent and isozyme specific inhibitors of enzymes
(particularly those harboring surface exposed histidine residues).5
Just how effective can targeting of the surface exposed histidine
residues be in designing an isozyme specific inhibitor? In this
pursuit, it was realized that matrix metalloproteinases (MMPs) are
involved in the pathogenesis of various human diseases, such as
cancer, arthritis, atherosclerosis, aneurism, periodontal disease,
skin ulceration, gastric ulcer, liver fibrosis, etc., and some of the
isozymes (e.g., MMP-9) harbor several patches of the surface
exposed histidine residues.6 Although both carbonic anhydrases
(CAs) and matrix metalloproteinases (MMPs) are Zn2+ containing
enzymes, they exhibit different selectivities for ‘‘zinc binding
{ Electronic supplementary information (ESI) available: synthetic details
and inhibition studies. See http://www.rsc.org/suppdata/cc/b5/b501780g/
*sanku.mallik@ndsu.edu (Sanku Mallik)
Fig. 1 Structures of differently branched IDA-Cu2+ containing
dk.srivastava@ndsu.edu (D. K. Srivastava)
conjugates.
This journal is ß The Royal Society of Chemistry 2005
Chem. Commun., 2005, 2549–2551 | 2549

View Article Online
concentration of an inhibitor required to inhibit 50% of the
enzyme activity), which are summarized in Table 1. An
examination of the data in Table 1 revealed that the IC50 value
of the inhibitors decreased by about 50 fold upon increase in the
number of IDA-Cu2+ moieties from one (conjugate-1) to four
(conjugate-4). Since the difference in the IC50 value between
conjugate-2 and conjugate-3 was about 2 fold, we regarded these
inhibitors to be nearly equally potent.
Based on literature data, it is known that at neutral pH, IDA-
Cu2+ preferentially interacts with the imidazole group of the
histidine residues.8 Such an interaction is explicitly corroborated by
a recent NMR spectroscopic study.8 Hence, it appears likely that
the inhibition of MMP-9 by the conjugates of Fig. 1 is due to the
interaction of their IDA-Cu2+ moieties with the enzyme’s histidine
residues. From the point of view of designing enzyme inhibitors, it
is noteworthy that the inhibitory potencies of the above conjugates
increase with an increase in the number of IDA-Cu2+ moieties.
Apparently, the origin of such a feature lies in the ‘‘multi-prong’’
attachment of the IDA-Cu2+ moieties (of the conjugates) to the
surface exposed histidine residues of the enzyme.
The question arose could we justify the multi-prong attachment
of conjugate-4 with the surface exposed histidine residues of
MMP-9 on the basis of the structural data of the protein. If so,
would conjugate-4 be immune to inhibiting other MMPs, which
do not possess similar (or highly abundant) patches of the surface
histidine residues. Such a demonstration would attest to the
selectivity of conjugate-4 for MMP-9. In this pursuit, we compared
the X-ray crystallographic structures of known MMPs from the
point of view of distribution of the surface exposed histidine
residues, and observed that the latter was markedly different
between MMP-9 and MMP-10 (Fig. 2).
Both these enzymes contain two Zn2+ atoms (one catalytic and
the other structural at nearly identical positions), and each of them
are picket-fenced by 3 histidine residues. The major difference
between these two enzymes is the number of surface histidine
residues, which is more abundant in the case of MMP-9 vis a` vis
MMP-10. It is clearly apparent that the corresponding histidine
residues (viz., H117, H118, H119, H441, and H432) of the MMP-9
structure are missing in MMP-10. Hence, the IDA-Cu2+ moieties
of conjugate-4 have a much higher propensity for binding to the
surface (harboring the histidine residues) of MMP-9 than to that
of MMP-10. Given this, we expected that conjugate-4 would serve
as a more potent inhibitor for MMP-9 than for MMP-10. To
probe this, we performed a comparative experiment for the MMP-
9 and MMP-10 catalyzed reactions in the presence of increasing
concentration of conjugate-4. As shown in Fig. 3, conjugate-4
exhibits a much pronounced inhibitory profile with MMP-9 as
compared to that with MMP-10. The inhibition of conjugate-4 for
the MMP-9 catalyzed reaction was judged to be competitive in
Fig. 2 Surface topologies of MMP-9 (top) and MMP-10 (bottom)
showing the surface exposed histidine residues. Created by the software
GRASP.10
Fig. 3 Inhibition of MMP-9 and MMP-10 by conjugate-4.
Table 1 IC50 values for the inhibition of MMP-9
nature. The solid smooth line is the best fit to the data for a Ki
value of 1.1 mM. Although we could not reliably determine the Ki
value of conjugate-4 for MMP-10, it appeared to be several orders
of magnitude higher than that for MMP-9.
Inhibitor
IC50/mM
Conjugate-1
Conjugate-2
Conjugate-3
Conjugate-4
Conjugate-5
Conjugate-4 (without Cu2+
Benzenesulfonamide
49 ¡ 7.3
3.8 ¡ 0.44
9.2 ¡ 1.3
1.1 ¡ 0.011
3.1 ¡ 0.40
79 ¡ 16
To determine the extent to which the benzenesulfonamide
moiety of conjugate-4 contributes to the overall inhibitory profile
of MMP-9, we synthesized conjugate-5 (Fig. 1). Note that the
latter conjugate differs from conjugate-4 only by the absence of the
)
.20 000
2550 | Chem. Commun., 2005, 2549–2551
This journal is ß The Royal Society of Chemistry 2005

View Article Online
active site pockets. Such inhibitors are mechanistically similar to
the binding of selective ‘‘antibodies’’ or charged ‘‘nanoparticles’’ at
the active site entrance of the enzymes.9 Like our multi-prong
conjugates, these ‘‘macromolecular’’ species interfere in the facile
entrance and exit of substrates and reaction products, respectively,
from the interacting enzyme sites, causing the inhibitory effects.
We surmise that the inhibitory potencies of multi-prong
conjugates can be significantly increased by incorporating reason-
ably high affinity (active site directed) ligands (e.g., hydroxamate
derivative in the case of MMP-9) in the overall structure.
The major advantage of such inhibitor designs would be the
formulation of drugs, which would selectively inhibit the
pathogenic isozymes (e.g., selected MMP isozymes) without
affecting their physiologically desirable counterparts. The lack of
such selectivity has been one of the major failures in developing
MMP inhibitors as potential drugs for the treatment of various
human diseases.3
Fig. 4 Cartoon showing the blocking of the active site pocket of MMP-9
by a multi-prong IDA-Cu2+ conjugate.
benzenesulfonamide moiety. The inhibition data (IC50 values)
reveal that conjugate-5 is only 3 fold less potent than conjugate-4
(Table 1). Hence, the benzenesulfonamide moiety of conjugate-4
makes a minuscule energetic contribution to the interaction with
MMP-9. Clearly the binding affinity of the IDA-Cu2+ containing a
four-prong conjugate to MMP-9 is not significantly enhanced by
the presence of the active site binding group, benzenesulfonamide.
In other words, the multi-prong interactions of the IDA-Cu2+
moieties of conjugate-4 or conjugate-5 with the surface exposed
histidine residues of MMP-9 are adequate to preclude the
accessibility of the substrate to the active site of the enzyme,
leading to its inhibition. Since the histidine residue on the surface
of MMP-10 is sparse, the inhibition of the enzyme by conjugate-4
is not so pronounced (as shown by the straight line of Fig. 3). We
propose that the complementary interaction between the IDA-
Cu2+ moieties of multi-prong conjugates and the surface exposed
histidine residues of MMP-9 form a ‘‘claw’’ like structure across
the active site pocket of the enzyme, as schematized in Fig. 4.
To further probe the validity of the multi-prong attachment
model of Fig. 4, we investigated the effectiveness of Cu2+ devoid
conjugate-4 on the MMP-9 catalyzed reaction. As shown in
Table 1, in the absence of Cu2+, conjugate-4 functions as a fairly
weak inhibitor with an IC50 value of 79 ¡ 16 mM. However, we
could not accurately determine the IC50 value of ‘‘free’’
benzenesulfonamide as the latter did not show inhibition up to
10 mM concentration. At higher concentrations, the enzyme
started precipitating due to the solubility problem. Hence, the IC50
value of benzenesulfonamide must be .20 mM (Table 1). Clearly,
the Cu2+ of conjugate-4 (and possibly of the other conjugates
shown in Fig. 1 as well) serves as the major determinant for
inhibition of the MMP-9 catalyzed reaction. We further
investigated the influence of imidazole on the inhibitory profile
of MMP-9, and found that 2 mM imidazole could easily abolish
the inhibitory effect of 5.5 mM conjugate-4 (see ESI{). Clearly,
imidazole competitively abolishes the interaction of IDA-Cu2+ of
conjugate-4 with the surface exposed histidine residues of the
enzyme. We are currently in the process of building a computer
graphic model of the physical interaction between conjugate-4 and
the surface exposed histidine residues of MMP-9, and we will
report these findings subsequently.
Research was supported by the NIH grants 1R15 HL077201-01
to DKS and 1R01 GM 63204-01A1 to SM. We thank Mr. Susmit
Sarkar for drawing the cartoon of Fig. 4.
Abir L. Banerjee, Shakila Tobwala, Manas K. Haldar,
Michael Swanson, Bidhan C. Roy, Sanku Mallik* and D. K. Srivastava*
Department of Chemistry, Biochemistry and Molecular Biology, North
Dakota State University, Fargo, ND-58105.
E-mail: sanku.mallik@ndsu.edu; dk.srivastava@ndsu.edu;
Fax: +1 701-231-8831; Fax: +1 701-231-8324; Tel: +1 701-231-8829
Tel: +1 701-231-7831
Notes and references
1 N. Kley, I. Ivanov and S. Meier-Ewert, Pharmacogenomics, 2004, 5,
395–404.
2 A. M. Davis, S. J. Teague and G. J. Kleywegt, Angew. Chem., Int. Ed.,
2003, 42, 2718–2736; S. J. Teague, Nat. Rev., 2003, 2, 527–541.
3 L. M. Coussens, B. Fingleton and L. M. Matrisan, Science, 2002, 295,
2387–2392; C. M. Overall and C. L. Otin, Nature Rev., 2002, 2, 657–672;
C. T. Supuran, A. Casini and A. Scozzafava, Med. Res. Rev., 2003, 23,
535–558.
4 D. A. Erlanson, J. A. Wells and A. C. Braisted, Annu. Rev. Biophys.
Biomol. Struct., 2004, 33, 199–223; S. R. Adam, R. E. Campbell,
L. A. Gross, B. R. Martin, G. K. Walkup, Y. Yao, J. Llopis and
R. Y. Tsien, J. Am. Chem. Soc., 2002, 124, 6065–6076; P. S. Portoghese,
D. L. Larson, L. M. Sayre, C. B. Yim, G. Ronsisavalle, S. W. Tam and
A. E. Takemori, J. Med. Chem., 1986, 29, 1855–1861.
5 A. L. Banerjee, M. Swanson, B. C. Roy, X. Jia, M. K. Haldar, S. Mallik
and D. K. Srivastava, J. Am. Chem. Soc., 2004, 126, 10875–83;
B. C. Roy, A. L. Banerjee, M. Swanson, X. Jia, M. K. Haldar, S. Mallik
and D. K. Srivastava, J. Am. Chem. Soc., 2004, 126, 13206–13207.
6 M. Egeblad and Z. Werb, Nat. Rev. Cancer, 2002, 2, 161–174;
S. Rowsell, P. Hawtin, C. A. Minshull, H. Jepson, S. M. Brockbank,
D. G. Barratt, A. M. Slater, W. L. McPheat, D. Waterson,
A. M. Henney and R. A. Pauptit, J. Mol. Biol., 2002, 319, 173–81.
7 H. Weingarten and J. Feder, Anal. Biochem., 1985, 147, 437–440.
8 D. W. Pack and F. H. Arnold, Chem. Phys. Lipids, 1997, 86, 135–52;
D. W. Pack, G. Chen, K. M. Maloney, C. T. Chen and F. H. Arnold,
J. Am. Chem. Soc., 1997, 119, 2479–2487; M. Nomura, T. Kobayashi,
T. Kohno, K. Fujiwara, T. Tenno, M. Shirakawa, I. Ishizaki,
K. Yamamoto, T. Matsuyama, M. Mishima and C. Kojima, FEBS
Lett., 2004, 566, 157–61.
9 K. M. Geoge, T. Schule, L. E. Sandoval, L. L. Jennings, P. Taylor and
C. M. Thompson, J. Biol. Chem., 2003, 278, 45512–8; N. O. Fischer,
A. Verma, C. M. Goodman, J. M. Simard and V. M. Rotello, J. Am.
Chem. Soc., 2003, 125, 13387–91.
10 A. Nicholls, K. Sharp and B. Honig, Proteins: Struct., Funct. Genet.,
1991, 11, 281–296.
The experimental outcome of the present investigation provides
a prototype of designed enzyme inhibitors, which blocks the
accessibility of the enzyme’s active sites without fitting into the
This journal is ß The Royal Society of Chemistry 2005
Chem. Commun., 2005, 2549–2551 | 2551