N-Hydroxyurea as zinc binding group in matrix metalloproteinase inhibition: Mode of binding in a complex with MMP-8

Article abstract of DOI:10.1016/j.bmcl.2005.09.057

The first crystallographic structure of an N-hydroxyurea inhibitor bound into the active site of a matrix metalloproteinase is reported. The ligand and three other analogues were prepared and studied as inhibitors of MMP-2, MMP-3, and MMP-8. The crystal s

Full text of DOI:10.1016/j.bmcl.2005.09.057

Bioorganic & Medicinal Chemistry Letters 16 (2006) 20–24
N-Hydroxyurea as zinc binding group in matrix metalloproteinase
inhibition: Mode of binding in a complex with MMP-8
Cristina Campestre,a Mariangela Agamennone,a Paolo Tortorella,a, Serena Preziuso,a
Alessandro Biasone,a Enrico Gavuzzo,b Giorgio Pochetti,b Fernando Mazza,c
Oliver Hiller,d Harald Tschesche,d Valerio Consalvie and Carlo Gallinaa,*
a
`
Dipartimento di Scienze del Farmaco, Universita degli Studi ÔG. dÕAnnunzioÕ, Chieti, Italy
bIstituto di Cristallografia del CNR, Montelibretti, Roma, Italy
c
`
Dipartimento di Chimica, Ingegneria Chimica e Materiali, Universita degli Studi, LÕAquila, Italy
dFakulta¨t fu¨r Chemie, Universita¨t Bielefeld, Germany
`
Dipartimento di Scienze Biochimiche, Universita La Sapienza, Roma, Italy
e
Received 19 May 2005; revised 20 September 2005; accepted 22 September 2005
Available online 18 October 2005
Abstract—The first crystallographic structure of an N-hydroxyurea inhibitor bound into the active site of a matrix metalloproteinase
is reported. The ligand and three other analogues were prepared and studied as inhibitors of MMP-2, MMP-3, and MMP-8. The
crystal structure of the complex with MMP-8 shows that the N-hydroxyurea, contrary to the analogous hydroxamate, binds the
catalytic zinc ion in a monodentate rather than bidentate mode and with high out-of-plane distortion of the amide bonds.
Ó 2005 Elsevier Ltd. All rights reserved.
Matrix metalloproteinases are a class of proteolytic en-
zymes that use an electrophilic zinc ion to catalyze
hydrolysis of the peptide bond.1 They are involved in
many important physiological processes. Over-regula-
tion of their activity, however, results in uncontrolled
degradation of the extracellular matrix in diseases such
as cancer,2 arthritis,3 and multiple sclerosis.4 Design
and synthesis of low molecular weight MMP inhibitors
have thus become an important target for many medic-
inal chemists and pharmaceutical companies.5
selectivity against various MMPs, only a small set of
zinc binding groups (ZBGs) have been identified and
employed in the design of MMP inhibitors. They in-
clude, in an arbitrary, but generally accepted, order of
efficacy,6 hydroxamate > reversed hydroxamate > thio-
late > phosphinate > carboxylate and phosphonate. An
interesting series of eleven new heterocyclic zinc chela-
tors, reminiscent of barbiturate ligands,7 have been
described,8a and the importance of novel ZBGs, togeth-
er with elucidation of their mode of binding, has been
further emphasized.8b
MMP inhibitors generally incorporate a substrate-like
fragment that can be accommodated in at least one of
the subsites of the enzyme active site and a functional
group capable of binding the catalytic zinc ion. While
a practically endless variability of the peptide or pepti-
domimetic fragment allows modulation of potency and
The hydroxamate9 function appears to be ideally suited
as ZBG for MMP inhibitors since it not only forms five-
membered chelates with optimal zinc–oxygen distances,
but even establishes two strong hydrogen bonds with the
protein (Fig. 1a).10 Hydroxamic acids, however, are fre-
quently affected by rapid excretion, low oral bioavail-
ability, and in vivo hydrolysis.11 The N-hydroxyurea
group, incorporating the CO–NH–OH determinant nec-
essary for zinc chelation can, in principle, retain all the
binding interactions of the hydroxamate. It has been
therefore evaluated against enzymes containing a cata-
lytic zinc ion such as carboxypeptidase A,12 carbonic
anhydrase,13 and MMPs by Abbott researchers14 and
Keywords: MMPs inhibitors; Zinc binding groups; N-Hydroxyurea
mode of binding; N-Hydroxyurea inhibitors; X-ray crystal structure;
Molecular modeling.
*
Corresponding author. Tel.: +39 0871 3555373; fax: +39 0871
3555267; e-mail: cgallina@unich.it
`
Present address: Dipartimento Farmaco-Chimico, Universita degli
Studi, Bari, Italy.
0960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.09.057

C. Campestre et al. / Bioorg. Med. Chem. Lett. 16 (2006) 20–24
21
One reason for the decrease in binding affinity of N-
hydroxyureas with respect to hydroxamate analogues
is the difference in conformational preferences. In the
solid state, N-hydroxyureas bearing at least one hydro-
gen atom on N3 (Fig. 1b) adopt a trans N1–CO amide
bond conformation,17 unproductive for zinc chelation.
This trans conformation is stabilized by intramolecular
hydrogen bonding in N-hydroxyureas,17 while hydroxa-
mic acids prefer, in general, the cis N–CO amide bond
conformation. Furthermore, the partial sp2 character
of the N-hydroxyurea N3–CO bond, with its implica-
tions on the geometry and conformational preferences,
strongly decreases the chain flexibility required for the
optimal alignment of the ligand.
Figure 1. (a) Key bonding interactions of batimastat hydroxamate
˚
function as found in the complex with MMP-8 (distances in A). (b)
Preferred conformation of N-hydroxyureas bearing a hydrogen atom
on N3 in the solid state.
by ourselves (unpublished results), providing rather
weak inhibitors. This ZBG, however, may be interesting
under the pharmacokinetic and toxicity profile, since
several 5-lipoxygenase N-hydroxyurea inhibitors with
improved oral bioavailability and slower in vitro metab-
olism, with respect to hydroxamates, were reported.15
On the basis of these considerations, N3 methylation of
N-hydroxyurea ligands was expected to improve binding
affinity for MMPs by hampering the trans N1–CO
amide bond conformation, balancing the stability of
the cis–trans conformers of the N3–CO amide bond,
and lowering the rotational barriers for their intercon-
version. The N3-methyl derivative 2b of the N-hydroxy-
urea 2a, and the new analogues 3a and 3b, based on a
flexible biphenyl ether hydrophobic moiety and a short-
er spacer, were synthesized (Schemes 1 and 2) and tested
against MMP-2, MMP-3, and MMP-8 (Table 1). Only
the N-methyl derivatives 2b and 3b allowed determina-
tion of their IC50, owing to the increase of their solubil-
ity in the assay conditions. Available data show that
methylation of 2a and 3a increased the binding affinity
for MMP-2 and MMP-3, respectively. The IC50 against
MMP-8, however, are in the range of millimolar and are
too high to allow evaluation of the effect of methylation.
To further investigate the behavior of this ZBG, we
focused our attention on the described16 nonpeptide
hydroxamate 1 and the structurally related N-hydroxy-
urea 2a14 (Fig. 2). They were reported to bind in the
MMP-3 active site, accommodating the lipophilic biaryl
moiety into the large S10 hydrophobic pocket, and ligat-
ing the catalytic zinc ion by the hydroxamate or
N-hydroxyurea group, respectively. The reported IC50
of 2a against MMP-3 was 80 lM,14 more than 20-fold
higher than that of the hydroxamate analogue 1. The
hydroxamate containing one less methylene in the linker
between N-hydroxyurea and the biphenyl group16 was
some 100-fold more potent than 1 and would have been
a much more attractive model for the present study, but
the corresponding N-hydroxyurea presented serious
problems of stability and synthesis.
Although none of the N-hydroxyureas 2a–3b attains an
IC50 lower than 58 lM, it is interesting to note that 2b
shows a 20-fold increase of potency by going from
MMP-8 to MMP-2. Rationalization of these results
would require some knowledge on the behavior of
N-hydroxyurea as ZBG. To shed some light on this mat-
ter, lacking adequate amounts of MMP-2 for the cocrys-
tallization, we determined the crystal structure18 of the
complex of 2b with MMP-8.
The main binding interactions of 2b in the active site of
MMP-8 are reported in Figure 3. The lipophilic biphe-
nyl substituent occupies the hydrophobic primary spec-
ificity pocket S10. The N-methyl group is involved in
favorable hydrophobic interactions with the Ile159 side
Figure 2. N-Hydroxyureas structurally related to the hydroxamate 1
studied as MMP inhibitors.
Scheme 1. Reagents and conditions: (a) BrCH2CH2OH, K2CO3, DMF, 90 °C, N2; (b) CBr4, PPh3, THF, N2; (c) NaN3, Bu4NI, CH3CN, reflux;
(d) PPh3, o-dichlorobenzene, reflux; H2O/acetic acid, reflux; (e) phenyl-N-hydroxycarbamate, pyridine, N2; (f) KI, CH3NH2, THF, 40 °C;
(g) p-nitrophenyl-N-hydroxycarbamate, CH3CN, N2.

22
C. Campestre et al. / Bioorg. Med. Chem. Lett. 16 (2006) 20–24
Scheme 2. Reagents and conditions: (a) BH3, THF, N2; (b) phenyl-N-hydroxycarbamate 11a, pyridine, N2; (c) p-nitrobenzenesulfonyl chloride,
pyridine, CH2Cl2; (d) CH3I, K2CO3, DMF; (e) HSCH2COOH, LiOH, DMF; (f) KI, CH3NH2, THF, 40 °C; (g) p-nitrophenyl-N-hydroxycarbamate,
CH3CN, N2.
a
Table 1. IC50 (lM), or percent inhibition, of N-hydroxyureas 2a–3b
against MMP-2, MMP-3, and MMP-8, at pH 7.5
retention of the binding interactions proper of hydroxa-
mate, are beyond the extensive flexibility of the MMP
active site loop.23 Thus, the N-hydroxyurea is forced
to bend and to ligate as monodentate ZBG. The
O@C–N1–O torsion angle (ꢀ122°) highly deviates from
planarity, and the two oxygen atoms are closer to the
trans rather than cis amide bond conformation, adopted
by hydroxamate in zinc chelation. The amide N3–CO
bond is also significantly distorted from planarity as
shown by the high value of the O@C–N3–Me torsion
angle (ꢀ43°). The presence of three non-H substituents
on N3 allows evaluation of the valence angles of this
atom. Their sum (325°) denotes a substantial sp3 charac-
ter of the nitrogen, which is in accordance with the high
value of the out-of-plane distortion (43°) of its amide
bond. The high value (58°) of the out-of-plane distortion
of the N1–CO amide bond indicates negligible amide
resonance and suggests substantial sp3 character also
for the N1 nitrogen.
Compound
MMP-2
MMP-3
MMP-8
1
NDb
>120d
58
>380d
685
3.4c
NDb
>120d
1200
>380d
1000
2a
2b
3a
3b
22% at 120e,f
200
>380d
380
a Replicate determinations indicate standard deviations less than 20%.
b Not determined.
c Ref. 16, pH 6.5.
d No inhibition observed at the reported concentration, upper solu-
bility limit in the assay conditions.
e Percent inhibition at the reported concentration.
f IC50 80 lM reported in Ref. 14.
Two hypothetical monodentate N-hydroxyurea zinc
interactions in the active site of carboxypeptidase A12
and carbonic anhydrase13 were previously reported
and the N-hydroxyurea group is proposed to bind in
the deprotonated form, in accordance with the assump-
tion of a value of 6.16 for its pKa. An experimental pKa
value of 9.7 has been however reported24 for H2N–CO–
NHOH, and calculated pKa values, in the 10–11 range,
are easily available25 for several N3-alkyl-N1-hydroxyu-
reas. In accordance with its weak acidity, the N-hy-
droxyurea group of the MMP-8:2b complex binds in
the protonated form (Fig. 2, N1–O . . .ꢀOOC–Glu198
Figure 3. Schematic representation of the main bonding interactions of
˚
the N-hydroxyurea 2b in the MMP-8 active site (distances in A).
˚
chain and the ethereal oxygen is H-bonded to Leu160
NH. The inhibitor hydroxyl is coordinated by the cata-
distance 3.3 A).
˚
lytic zinc ion (2.3 A) and forms an H-bond with the
Glu198 carboxylate, whereas the carbonyl oxygen is
It was also interesting to compare the mode of binding
of the N-hydroxyurea 2b and the reference hydroxa-
mate 1 containing an sp3 methylene group in place of
the N3-CH3 group of 2b. To this purpose, hydroxa-
mate 1 was docked into the MMP-8 binding site
(Fig. 4) using the program FLO+/QXP.26 The simulat-
ed docking conformation of 1 (yellow carbon atoms)
retains both zinc chelation and the two H-bonds typi-
cal of the hydroxamate ZBG, although the bisaryloxy
group is more deeply buried into the S10 pocket and
the H bond between the ethereal oxygen and Leu160
NH is lost. In the MMP-8:2b complex, both the high
out-of-plane distortions, involving N1 and N3 amide
bonds, and the pyramidal character of N3 show that
the presumably planar preferred conformation of the
N-hydroxyurea17 is forced to bend, in the active site
of MMP-8, in order to coordinate the catalytic zinc
˚
turned toward the solvent, 4.8 A away from the zinc
ion. The inhibitor NH forms a water mediated H-bond
with the Ala163 CO group. The bridged water is further
bonded to Ala163 NH and Glu198 carboxylate. Crystal
structures of native MMP-219 and MMP-320 contain a
water molecule coordinated by the catalytic zinc ion
and H-bonded to the conserved Glu carboxylate, and
all MMPs, presumably, bind water prior to substrate
or inhibitor bonding. This water molecule may be com-
pletely displaced upon inhibitor binding,6 or retained as
bridged water, in some complexes with hydroxamates21
and carboxylates.22
Conformational changes required to accommodate the
N-hydroxyurea ZBG in its planar conformation, with

C. Campestre et al. / Bioorg. Med. Chem. Lett. 16 (2006) 20–24
23
Figure 6. Superimposition between the calculated geometry of the
adduct formed by MMP-2 (orange) with the inhibitor 2b (light blue
carbon atoms) and the crystallographic pose of the same inhibitor
(green carbon atoms) in the MMP-8 (yellow). Only the relevant section
of the active site is displayed.
Figure 4. Superimposition of the crystallographic pose of 2b (green
carbon atoms) with the hydroxamate 1 (yellow carbon atoms) as
docked into the MMP-8 binding site. MMP-8 is represented as a solid
ribbon.
Leu164 NH. In the MMP-2:2b complex (Fig. 6),
the loop defining the S10 pocket is moved away about
˚
1 A from the biphenyl internal ring, with respect to
ion by the hydroxyl oxygen and simultaneously insert
the biphenyl group into the S10 pocket.
MMP-8:2b. While the two rings of the biphenyl in the
MMP-8:2b complex are forced in a practically eclipsed
conformation (dihedral angle 9°), they can be better
accommodated, in the wider MMP-2 active site, with a
dihedral angle of 47°. The ensuing release of the torsion-
al strain probably contributes to the observed increase
in affinity for MMP-2. To further improve N-hydroxyu-
reas as MMP inhibitors, work is in progress to design
and test ligands that may favor zinc chelation by
the N-hydroxyurea group in its preferred planar
conformation.
Although N-hydroxyurea group was reported27 to form
chelation complexes with Fe(III) and Cu(II) metal ions,
it seems to be less effective than hydroxamate for zinc
chelation in MMP inhibitors. The monodentate zinc
coordination observed in the MMP-8:2b complex is
probably also conserved in the active site of MMP-2
and MMP-3. When 2b was minimized26b into the active
site of these enzymes, maintaining the experimentally
observed zinc coordination geometry, the simulated
complexes could explain the increased affinity for
MMP-2 and MMP-3. The MMP-3:2b complex (Fig. 5)
is stabilized by a further H-bond of the ethereal oxygen
with Ala165 NH, in addition to the usual H-bond with
Acknowledgments
This work was supported by Italian MIUR and by the
Deutsche Forschungsgemeinschaft, Germany, Grant
Ts 8/37-3 (H.T.). We are grateful to Prof. S. Fornarini
`
(Universita di Roma ÔLa SapienzaÕ) for the FT-ICR
mass spectrometric data.
Supplementary data
Chemistry; crystallographic details; MMP inhibition;
molecular modeling. Supplementary data associated
with this article can be found, in the online version, at
doi:10.1016/j.bmcl.2005.09.057.
References and notes
Figure 5. Calculated geometry of 2b (green carbon atoms) minimized
in the MMP-3 active site. Two H-bonds connecting 2b ethereal oxygen
with Ala165 and Leu164 are shown as green dotted lines. The rest of
the enzyme is represented as a solid ribbon.
1. Woessner, J. F.; Nagase, H. Matrix Metalloproteinases
and TIMPS; Oxford University Press: Oxford, 2000.

24
C. Campestre et al. / Bioorg. Med. Chem. Lett. 16 (2006) 20–24
2. (a) Yip, D.; Ahmad, A.; Karapetis, C. S.; Hawkins, C. A.;
16. Hajduk, P. J.; Sheppard, G.; Nettesheim, D. G.; Olejnic-
zak, E. T.; Shuker, S. B.; Meadows, R. P.; Steinman, D.
H.; Carrera, G. M.; Marcotte, P. A.; Severin, J.; Smith,
H.; Gubbins, E.; Simmer, R.; Holzmann, T. F.; Morgan,
D. W.; Davidsen, S. K.; Summers, J. B.; Fesik, S. W.
J. Am. Chem. Soc. 1997, 119, 5818.
Harper, P. G. Invest. New Drug 1999, 17, 387; (b) Nelson,
A. R.; Fingleton, B.; Rothemberg, M. L.; Matrisian, L. M.
J. Clin. Oncol. 2000, 18, 1135.
3. (a) Clark, I. M.; Rowan, A. D.; Cawston, T. E. Expert
Opin. Anti-Inflammatory Immunomodulatory Invest. Drugs
2000, 2, 16; (b) Bottomley, K. M.; Johnson, W. H.;
Walter, D. S. J. Enzyme Inhib. 1998, 13, 79.
17. Bryde Nielsen, B.; Frydenvang, K.; Kjoller Larsen, I. Acta
Chrystallogr. C 1993, 49, 1018.
4. Leppert, D.; Lindberg, R. L. P.; Kappos, L.; Leib, S. L.
Brain Res. Rev. 2001, 36, 249.
5. (a) Leung, G.; Abbenante, G.; Fairlie, D. P. J. Med.
Chem. 2000, 43, 305; (b) Skiles, J. W.; Gonnella, N. C.;
Jeng, A. Y. Curr. Med. Chem. 2004, 11, 2911.
6. Babine, R. E.; Bender, S. L. Chem. Rev. 1997, 97, 1359.
7. Brandstetter, H.; Grams, F.; Glitz, D.; Lang, A.; Huber,
R.; Bode, W.; Krell, H.; Engh, R. A. J. Biol. Chem. 2001,
276, 17405.
8. (a) Puerta, D. T.; Lewis, J. A.; Cohen, S. M. J. Am. Chem.
Soc. 2004, 126, 8388; (b) Puerta, D. T.; Cohen, S. M. Curr.
Top. Med. Chem. 2004, 4, 1551.
18. Atomic coordinates can be retrieved from the Brookhaven
Protein Databank (http://rcsb.org/pdb). PDB ID code
1ZP5. For experimental conditions and statistics, see
Supplementary Data section.
19. Dhanaraj, V.; Williams, M. G.; Ye, Q. Z.; Molina, F.;
Johnson, L. L.; Ortwine, D. F.; Pavlovsky, A. G.; Rubin,
J. R.; Skeean, R. W.; White, A. D.; Humblet, C.; Hupe, D.
J.; Blundell, T. L. Croat. Chem. Acta 1999, 72, 575.
20. Pavlovsky, A. G.; Williams, M. G.; Ye, Q. Z.; Ortwine, D.
F.; Purchase 2nd, C. F.; White, A. D.; Dhanaraj, V.; Roth,
B. D.; Johnson, L. L.; Hupe, D. J.; Humblet, C.; Blundell,
T. L. Protein Sci. 1999, 8, 1455.
9. (a) Muri, E. M. F.; Nieto, M. J.; Sindelar, R. D.;
Williamson, J. S. Curr. Med. Chem. 2002, 9, 1631; (b)
Levin, J. I. Curr. Top. Med. Chem. 2004, 4, 1289.
10. Grams, F.; Crimmin, M.; Hinnes, L.; Huxley, P.; Pieper,
M.; Tschesche, H.; Bode, W. Biochemistry 1995, 34, 14012.
11. Sing, J.; Conzentini, P.; Cubdy, K.; Gainor, J.; Gilliam, C.
L.; Gordon, T. D.; Johnson, J. A.; Schneider, E. D.; Wahl,
L. C.; Whipple, D. A. Bioorg. Med. Chem. Lett. 1995, 5, 337.
12. Chung, S. J.; Kim, D. H. Bioorg. Med. Chem. 2001, 9, 185.
13. Scozzafava, A.; Supuran, C. Bioorg. Med. Chem. 2003, 11,
2241.
14. Michaelides, M. R.; Dellaria, J. F.; Gong, J.; Holms, J. H.;
Bouska, J. J.; Stacey, J.; Wada, C. K.; Heyman, H. R.;
Curtin, M. L.; Guo, Y.; Goodfellow, C. L.; Elmore, I. B.;
Albert, D. H.; Magoc, T. J.; Marcotte, P. A.; Morgan, D.
W.; Davidsen, S. K. Bioorg. Med. Chem. Lett. 2001, 11,
1553.
15. Stewart, A. O.; Bhatia, P. A.; Martin, J. G.; Summers, J.
B.; Rodriques, K. E.; Martin, M. B.; Holms, J. H.; Moore,
J. L.; Craig, R. A.; Kolasa, T.; Ratajczyk, J. D.;
Mazdiyasni, H.; Kerdesky, F. A. J.; DeNinno, S. L.;
Maki, R. G.; Bouska, J. B.; Young, P. R.; Lanni, C.; Bell,
R. L.; Carter, G. W.; Brooks, C. D. W. J. Med. Chem.
1997, 40, 1955.
21. (a) Browner, M. F.; Smith, W. W.; Castelhano, A. L.
Biochemistry 1995, 34, 6602; (b) Cheng, M.; De, B.;
Almstead, N. G.; Pikul, S.; Dowty, M. E.; Dietsch, C. R.;
Dunaway, C. M.; Gu, F.; Hsieh, L. C.; Janusz, M. J.;
Taiwo, Y. O.; Natchus, M. G.; Hudlicky, T.; Mandel, M.
J. Med. Chem. 1999, 42, 5426.
22. Gavuzzo, E.; Pochetti, G.; Mazza, F.; Gallina, C.; Gorini,
B.; DÕAlessio, S.; Pieper, M.; Tschesche, H.; Tucker, P. A.
J. Med. Chem. 2000, 43, 3377.
23. (a) Rush, T. S., III; Powers, R. Curr. Top. Med. Chem.
2004, 4, 1311; (b) Rao, B. G. Curr. Pharm. Des. 2005, 11,
295.
24. Remko, M.; von der Lieth, C. W. Struct. Chem. 2004, 15,
285.
25. Calculated using Advanced Chemistry Development
(ACD/Labs) Software Solaris V4.67 (Ó 1994–2005 ACD/
Labs) and available through SciFinder Scholar by the
American Chemical Society.
26. (a) McMartin, C.; Bohacek, R. S. J. Comput. Aided Mol.
Des. 1997, 11, 333; (b) Detailed settings for QXP docking
and minimizations are given in Supplementary Data.
27. Harmon, R. E.; Dabrowiak, J. C.; Brown, D. J.; Gupta, S.
K.; Herbert, M.; Chitharanjan, D. J. Med. Chem. 1970,
13, 577.