Homology modeling of gelatinase catalytic domains and docking simulations of novel sulfonamide inhibitors.

Three-dimensional models for the catalytic domain of gelatinases (MMP-9 and -2) have been constructed based on the X-ray crystal structure of MMP-3. Conformations of the loop segment which forms the bottom half of the S1' subsite but shows conformational diversity among the crystal structures of other MMPs have been explored by simulated annealing of each gelatinase model complexed with two highly potent "probe" inhibitors. Representative catalytic domain models have been selected for each gelatinase from the set of generated conformations based on shape complementarity of the loop to the probe inhibitors. The single model selected for MMP-9 was utilized to explain the structure-activity relationship of our novel sulfonamide inhibitors. Molecular dynamics (MD) simulations of the complex models revealed important features of the binding mechanism of our inhibitors: (i) the ligand carboxylate group coordinating to the catalytic zinc ion and hydrogen bonding to the Glu219 side chain, (ii) one of the sulfonyl oxygens forming hydrogen bonds with the main chain NHs (Leu181 and Ala182), (iii) the sulfonyl substituent making extensive hydrophobic contact with the S1' subsite. The gauche conformation exclusively adopted by the sulfonamide C-N-S-C torsion plays an important role in achieving the third binding feature by properly directing the substituent into the S1' subsite. Improvement of the inhibitory activity according to straight elongation of the sulfonyl substituent was attributed to an increase of the hydrophobic contact between the substituent and the S1' subsite. Structural modifications which alter the straight shape of the substituent lead to deterioration of the activity. On the other hand, the two candidate models selected for MMP-2 differ in the bottom shape of the S1' subsite: one with a channel-like subsite and the other with a pocket-like subsite resembling that of the MMP-9 model. The bottom shape was experimentally probed by chemical synthesis of inhibitors having elongated sulfonyl substituents whose terminal alkyl groups were shown by MD simulations to protrude from the S1' subsite bottom into the solvent. Gelatinase assays of these inhibitors showed that elongation of the substituent significantly reduces activity against MMP-9 while retaining activity against MMP-2, consequently increasing the selectivity between MMP-2 and -9. The results confirm that MMP-9 has a pocket-like S1' subsite with a floorboard and MMP-2 has a channel-like S1' subsite.

Multiple binding modes for the receptor-bound conformations of cyclic AII agonists.

Angiotensin II, Asp-Arg-Val-Tyr-His-Pro-Phe, binds its receptor with a postulated turn centered at residue four. Analogs of angiotensin II which contain a disulfide bridge between the side chains of residues 3 and 5 retain significant activity consistent with this hypothesis. Incorporation of 4-mercaptoproline residues, a hybrid, or chimeric amino acid which combines the properties of proline and homocysteine, into either of these positions with analogous disulfide bridges allows retention of high affinity for the receptor. These more highly constrained bicyclic systems give new insight into the details of molecular recognition of residues 3-5 of angiotensin by the receptor. Retention of activity by the antiparallel dimer of [Sar1,Cys3,5]-AII in which the peptide backbone is held in an extended conformation was unexpected. Analysis of the conformational constraints imposed in these active analogs suggests that AII agonists bind to their receptor with different backbone conformations in the region of the central tyrosine residue.

The utility of side-chain cyclization in determining the receptor-bound conformation of peptides: cyclic tripeptides and angiotensin II.

The effect of side-chain cyclization on accessible backbone conformations of tripeptides, X-Ala-Y (X and/or Y = Cys, Hcy (Hcy: homocysteine), cis 4-mercaptoproline (MPc), and trans 4-mercaptoproline (MPt)), was elucidated using two variants of systematic conformational search. In addition to cyclization through a disulfide bond, the thioether (-S-CH2-) and amide (-CO-NH-) side-chain analogues of Cys-Ala-Cys and Hcy-Ala-Hcy were evaluated. The number of valid backbone conformations and the allowed phi, psi space were evaluated for each compound, and the ability of the cyclic tripeptides to accommodate beta-turn conformations was examined in order to assess the value of cyclization in limiting conformational freedom. Based on the number of conformations, cyclization was highly effective in reducing the backbone degree of freedom: in order of decreasing number of conformations, Ala-Ala-Ala 1 >> Hcy-Ala-Hcy 2 >> Cys-Ala-Hcy 3 approximately equal to Hcy-Ala-Cys 4 >> MPc-Ala-Hcy 5, 7 > Cys-Ala-Cys 6 > MPc-Ala-Cys 8 > Hcy-Ala-MPt 9 > Cys-Ala-MPt 10 approximately equal to MPc-Ala-MPt 11. Although Hcy-Ala-Hcy 2 had the greatest number of conformations of the cyclic peptides studied, it was still greatly constrained relative to its linear analogue 1. The bicyclic ring system introduced by MP was even more effective in constraining the cycle, having greater impact at position 3 than at position 1. Under the conditions of the study, cyclization of MP-containing analogues could be effected only with the cis isomer (MPc) at position 1 and/or the trans isomer (MPt) at position 3. Sterically allowed conformations of Ala2 for the cyclic tripeptides 2-4 were generally similar to those of the linear tripeptide 1, while those of Cys-Ala-Cys 6 and MPc-Ala-Hcy 7 were restricted to a smaller region of phi 2, psi 2 space: the right- and left-handed alpha-helical conformation and the beta-conformation. This trend was even more pronounced for Hcy-Ala-MPt 9, Cys-Ala-MPt 10, and MPc-Ala-MPt 11, in which Ala2 was severely restricted to a very small region of phi, psi space: the left-handed alpha-helical conformation for 9-11, plus the beta conformation for 9. This suggests that MP at the 3-position is incompatible with a right-handed alpha-helical conformation at position 2.(ABSTRACT TRUNCATED AT 400 WORDS)