Background: Matrix metalloproteinases (MMPs) contribute to various physiological and
pathophysiological processes. An imbalance in MMP activity causes pathological conditions including
inflammatory diseases, cancer, and cardiovascular diseases. Each MMP member has many 3D structures
available; therefore, selecting one structure for virtual screening becomes challenging.
Methods: In this study, we used the cross-docking approach to rank the available MMP structures for their probable successful
performance in virtual screening. To determine structures that would offer best average RMSD (root mean square
deviation), we performed cross-docking studies on 123 holo (protein–ligand) structures of seven MMP enzyme groups
(MMP-1, MMP-3, MMP-7, MMP-8, MMP-9, MMP-12, and MMP-13).
Results: MMP enzymes with more flexible residues had fewer structures with RMSD < 2.0 A. Further, same resolution
and binding affinities, difference in ligand size, and chemotype of the co-crystalized ligand were parameters that could
greatly affect the corresponding cross-dock results and the calculated average RMSD for the structures. Four of the six
best MMP-12 receptors, which were identified using the average RMSD metric, had the highest EF1% (emrichment factor)
in the retrospective enrichment study.
Conclusion: According to the enrichment results, structures with lower average RMSD have a high probability of being
appropriate candidates. These study findings will help in receptor selection for an MMP virtual screening protocol and
lead to better enrichment of MMP inhibitors.