Due to the conserved nature of protease active sites, it has been difficult to develop protease inhibitors that are both potent and specific. The vast majority of both small molecule and naturally-occuring protein protease inhibitors target the active site, and while potent, are often promiscuous. With the goal of developing novel classes of protease inhibitors that would be both potent and specific, a phage-displayed antibody library was panned against the cancer-associated protease membrane-type serine protease 1 (MT-SP1). Two inhibitors were discovered from the screen, and their mechanism of inhibition was further studied to understand the basis of their potency and specificity.
A number of kinetic and biochemical experiments revealed that while the two inhibitors, E2 and S4, bound to the enzyme with different binding mechanisms, they were both competitive inhibitors of both small-molecule and macromolecular MT-SP1 substrates. Mutational analysis by alanine scanning of the surface loops surrounding the protease active site revealed that the antibodies made numerous contacts with these loops. The surface loops are sites of natural diversity among closely related serine proteases, and created the epitope that allowed for antibody specificity. Taken together, these results suggest that these antibody inhibitors gain a measure of potency by binding in the protease active site, and additional potency and specificity through interactions with the protease surface loops. These strategies could be adapted in efforts to develop specific inhibitors against other proteases.
A structure of the Fab construct of E2 in complex with the protease domain of MT-SP1 was determined by x-ray crystallography. The structure revealed that the inhibitor has a novel mechanism of inhibition; it gains potency and specificity through interactions with the protease surface loops, and inhibits by binding a very large H3 loop in the active site in a catalytically non-competent manner.
The results here illustrate that antibodies can be effective inhibitors because they can bind to features that traditional protease inhibitors do not. These lessons can be used to guide the development of molecules that can specifically inhibit any protease.