Abstract 1915 Allosteric inhibition of serine proteases: a case study with the pro-oncogenic enzyme mesotrypsin

Mesotrypsin is an S1 family serine protease which promotes invasiveness of several human cancers. Other S1 family members play important roles in digestion, coagulation, and other critical biological processes. The active sites of these enzymes are highly similar, and selective targeting of specific serine proteases via conventional orthosteric small molecules is a technical challenge that is often not achieved. Allosteric sites tend to proffer greater variability than active sites to allow for selective targeting among similar proteins. However, until recently, there had been no reported allosteric site for the regulation of mesotrypsin activity. Our group has recently solved an x-ray crystal structure of unliganded mesotrypsin that revealed a loop of residues obscuring the active site, identified as an autoinhibitory loop. The structure also exhibited an allosteric binding pocket in this novel conformation which possessed two residues that differed among the three human trypsin isoforms. We hypothesized that the loop might be stabilized in the autoinhibitory conformation via a small molecule ligand targeting the exposed allosteric binding site. We further posited that the allosteric site might permit selective inhibition of mesotrypsin, due to the sequence differences lining the pocket. To explore this hypothesis, we subjected the mesotrypsin allosteric binding site and altered morphology of the active site to high-throughput virtual screening, resulting in collectively 58 hit compounds that met multiple cutoff criteria. The hits were then tested for inhibitory capability towards mesotrypsin using enzyme kinetics with a chromogenic substrate. Four of the compounds demonstrated mesotrypsin inhibition and were similarly tested against cationic trypsin and anionic trypsin. A single compound displayed approximately 5-fold selectivity for mesotrypsin over the other trypsins, to date the most selective small molecule targeting mesotrypsin. Additionally, we evaluated the compound selectivity against various coagulation pathway serine proteases that represent clinically relevant off-targets. The results demonstrated a strong preference for mesotrypsin. Our approach for selective allosteric inhibition of mesotrypsin may represent a paradigm shifting mode for targeting other S1 family serine proteases. Examination of publicly available structures of serine proteases indicates conformational heterogeneity the 220 loop, analogous to the mesotrypsin autoinhibitory loop. There appears to be a single well-defined uninhibited conformation that is poised for substrate access and catalytic activity. However, there are multiple autoinhibited conformations with differing levels of occlusion of the active site observed in different enzymes. Structural profiling of these autoinhibited conformations reveals prospective allosteric binding pockets with differential morphologies, suggesting potential for selective inhibition. Taken together, our findings support our hypothesis that selective inhibition of mesotrypsin, and perhaps other serine proteases, may be achieved by exploiting a cryptic allosteric binding site, locking the enzymes in an inactive conformation. The authors would like to acknowledge funding support under NIH R01GM144393, as well as synchrotron access from Lawrence Berkeley National Laboratory under ALS-11671.