Generating N-TINP2 with multiple mutations for enhanced affinity to MMP9

Protein-protein interactions (PPIs) play vital roles in diverse biological processes. Hence, understanding of PPIs is important for development of strong interactions between biological targets and protein-based therapeutics.
So far, studies were mainly focused on single mutations of a given protein to enhance target affinity. We have recently developed an alternative approach, using affinity screens of protein library followed by deep sequencing and machine learning analysis to predict the effect multiple mutations confer on target affinity. To test our approach, we identified several variants that were previously predicted from performing yeast surface display (YSD) affinity screens of an multiple mutation library of N terminal tissue inhibitor of metalloproteinases 2 (N-TIMP2) against its cognate receptor Matrix MetaloProteainase-9 (MMP9). We then generated the variants in two different formats, in YSD and as purified soluble proteins, and tested their affinities to MMP9. Our results show that our approach for predicting affinities could accurately predict affinities determined from the YSD analysis, the same method used for the library screening process. Our method for predicting affinity could then be used to predicts PPIs in other protein complexes.