Probing the effect of all possible mutations on protein binding affinity without the need of protein purifications

Quantifying the effects of various mutations on binding affinity is crucial for understanding the evolution and mechanism of protein-protein interactions (PPIs). In our study, the binding landscape of N-TIMP2/MMP-9 complex is investigated by a novel approach that combines a FACS sort separated by different affinities, next generation sequencing (NGS) of the sorted populations and an extensive analysis of the resulting NGS data. Herein, we are the first to demonstrate that NGS-based data can be used to obtain quantitative data for binding affinity of all variants with single mutations in the binding interface. An important advance in our research is the use of K_i values of several purified mutants to scale the NGS data for the N-TIMP2/MMP-9 complex. In fact, in all previous experiments, the linear correlation between the NGS results and the K_i values was only reported for a small range of K_i values (with ‘shallow’ affinity landscapes), and for complexes where protein concentration used is different than the PPI K_i. In our work, these biases are removed by using two mixed libraries, with high and low binding affinity toward MMP-9, generating one library with a wide affinity range to the target. The high affinity library is based on high affinity N-TIMP2 variants to MMP-9. The low affinity library is based on a low affinity N-TIMP2 variant, namely Ala-N-TIMP2, resulting in N-TIMP2/MMP-9 complex having its K_i value similar to the MMP concentration used in FACS sorting. Therefore, we predict that the resulting binding affinity landscape would correlate well with findings from previous studies and independent approaches and for complexes with ‘steep’ affinity landscapes, where single mutations may result in a significant change in binding affinity.