Rational design of high-affinity variants of interferon-gamma receptor 1

Pavel Mikulecký, Jiří Černý, Lada Biedermannová, Peter Šebo, Bohdan Schneider

Institute of Biotechnology AS CR, v.v.i., Vídeňská 1083, CZ-142 20 Prague, Czech Republic. pavel.mikulecky@img.cas.cz, bohdan@img.cas.cz

Protein-protein interactions play an essential role in many biological processes and our goal is to elucidate specificity of the interactions between human interferon gamma receptor 1 (hIFNgR1) and its natural ligand interferon gamma (hIFNg), an important protein of innate immunity. We searched for mutations within the receptor molecule to increase its affinity to hIFNg by computational analysis of existing crystal structures of hIFNg/hIFNgR1 (PDB IDs 1fg9 and 1fyh) and modeling using empirical force field implemented in the software FoldX. All selected hIFNgR1 variants were expressed in Escherichia coli and successfully purified to homogeneity. Kinetics of interactions between hIFNg and hIFNgR1-wild type as well as mutants was determined by surface plasmon resonance (SPR). The first set of mutations was designed on the interface of the hIFNg/hIFNgR1 complex and according to SPR measurements the affinity of most receptor variants had virtually the same affinity as wild-type receptor, a few had affinity slightly decreased, but a few variants bound hIFNg with significantly higher affinity. The second set of mutations of hIFNgR1 included residues that were not directly involved in binding of hIFNg and were supposed to fill up cavities inside the receptor molecule. SPR measurements showed that the affinities of these new single mutants were practically unchanged but enhanced in combination with the previously tested mutation at the interface. Our results indicate that rational design by relatively simple and accessible computational methods is capable of predicting hIFNgR1 variants with significantly increased affinity. These new high-affinity binders help in better understanding of forces governing protein-protein interactions and could be developed into a new diagnostic tool.

Support from grant P305/10/2184 from the Czech Science Foundation is greatly acknowledged. All authors are supported by the institutional grant AV0Z50520701.