Oligomerisation status and evolutionary conservation of interfaces of protein structural domain superfamilies

Abstract

Protein–protein interactions are important in carrying out many biological processes and functions. These interactions may be either permanent or of temporary nature. Several studies have employed tools like solvent accessibility and graph theory to identify these interactions, but still more studies need to be performed to quantify and validate them. Although we now have many databases available with predicted and experimental results on protein–protein interactions, we still do not have many databases which focus on providing structural details of the interacting complexes, their oligomerisation state and homologues. In this work, protein–protein interactions have been thoroughly investigated within the structural regime and quantified for their strength using calculated pseudoenergies. The PPCheck server, an in-house webserver, has been used for calculating the pseudoenergies like van der Waals, hydrogen bonds and electrostatic energy based on distances between atoms of amino acids from two interacting proteins. PPCheck can be visited at http://caps.ncbs.res.in/ppcheck/. Based on statistical data, as obtained by studying established protein–protein interacting complexes from earlier studies, we came to a conclusion that an average protein–protein interface consisted of about 51 to 150 amino acid residues and the generalized energy per residue ranged from −2 kJ mol⁻¹ to −6 kJ mol⁻¹. We found that some of the proteins have an exceptionally higher number of amino acids at the interface and it was purely because of their elaborate interface or extended topology i.e. some of their secondary structure regions or loops were either inter-mixing or running parallel to one another or they were taking part in domain swapping. Residue networks were prepared for all the amino acids of the interacting proteins involved in different types of interactions (like van der Waals, hydrogen-bonding, electrostatic or intramolecular interactions) and were analysed between the query domain-interacting partner pair and its remote homologue-interacting partner pair. We found that, in exceptional cases, homologous proteins belonging to the same superfamily, but with remote sequence similarity, can share similar interfaces.