Structural network comparison of domains across remote homologues arising from different domain architectures

The homology of protein domains indicates an evolutionary relationship between them. Remote homologues, on the other hand, share very low sequence similarity while maintaining a highly similar structural topology. These similar structures can be compared using several metrices, but a comparison of networks built from amino acid residues to mimic non-covalent interactions provides a better understanding of their similarities or minute differences. Comparison of structurally related domains are studied to some extent, while the impact of domain combinations is poorly studied. In this study, several types of pairs of domains, arising from different multi-domain proteins were compared to understand the effect of cognate domain on the structural architecture of remote homologues. Several network parameters were compared among these domains. We found that an interacting domain of unrelated structure exerts higher structural and evolutionary pressure on the remote homologues, leading to structural embellishments and functional modifications. This exerted change is mediated by local rearrangements of the amino acid residues in the structure. Case studies were conducted to understand these effects. These results will assist in re-defining the homology among domains, which would rather help in better functional annotations and structural modelling.