Abstract
Early-stage evolutionary development of the universal genetic code remains a fundamental, open problem. One of the possible scenarios suggests that the code evolved in response to direct interactions between peptides and RNA oligonucleotides in the primordial environment. Recently, we have revealed a strong matching between base-binding preferences of modern protein sequences and the composition of their cognate mRNA coding sequences. These results point directly at the physicochemical foundation behind the code's origin, but also support the possibility of direct complementary interactions between proteins and their cognate mRNAs, especially if the two are unstructured. Here, we analyze molecular-surface mapping of knowledge-based amino-acid/nucleobase interaction preferences for a set of complete, high-resolution protein structures and show that the connection between the two biopolymers could remain relevant even for structured, folded proteins. Specifically, protein surface loops are strongly enriched in residues with a high binding propensity for guanine and cytosine, while adenine- and uracil-preferring residues are uniformly distributed throughout protein structures. Moreover, compositional complementarity of cognate protein and mRNA sequences remains strong even after weighting protein sequence profiles by residue solvent exposure. Our results support the possibility that protein/mRNA sequence complementarity may also translate to cognate interactions between structured biopolymers.