Investigation on interactions of non-canonical RNAs with proteins

Non-canonical RNAs (nclRNAs), characterized by unique structural and functional features, are increasingly recognized in RNA-protein interaction networks. Here we have analysed the interfacial properties of non-canonical RNA-protein complexes (ncRPCs), using a non-redundant dataset of 26 three-dimensional structures, and classified them based on the different RNA types. The analysis is also carried out on canonical RNA-protein complexes (cRPCs), and the findings are compared. We find that ncRPCs exhibit wider variation and greater average interface area, with pseudoknots forming the largest interfaces. Interface composition highlights that guanine and uracil are dominant in ncRPCs, supporting flexible non-Watson-Crick interactions, while cytosine and guanine are more prevalent in cRPCs. Lysine, tyrosine, glutamine and proline are more abundant in ncRPCs compared to cRPCs, enabling electrostatic and π-stacking interactions with irregular RNA geometries. In contrast, arginine, phenylalanine, tryptophan, and cytosine required for stabilizing structured RNA folds, are more prevalent in cRPCs than in ncRPCs. The study of interfacial interactions shows that ncRPCs form significantly more π-interactions and direct hydrogen bonds than cRPCs. On the contrary, frequency of salt bridges in ncRPCs is less as compared to cRPCs. Structural parameters such as normalized Shannon entropy and normalized base-pair distance confirm greater conformational flexibility in ncRPCs. Significant differences are observed in the interfacial physicochemical properties of ncRPCs and cRPCs, emphasizing how nclRNAs differ from typical canonical RNAs while interacting with proteins. The findings from this study may serve as a basis for new nclRNA-targeted therapeutic approaches and for designing computational tools for predicting and modelling nclRNA-protein interactions.