The signature of the five-stranded vRRM fold defined by functional, structural and computational analysis of the hnRNP L protein

The RNA recognition motif (RRM) is the far most abundant RNA binding domain. In addition to the typical β1α1β2β3α2β4 fold, various sub-structural elements have been described and reportedly contribute to the high functional versatility of RRMs. The heterogeneous nuclear ribonucleoprotein L (hnRNP L) is a highly abundant protein of 64kDa comprising four RRM domains. Involved in many aspects of RNA metabolism, hnRNP L specifically binds to RNAs containing CA-repeats or CA-rich clusters. However, a comprehensive structural description of hnRNP L including its sub-structural elements is missing. Here, we present the structural characterization of the RRM domains of hnRNP L and demonstrate their function in repressing exon 4 of SLC2A2. By comparison of the sub-structural elements between the two highly similar paralog families of hnRNP L and PTB, we defined signatures underlying interacting C-terminal coils (ICC), the RRM34 domain interaction and RRMs with a C-terminal 5th β-strand, a variation we denoted vRRMs. Furthermore, computational analysis revealed new putative ICC-containing RRM families and allowed to propose an evolutionary scenario explaining the origins of the ICC and 5th β-strand sub-structural extensions. Our studies provide insights of domain requirements in hnRNP L mediated alternative splicing and molecular descriptions for the sub-structural elements. In addition, the analysis presented may help to classify other abundant RRM extensions and to predict structure-function relationships.