An internal loop region is responsible for inherent target specificity of bacterial cold-shock proteins

Cold-shock proteins (Csps), of around 70 amino acids, share a protein fold for the cold-shock domain (CSD) that contains RNA-binding motifs, RNP1 and RNP2, and constitute one family of bacterial RNA-binding proteins. Despite similar amino acid composition, Csps have been shown to individually possess inherent specific functions. Here, we identify the molecular differences in Csps that allow selective recognition of RNA targets. Using chimeras and mutants of Escherichia coli CspD and CspA, we demonstrate that Lys43-Ala44 in an internal loop of CspD, and the N-terminal portion with Lys4 of CspA, are important for determining their target specificities. Pull-down assays suggest that these distinct specificities reflect differences in the ability to act on the target RNAs rather than differences in binding to the RNA targets. A phylogenetic tree constructed from 1,573 Csps reveals that the Csps containing Lys-Ala in the loop form a monophyletic clade, and the members in this clade are shown to have target specificities similar to E. coli CspD. The phylogenetic tree also finds a small cluster of Csps containing Lys-Glu in the loop, and these exhibit a different specificity than E. coli CspD. Examination of this difference suggests a role of the loop of CspD-type proteins in recognition of specific targets. Additionally, each identified type of Csp shows a different distribution pattern among bacteria. Our findings provide a basis for subclassification of Csps based on target RNA specificity, which will be useful for understanding the functional specialization of Csps.