Determinants of RNA recognition by the FinO domain of theEscherichia coliProQ protein

Abstract

The regulation of gene expression by small RNAs in Escherichia coli depends on RNA binding proteins Hfq and ProQ, which bind mostly distinct RNA pools. To understand how ProQ discriminates between RNA substrates, we compared its binding to six different RNA molecules. Full-length ProQ bound all six RNAs similarly, while the isolated N-terminal FinO domain (NTD) of ProQ specifically recognized RNAs with Rho-independent terminators. Analysis of malM 3ʹ-UTR mutants showed that tight RNA binding by the ProQ NTD required a terminator hairpin of at least two base pairs preceding an 3ʹ oligoU tail of at least four uridine residues. Substitution of an A-rich sequence on the 5ʹ side of the terminator to uridines strengthened the binding of several ProQ-specific RNAs to the Hfq protein, but not to the ProQ NTD. Substitution of the motif in the malM-3ʹ and cspE-3ʹ RNAs also conferred the ability to bind Hfq in E. coli cells, as measured using a three-hybrid assay. In summary, these data suggest that the ProQ NTD specifically recognizes 3ʹ intrinsic terminators of RNA substrates, and that the discrimination between RNA ligands by E. coli ProQ and Hfq depends both on positive determinants for binding to ProQ and negative determinants against binding to Hfq.RNA-binding proteins are important contributors to major life processes, including the regulation of gene expression by RNAs (1). In many bacteria, the prominent role played by trans-encoded base-pairing small RNAs (sRNAs) in regulating gene expression requires a matchmaker protein called Hfq, which assists sRNA in pairing to complementary sequences in target mRNAs and affects sRNA stability (2-6). Global identification of new RNA/protein interactions has been enabled by deep-sequencing approaches (7), such as Grad-seq, which uses glycerol gradients to identify novel RNA/protein complexes (8), CLIP-seq, which uses crosslinking to define protein binding sites in the transcriptome (9), and RIL-seq which identifies protein-dependent RNA-RNA interactions (10). Recent studies using these approaches showed that another protein, named ProQ, is a global RNA binding protein in Escherichia coli and Salmonella enterica (11-13), and is involved in sRNA interactions with other RNA molecules (12,14,15). The pool of RNA ligands of ProQ is mostly distinct from that of Hfq and contains more mRNAs than sRNAs (11,12).While ProQ was originally discovered in a search for genes that affect proline transport (16), this protein contributes to several physiological processes in E.coli and S. enterica (17)including DNA metabolism (13,14), bacterial virulence (15), and adaptation to osmotic stress (12) and resource limitation (18). ProQ is active as a monomer (19), and it belongs to the FinO-domain family with representatives in numerous γ-proteobacteria (17,20). Other members of this family include the F-like plasmid FinO protein (21,22), Legionella pneumophila RocC protein (23), and S. enterica pCol1B9 plasmid-encoded FopA protein (J.Vogel, personal communication), which each interact with few RNAs...