Thermodynamic modeling of RsmA - mRNA interactions capture novel direct binding across the Pseudomonas aeruginosa transcriptome

Pseudomonas aeruginosa (PA) is a ubiquitous, Gram-negative, bacteria that can attribute its survivability to numerous sensing and signaling pathways; conferring fitness due to speed of response. Post-transcriptional regulation is an energy efficient approach to quickly shift gene expression in response to the environment. The conserved post-transcriptional regulator RsmA is involved in regulating translation of genes involved in pathways that contribute to virulence, metabolism, and antibiotic resistance. Prior high-throughput approaches to map the full regulatory landscape of RsmA have estimated a target pool of approximately 500 genes; however, these approaches have been limited to a narrow range of growth phase, strain, and media conditions. Computational modeling presents a condition-independent approach to generating predictions for binding between the RsmA protein and highest affinity mRNAs. In this study, we improve upon a two-state thermodynamic model to predict the likelihood of RsmA binding to the 5′ UTR sequence of genes present in the PA genome. Our modeling approach predicts 1043 direct RsmA-mRNA binding interactions, including 457 novel mRNA targets. We then perform GO term enrichment tests on our predictions that reveal significant enrichment for DNA binding transcriptional regulators. In addition, quorum sensing, biofilm formation, and two-component signaling pathways were represented in KEGG enrichment analysis. We confirm binding predictions using in vitro binding assays, and regulatory effects using in vivo translational reporters. These reveal RsmA binding and regulation of a broader number of genes not previously reported. An important new observation of this work is the direct regulation of several novel mRNA targets encoding for factors involved in Quorum Sensing and the Type IV Secretion system, such as rsaL and mvaT. Our study demonstrates the utility of thermodynamic modeling for predicting interactions independent of complex and environmentally-sensitive systems, specifically for profiling the post-transcriptional regulator RsmA. Our experimental validation of RsmA binding to novel targets both supports our model and expands upon the pool of characterized target genes in PA. Overall, our findings demonstrate that a modeling approach can differentiate direct from indirect binding interactions and predict specific sites of binding for this global regulatory protein, thus broadening our understanding of the role of RsmA regulation in this relevant pathogen.