Tracking the targets of non-coding regulatory small RNAs made easier
A simple method for defining interactions between the small RNA regulators and various transcripts opens the way for dissecting post-transcriptional regulatory networks
It did not take long for prokaryotic molecular biologists to recognize that the amounts of individual proteins made in a cell is not only determined by the levels of the corresponding mRNAs but also by the efficiency of their translation and stability of the transcript. Both of these parameters are influenced by a class of regulatory factors, the non-coding small RNAs (sRNAs), that exert their regulatory functions by base pairing with target mRNAs. In principle, this mode of regulation is analogous to the regulation of gene expression in higher organisms mediated by microRNAs.Since sRNAs can base pair with multiple mRNAs, often involving short, often interrupted regions of complementarity, direct targets of their regulation have been difficult to identify by computational methods. This hindered our appreciation of the complex regulatory networks, controlling the responses of bacteria to external stimuli, ranging from adaptation to new environments and stress response, to innate defenses during host infection. Several laboratories, including ours, have devised experimental approaches to study direct interactions between sRNAs and their targets. The sRNA target identification method developed by our group was published last week in Nature Microbiology (ID 10.1038/nmicrobiol.2016.239) and is based on a simple concept that RNA ligase from bacteriophage T4, will preferentially ligate sRNAs base-paired with mRNAs at their respective 3’ and 5’ ends. These sRNA/mRNA chimaeras can be identified following the enrichment for the sRNAs using complementary oligonucleotides and sequencing of the cDNAs by next generation sequencing platforms (RNA-seq). The method is referred to as GRIL-seq, an acronym for Global sRNA Target Identification by Ligation and Sequencing.The outline of the procedure is shown in the illustration below.
This simple technology took over a year to optimize, in part because of our concern about the presence of triphosphates at 5’ ends of primary transcripts since these cannot be ligated to 3’ hydroxyl ends. Numerous attempts to isolate sRNA/mRNA complexes from lysed cells, remove the two terminal phosphates with tobacco acid phosphatase followed by ligation with T4 RNA ligase, proved to be a fruitless, labor-intensive (and expensive) undertaking. However, our concerns about the adverse effects of 5’ triphosphates on the ligation reaction was misplaced, since the degradation of mRNAs (and sRNAs) by cellular RNAses, as part of normal transcript turnover, is preceded in bacteria by the removal of the pyrophosphate from the 5´ end of triphosphorylated RNA, leaving a 5´ monophosphate RNA, by the enzyme RNA 5´ pyrophosphohydrolase (RppH). This made it possible to move from in vitro to in vivo ligation, by simply expressing the T4 RNA ligase gene (t4rnl1) in bacteria for a short time and identifying chimaeras by RNA-seq.Indeed this was confirmed by monitoring the formation of ligation products between selected sRNAs controlling known mRNA targets, and demonstrating that chimaeras can be only detected in wild-type cells expressing the RppH protein and not in an rppH mutant.
We applied GRIL-seq to the analysis of global iron regulation, via iron-responsive sRNAs PrrF1 and RyhB in Pseudomonas aeruginosa and in Escherichia coli, respectively. The method should be widely applicable to any microorganisms capable of expressing a recombinant RNA ligase and their genome contains a gene encoding an enzyme with the RppH-like pyrophosphatase activity. In some organisms, this may require use of specific expression constructs and codon optimization for RNA ligase and conceivably, ectopic expression of RppH. As we demonstrate in this paper, the ability to detect direct physical interactions between sRNAs and their target transcripts can lead to new discoveries about the mechanism of global gene regulation. Our work specifically highlights the role of the so-called “RNA sponges”, where an sRNA is sequestered from its target by base paring with another decoy transcript. In our Nature Microbiology paper (in Supplementary Note 3), we used GRIL-seq to identify a novel sRNA sponge, derived from processing of the catalase (katA) mRNA, giving a stable 3’ fragment. The fragment is capable of base pairing with PrrF1, thus inactivating its negative regulatory effect on targeted mRNAs. Therefore GRIL-seq may serve as a easy-to-use tool not only for accurate cataloguing of mRNAs directly controlled by individual sRNAs (the sRNA regulon) but conceivably, play an important role in uncovering novel features of sRNA-mediated post-transcriptional regulation.