By Tom Smith, Eneko Villanueva, Rayner Queiroz, Gavin Thomas & Kathryn Lilley

“There is no such thing as a new idea. It is impossible. We simply take a lot of old ideas and put them into a sort of mental kaleidoscope. We give them a turn and they make new and curious combinations.”
(Mark Twain)

If you work on RNA, it’s likely that you will have used acid guanidinium thiocyanate-phenol-chloroform (AGPC), commonly referred to as the “TRIzol protocol” after the branded reagent, to extract RNA from your samples. Under acidic conditions, denatured protein are soluble in organic solvents, whereas RNA is soluble in aqueous solutions. Mixing cell lysates with AGPC, followed by centrifugation therefore yields two phases: an upper aqueous phase containing RNA and a lower organic phase containing protein. Protocols for this method usually caution against taking any material from the interface between the phases on the basis that it may contain DNA, lipids or carbohydrates: “here be dragons”. We had the idea to flip the method on its head and instead extract protein:RNA complexes from the AGPC interface, on the basis that the two components will “pull” towards the opposite phases, leaving the complex at the interface. To capture RNA-bound protein, we first had to fix their interactions in place as non-covalent interactions are lost when the protein is denatured.  We chose to do this with 254 nm UV, which is commonly used to crosslink RNA and protein at zero distance, without the need to incorporate modified nucleotides such as 4-thiouridine.

The first conundrum we wanted to address was how to extract a clean interface without any free RNA or protein. Following the KISS principle (Keep it short and simple), we repeated the AGPC phase separation multiple times and observed that 3 rounds are optimal to enrich the RNA:protein complexes. Isolation RNA or protein from the cleaned interface is also remarkably simple: Using RNase digestion, one can degrade the RNA component so that in a final round of AGPC phase separation, the protein will return to the organic phase. Similarly, digestion of the protein with Proteinase K returns the full length RNA to the aqueous phase. We have called our method Orthogonal Organic Phase Separation (OOPS) to reflect the capacity to extract both protein and RNA from the enriched RNA:protein complexes [1]. The simplicity of OOPS means it is highly reproducible and yields far more material than current methods to enrich RNA-bound protein.

To our surprise, it turns out we were not the first to use UV crosslinking to enrich RNA:product complexes at an aqueous:organic phase interface. After we developed our method, we discovered that the extraction of RNA:protein complexes from a phenol interface was developed in the 1970s and Wagenmaker et al used this method to study the efficiency of RNA-protein UV crosslinking in 1980 [2]. Whilst the use of UV to crosslink RNA and protein has become commonplace, the potential to extract the resultant RNA-protein complexes from the interfaces appears to have been missed. An even bigger surprise came when we presented our data and discovered there were 2 other groups simultaneously pursuing the same idea, although there are important differences in the final protocols designed by each group [3-4]. Clearly this was an approach that was ripe for re-invention and adaptation in the age of high-throughput quantification techniques. After communicating with the other two groups, and to ensure that each independent conception was recognized, we coordinated submission of our manuscript pre-prints to BioRxiv on the same day and also coordinated the publicity of the 3 approaches using social media.

One obvious application of OOPS is to characterise proteins interacting with RNA (RNA binding proteins, RBPs). In this respect, OOPS has two major advantages over the major method used to identify RNA binding proteins, RNA interactome capture (RIC) [5]. Firstly, much less material is required due to the aforementioned higher yield of OOPS. Secondly, RIC is limited to the study of proteins that bind polyadenylated-RNA since it uses oligo(dT) to purify RNA:protein adducts. It is therefore not applicable for the study of proteins which bind non-coding RNAs, including lincRNAs, and those species with short and/or infrequent mRNA polyadenylation, including many microorganisms. Since OOPS can capture all RNAs, we used it to obtain the first RBPome of the model bacterium, Escherichia coli. The 364 proteins captured cover all aspects of RNA biology in E. coli, including regulation of transcription termination, RNA degradation, and RNA stability through polyA-polymerase action. As expected, we identified many of the proteins which function in translation through their interaction with rRNAs, tRNAs, mRNAs, sRNAs and other non-coding RNAs. These comprised proteins involved in the maturation of the ribosome, including those that post-transcriptionally modify rRNA, ribosomal subunit proteins that interact with rRNA in the mature ribosome, proteins which post-transcriptional modify tRNA, aminoacyl-tRNA synthetases which activate tRNA, and translation factors. We also captured Ffh and FtsY which complex with 4.5 S RNA to form the recognition particle (SRP) and further proteins that interact with small regulatory RNAs (sRNAs) including Hfq and ProQ.

In addition to these known interactors, we discovered 234 potentially novel RBPs, suggesting OOPS could yield new findings when applied to species which have been neglected by previous methods. For example we identified glycolytic enzymes previously shown to interact with RNA in mammals, and proteins with intriguing subcellular distributions such as MreB, MinD and SecA, which could conceivably coordinate RNA localisation, a process that has been proposed recently to occur in microbes, but which has not been widely studied. Of course, confirming the function of this proteins when interacting with RNA will require further experimentation, but we are confident that we have uncovered many new protein-RNA interactions. OOPS could be immediately applied to other model systems and important human pathogens to further our understanding of pathogen cellular biology and host-pathogen interactions. We are currently using OOPS as a start point for multiple downstream applications which we hope will aid comprehensive and dynamic studies of RNA-protein interactions in all species.

1. Queiroz, R.M.L., Smith T., Villanueva, E. et al., 2019. Comprehensive identification of RNA–protein interactions in any organism using orthogonal organic phase separation (OOPS). Nature Biotechnology. Corrected proof available on line.

2. Wagenmakers, A.J.M., Reinders, R.J. & Van Venrooij, W.J., 1980. Cross‐linking of mRNA to Proteins by Irradiation of Intact Cells with Ultraviolet Light. European Journal of Biochemistry.112(2), pp.323–330.

3. Trendel, J. et al., 2018. The Human RNA-Binding Proteome and Its Dynamics during Translational Arrest. Cell. In press. Corrected proof available on line

4. Urdaneta, E.C. et al., 2018. Purification of Cross-linked RNA-Protein Complexes by Phenol-Toluol Extraction. bioRxiv, p.333385.

5. Castello, A. et al., 2012. Insights into RNA Biology from an Atlas of Mammalian mRNA-Binding Proteins. Cell, 149(6), pp.1393–1406.

This Behind the Paper was first posted on the Nature Research Microbiology Community.