Abstract:
A large portion of the human proteome (~2000 proteins) binds to RNA. RNA-protein interactions influence the functions of all RNA species, e.g. messenger RNA (mRNA), transfer RNA (tRNA), small nuclear RNA (snRNA), and long non-coding RNA (lncRNA) and in all cellular compartments, e.g. the nucleus, the cytoplasm, and in organelles like mitochondria and endosomes. An RNA’s fate is dictated by these interactions throughout its entire life span from biogenesis and processing over transport and subcellular localization to the modulation of the biological activity and ultimately its degradation. During these processes, accumulation effects of RNAs and proteins can occur that lead to the formation of aggregates. These fulfill important biological functions like gene expression and splicing regulation, but can also be a source of cellular toxicity by aberrant protein recruitment e.g. in triplet repeat expansion disorders like myotonic dystrophy. Moreover, pathogenic RNA identification by pattern recognition receptors is a key element of the innate immunity. Together, this highlights the importance of RNA-protein interactions on a molecular level as well as for the whole organism. Hence it is crucial to understand RNA-protein interactions in a scenario as lifelike as possible.
Currently, interactions between RNAs and proteins and the protein composition of their aggregates are typically identified by enrichment assays or proximity biotinylation and subsequent mass spectrometry (MS). These assays are often limited to highly abundant RNAs, require large input amounts and/or genetic manipulation of the target cells. Applying proximity biotinylation in situ – as described here by the FISH ID protocol – removes these limitations. On top of that, it enables to identify an RNA’s interactome under different cellular states such as the absence or presence of oxidative stress and with sub-compartment specificity. Here, a recombinant biotin ligase was recruited to the RNA-of-interest (ROI) by target-specific chemically modified antisense probes that are bound covalently by the enzyme’s tag.
Formation of the previously mentioned intracellular aggregates has also been observed after administration of chemically modified (RNA-like) therapeutic antisense oligonucleotides (ASOs). Interactions between proteins and ASOs are on the one hand responsible for their function, e.g. for cellular uptake and recruitment of effector proteins. On the other hand, the toxicity of an ASO drug is often connected to its protein interactions. Most of these interactions were identified by pulldown assays coupled to MS as well. The obtained interactome is independent from the subcellular localization of proteins and oligonucleotides. Hence it cannot reflect the intracellular interactions and does not allow any connection to other drug properties like efficacy and toxicity. This was overcome by the development of in situ ASO identification (isASO-ID): A purified biotin ligase construct was redirected to the ASO-of-interest by a coupling moiety. Proximal proteins were identified by MS after biotinylation and streptavidin enrichment. This allowed deciphering the interactomes with respect to chemical modifications, ASO concentration, the uptake pathway and under additional stress induction. Ultimately, the assay was applied to uncover the molecular basis of toxicity for an emerging class of ASO drugs: ADAR-recruiting ASOs. They impaired regular RNA processing upon their interaction with RNA processing enzymes. A better understanding of this will pave the way towards the engineering of safer ADAR-recruiting ASOs.
Taken together, FISH-ID and isASO-ID allow de novo identification of protein-RNA interactions in the cellular and thus – compared to existing methods – more relevant environment. They promise to be readily transferable to other cell types and will help to understand the roles of endogenous RNAs in disease and impact the development of safe and efficient ASO therapeutics, respectively.
RNA