DISCUSSION
In this article, we present hnRNPA1 recruiting oligonucleotides (AROs), a new tool for mRNA knockdown consisting of a short targeting oligonucleotide and a hnRNPA1 recruiting loop. Through a series of RT-qPCR experiments targeting KRT14 and TBK1, we have shown that AROs expressed from plasmids are able to knock down the expression of two independent genes in a targeted fashion. We have shown that both the antisense targeting RNA and the hnRNPA1 recruiting loop must be present for an efficient knockdown. We have also shown that AROs function in an hnRNPA1 dependent manner. Finally, we have shown that AROs are functional in multiple cell lines including human keratinocytes (HaCat) and squamous cell carcinoma cells (SCC9) and can affect a loss-of-expression functional phenotype.
Our results suggest that the location of the sequence targeted by an ARO on the exon of interest does not have a large influence on knockdown efficiency. We initially hypothesized that the targeting location would have a bigger impact on ARO function. Specifically, we expected that it would be necessary to direct AROs to splice junctions or to ASF/SF2 motifs, as preventing binding of SR splice enhancer proteins is a known mechanism of hnRNPA1 splice site repression (Jean-Philippe et al., 2013; Kashima et al., 2007). One possible explanation is the cooperative recruitment of hnRNPA1 molecules along RNA in the 3’ to 5’ direction (Jean-Philippe et al., 2013). This might mean that an ARO binding at a downstream exonic sequence is sufficient to suppress SR protein recruitment at upstream regions. Another possibility is that the exon we choose was too small to see location specific effects. Larger exons with more space between splice sites should be tested in the future to help answer this question.
A major advantage of AROs over most knockdown technologies is the ease with which they can transition from promotor-based expression (i.e. from a plasmid or lentivirus construct) to RNA oligonucleotides that can be directly delivered to cells or organisms. Here we have shown the utility of AROs as a potential therapeutic by synthesizing single-stranded RNA AROs and showing that they are able to achieve similar levels of knockdown as their plasmid counterparts. By using ssRNA AROs to knock down KRT14 in SCC9, a squamous cell carcinoma line used as a model for epithelial-mesenchymal transition studies in head and neck cancer, we demonstrate that ARO-based knockdown resulted in a biologically relevant phenotype – the decreased ability of carcinoma cells to migrate in a cell invasion assay.
An unexpected finding in our RNA oligonucleotide experiments was that while the double ARL ssRNA (ARL-Oligo-ARL) worked to some degree at high concentrations, the single ARL (ARL-Oligo) achieved significantly better gene knockdown at lower concentrations. This contrasted with our plasmid experiments which found that double ARL AROs performed best. A possible explanation for this may be that the additional size or complexity of double ARL oligonucleotides may have lessened their ability to make it to the nucleus. Another possibility is that complementary regions between the two ARLs caused secondary structure leading to double-stranded portions of the construct being degraded in the cytoplasm. This would be less likely to occur in the plasmid construct as those AROs would be produced in the nucleus and would presumably have more opportunity to affect pre-mRNA before being exported and degraded.
We hypothesize that AROs knock down gene expression by causing errors in pre-mRNA splicing followed by nonsense-mediated decay. However, it is important to point out that there are several plausible alternative hypotheses. For example, hnRNPA1 has been shown to bind to the 3’ UTR of some genes, and this can lead to mRNA deadenylation and degradation (Geissler et al., 2016). This may be why binding along any point of an exon in pre-RNA seems to be effective. While more work will need to be done to determine the specific mechanism of action of AROs to better understand and improve their function, their utility as biological tools in their current form remains substantial.
AROs represent a novel paradigm for recruiting RNA binding proteins to pre-mRNAs via synthetic constructs. Here we have shown that hnRNPA1 can be recruited via a short targeting oligonucleotide to knock down transcript levels of a target gene. While more work will need to be done to understand the exact mechanism of how AROs function, the work here illustrates their potential as a novel, biologically functional way to knock down target transcripts through the redirection of endogenous cellular machinery. Because RNA binding proteins are ubiquitous and abundant across cell types, their recruitment to targeted pre-mRNAs may be a useful way to modulate RNA production, modification, or degradation. The diverse array of functions they perform and their directable nature make them an ideal source to mine for molecular biology tools. The ability to express AROs, and similarly designed molecules, from a Pol II promoter also has substantial advantages. AROs could be placed under the control of cell or tissue specific promotors and used in tissue specific studies or placed behind activatable promotors.
AROs can both be expressed by constructs or used directly as ssRNA molecules with little modification. This dual functionality puts them in a unique position to bridge the gap from high-throughput screens to independently functional molecules. We believe AROs fill an important niche in the scientific toolbox and we hope the work here can serve as a template for future constructs that make use of the potential of RNA binding proteins as tools for the modulation of RNA homeostasis.