synthesis. Through quantitative-based mass-spectrometry approaches using SILAC (stable isotope labeling with amino acids in cell culture), we now know that while a single miRNA can repress several hundreds of proteins, overexpression of miRNAs in mammalian cell lines cause mostly mild repression and rarely exceeds fourfold (Baek et al., 2008; Lim et al., 2005; Selbach et al., 2008). In stark contrast, some miRNAs (lsy-6, let-7, lin-4, etc.) elicit near-complete repression of target mRNAs to ensure proper development (Johnston and Hobert, 2003; Moss et al., 1997; Slack
136 et al., 2000).Thus, some miRNAs function as molecular switches while other miRNAs only act as fine-tuners of gene expression. For several years now, research groups have focused on understanding how individual miRNAs can differ in their overall silencing output. As the core component of miRISC, the Argonaute protein is necessary for any silencing to occur (reviewed in (Peters and Meister, 2007)). An interpretation in the field is that Argonaute proteins act as a molecular scaffold. Depending on the cell type, developmental stage, or alternative environmental conditions, the core miRISC may interact with different cofactors. Each of these cofactors can have its own activity or interact with other proteins that increase miRNA-mediated silencing activity. For instance, in Chapter 4, we showed that miRISC cofactor, NHL-2 increases the rate of miRNA-mediated deadenylation of target mRNAs. Our data suggest that the interaction of NHL-2 with CGH-1 leads to the reorganization of RNPs that accelerate deadenylase activity. With such an interpretation in mind, the absence, substitution, or post-translational modifications of a cofactor might affect miRNA activity. Could any one of these reasons explain why in Chapter 3, GYF-1 does not impact the function of lsy-6 and lin-4 miRNAs? While we know that GYF-1 is expressed in all the stages of C. elegans development, we do not know if GYF-1 is expressed in all cells or if GYF-1 undergoes post-translational modifications or substitution by another abundant cofactor in specific cell types. We now have the tools to address these questions. Affinity purification of lsy-6 or lin-4 miRNAs combined with proteomics can be used to investigate the composition of miRISC.
Early on, the identification of cofactors of miRISC relied primarily on classical forward genetic screens. But redundancy within proteins and miRNA family members can often mask the effect of loss of a single gene. Thus, reverse genetic screens using RNAi in sensitized miRNA backgrounds contributed to identifying proteins that regulate miRNA activity (Hurschler et al., 2011; Nolde et
137 al., 2007; Parry et al., 2007; Rausch et al., 2015). While RNAi is an easy and rapid approach, knockdown efficiencies can vary between experiments and sometimes fail to produce any phenotype even in sensitized miRNA backgrounds. This could be why previous genetic screens using RNAi (Parry et al., 2007; Rausch et al., 2015) failed to identify GYF-1 as a regulator of miRNA activity. An alternative approach has been to look at proteins associated with the Argonaute, and then through either genetics or knockdown experiments, uncover the function of miRNA-associated proteins (Chu and Rana, 2006; Hammell et al., 2009b; Kakumani et al., 2020).
This was seen in chapter 3, by comparing proteomic datasets of key components involved in the miRNA pathway, we identified several uncharacterized proteins beyond GYF-1. We have now begun to expand the characterization of other attractive candidates, including STAM-1 (See Appendix A5.1 for a complete list) obtained through this proteomics-based discovery approach.
The C. elegans STAM-1 (ortholog of human STAM and STAM2) is an accessory protein of the endosomal sorting complex required for transport (ESCRT) (Rusten et al., 2012), and might perhaps play a role in vesicular-based transport of miRNAs between subcellular compartments.
Preliminary studies indicate that STAM-1 is important for lsy-6 miRNA activity (See Appendix A5.2). Biochemical and genetic assays employed in chapters 3 and 4, along with immunofluorescence procedures, could thus be used to study the underlying mechanisms regulating the subcellular localization of miRNAs and their biological significance. Similar experiments can be performed in other model systems and ask if our findings in C. elegans are conserved and relevant to organism development and disease.
The past decade has seen the emergence of CRISPR/Cas9 tools augment our ability to precisely edit genomes and examine the function of a gene in its endogenous context. Indeed, this allowed us to engineer null alleles of gyf-1 and express various tagged versions of GYF-1 and NHL-2.
138 Also, targeted point mutations engineered in gyf-1 and nhl-2 loci uncovered the contribution of individual molecular determinants to miRNA function. Moving forward, an exciting application of CRISPR/Cas9 technology would be to identify biologically relevant targets of miRNAs and other RBPs. Methods such as ribosome profiling or CLIP have been widely used to this end.
However, these approaches generate a long list of candidate genes with false positives and eventually require in vivo validation. Instead, a recent work used a CRISPR screen in C. elegans and identified a single target gene (egl-1) alone sufficient to induce embryonic lethality observed in miR-35-42 family mutants (Yang et al., 2020). It would be of great interest to extend this method to identify biologically relevant targets of other miRNAs, GYF-1, NHL-2, and other RBPs.
Additionally, previously validated miRNA targets through reporter assays or overexpression experiments should be revisited, and this time validated with the use of CRISPR. This would help us to confirm previous candidates and possibly to identify functionally relevant miRNA targets.