Molecular and developmental analysis of non-coding RNA metabolism in" C. elegans" : the exoribonuclease XRN2 and the RNA- binding proteins SART-3 and USIP-1

About three quarters of a eukaryotic genome are transcribed into RNA. However, only <2% of these transcripts are translated into protein while the bulk of transcripts execute their biological function as RNA. Non-protein coding RNAs (ncRNAs) associate with proteins in ribonucleoprotein particles (RNPs) to regulate gene expression at various stages thereby greatly increasing the functional complexity of the genome. Nonetheless, the function and mode of action of the vast majority of ncRNAs is unknown and even in well studied examples little is known about the post-transcriptional regulation of ncRNAs themselves. In the thesis at hand, I explored the molecular and developmental functions of proteins implicated in the metabolism of ncRNAs, namely the miRNA-degrading enzyme XRN2 and the U6 snRNA-interacting proteins SART-3 and USIP-1. The XRN2 project was a collaboration with Takashi Miki and Hannes Richter.
XRN2 project
XRN2 is a conserved 5’-to-3’ exoribonuclease involved in various pathways including transcription termination and processing of precursor forms of rRNAs and snoRNAs. Our lab had established a function of XRN2 in the turnover of mature miRNAs, however, whether XRN2 targets all or specific miRNAs in vivo remained unclear. Although XRN2 substrates have extensively been characterized, the developmental function of XRN2 is essentially unexplored. Moreover, knowledge of co-factors regulating XRN2 function beyond transcription termination is scarce in multicellular organisms. In order to elucidate the developmental role of XRN2, we characterized an xrn-2 null and xrn-2 temperature-sensitive mutant. We found that XRN2 is essential during several stages of C. elegans development, including embryogenesis, and that only specific miRNAs are affected by XRN2 in vivo. Co-immunoprecipitations identified PAXT-1 (PArtner of XRN-Two 1) as a tight interaction partner of XRN2. paxt-1 depletion enhanced the xrn-2ts mutant phenotype and a paxt-1 null mutant slowed-down miRNA degradation in vivo, similar to XRN2 inactivation. These observations, as we showed, are due to a stabilizing effect of PAXT-1 on XRN2. Truncation mutants of PAXT-1 revealed a conserved N-terminal domain of unknown function, DUF3469, sufficient for XRN2 binding. We were excited to discover that human proteins containing DUF3469 were also able to bind to XRN2. Hence, we renamed DUF3469 to XRN2-binding domain (XTBD). Collectively, we identified PAXT-1 as an essential interaction partner of XRN2 in C. elegans and established a protein domain (XTBD) that serves as a binding platform for XRN2 beyond C. elegans.
Finally, the laboratory of Dr. Martin Simard found that the scavenger decapping enzyme DCS-1 interacts with the exonuclease XRN1, a paralogue of XRN2, to promote miRNA degradation in C. elegans. Collaborating on their project, I evaluated the subcellular localization of XRN1 and XRN2 in C. elegans and provided tools useful to their experiments such as an XRN1 antibody. This collaborative work has been published and can be found in section 7.
SART-3 project
The human protein SART3 and its yeast homolog Prp24 have previously been implicated in spliceosome assembly, namely the association of the U4 and U6 snRNP into the U4/U6 di-snRNP complex. Additionally, a physical interaction of SART3 with the Argonaute proteins AGO1 and AGO2 had been reported, suggesting an involvement of SART3 in the miRNA pathway. However, a putative function of SART3 in the miRNA pathway remained to be established. In order uncover such a function and to shed light on the so far largely neglected systemic role of SART3 in a multicellular context, I investigated its C. elegans homolog SART-3. Co-immunoprecipitations of SART-3 revealed an interaction with a previously uncharacterized putative terminal uridylyl transferase (TUTase), whereas I could not verify an interaction between SART-3 and AGO1/AGO2. It is known that SART3 binds specifically to the U6 snRNA which contains a post-transcriptionally elongated uridine (U)-tail essential for spliceosome assembly. Therefore it was appealing to assume that this U-tail is polymerized by the identified TUTase. Subsequent analyses unveiled an interaction between the TUTase and U6 snRNA, which hence was renamed to U Six snRNA Interacting Protein 1 (USIP-1). It appeared that USIP-1 binds to a U6 snRNA species that is devoid of Lsm proteins suggesting a role for USIP-1 early in spliceosome assembly. Moreover, knock-down of sart-3 in a usip-1 null mutant background led to a synthetic, embryonic lethal phenotype. This phenotype was rescued by transgenic expression of wild-type USIP-1. Although formal demonstration of TUTase-activity for USIP-1 is lacking, the synthetic lethality was not rescued by a supposedly catalytically inactive version of USIP-1. In sum, I established a physical and functional interaction between two previously uncharacterized proteins in C. elegans, SART-3 and USIP-1, and explored their developmental phenotypes.