Mapping targets for small nucleolar RNAs in yeast

Recent analyses implicate changes in the expression of the box C/D class of small nucleolar RNAs (snoRNAs) in several human diseases.

Systematic mapping of snoRNA-target binding provides a catalogue of high-confidence binding sites and indicates numerous potential regulatory interactions.

Here we report the identification of potential novel RNA targets for box C/D snoRNAs in budding yeast, using the approach of UV crosslinking and sequencing of hybrids (CLASH) with the snoRNP proteins Nop1, Nop56 and Nop58. We also developed a bioinformatics approach to filter snoRNA-target interactions for bona fide methylation guide interactions.

We recovered 241,420 hybrids, out of which 190,597 were classed as reproducible, high energy hybrids. As expected, the majority of snoRNA interactions were with the ribosomal RNAs (rRNAs). Following filtering, 117,047 reproducible hybrids included 51 of the 55 reported rRNA methylation sites. The majority of interactions at methylation sites were predicted to guide methylation. However, competing, potentially regulatory, binding was also identified. In marked contrast, following CLASH performed with the RNA helicase Mtr4 only 7% of snoRNA-rRNA interactions recovered were predicted to guide methylation. We propose that Mtr4 functions in dissociating inappropriate snoRNA-target interactions. Numerous snoRNA-snoRNA interactions were recovered, indicating potential cross regulation. The snoRNAs snR4 and snR45 were recently implicated in site-directed rRNA acetylation, and hybrids were identified adjacent to the acetylation sites. We also identified 1,368 reproducible snoRNA-mRNA interactions, representing 448 sites of interaction involving 39 snoRNAs and 382 mRNAs. Depletion of the snoRNAs U3, U14 or snR4 each altered the levels of numerous mRNAs. Targets identified by CLASH were over-represented among these species, but causality has yet to be established. Conclusions: Systematic mapping of snoRNA-target binding provides a catalogue of high-confidence binding sites and indicates numerous potential regulatory interactions.