SnoRNA processing

Abstract: Synthesis of small nucleolar RNA protein complexes (snoRNPs) occurs in several stages, which are almost certainly overlapping (for recent summaries, see references 66 and 81). The major phases include (i) synthesis of the snoRNA and protein components, (ii) assembly of the snoRNP, and (iii) movement of the particle to the nucleolar complex, perhaps by way of Cajal bodies (22, 45, 54, 58). In metazoans, most snoRNAs are derived from introns of protein genes, and all of the assembly steps are believed to occur in the nucleoplasm. In Saccharomyces cerevisiae, most snoRNAs are transcribed from independent genes; however, a few are encoded in introns (81). Processing of the pre-snoRNAs in yeast is coupled to mRNA splicing (50, 53), and this situation likely pertains to metazoans as well. Successful progression of the snoRNA transcript through each biosynthetic stage depends on the box C/D and H/ACA sequences used to classify the two large families of snoRNPs (see, e.g., references 8, 13, 14, 19, 27, 38, 39, 44, 45, 58, 64, and 79). In vivo function studies have shown that the binding of proteins to the motifs defined by the box elements is essential for proper processing of the snoRNA (end formation) and its localization to the nucleolus, presumably as a nascent snoRNP (reviewed in references 35, 66, and 81). Current models propose that protein binding to these motifs may be the first step in snoRNP assembly. Notably, the box elements are also intimately involved in the two types of nucleotide modification reaction mediated by the snoRNPs, i.e., ribose methylation by the C/D snoRNPs and pseudouridine formation by the H/ACA snoRNPs (4–6, 18, 34, 47, 66). The final list of factors involved in snoRNP synthesis and function will probably be quite extensive. For example snoRNA processing involves several nucleases, some of which also participate in the processing of rRNA, other small RNAs, and pre-mRNA (72; see also references 2, 15, 16, and 71). For the snoRNP particles, four common proteins that are specific for the C/D class and another four that are specific for the H/ACA class have been identified. The core C/D proteins in Saccharomyces cerevisiae include Nop1p, Nop56p, Nop58p, and Snu13p (reviewed in references 35, 52, and 73; see also references 19, 21, 36, 59, 68, 80, and 83). Orthologs of these proteins have been demonstrated to exist in other species as well, including humans (41, 59). Remarkably, several (but not all) archaeal organisms contain C/D-like guide RNAs and snoRNP-like proteins of both classes (20, 49, 77). In addition to the common proteins, some snoRNPs such as U3 have unique proteins (references 17 and 74 and citations therein). The C/D snoRNP protein Nop1p is a strong candidate for a snoRNP-based methyltransferase, as point mutations in a methylase-like domain disrupt methylation of rRNA globally (67, 76). The Snu13 protein was recently determined to bind specifically to the C/D motif (80). Surprisingly, this protein, known as 15.5 kDa in human cells (48), is also an integral component of the U4/U6.U5-splicing snRNP (65) and binds to a C/D-like motif in U4 (75, 80). It remains to be seen if the presence of Snu13/15.5 kDa in both snoRNPs and snRNPs reflects a mechanistic link between splicing and snoRNP synthesis or simply occurs by coincidence. Proteins common to the H/ACA family of snoRNAs have been characterized best in S. cerevisiae and include Cbf5p, Gar1p, Nhp2p, and Nop10p (8, 23, 25, 37, 78). As for the C/D core proteins, the H/ACA snoRNP proteins are also evolutionarily conserved (54, 77). Cbf5p is almost certainly the universal H/ACA-associated Ψ synthase, based on the observation that point mutations in known pseudouridine synthase motifs abolish activity in vivo (84). Gar1p has RNA binding activity (6) but is not known to be motif specific. With the goal of discovering proteins that act early in snoRNP biogenesis, one of our laboratories identified four mouse proteins that associate with a model C/D snoRNA in vitro (46). Two of the proteins are orthologs of yeast Nop56 and Nop58/Nop5p, which were originally identified by genetic strategies (21). The other two proteins, called p50 and p55, are novel and are related to each other at the levels of 42% identity and 68% similarity (46). They are strongly conserved among the Eukarya and many, but not all, Archaea; interestingly, the Archaea contain a single ortholog. Each protein has distinguishing Walker A and B motifs, which suggest ATP/GTP binding and nucleoside triphosphatase (NTPase) functions. Additional sequence relatedness occurs with bacterial DNA helicase RuvB in N- and C-terminal regions that together comprise about 50% of the overall sequence (32). Neither protein has been linked to snoRNAs previously; however, both have been characterized in other contexts. Both proteins have been reported to occur in the nucleoplasm (10, 40, 42, 46), suggesting that they may play a role in an early stage of snoRNP biogenesis. Studies by others have implicated the novel snoRNA-related proteins in chromatin remodeling and transcription through association with components involved in those processes or stimulation of gene expression (see Discussion). Most relevant to the present study, the p50 and p55 proteins have been shown to associate with each other in yeast and human cells (9, 31, 40, 57, 82) and proteins from rat and human cells exhibit DNA helicase activities of opposite polarity in vitro (28, 31). Historically, the proteins were discovered independently by groups studying different processes in different organisms. Accordingly, they have several names. The mouse protein, called p50 in the previous snoRNA report, had not been described at that time. Subsequent orthologs were identified and given an assortment of names, including Tip49b (TATA-box binding protein [TBP]-interacting protein 49b), ECP-51 (erythrocyte cytosol protein 51), TIP48 (trans-activation domain-interacting protein 48), Reptin52 (repressing pontin 52), TAP54-β (Tip60-associated protein of 54 kDa in humans), P47 (47-kDa protein), scTip49b (S. cerevisiae Tip49b), scRUVBL2 (S. cerevisiae RUVBL2), Rvb2 (RuvB-like protein 2), and Tih2 (Tip49-homology protein 2) (9, 24, 28, 31, 40, 55, 57, 61, 82). The mouse protein called p55 in the snoRNA binding study was first discovered in rats and was called Tip49 (32). Orthologs of this protein have since been reported as well, and various names have been assigned including Tip49 (TBP-interacting protein 49), Pontin52 (Pont meaning bridge and 52 referring to kilodaltons), ECP-54 (erythrocyte cytosol protein 54), NMP238 (nuclear matrix protein 238), RUVBL1 (Ruv B-like protein 1), TIP49 (transactivation domain interacting protein 49), TAP54-alpha (Tip60-associated protein of 54 kDa), p50 (50-kDa protein), Tip49a (TBP-interacting protein 49a), scTip49a (S. cerevisiae Tip49a), scRUVBL1 (S. cerevisiae RuvB-like protein 1), Rvb1 (RuvB-like protein 1), and Tih1 (Tip49 homology protein 1) (9, 10, 26, 28, 31, 33, 40, 55, 57, 61, 82). In this report, which features the yeast p50 variant, we refer to the protein as Rvb2p, as this is the name adopted in the S. cerevisiae Genome Database. The mouse ortholog and other orthologs are referred to simply as p50. To gain insight into the relationship of p50 with the snoRNAs, we carried out a genetic study of the yeast protein. Depletion and mutation analyses revealed an essential role in the accumulation of the box C/D snoRNAs and, surprisingly, the box H/ACA snoRNAs as well. Depletion also caused core proteins from each snoRNP family to accumulate in the nucleoplasm. Finally, point mutations revealed that snoRNA production requires the phylogenetically conserved ATP/GTP-binding and hydrolysis motifs. We discuss these results in the context of the snoRNP literature and reports that link these proteins with a variety of other processes.