A small nucleolar RNP protein is required for pseudouridylation of eukaryotic ribosomal RNAs
1997; Springer Nature; Volume: 16; Issue: 15 Linguagem: Inglês
10.1093/emboj/16.15.4770
ISSN1460-2075
Autores Tópico(s)Peptidase Inhibition and Analysis
ResumoArticle1 August 1997free access A small nucleolar RNP protein is required for pseudouridylation of eukaryotic ribosomal RNAs Cécile Bousquet-Antonelli Cécile Bousquet-Antonelli Laboratoire de Biologie Moléculaire Eucaryote du CNRS, Université Paul Sabatier, 118 route de Narbonne, 31062 Toulouse, Cedex, France Search for more papers by this author Yves Henry Yves Henry Laboratoire de Biologie Moléculaire Eucaryote du CNRS, Université Paul Sabatier, 118 route de Narbonne, 31062 Toulouse, Cedex, France Search for more papers by this author Jean-Paul Gélugne Jean-Paul Gélugne Laboratoire de Biologie Moléculaire Eucaryote du CNRS, Université Paul Sabatier, 118 route de Narbonne, 31062 Toulouse, Cedex, France Search for more papers by this author Michèle Caizergues-Ferrer Michèle Caizergues-Ferrer Laboratoire de Biologie Moléculaire Eucaryote du CNRS, Université Paul Sabatier, 118 route de Narbonne, 31062 Toulouse, Cedex, France Search for more papers by this author Tamás Kiss Corresponding Author Tamás Kiss Laboratoire de Biologie Moléculaire Eucaryote du CNRS, Université Paul Sabatier, 118 route de Narbonne, 31062 Toulouse, Cedex, France Search for more papers by this author Cécile Bousquet-Antonelli Cécile Bousquet-Antonelli Laboratoire de Biologie Moléculaire Eucaryote du CNRS, Université Paul Sabatier, 118 route de Narbonne, 31062 Toulouse, Cedex, France Search for more papers by this author Yves Henry Yves Henry Laboratoire de Biologie Moléculaire Eucaryote du CNRS, Université Paul Sabatier, 118 route de Narbonne, 31062 Toulouse, Cedex, France Search for more papers by this author Jean-Paul Gélugne Jean-Paul Gélugne Laboratoire de Biologie Moléculaire Eucaryote du CNRS, Université Paul Sabatier, 118 route de Narbonne, 31062 Toulouse, Cedex, France Search for more papers by this author Michèle Caizergues-Ferrer Michèle Caizergues-Ferrer Laboratoire de Biologie Moléculaire Eucaryote du CNRS, Université Paul Sabatier, 118 route de Narbonne, 31062 Toulouse, Cedex, France Search for more papers by this author Tamás Kiss Corresponding Author Tamás Kiss Laboratoire de Biologie Moléculaire Eucaryote du CNRS, Université Paul Sabatier, 118 route de Narbonne, 31062 Toulouse, Cedex, France Search for more papers by this author Author Information Cécile Bousquet-Antonelli1, Yves Henry1, Jean-Paul Gélugne1, Michèle Caizergues-Ferrer1 and Tamás Kiss 1 1Laboratoire de Biologie Moléculaire Eucaryote du CNRS, Université Paul Sabatier, 118 route de Narbonne, 31062 Toulouse, Cedex, France *Corresponding author. E-mail: [email protected] The EMBO Journal (1997)16:4770-4776https://doi.org/10.1093/emboj/16.15.4770 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info Eukaryotic rRNAs possess numerous post-transcriptionally modified nucleotides. The most abundant modifications, 2′-O-ribose methylation and pseudouridylation, occur in the nucleolus during rRNA processing. The nucleolus contains a large number of small nucleolar RNAs (snoRNAs) most of which can be classified into two distinct families defined by conserved sequence boxes and common associated proteins. The C and D box-containing snoRNAs are associated with fibrillarin, and most of them function as guide RNAs in site-specific ribose methylation of rRNAs. The nucleolar function of the other class of snoRNAs, which share box H and ACA elements and are associated with a glycine- and arginine-rich nucleolar protein, Gar1p, remains elusive. Here we demonstrate that the yeast Saccharomyces cerevisiae Gar1 snoRNP protein plays an essential and specific role in the overall pseudouridylation of yeast rRNAs. These results establish a novel function for Gar1 protein and indicate that the box H/ACA snoRNAs, or at least a subset of these snoRNAs, function in the site-specific pseudouridylation of rRNAs. Introduction In eukaryotic cells, synthesis and maturation of cytoplasmic rRNAs take place in the nucleolus (Hadjiolov, 1985). The 18S, 5.8S and 25/28S rRNAs are synthesized as a single precursor RNA (pre-rRNA) which contains long external and internal transcribed spacer sequences that are removed by complex nucleolytic reactions (Eichler and Craig, 1995; Venema and Tollervey, 1995; Sollner-Webb et al., 1996). Before or during processing, many nucleotides in the rRNA-coding regions of pre-rRNA are methylated at the 2′-O-hydroxyl position, and numerous uridine residues are converted into pseudouridines (Maden, 1990; Eichler and Craig, 1995; Ofengand et al., 1995). Since these modified nucleotides are confined to the universal core regions of rRNAs and their positions show remarkable evolutionary conservation, they are believed to contribute to the function of rRNAs (Maden, 1990; Lane et al., 1995; Ofengand et al., 1995). The nucleolus contains a large number of small nucleolar RNAs (snoRNAs) (Filipowicz and Kiss, 1993; Maxwell and Fournier, 1995; Sollner-Webb et al., 1996; Tollervey and Kiss, 1997). Almost 100 distinct snoRNA species have been identified in mammalian and yeast cells (Maxwell and Fournier, 1995; Balakin et al., 1996; Kiss-László et al., 1996; Nicoloso et al., 1996; Tycowski et al., 1996a; Ganot et al., 1997b). Many snoRNAs, such as U3 (Kass et al., 1990; Savino and Gerbi, 1990; Hughes and Ares, 1991), U8 (Peculis and Steitz, 1993), U14 (Li et al., 1990), U22 (Tycowski et al., 1994), snR30 (Morrissey and Tollervey, 1993) and 7-2/MRP RNA (Schmitt and Clayton, 1993; Chu et al., 1994; Lygerou et al., 1996) have been demonstrated to play essential roles in the nucleolytic processing of pre-rRNA. Recently, it was shown that a large subset of snoRNAs containing conserved C and D/D′ boxes and displaying extensive complementarities to rRNAs function as guide RNAs in the site-specific 2′-O-ribose methylation of pre-rRNA (Cavaillé et al., 1996; Kiss-László et al., 1996; Tycowski et al., 1996b). Formation of a long (10–21 nucleotide) uninterrupted snoRNA–rRNA duplex in the immediate vicinity of the D or D′ box of the snoRNA specifies the target nucleotide for ribose methylation (Cavaillé et al., 1996; Kiss-László et al., 1996). The methylation guide snoRNAs, like other C and D box-containing snoRNAs (Tyc and Steitz, 1989; Tycowski et al., 1993), are associated with an evolutionarily conserved nucleolar protein, fibrillarin (Kiss-László et al., 1996; Tycowski et al., 1996a). Consistent with this, the yeast equivalent of fibrillarin, Nop1p, which is specifically associated with all yeast box C/D snoRNAs (Ganot et al., 1997b), is fundamental to the global ribose methylation of rRNAs (Tollervey et al., 1993). To date, no factors required for site-specific pseudouridylation of eukaryotic rRNAs have been identified. In addition to the family of box C/D snoRNAs, the nucleolus of yeast and mammalian cells contains another distinct set of snoRNAs which are defined by two conserved sequence elements, boxes H and ACA (Balakin et al., 1996; Ganot et al., 1997b). The box H/ACA snoRNAs also share a common secondary structure; they fold into two hairpin structures that are connected and followed by short single-stranded regions containing the H and ACA elements, respectively (Ganot et al., 1997b). The nucleolar function of this class of snoRNAs is largely unknown. All yeast box H/ACA snoRNAs tested, with the exception of snR30 which is required for 18S rRNA processing (Morrissey and Tollervey, 1993), are dispensable for viability (Maxwell and Fournier, 1995; Balakin et al., 1996, and references therein). In yeast, all box H/ACA snoRNAs are associated with an essential glycine- and arginine-rich nucleolar protein, Gar1p (Balakin et al., 1996; Ganot et al., 1997b). Although the Gar1 protein is required for the accumulation of mature 18S rRNA (Girard et al., 1992, 1993), its nucleolar function as a component of numerous non-essential snoRNP particles is largely unknown. In this study, we have investigated the possible involvement of the Gar1 snoRNP protein in post-transcriptional modification of rRNAs in the yeast Saccharomyces cerevisiae. We show that depletion of Gar1 protein, in addition to inhibiting 18S rRNA production, also prevents the global pseudouridylation of 35S and 25S rRNAs, but not of tRNAs. Genetic restoration of Gar1p production re-establishes pseudouridylation of rRNAs. These results identify a novel cellular function for Gar1p and, more importantly, indicate that the Gar1p-associated box H/ACA snoRNAs, or a subset of these snoRNAs, function in the site-specific pseudouridylation of rRNAs. Results Construction of a yeast strain carrying a thermosensitive allele of the GAR1 gene The recent notion that the majority of fibrillarin-associated snoRNAs function as guide RNAs in site-specific ribose methylation of pre-rRNA raised the intriguing hypothesis that the other major group of snoRNAs, the Gar1p-associated box H/ACA snoRNAs, might be involved in formation of ribosomal pseudouridines (Ganot et al., 1997b). This idea is reinforced by the observation that all yeast box H/ACA snoRNAs tested, with the exception of snR30 (Morrissey and Tollervey, 1993), are dispensable for viability (Maxwell and Fournier, 1995; Balakin et al., 1996). Since yeast Gar1p interacts specifically and intimately with all box H/ACA snoRNAs (Balakin et al., 1996; Ganot et al., 1997b), the above hypothesis leads to the prediction that in a yeast strain deficient in functional Gar1p, the overall pseudouridylation of rRNAs should be inhibited while ribose methylation should remain unaffected. To assay for rRNA pseudouridylation in a yeast strain lacking functional Gar1p, we made use of the gar1.1 temperature-sensitive allele. This allele codes for a mutant protein, Gar1.1p, that is degraded rapidly at the non-permissive temperature of 37°C. As compared with the wild-type protein, the mutant Gar1.1p carries six amino acid substitutions (Figure 1). The two extra amino acids in the N-terminal region of Gar1.1p, an aspartic acid (D6) and a proline (P7) resulting from the introduction of a new restriction site (see Materials and methods), do not contribute to the temperature-sensitive phenotype of Gar1.1p (data not shown). A yeast strain (gar1::LEU2) lacking the endogenous wild-type GAR1 gene was complemented with a centromeric vector bearing the gar1.1 allele. Western blot analysis was used to monitor the accumulation of intracellular Gar1.1p in the gar1.1 strain grown at the permissive (25°C) or non-permissive (37°C) temperature (Figure 2A). The amount of Gar1.1p was strongly reduced after 3 h growth at 37°C, whereas the levels of the nucleolar Nop1 snoRNP protein (Schimmang et al., 1989; Maxwell and Fournier, 1995), and the cytoplasmic Srp21p, a protein component of yeast signal recognition particle (Brown et al., 1994), remained unaffected. Since accumulation of the mRNA encoding the mutant Gar1.1p was not altered by shifting the gar1.1 strain to non-permissive temperature (data not shown), we concluded that the instability of Gar1p at 37°C is due to conformational changes in the protein structure. Figure 1.Schematic structure of the Gar1.1 protein encoded by the gar1.1 allele. The mutant protein is represented as a box ranging from position 1 (N-terminus) to position 207 (C-terminus). The glycine- and arginine-rich GAR domains (Girard et al., 1992) are indicated by shaded boxes. The amino acid substitutions and their corresponding positions within the mutant Gar1.1p and the wild-type Gar1p are indicated. Introduction of a new BamHI site into the GAR1 ORF resulted in the addition of an aspartic acid and a proline between positions five and six of the wild-type protein. Consequently, the mutant protein is 207 rather than 205 (Girard et al., 1992) amino acids long. Download figure Download PowerPoint Figure 2.Characterization of the gar1.1 thermosensitive allele of the yeast S.cerevisiae GAR1 gene. (A) Expression of the gar1.1 allele. Accumulation of the Gar1.1, Nop1 and Srp21 proteins in the gar1.1 yeast strain grown for 3 h at either 37 (lane 1) or 25°C (lane 2) was measured by Western blot analyses. The same blot was reacted with anti-GAR1, anti-SRP21 or anti-NOP1 antibodies. Molecular size markers are indicated on the left. (B) Pulse labelling of cellular RNAs. The gar1.1 strain was grown at 25°C in minimal medium to mid-exponential phase. Cells were washed and resuspended in rich medium containing glucose but lacking inorganic phosphate, and were grown for 3 h at either 25 or 37°C. The cultures were labelled with [32P]orthophosphate for 15 min before total RNA was extracted and analysed by electrophoresis on a 1.2% agarose–formaldehyde gel. The positions of the rRNA species are indicated. Download figure Download PowerPoint Depletion of Gar1p inhibits pseudouridylation of 35S and 25S rRNAs To test for rRNA pseudouridylation, the gar1.1 strain was grown on rich medium without inorganic phosphate for 3 h either at the permissive (25°C) or non-permissive (37°C) temperature before addition of [32P]orthophosphate. After 15 min of labelling, total RNAs were extracted and analysed by gel electrophoresis (Figure 2B). Consistent with previous studies on Gar1p-depleted yeast cells (Girard et al., 1992), the gar1.1 strain impaired production of mature 18S rRNA at non-permissive temperature (lane 1). The 35S, 25S rRNAs and tRNAs were electroeluted and their pseudouridine content was determined by RNase T2 digestion followed by two-dimensional thin-layer chromatography (TLC) and PhosphorImager quantification (Figure 3). At the non-permissive temperature that does not support accumulation of the Gar1 protein, the global pseudouridylation of 35S and 25S rRNAs was almost completely abolished (Figure 3A and B). While the wild-type yeast 18S and 25S rRNAs are known to contain 13 and 30 pseudouridine residues, respectively (Ofengand et al., 1995), the 35S pre-rRNA and the 25S rRNA obtained from cells grown at 37°C were calculated to carry only 2–4 and 1–3 pseudouridines, respectively. At 25°C, pseudouridylation of the 35S pre-rRNA and 25S rRNA was much more efficient (Figure 3D and E), although the measured pseudouridine content of the 25S rRNA (20–25 residues per molecule) never reached the wild-type levels. This indicates that the mutant Gar1.1p cannot fully support rRNA pseudouridylation even at permissive temperature. The fact that the 35S pre-rRNA at 25°C contained ∼30–35 pseudouridine residues confirms that most or all conversions of uridine residues into pseudouridines occur already on the 35S primary transcript (Brand et al., 1979). Figure 3.Depletion of Gar1p inhibits pseudouridylation of rRNAs. In vivo labelled 35S pre-rRNA, 25S rRNA and tRNA originating from the gar1.1 temperature-sensitive strain grown at permissive (25°C) or restrictive (37°C) temperature were isolated by electroelution from a preparative gel similar to the one shown in Figure 1C. The purified RNAs were digested by RNase T2 and analysed by two-dimensional TLC. Spots corresponding to the Ap, Gp, Cp and Up residues are indicated in (A). The position of the pseudouridine 3′-monophosphate residue is indicated in each panel (Ψp). Weak spots represent 2′-O-methylated dinucleotides resistant to RNase T2 digestion (Keith, 1995). The diagram indicates the relative pseudouridine contents of 35S (open columns) and 25S (closed columns) rRNAs obtained from the gar1.1 strain grown at either 25 or 37°C and from a control wild-type strain grown at 25°C. The levels of pseudouridines and uridines were measured by PhosporImager quantification and the percentage of uridine residues converted into pseudouridines was calculated. Values represent means ± SEM from at least three independent experiments. Yeast 35S and 25S rRNAs contain 1847 and 868 uridines, respectively. Download figure Download PowerPoint Pulse–chase labelling of pre-rRNAs in yeast cells lacking a functional GAR1 gene unambiguously demonstrated that the Gar1 protein is not required for global 2′-O-ribose methylation of pre-rRNA (Girard et al., 1992). Consistent with this observation, quantification of ribose-methylated nucleotides in 25S and 35S rRNAs by P1 nuclease digestion followed by two-dimensional TLC analyses showed that the 2′-O-ribose methylation of pre-rRNA was not altered in the gar1.1 strain grown at the non-permissive temperature (data not shown). Hence, weak spots representing ribose-methylated dinucleotides resistant to RNase T2 digestion (Smith and Dunn, 1959) were always detected during the RNase T2–TLC analyses of 25S or 35S rRNAs obtained from the gar1.1 or its derivative strains (Figures 3 and 4C). We also tested the possible involvement of Gar1 protein in pseudouridylation of tRNAs (Figure 3C and F). In marked contrast to rRNAs, the global pseudouridylation of tRNAs was not affected by shifting the gar1.1 strain to non-permissive temperature, indicating that the Gar1 protein is required specifically for rRNA pseudouridylation. Figure 4.Genetic restoration of normal levels of Gar1p protein in the thermosensitive gar1.1 strain restores pseudouridylation of 25S rRNA at non-permissive temperature. (A) Western blot analysis. Yeast strains lacking the intact chromosomal GAR1 gene, but expressing plasmid-born wild-type Gar1 (GAR1 strain) or thermosensitive Gar1.1 (gar1.1 strain) proteins, were transformed with a centromeric vector (pGAL) carrying the wild-type GAR1 gene under the control of the GAL1-10 promoter, resulting in strains GAR1+pGAL and gar1.1+pGAL. Proteins extracted from the GAR1+pGAL and gar1.1+pGAL strains grown either in glucose-containing medium for 3 h at 25 (lanes 1 and 4) or 37°C (lanes 2 and 5), or in glucose-containing medium for 3 h and for an additional 8 h in galactose-containing medium at 37°C (lanes 3 and 6) were analysed by Western blotting. To facilitate cloning, we have introduced linker sequences into the GAR1 ORF. As a consequence, the Gar1 protein encoded by pGAR1 is 23.5 kDa and the Gar1.1 protein encoded by the gar1.1 allele on pGAR1.1 is 21.7 kDa. The pGAL vector contains an unmodified GAR1 ORF coding for a 21.5 kDa protein. For other details, see legend to Figure 1B. (B) Denaturing agarose gel electrophoresis of in vivo labelled total RNAs extracted from the GAR1+pGAL and gar1.1+pGAL strains grown under the conditions described above. Pulse labelling, extraction and separation of cellular RNAs were performed as described for the gar1.1 strain in the legend to Figure 1. (C) Analysis of the pseudouridine content of 25S rRNAs obtained from the GAR1+pGAL or gar1.1+pGAL strains. In vivo labelled 25S rRNAs eluted from preparative gels similar to the one shown in (B) were hydrolysed by RNase T2 and analysed by two-dimensional TLC. Culture conditions are indicated above the panels. The position of the pseudouridine 3′-monophosphate residue is indicated in each panel (Ψp). Download figure Download PowerPoint Expression of wild-type Gar1p restores rRNA pseudouridylation at non-permissive temperature To demonstrate unambiguously the specific involvement of the Gar1 snoRNP protein in pre-rRNA pseudouridylation, we tested whether genetic restoration of wild-type Gar1p production in the gar1.1 strain at non-permissive temperature would restore normal levels of rRNA pseudouridylation. For this purpose, we constructed a centromeric vector, called pGAL, where the wild-type GAR1 open reading frame (ORF) was placed under the control of the GAL1-10 promoter. This plasmid was introduced into the gar1::LEU2 strain complemented with plasmids bearing either the mutant gar1.1 or the wild-type GAR1 gene. The resulting gar1.1+pGAL and GAR1+pGAL strains were grown on glucose-containing medium for 3 h at 25 or 37°C. Half of the cultures grown at 37°C were resuspended in galactose-containing medium and grown for an additional 8 h at 37°C to induce expression of the wild-type GAR1 gene from the GAL1-10 promoter. Accumulation of the Gar1 proteins and the state of rRNA pseudouridylation was monitored for each sample as described above for the gar1.1 strain. The strong reduction in Gar1p levels observed in the gar1.1+pGAL strain grown on glucose at 37°C (Figure 4A, lane 5) was correlated with an inhibition of 18S rRNA production (Figure 4B, lane 5) and a strong reduction of 25S rRNA pseudouridylation (Figure 4C, panel e). Shifting of the gar1.1+pGAL strain to galactose medium at 37°C restored normal levels of Gar1p (Figure 4A, lane 6), production of 18S rRNA (Figure 4B, lane 6) as well as pseudouridylation of 25S rRNA (Figure 4C, panel f). In the control GAR1+pGAL strain that expresses the wild-type GAR1 gene constitutively (Figure 4A, lanes 1–3), processing (Figure 4B, lanes 1–3) and pseudouridylation (Figure 4C, panels a–c) of rRNAs remained unaffected, regardless of growth conditions. These results demonstrate that Gar1p plays a specific and fundamental role in the overall pseudouridylation of yeast rRNAs. Pseudouridylation of pre-rRNA is independent of 18S rRNA processing In addition to the Gar1 (Girard et al., 1992) and Nop1 (Tollervey et al., 1991) snoRNP proteins, several Nop1p-associated snoRNAs (U3, U14, U22) and at least one Gar1p-associated snoRNA, snR30, are required for the production of 18S rRNA (Maxwell and Fournier, 1995; Venema and Tollervey, 1995; Sollner-Webb et al., 1996; Tollervey and Kiss, 1997). Since processing events in the 5′ external transcribed spacer (ETS) and internal transcribed spacer 1 (ITS1) regions of pre-rRNA are coupled, the snoRNPs essential for mature 18S rRNA production are believed to assemble into a large multi-RNP complex, called the processome (Fournier and Maxwell, 1993; Maxwell and Fournier, 1995; Venema and Tollervey, 1995). In principle, it is imaginable that formation of an active processome on the pre-rRNA may be crucial for pseudouridylation of pre-rRNA, and inhibition of ribosomal pseudouridine formation in the Gar1p-depleted strain may reflect a general property of mutants impaired in 18S production. To exclude this possibility, we checked rRNA pseudouridylation in yeast strains depleted of snR30 and U3 snoRNAs. We used strains D190 (Morrissey and Tollervey, 1993) and JH88 (Hughes and Ares, 1991), in which the disrupted chromosomal snR30 or U3a/U3b genes are complemented with the snR30 and U3b genes placed under the control of the GAL1-10 promoter, respectively. In the D190 and JH88 strains, after 18 h growth on glucose-containing medium, snR30 and U3 snoRNAs were hardly detectable, and accumulation of 18S rRNA was completely abolished (Hughes and Ares, 1991; Morrissey and Tollervey, 1993; and data not shown). However, as compared with the control cells grown on galactose-containing medium (Figure 5A and C), pseudouridylation of 35S pre-rRNAs in the snR30- (Figure 5B) or U3-depleted (Figure 5D) cells was globally unaffected. Similar results were obtained for 25S rRNAs as well (data not shown). These results prove that pseudouridylation of pre-rRNA is independent of the processing of 18S RNA and, more importantly, demonstrate that inhibition of rRNA pseudouridylation is restricted to strains deficient in GAR1 function. Figure 5.Depletion of snR30 and U3 snoRNAs has no effect on the global pseudouridylation of 35S pre-rRNA. Yeast strains D190 and JH88 were grown for 18 h on rich medium lacking inorganic phosphate and supplemented either with galactose or with glucose in order to deplete cells of snR30 (D190) or U3 (JH88) snoRNAs. Cells were labelled with [32P]orthophosphate for 15 min and total RNA was extracted. Purified 35S rRNAs were digested with RNase T2 and analysed by two-dimensional TLC. For other details see the legend to Figure 3. Download figure Download PowerPoint Discussion Recently, it became evident that the great majority of snoRNPs fall into two distinct structurally and probably functionally well-defined families (Balakin et al., 1996; Ganot et al., 1997b). Many snoRNAs share two conserved sequence elements, boxes C and D, and are associated with a common antigen, fibrillarin. Some fibrillarin-associated snoRNAs are required for processing of rRNAs, but most of them function as guide RNAs in site-specific ribose methylation of pre-rRNA. The other major group of snoRNAs share an evolutionarily conserved 'hairpin–hinge–hairpin–tail' secondary structure and two conserved sequence elements positioned in the single-stranded hinge (box H) and tail (box ACA) regions of the snoRNA (Balakin et al., 1996; Ganot et al., 1997b). In yeast, all box H/ACA snoRNAs are associated with Gar1p, which is an abundant nucleolar protein required for accumulation of mature 18S rRNA (Girard et al., 1992). In this study, using a thermosensitive allele of the GAR1 gene, we have demonstrated that Gar1p plays a specific and fundamental role in global pseudouridylation of yeast rRNAs. These results establish a novel function for Gar1p and identify the first factor involved in rRNA pseudouridylation. The fact that an snoRNP protein common to all box H/ACA snoRNPs is required for global pseudouridylation of rRNAs strongly suggests that the box H/ACA snoRNAs, or at least a subset of these RNAs, function in the site-specific pseudouridylation of pre-rRNA. Since the overwhelming majority of the intracellular pool of yeast Gar1 snoRNP protein is associated with snoRNPs or higher order nucleolar structures (our unpublished data), it seems unlikely that the observed function of Gar1p in pre-rRNA pseudouridylation is independent of snoRNA association. This leads to the conclusion that selection of specific uridine residues in the pre-rRNA and/or isomerization of the selected uridines to pseudouridines is mediated by box H/ACA snoRNPs. Indeed, recent genetic depletion and reconstitution experiments on two yeast box ACA snoRNAs, snR5 and snR36, proved that these snoRNAs function in the site-specific pseudouridylation of pre-rRNA (Ganot et al., 1997a). The yeast snR5 is required for the pseudouridylation of 25S rRNA at positions Ψ1003 and Ψ1123, and snR36 plays a crucial and specific role in the formation of Ψ1185 in the 18S rRNA. In the same study, inspection of mammalian, yeast and Tetrahymena box H/ACA snoRNAs revealed a general model by which box H/ACA snoRNAs can specifically select ribosomal pseudouridylation sites. According to this model, the pseudouridylation guide snoRNAs, similarly to the 2′-O-ribose methylation guide snoRNAs, function through direct Watson–Crick interactions with pre-rRNA sequences. Single-stranded sequences in an internal loop structure of the 5′- or 3′-terminal hairpin of box H/ACA snoRNAs select the rRNA sequences that precede and follow the target uridine residue (Ganot et al., 1997a). Yeast strains depleted of Gar1p impair both processing of 18S rRNA (Figure 2B; Girard et al., 1993) and formation of ribosomal pseudouridines (Figure 3 and 4C). At the moment, it is unclear whether inhibition of 18S rRNA production in the gar1.1 strain at restrictive temperature is due to the lack of global rRNA pseudouridylation. Pseudouridylation of eukaryotic rRNAs occurs predominantly on the primary rRNA transcripts before nucleolytic processing (Jeanteur et al., 1968; Brand et al., 1979; this study). This raised the intriguing idea that pseudouridine residues in the pre-rRNA may facilitate the processing of rRNAs (reviewed in Maden, 1990; Eichler and Craig, 1995). Our results show that pseudouridylation of pre-rRNA is not required for the efficient and correct processing of 25S rRNA (Figure 2B, lane 1, and Figure 4B, lane 5). However, the importance of ribosomal pseudouridines for 18S processing is less well understood. It is conceivable that depletion of Gar1p may hinder assembly of functional box H/ACA snoRNPs involved in the synthesis of ribosomal pseudouridines that are essential for 18S formation. According to an alternative explanation, some box H/ACA snoRNPs may function directly or indirectly in the nucleolytic processing of 18S rRNA. Supporting this scenario, genetic depletion of yeast snR30, the only known essential box H/ACA snoRNA that lacks apparent pseudouridine recognition motifs (Ganot et al., 1997a), disrupts 18S production (Morrissey and Tollervey, 1993), but does not affect the global pseudouridylation of pre-rRNA (Figure 5). Of course, we cannot rule out the formal possibility that snR30, by using a mechanism distinct from the majority of box H/ACA snoRNAs, functions in the formation of a specific pseudouridine residue that plays a fundamental role in the maturation or accumulation of yeast 18S rRNA. Consistent with the notion that pre-rRNA pseudouridylation happens before nucleolytic processing of pre-rRNA, correct processing of pre-rRNA or formation of a pre-rRNA–multi-snoRNP complex competent in rRNA processing does not seem to be crucial for global pseudouridylation of pre-rRNA (Figure 5). Moreover, depletion of yeast Nop1 snoRNP protein, although it abolishes both the formation of mature 18S rRNA and the 2′-O-methylation of pre-rRNA (Tollervey et al., 1993), does not inhibit the overall pseudouridylation of pre-rRNA (our unpublished data). This further corroborates that site-specific ribose methylation and pseudouridylation of pre-rRNA are independent of each other and rely on different sets of nucleolar factors. Consistent with this, we recently have shown that yeast Nop1p, contrary to previous reports (Schimmang et al., 1989; Balakin et al., 1996), is specifically and exclusively associated with box C/D snoRNAs while the box H/ACA snoRNAs are specifically complexed with the Gar1 snoRNP protein (Balakin et al., 1996; Ganot et al., 1997b). Recently, it was discovered that all known pseudouridine synthases can be grouped into four distinct protein families (Koonin, 1996). Three
Referência(s)