Characterization of Yeast Protein Deg1 as Pseudouridine Synthase (Pus3) Catalyzing the Formation of Ψ38 and Ψ39 in tRNA Anticodon Loop

Abstract

The enzymatic activity of yeast gene product Deg1 was identified using both disrupted yeast strain and cloned recombinant protein expressed in yeast and in Escherichia coli. The results show that the DEG1-disrupted yeast strain lacks synthase activity for the formation of pseudouridines ⌿ 38 and ⌿ 39 in tRNA whereas the other activities, specific for ⌿ formation at positions 13, 27, 28, 32, 34, 35, 36, and 55 in tRNA, remain unaffected. Also, the His 6 -tagged recombinant yeast Deg1p expressed in E. coli as well as a protein fusion with protein A in yeast display the enzymatic activity only toward ⌿ 38 and ⌿ 39 formation in different tRNA substrates. Therefore, Deg1p is the third tRNA:pseudouridine synthase (Pus3p) characterized so far in yeast. Disruption of the DEG1 gene is not lethal but reduces considerably the yeast growth rate, especially at an elevated temperature (37°C). Deg1p localizes both in the nucleus and in the cytoplasm, as shown by immunofluorescence microscopy. Identification of the pseudouridine residues present (or absent) in selected naturally occurring cytoplasmic and mitochondrial tRNAs from DEG1-disrupted strain points out a common origin of ⌿ 38 -and ⌿ 39 -synthesizing activity in both of these two cellular compartments. The sensitivity of Pus3p (Deg1p) activity to overall three-dimensional tRNA architecture and to a few individual mutations in tRNA was also studied. The results indicate the existence of subtle differences in the tRNA recognition by yeast Pus3p and by its homologous tRNA:pseudouridine synthase truA from E. coli (initially called hisT or PSU-I gene product).The modified nucleoside pseudouridine (5-(␤-D-ribofuranosyl)uracil, abbreviated as ⌿), 1 is found very frequently in all kinds of RNA from eubacteria, archaea, and eukaryotes (1). It is present in all transfer RNAs (2), large and small subunits of ribosomal RNA (3, 4), most small nuclear RNAs (5, 6), and selected small nucleolar RNAs (7).The numerous ⌿ residues in RNA are produced by a family of enzymes (pseudouridine synthases, RNA:⌿ synthases). These enzymes act on specific uridine residues of the RNA molecules, but still very little is known about the number of these enzymes in a given cell as well as their mechanism and their RNA recognition mode.In yeast tRNA, formation of pseudouridines at positions 13, 32, and 55 is catalyzed by three distinct enzymes (8), whereas in the same yeast cell one single enzyme (Pus1p) is responsible for ⌿ formation at positions 27, 34, and 36 (9). Several distinct pseudouridine synthase activities acting on eukaryotic small nuclear RNAs were also detected in crude cell extracts (5); however, none of these enzymes has been identified so far. From recent works on rRNA maturation, it appears that most (if not all) of the ⌿ in eukaryotic rRNA is probably synthesized by a single (or at least very few) rRNA:⌿ synthase(s); in this latter case the enzyme(s) is(are) guided to the different target uridines within the rRNA by a huge family of diverse small nucleolar ribonucleoproteins present in ...

Cbf5p, a potential pseudouridine synthase, and Nhp2p, a putative RNA-binding protein, are present together with Gar1p in all H BOX/ACA-motif snoRNPs and constitute a common bipartite structure

“…Highly conserved homologues of Cbf5p share a high level of sequence similarity with known ⌿ synthases+ A: Amino acid alignment of the S. cerevisiae Cbf5p (accession no+ L12351), S. pombe Cbf5p (accession no+ Z97210), rat NAP57 (accession no+ Z34922), C. elegans (accession no+ Z92803), Drosophila (accession no+ AF017230), and Methanococcus (accession no+ U67472) Cbf5p/NAP57 protein homologues+ The sequences of the E. coli (accession no+ P09171) and Haemophilus influenza (accession no+ P45142) TRUB ⌿ synthases are also included as a comparison+ Identical residues are indicated in white on black boxes+ Conserved residues are indicated by grey boxes and are grouped as in Figure 5+ The boxed regions represent the two conserved regions found in the majority of ⌿ synthases+ Amino acid positions are indicated on the left+ Dashed lines indicate spaces inserted to align the amino acid sequences+ Only the homologous sequences have been included in this alignment+ B: Amino acid alignment of the highly charged C-terminal domains of Cbf5p homologues from S. cerevisiae, S. pombe, K. lactis (accession no+ AF008563), Dictyostelium discoidium (accession no+ C23692), rat, human (accession no+ U59151), Drosophila, and C. elegans+ KKE or KKD amino acid repeats are indicated in white on black boxes+ Other charged amino acids are indicated by grey boxes+ region of the snoRNA, and hence resemble the twodomain structure seen using the electron microscope (Fig+ 9)+ The extremely high degree of sequence identity seen between Cbf5p and other known ⌿ synthases suggests that this protein is indeed the enzyme activity responsible for snoRNA-directed modification (Koonin, 1996)+ While preparing this article, a paper describing Cbf5p as an integral H/ACA snoRNP protein involved in snoRNA biogenesis, 18S rRNA processing, and ⌿ synthesis was published (Lafontaine et al+, 1998)+ This suggests that the H/ACA snoRNPs are individual ⌿ synthases with the RNA component functioning in substrate recognition through direct base-pairing interactions+ The ⌿ synthases studied so far do not contain an RNA subunit and recognize the substrate through protein-RNA interactions (Kammen et al+, 1988;Nurse et al+, 1995;Wrzesinski et al+, 1995a,b;Simos et al+, 1996;Lecointe et al+, 1998)+ Hence, the H/ACA snoRNPs represent a novel type of modification enzyme+ In comparing the amino acid sequence of Cbf5p with known ⌿ synthases it is clear that the longer snoRNP protein contains essential highly conserved domains not present in the prokaryotic modification enzymes+ If one considers the fact that Cbf5p is essential for ribosome biogenesis and is proposed to interact with several other nucleolar proteins (Meier & Blobel, 1994;Cadwell et al+, 1997), this suggests that Cbf5p has other roles in the structure and function of this class of snoRNPs in addition to catalyzing ⌿ synthesis+ FIGURE 8. Electron microscopy of yeast snR30 and snR42 snoRNP complexes+ Purified snR30 and snR42 snoRNP complexes were negatively stained with uranyl formate+ Three common classes of (A) snR30 snoRNP and (B) snR42 snoRNP images+ For each row five examples of each class are shown of both the snR30 and snR42 complexes+ The corresponding rows in A and B show the respective particle in a similar projection form+ The left and mid...…”

“…The rRNA ⌿ synthase has yet to be identified; however, it is entirely possible that the protein responsible for the enzymatic activity is an integral component of the H/ACA snoRNPs+ The prokaryotic tRNA and rRNA ⌿ synthases do not use a guide RNA to recognize specific RNA substrates+ Multiple tRNA ⌿ synthases exist in prokaryotes that recognize either a single site or several sites with a similar structure (Kammen et al+, 1988;Nurse et al+, 1995;Wrzesinski et al+, 1995a,b;Simos et al+, 1996;Lecointe at al+, 1998)+ Therefore, the use of a guide RNA to direct rRNA ⌿ synthesis in eukaryotes is quite unique (Ganot et al+, 1997a;Ni et al+, 1997)+ It has recently been demonstrated that Gar1p, a protein common to all H/ACA snoRNPs, is required for ⌿ formation in yeast (Bousquet-Antonelli et al+, 1997)+ It is, however, highly unlikely that this protein is the modification enzyme, because Gar1p has little homology to known ⌿ synthases and is therefore postulated to function as an accessory protein in this event (Bousquet-Antonelli et al+, 1997)+ One possible candidate for the ⌿ synthase is Cbf5 (mammalian homologue NAP57; Meier & Blobel+, 1994), a nucleolar protein involved in ribosome biogenesis that shows a remarkably high degree of sequence similarity with known ⌿ synthases (Koonin, 1996;Cadwell et al+, 1997)+ Indeed, while we were preparing this article, a paper describing Cbf5p as an integral H/ACA snoRNP protein involved in snoRNA biogenesis, 18S rRNA processing, and ⌿ synthesis was published (Lafontaine et al+, 1998)…”

The yeast Saccharomyces cerevisiae YDL112w ORF encodes the putative 2′-O-ribose methyltransferase catalyzing the formation of Gm18 in tRNAs

“…The presence of an ;1,280-amino-acid-long N-terminal extension in Trm3p, as compared with spoU and other prokaryotic homologs, is intriguing+ In contrast, the four pseudouridine synthases that modify yeast tRNAs are only moderately longer than their E. coli homologs, truA and truB (Simos et al+, 1996;Becker et al+, 1997a;Lecointe et al+, 1998)+ Given that spoU and Trm3p appear to recognize the same three-dimensional architecture of mature tRNAs in E. coli or in yeast, the long N-terminal domain of Trm3p seems unlikely to be directly involved in the methylase catalytic activity or in substrate recognition+ The intracellular location of Trm3p remains unknown+ However, it is noteworthy that Pus1p, a tRNA pseudouridine synthase with an introndependent activity, is a nuclear protein interacting genetically with nuclear pore protein Nsp1p (Simos et al+, 1996)+ Likewise, splicing of tRNA may be coupled to tRNA export through the nuclear pore complex, given that the tRNA endonuclease appears to be an integral nuclear membrane protein (Peebles et al+, 1983) and that the tRNA ligase is localized close to nuclear pores (Clark & Abelson, 1987)+ Moreover, Trm1p that catalyzes formation of tRNA N2, N2-dimethylguanosine localizes at or near the nuclear membrane (Rose et al+, 1992(Rose et al+, , 1995+ Since formation of Gm18, like a few other tRNA modifications, occurs only after intron excision, an association of Trm3p with the transport machinery and/or the nuclear pore complex through its N-terminal domain may be envisioned+ Alternatively, Trm3p could reside in the cytoplasm and its long extension at the N-terminus could be required for its assembly into a macromolecular complex in much the same way as the aminoacyl-tRNA synthetase complexes in higher eukaryotes (Mirande, 1991;Kisselev & Wolfson, 1994)+…”

“…Formation of a given type of nucleotide modification at multiple sites of an RNA molecule frequently involves several distinct modifying activities, but some enzymes can modify different positions within an RNA substrate (Simos et al+, 1996;Lecointe et al+, 1998), and a dual substrate specificity can even be envisioned (Wrzesinski et al+, 1995)+ Analysis of TRM3-disrupted yeast strain tRNAs and in vitro modification experiments of tRNA transcripts indicate that Trm3p has a very narrow substrate and site specificity, corresponding to the exclusive formation of Gm18 in all yeast cytoplasmic tRNAs naturally bearing this methylation+ Several instances of a single nuclear gene encoding an enzyme modifying both cytoplasmic and mitochondrial tRNAs have been reported (Hopper et al+, 1982;Becker et al+, 1997a;Lecointe et al+, 1998), and the possibility that Trm3p is also responsible for Gm18 formation in mitochondrial tRNAs cannot be ruled out+ In E. coli RNA, a total of seven 29-O-ribose-methylated nucleotides have been identified, including two other tRNA ribose methylations in addition to Gm18, Um32, and Cm32, the formation of which might be catalyzed by the same enzyme different from spoU+ As for the four rRNA ribose methylations, one of which is on a guanosine (Smith et al+, 1992), they do not depend upon spoU but are thought to each involve a distinct methylase (Gustafsson et al+, 1996)+ Accordingly, we have detected 6 ORFs encoding putative RNA ribose methylases in the E. coli genome (Table 1)+ While a pseudouridine synthase capable of forming c32 in tRNA Phe in vitro also catalyzes specific formation of c746 in E. coli 23S rRNA (Wrzesinski et al+, 1995), our data do not provide evidence for a dual substrate specificity of Trm3p+ Since sites of cytoplasmic rRNA ribose methylations are exclusively specified by the RNA duplex structure involving a cognate small nucleolar RNA guide (Cavaillé et al+, 1996;Kiss-Laszlo et al+, 1996;Nicoloso et al+, 1996;Bachellerie & Cavaillé, 1997) the involvement of only a few, if not a single rRNA ribose methylase, devoid of any site specificity and obviously distinct from Trm3p, seems likely+ As for snRNAs, which contain phylogenetically conserved ribose-methylated nucleotides (including several guanosines) in vertebrates, their ribose methylation patterns in S. cerevisiae are not known (Massenet et al+, 1998)+ However, given that Trm3p activity is strongly dependent in vitro on the three-dimensional conformation of mature tRNA, as discussed below, its involvement in snRNA ribose methylation seems unlikely+ Finally, yeast mRNAs are devoid of ribose methylations adjacent to the 59 cap, in contrast to vertebrate mRNAs (Narayan & Rottman, 1992)+…”

“…Most cellular RNAs undergo a number of nucleoside modifications, the biological role of which is still often elusive (Agris, 1996)+ However, their importance has been clearly established in a few cases, essentially for tRNAs that contain the greatest variety and abundance of modified nucleosides (Björk, 1995), several of which appear to be involved in the optimization of tRNA function, not only in translation but also in other processes involving tRNAs, ranging from cell-cycle control to priming of retroviral replication (Persson, 1993;Björk, 1995;Fossé et al+, 1998)+ The two most prevalent types of nucleoside modifications in prokaryotic and eukaryotic RNAs are pseudouridylations and 29-O-ribose methylations, the precise role(s) of which remain(s) generally unknown, although they can both fine-tune the threedimensional folding of RNA and its interactions with ligands + Only a few RNA-modifying enzymes have been characterized as yet, particularly in eukaryotic cells+ Formation of the numerous pseudouridines in yeast tRNAs is catalyzed by several tRNA:pseudouridine synthases, four of which have been recently identified (Simos et al+, 1996;Becker et al+, 1997a;Lecointe et al+, 1998)+ Paradoxically, formation of the much larger number of pseudouridines present in yeast cytoplasmic rRNAs seems to involve a single pseudouridine synthase, Cbf5p (Lafontaine et al+, 1998), in line with the finding that each site of pseudouridine formation is selected by base-pairing of a specific small nucleolar RNA guide belonging to the box H-ACA family (Ganot et al+, 1997;Ni et al+, 1997;Smith & Steitz, 1997)+ As for the scores of cytoplasmic rRNA ribose methylations, also sitespecified by base-pairing of small nucleolar RNAs, belonging in this case to the box C/D family (Kiss-Laszlo et al+, 1996;Nicoloso et al+, 1996;Bachellerie & Cavaillé, 1997), the enzyme(s) catalyzing their formation remain(s) unknown+ Ribose methylations of eukaryotic RNAs also include five different positions in tRNAs (Sprintz et al+, 1998), nearly two dozen sites in vertebrate snRNAs (Massenet et al+, 1998), and the two 59cap-proximal nucleotides of mRNAs in higher eukaryotes (Narayan & Rottman, 1992)+ Methylases form a large family of enzymes that catalyze methyl transfer from the ubiquitous S-adenosyl-L-methionine (AdoMet) to either nitrogen, oxygen, or carbon atoms in DNA, RNA, proteins, and small molecules+ Although they use a common cofactor, methylases acting on different types of biological macromolecules exhibit very extensive sequence divergence+ Even among the DNA methyltransferases, which belong to different classes defined by the position methylated, primary structures are quite different+ Only the structural analysis revealed striking similarities pointi...…”