A functional homolog of a yeast tRNA splicing enzyme is conserved in higher eukaryotes and in Escherichia coli

Abstract

tRNA splicing in the yeast Saccharomyces cerevisiae requires an endonuclease to excise the intron, tRNA ligase to join the tRNA half-molecules, and 2'-phosphotransferase to transfer the splice junction 2'-phosphate from ligated tRNA to NAD, producing ADP ribose 1"-2" cyclic phosphate (Appr>p). We show here that functional 2'-phosphotransferases are found throughout eukaryotes, occurring in two widely divergent yeasts (Candida albicans and Schizosaccharomyces pombe), a plant (Arabidopsis thaliana), and mammals (Mus musculus); this finding is consistent with a role for the enzyme, acting in concert with ligase, to splice tRNA or other RNA molecules. Surprisingly, functional 2'-phosphotransferase is found also in the bacterium Escherichia coli, which does not have any known introns of this class, and does not appear to have a ligase that generates junctions with a 2'-phosphate. Analysis of the database shows that likely members of the 2'-phosphotransferase family are found also in one other bacterium (Pseudomonas aeruginosa) and two archaeal species (Archaeoglobus fulgidus and Pyrococcus horikoshii). Phylogenetic analysis reveals no evidence for recent horizontal transfer of the 2'-phosphotransferase into Eubacteria, suggesting that the 2'-phosphotransferase has been present there since close to the time that the three kingdoms diverged. Although 2'-phosphotransferase is not present in all Eubacteria, and a gene disruption experiment demonstrates that the protein is not essential in E. coli, the continued presence of 2'-phosphotransferase in Eubacteria over large evolutionary times argues for an important role for the protein.

Figure 1

The E. coli and mouse Tpt1 homologs have phosphotransferase activity. Partially purified phosphotransferase from yeast (Tpt1; A), E. coli (KptA; B), and mouse (mTpt1; C) were incubated with ligated tRNA^PHE substrate and 1 mM NAD, as indicated. Portions of the mixtures were subsequently incubated with calf intestinal phosphatase (CIP), or yeast cyclic phosphodiesterase (CPDase), as shown, and samples were applied to TLC plates to resolve products.

Figure 2

E. coli KptA protein transfers the 2′-phosphate from the substrate ligated tRNA. Different ³²P-labeled derivatives of the substrate were tested as the source of the transferred phosphate, with 1 mM NAD and 5-fold serial dilutions of KptA protein, starting from 125 units of activity. After incubation, 0.1 unit of calf intestinal phosphatase was added to each tube, followed by a further 15-min incubation, to convert contaminating nuclease products (but not Appr>p) to Pi, and samples were applied to TLC plates. Lanes d–h, ligated tRNA^PHE with a labeled 2′-phosphate (the standard substrate); i–k, labeled pre-tRNA as substrate; and l–n, mature tRNA (previously dephosphorylated with yeast Tpt1 protein). Lanes a–c, controls with ligated tRNA substrate and NAD: a, buffer control, followed by CIP buffer; b, Tpt1p, followed by CIP buffer; c, Tpt1p, followed by CIP.

Figure 3

Alignment of the phosphotransferase homologs. Phosphotransferases identified by biochemical and database searches are presented. The alignment is shaded to a 50% consensus by using macboxshade, with dark and light shading, indicating identical and similar residues, respectively. The human sequence is only a partial EST, and the beginning of the mouse sequence is absent from its EST. Numbers indicate the residue number for each homolog. E. coli, sp|P39380|YJII (ORF218); P. aeruginosa, gnl|PAGP|Contig581; S. pombe, EMB|Z99259|SPAC2C4; A. fulgidus #1,gb|AE001076; A. fulgidus #2, gb|AE000995; P. horikoshii, dbj|AB009470; A. thaliana, gb|AC002387; M. musculus, AA245980; A rice EST is not shown.

Figure 4

Phylogeny of the phosphotransferase homologs. An unrooted phylogram, based on the Neighbor-Joining method, of the various phosphotransferases is presented, using the alignment in Fig. 3. Numbers next to each branch represent bootstrap values as a percentage of 1,000 trials. The branch lengths represent actual amino acid divergences based on the scale shown (0.05 denoting 5% amino acid divergence).