Identification of a novel gene encoding a flavin-dependent tRNA:m5U me
http://www.100md.com
《核酸研究医学期刊》
CNRS, Enzymology and Structural Biochemistry Laboratory 1 avenue de la Terrasse, F-91198, Gif-sur-Yvette, France 1INSERM Avenir group, Institute of Genetics and Microbiology, CNRS, University Paris XI Orsay, F-91405, France
*To whom correspondence should be addressed. Tel: +33 1 69823468; Fax: +33 1 69823129; Email: henri.grosjean@lebs.cnrs-gif.fr
ABSTRACT
Formation of 5-methyluridine (ribothymidine) at position 54 of the T-psi loop of tRNA is catalyzed by site-specific tRNA methyltransferases (tRNA:m5U-54 MTase). In all Eukarya and many Gram-negative Bacteria, the methyl donor for this reaction is S-adenosyl-L-methionine (S-AdoMet), while in several Gram-positive Bacteria, the source of carbon is N5, N10-methylenetetrahydrofolate (CH2H4folate). We have identified the gene for Bacillus subtilis tRNA:m5U-54 MTase. The encoded recombinant protein contains tightly bound flavin and is active in Escherichia coli mutant lacking m5U-54 in tRNAs and in vitro using T7 tRNA transcript as substrate. This gene is currently annotated gid in Genome Data Banks and it is here renamed trmFO. TrmFO (Gid) orthologs have also been identified in many other bacterial genomes and comparison of their amino acid sequences reveals that they are phylogenetically distinct from either ThyA or ThyX class of thymidylate synthases, which catalyze folate-dependent formation of deoxyribothymine monophosphate, the universal DNA precursor.
INTRODUCTION
Transfer RNAs in all living organisms contain a number of nucleosides that are post-transcriptionally modified on the base and/or the 2'-hydroxyl group of the ribose (1). One such common modified nucleoside is 5-methyluridine (m5U, also designated T for ribothymidine). This C5-methylated uridine is invariably found at position 54, in the so-called T-psi loop of tRNA of almost all Bacteria and Eukarya (2). In thermophilic Bacteria, such as Thermus thermophilus, it is further hypermodified to a 2-thio-derivative , while in certain Eukarya, a 2'-O-methyl-derivative is occasionally found .
Site-specific methylation of U-54 in Escherichia coli tRNA is catalyzed by tRNA:m5U-54 methyltransferase (EC.2.1.1.35). This enzyme, initially designated RUMT for RNA uridine methyltransferase, was the first RNA modification enzyme discovered that acts at the polynucleotide level (4,5). This enzyme is also called TrmA (tRNA methyltransferase A), and a gene trmA encoding this enzyme was first identified in E.coli (6,7). From the standpoint of mechanism and specificity, the tRNA:m5U-54 methyltransferase of E.coli is one of the best characterized RNA modification enzymes . In the majority of RNA methyltransferases studied so far , RUMT uses S-adenosylmethionine (S-AdoMet) as the methyl donor. Automated bioinformatic approaches have included all trmA and TRM2 homologs in the same cluster of orthologous genes . This cluster contains a superfamily of S-AdoMet-dependent RNA:m5U MTases that are specific not only for uridine at position 54 of tRNA, but also paralogs that function in uridine methylation in other RNAs .
Earlier studies have indicated that not all bacterial tRNA:m5U-54 MTases use S-AdoMet as methyl donor. For example, in Enterococcus faecalis (formerly Streptococcus faecalis) and Bacillus subtilis, it was reported that the carbon donor of the methyl group is N5, N10-methylenetetrahydrofolate (CH2H4folate) . The first indication for this came from an observation that bulk tRNAs isolated from folate-deprived E.faecalis cells lacked m5U-54 in their T-psi loop (15). Moreover, in B.subtilis and Micrococcus lysodeikticus, trimethoprim, a specific inhibitor of bacterial dihydrofolate reductase, inhibits formation of m5U-54 in vivo (16), indicating that in these Gram-positive bacteria, the carbon source used in tRNA methylation derives from the folate pool. The results of these in vivo studies were later confirmed by demonstrating that in vitro activity of purified tRNA: m5U-54 MTase of E.faecalis not only requires CH2H4folate but also reduced flavin adenine nucleotide (FADH2) (14,17,18), thus forming a distinct class of tRNA:m5U-54 MTases (EC.2.1.1.74). Strikingly, this observation is reminiscent of the enzymatic mechanism that has been described for the alternative flavin-dependent ThyX class of thymidylate synthases (EC.2.1.1.148) (19–21), but differs from the reaction catalyzed by a canonical thymidylate synthase ThyA, which uses CH2H4folate both as a carbon source and as a reductant . Moreover, in the case of ThyX catalysis, it has been recently demonstrated that a hydride from NAD(P)H is transferred, via a FAD cofactor to reduce the methylene group, to a methyl residue (23–26). The gene encoding the folate-dependent tRNA:m5U-54 MTase has not yet been identified. It is, therefore, not known whether these analogous folate-dependent methylation reactions, involved in RNA or DNA metabolism, are catalyzed by distantly related enzymes, possibly originating from the early RNA World, or, on the contrary, represent independent catalytic mechanisms.
Benefiting from large-scale microbial sequencing and structural genomics projects, we predicted that bacterial Gid proteins would correspond to a novel class of bacterial site-specific tRNA:m5U-54 MTases. This prediction was confirmed through genetic studies and biochemical analyses of tRNA molecules isolated from B.subtilis wild-type and mutants strains. In vitro characterization of the purified recombinant B.subtilis tRNA:m5U-54 MTase indicates that this protein alone is sufficient for tRNA methylation reaction. Our studies further indicate that despite the fact that thymidylate synthase ThyX and Gid proteins catalyze a similar methylation reaction, they lack detectable sequence, and probably structural similarity. Our analyses suggest that the enzymes methylating nucleotides in tRNA and DNA precursor using CH2H4folate and NAD(P)H/FAD as carbon donor and reductant, respectively, have independent evolutionary origins.
MATERIALS AND METHODS
Strains
B.subtilis strain BFS2838 carrying inactivated gidermR gene was kindly provided by S. Seror . E.coli strain GRB113 (metA, trmA5, zij-90::Tn10), encoding an inactivated TrmA protein was a kind gift from G. R. Bj?rk, Ume? University, Sweden. E.coli Sure? strain (e14–(McrA–) mcrCB-hsdSMR-mrr)171 endA1 supE44 thi-1 gyrA96 relA1 lac recB recJ sbcC umuC::Tn5 (Kanr) uvrC was purchased from Stratagene.
Construction of an N-terminal His6-tagged BsuGidA and BsuGid-overexpressing plasmids and purification of the corresponding recombinant proteins
The gidA (GIDA_BACSUB; P25812 ) and gid genes (GID_BACSU; P39815 ; renamed trmFO in this work) were amplified by PCR from B.subtilis strain 168 DNA, using Pfu DNA polymerase (Promega) and the following primers (sequence in small characters correspond to genome sequence): gidAfw (CGGGATCCatggggtatgaagcaggccaatac) and gidArev (TCCCCCGGGctactcggctatcttcgcaatgcg) or gidfw (CGGGATCCatgaaccaacaaacagtgaatgta) and gidrev (TCCCCCGGGctatattgttttcgaaattgtttg). The resulting 1893 or 1314 bp PCR products were then digested with BamHI and SmaI, respectively, and cloned into pQE80L to generate pQE80L-BsuGidA or pQE80L-BsuGid. To purify recombinant GidA and Gid proteins, pQE80L-BsuGidA was transformed into E.coli Sure? strain, and pQE80L-BsuGid was transformed into Sure? or GRB113 (trmA5) strain. Resulting strains were grown at 37°C in 500 ml of Luria–Bertani (LB) medium (Invitrogen) containing 100 μl/ml ampicillin until OD600 = 0.6. After induction of Gid or GidA protein expression by isopropyl ?-D-thiogalactopyranoside (IPTG) (VWR International, final concentration = 1 mM), the cultures were further grown at 37°C for 3 h. After harvesting the cells by centrifugation, the pellet were flash-frozen in liquid N2 and stored at –80°C. Frozen cells were thawn on ice and resuspended in 5 ml of lysis buffer (50 mM sodium phosphate, pH 7.6, 300 mM NaCl, 10% glycerol and 20 mM imidazole) containing 5 μl Protein Inhibitor Cocktail (PIC, Sigma) and 1.5 μl ?-mercaptoethanol. Cells were broken by 2 freeze (liquid N2)/thaw (37°C) cycles and ultrasonication. The lysate was centrifuged for 15 min at 10 000 g at 4°C. Supernatant was loaded onto 2 ml of Ni-NTA resin and washed with 25 ml of lysis buffer. Gid or GidA proteins were eluted with 10 ml elution buffer (same as lysis buffer, but containing 250 mM imidazole). Yellow fractions, containing the Gid or GidA protein, were pooled (to 3 ml of total volume) and dialyzed against 500 ml of 30 mM HEPES buffer, pH 7.5, containing 200 mM NaCl and 10% glycerol. Protein was aliquoted, flash-frozen in N2 and stored at –80°C. To measure any cofactor release from Gid, 5 μg of protein was diluted in 100 μl of distilled water and incubated for 5 min at 90°C. The sample was centrifuged at 10 000 g for 15 min. Absorption and fluorescence spectra of the obtained supernatant were measured.
Preparation of cell-free extracts
Cell-free extracts of B.subtilis strains 168 (wild-type) and BFS2838 (gidermR) were prepared from an exponentially growing cell culture at 37°C. After centrifugation and washing the cell pellet with lysis buffer (25 mM Tris–HCl buffer, pH 7.5, 10 mM MgCl2, 25 mM KCl and 2 mM DTT), it was resuspended in a 1.5 vol of lysis buffer containing 1% (v/v) of PIC, Sigma. An S10 cell-free extract was obtained after ultrasonication and centrifugation for 15 min at 10 000 g. Further centrifugation of supernatant for 1 h at 4°C resulted in S100 cell-free extracts. Cell-free extracts of E.coli strains pQE80L-BsuGid/GRB113 (trmA5) and pQE80L/GRB113 (trmA5) were prepared similarly as for the protein purification, except that they were grown at 37°C to an OD600 of 0.8 in 10 ml of liquid Luria Broth with 100 μl/ml carbenicillin. After the Gid protein induction, harvesting of cells, resuspension in 500 μl lysis buffer, cell disruption by ultrasonication and centrifugation, the S10 cell-free extract was produced.
Enzymatic activity assays
UTP-labeled yeast tRNAAsp transcript, used for determining the tRNA:m5U-54 MTase activity of TrmFO (Gid) protein, was prepared and purified as described elsewhere (27,28). A total of 50–100 fmol of -labeled tRNAAsp were incubated at 37°C in a 50 μl reaction mixture containing 40 mM N- piperazine-N--Na buffer (HEPES-Na, Sigma) at pH 7.0, 0.25 mM FAD, Fluka, 0.5 mM NADH (reduced nicotinamide adenine dinucleotide, Sigma), 1 mM reduced nicotinamide adenine dinucleotide phosphate (NADPH), 0.25 mM (6R)-N5,N10-CH2H4PteGlu-Na2 (methylenetetrahydrofolate, provided by Dr R. Moser, Merck-Eprova, AG, Switzerland), 5 mM DTT (Promega), 15 U of RNase inhibitor (Fermentas) and 10–25 μg of total protein of a B.subtilis or E.coli cell-free extract. At the end of the incubation period, modified tRNA was extracted and digested with nuclease P1 (Roche), the modified nucleotides were separated by 2D thin-layer chromatography (2D-TLC) and data were analyzed as described previously (29). Methylating activity of purified recombinant BsuGid (TrmFO) (1 μg per test) and BsuGidA protein (1 μg per test) were tested using the same experimental conditions as above. Activity of the MnmC enzyme on bulk tRNAs from B.subtilis strains 168 or BFS2838 (gidermR) was tested as follows: five microgram of purified recombinant MnmC protein (provided by Dr L. Droogmans, University of Brussels, Belgium) was added to 300 μl of a reaction mixture containing 50 mM Tris–HCl, pH 8.0, 20 mM NH4Cl, 62 μM -AdoMet (53 Ci/mol, Amersham) and 100 μg of bulk B.subtilis tRNAs. After 1 h incubation at 37°C, tRNA was recovered, digested by nuclease P1, and the resulting radiolabeled nucleotides were analyzed by 2D-TLC as described previously (30).
Isolation of tRNA and chromatographic analysis of tRNA hydrolysates
Bulk tRNAs of B.subtilis strains 168 and BFS2838 (gidermR) were obtained essentially as described previously (31), except that the tRNA deacylation step was omitted and a monoQ column (5 ml from Biorad) was used instead of DEAE-cellulose. For bulk tRNAs from E.coli GRB113 (trmA5), transformed by pQE80L-BsuGid or pQE80L (control experiment), cell cultures were first grown at 37°C in 200 ml of liquid Luria Broth in the presence of 100 μl/ml carbenicillin. At OD600 = 0.6, IPTG was added to the final concentration of 1 mM, and the cells were grown for additional 3 h at 37°C before to be collected in the cold by centrifugation and purified as above. Obtained purified bulk tRNAs were completely degraded to nucleosides with P1 nuclease and alkaline phosphatase (Sigma) and the resulting hydrolysates analyzed by high performance liquid chromatography (HPLC) on a Supelcosil LC18 column (Supelco) with Waters HPLC instrument, as described previously (32).
RESULTS
Comparative genomics identifies a candidate gene encoding a new family of flavin-dependent methyltransferases
An enzyme of two identical subunits of 58 kDa that catalyzes the site-specific formation of 5-methyluridine in position 54 (m5U-54) of tRNA using CH2H4folate as a source of one-carbon unit and a combination of coenzymes NAD(P)H/FAD as reductant, has been purified from E.faecalis (33). We attempted an identification of the gene encoding this activity (described under EC 2.1.1.74 ) based on the facts that a folate-dependent pathway for tRNA methylation exists in some Gram-positive Bacteria species , whereas an S-AdoMet-dependent enzyme is used instead in Eukarya , in gamma-proteobacteria and in a few Archaea .
Our primary searches used an updated version of COG database (http://www.ncbi.nlm.nih.gov/COG/), which currently consists of 4873 gene families (11). Using the phylogenetic distribution analysis tool of this database, we obtained a list of 155 COGs (3% of total number of families) that are present in B.subtilis and Bacillus halodurans, but absent in Archaea, Eukarya and gamma-proteobacteria (data not shown). We have no data for E.faecalis, M.lysodeikticus and G.stearothermophilus, as they are not included in the current data release. Next, among the 155 candidates, we searched for the presence of a characteristic ‘GXGXXG’ motif that is part of the conserved Rossman-fold found in a large number of FAD binding proteins . As a result, one COG family (COG1206) emerged as an evident protein family encoding a putative tRNA:m5U-54 MTase. These COG1206 proteins, also designated as Gid proteins (in reference to the B.subtilis protein) are: (i) currently annotated as ‘NADPH(FAD)-utilizing enzymes possibly involved in translation’, (ii) contain a readily identifiable FAD binding motif ‘GXGXXG’ (in fact G-X-G-L-A-G--E-X-A, see details below) and (iii) their molecular weight is 50 kDa, in close agreement with the 58 kDa determined on SDS–PAGE gels for the -subunit of the E.faecalis folate-dependent tRNA:m5U-54 MTase (33).
COG1206 proteins have a wide phylogenetic distribution
Systematic screening of >200 fully sequenced genomes (http://www.ncbi.nih.gov/genomes/lproks.cgi), using pattern-hit initiated BLAST algorithm (39) and B.subtilis Gid (GID_BACSU; P39815 ) as a query, identified 80 bacterial species containing a gid gene, whereas no hits were found in archaeal nor eukaryal genomes. This phylogenetic distribution of gid genes (Figure 1) is much wider than initially anticipated. In addition to the expected Gram-positive bacteria (Bacillales and Lactobacillales), a gene for Gid-like protein is also found in alpha-proteobacteria, delta-proteobacteria and cyanobacteria. Phylogenetic analyses of a subset of Gid orthologs, using neighbor-joining trees performed with ClustalX program (40), indicate that their phylogeny is congruent with species phylogeny, suggesting a relatively ancient bacterial origin for Gid proteins (Figure 1).
Figure 1 A phylogenetic tree based on a subset of Gid homologs retrieved by pattern-hit initiated BLAST algorithm. Clustal X was used for sequence alignments and phylogenetic trees were constructed using the neighbor-joining methods. GidA sequences from B.subtilis, T.thermophilus and Deinococcus radiodurans were used as an outgroup. Branch points with closed circles indicate a bootstrap support >90%. The shown topology was also supported by quartet puzzling with maximum likelihood analysis performed using Tree-Puzzle 5.1 program implemented at www.pasteur.fr (data not shown).
The Gid protein of B.subtilis is involved in m5U-54 formation in tRNA
To determine whether Gid proteins are involved in the biosynthesis of m5U-54, the presence of this methylated nucleoside was analyzed in B.subtilis tRNA isolated from a BFS2838 (gidermR) strain, lacking functional Gid protein (kindly provided by S. Seror, University of Paris XI). No obvious phenotype has been described for this B.subtilis strain (see http://locus.jouy.inra.fr/cgi-bin/dev/chiapell/strain_pheno_old.pl?STRAIN=BFS2838). Bulk tRNAs from the mutant strain and the corresponding wild-type strain B.subtilis 168 were extracted, and their nucleoside contents were analyzed by HPLC as described in Materials and Methods. Results in Figure 2A and B clearly demonstrate that m5U is absent in the tRNA from the gidermR mutant, whereas tRNA of the wild-type strain 168 contains the m5U modification. The small residual peak eluting at the same position as m5U in Figure 2B was identified as inosine through its characteristic UV absorbance spectrum. The maximum wavelength for inosine is at 250 nm (Figure 2D), compared with 267 nm for 5-methyluridine (Figure 2C).
Figure 2 B.subtilis BFS2838 (gidermR) strain lacks m5U modification in the tRNA. Bulk tRNAs from wild-type B.subtilis 168 (A) or the gidermR mutant BFS2838 (B) were isolated, completely digested to nucleosides by nuclease P1 and alkaline phosphatase and analyzed by HPLC (see Materials and Methods). Alternatively, T7 polymerase transcripts of yeast wild-type tRNAAsp, uniformly labeled with UTP were incubated for 1 h at 37°C with an S100 cell extract from 168 or BFS2838 . After incubation, bulk tRNAs were completely digested to monophosphate nucleosides by nuclease P1 and analyzed by 2D-TLC. Radiolabeled compounds were detected and quantified using PhosphoImager detector. (C) The spectrum analysis of the HPLC fraction corresponding to m5U (A). (D) The spectrum of the small peak detected in (B), corresponding to inosine. Nature of the modified nucleosides in chromatography peaks was determined by comparison with the standards (55).
The absence of C5-methylation activity for U-54 in gidermR mutant strain was further confirmed by testing the corresponding methylation activity in cell extracts. Thus, -radiolabeled T7-transcript of a synthetic yeast tRNAAsp gene was used as substrate, and incubations were performed in the presence of CH2H4folate, NADH/NADPH and FAD as indicated in Materials and Methods. After incubation, the tRNA was digested into 5'-monophosphate nucleosides and the hydrolysate was analyzed by 2D-TLC. The radiolabeled spots, corresponding to -labeled UMP-derivatives were detected by autoradiography. As shown in Figure 2A and B (insets), while the wild-type cell extract was able to catalyze the formation of m5U-54 in vitro, the extract from the gidermR mutant strain did not catalyze such a methylation reaction, thus indicating that the gid gene product is involved in the production of m5U-54 in tRNA.
The Gid protein of B.subtilis is sufficient for methylation of U-54 in E.coli tRNA in vivo
To investigate whether the B.subtilis Gid protein alone can substitute for the S-AdoMet-dependent E.coli TrmA protein for the formation of m5U-54 in vivo, we cloned the B.subtilis gid into an E.coli expression vector pQE80L, under the control of an IPTG inducible promoter. The resulting plasmid, pQE80L-BsuGid was transformed into an E.coli strain GRB113, carrying trmA5 mutation. This E.coli strain grows normally in LB medium but completely lacks S-AdoMet-dependent tRNA:m5U-54 MTase activity (41). After gid expression for 3 h, the cells were collected by centrifugation and divided into two parts. Bulk tRNA was purified from one part, while the remaining cell pellet was used to prepare an S10 cell extract (see Materials and Methods). HPLC analysis of P1/alkaline phosphatase-treated bulk tRNA hydrolysate demonstrated the presence of m5U nucleoside in the E.coli trmA5 strain transformed by pQE80L-BsuGid (Figure 3A), while in the control E.coli mutant strain, transformed by unmodified expression vector, no m5U was detectable (Figure 3B). As described above with B.subtilis bulk tRNAs (Figure 2), we verified that the very small peak migrating at the position expected for m5U in the HPLC analysis of tRNA hydrolysate of the control strain corresponds to inosine (see UV-spectrum in Figure 3D, compare with Figure 3C for m5U). In parallel, the S10 cell extract was incubated together with appropriate cofactors and labeled yeast tRNAAsp transcript. The P1-hydrolyzate of the resulting modified tRNA was then analyzed by 2D-TLC. Results in Figure 3A and B (insets) indicate that m5U-54 in tRNAAsp is formed only when a cell-free extracts from the E.coli trmA5 strain transformed with pQE80L-BsuGid, confirming that the B.subtilis Gid protein efficiently modified tRNAs under the physiological conditions of E.coli cells.
Figure 3 Recombinant BsuGid protein catalyzes the formation of the m5U-54 modification in tRNA in vivo. HPLC analysis performed with bulk tRNAs purified from E.coli GRB113 (trmA5) transformed with pQE80L-BsuGid (A), or with the ‘empty plasmid’ pQE80L (B). Insets in (A) and (B) correspond to autoradiograms of tRNA hydrolysates resulting from in vitro enzymatic activity tests performed with cell-free extracts of E.coli GRB113 (trmA5) strain transformed with pQE80L-Bsu Gid or with pQE80L . As in Figure 2, (C) shows the spectrum analysis of the HPLC fraction corresponding to m5U (A). (D) The spectrum of the small peak detected in (B), which corresponds to inosine.
Purified recombinant BsuGid protein catalyzes the in vitro formation of m5U-54 in tRNA
The B.subtilis Gid protein was tagged with six histidine residues at the N-terminus, and purified to near homogeneity through affinity chromatography, either from E.coli Sure? strain (Figure 4A) or from E.coli trmA5 strain, both transformed with pQE80L-BsuGid. The purified protein is yellow and elutes from an S-200 gel filtration column at 85 kDa (data not shown), suggesting that the functional form of the enzyme may be a homodimer. Heating of the protein at 90°C releases yellow cofactor that has absorption (data not shown) and fluorescence spectra (Figure 4B) characteristic for oxidized flavins. This cofactor is likely FAD that was present in 0.8 mol per 1 mol of Gid from Myxococcus xanthus . Qualitative experiments indicated that the purified protein catalyzes the site-specific formation of m5U-54 in -labeled yeast tRNAAsp transcript (Figure 4C). Specific activity of the purified recombinant protein is low; nevertheless, this data reinforce observations obtained above by means of genetics. Some activity was observed without the addition of a carbon donor in the reaction mixture, suggesting that either a small amount of a carbon donor co-purifies with the enzyme or, alternatively, the purified enzyme contain tightly bound methylene or methyl intermediates. It is of note that the enzymatic test described here is highly sensitive, detecting even femtomolar amount of methylated uridine in the -radiolabeled substrate and has not been systematically optimized during this work.
Figure 4 (A) Electrophoretic analysis of purified recombinant BsuGid protein. An SDS–PAGE analysis was performed using 11% gels, stained with Coomassie blue. Lane 1, soluble proteins from sonicated total cell-free extract of E.coli pQE80L-BsuGid/SURE; lane 2, S10 fraction from E.coli pQE80L-BsuGid/SURE; lane 3, molecular weight markers; lane 4, proteins eluted from the immobilized metal ion adsorption chromatographic column. (B) Fluorescence spectrum of a cofactor released by heat denaturation from purified BsuGid protein. The observed emission maximum (after excitation at 450 nm) at 520 nm is typical for flavin nucleotides. (C) Time course of m5U-54 formation catalyzed by BsuGid. The molar ratio of m5U over total U in yeast tRNAAsp was evaluated over time at 37°C in the presence (closed circles) or absence (open circles) of CH2H4folate.
Taking together all the above information, we now propose to rename the Gid protein as TrmFO (FO for the folate) and the corresponding gene trmFO, in order to differentiate them from the conventional S-AdoMet-dependent TrmA enzyme and trmA gene.
TrmFO (Gid) and GidA proteins are two evolutionarily related families of proteins with distinct functions
TrmFO proteins of 50 kDa (designated Gid in Genome Data Banks) show 40% sequence similarity with another protein family referred to as GidA proteins (in reference to E.coli protein of 70 kDa) (Figure 5). The readily detectable sequence homology, together with the currently used name ‘small GidA’ for Gid proteins , has created confusion regarding the putative cellular functions of TrmFO. Our studies (see also below) have now revealed that in reality, paralogous TrmFO and GidA proteins are two distinct families of proteins that probably evolved from a common ancestor but acquired different, non-overlapping cellular functions during evolution.
Figure 5 Comparison of the amino acid sequences of Gid (TrmFO) and GidA proteins. GidA proteins are systematically longer than Gid proteins, having an additional C-terminal domain. Both proteins apparently bind FAD cofactors (42,47) (data not shown). The two proteins can be discriminated by a sequence motif partially overlapping with the FAD binding motif (‘motif 1’). An additional GidA-specific motif located at the C-terminus of the protein also distinguishes the two paralogs (‘motif 2’).
First, GidA proteins belong to a different cluster of orthologous genes (COG0445) and, in contrast to TrmFO proteins (belonging to COG1206), they are present in mitochondria of Eukarya and in Bacteria, with the exception of high GC% Gram-positive bacteria, such as Mycobacterium and Corynebacterium species (http://string.embl.de/).
Second, consensus motifs implicated in flavin binding are slightly different in the 64 TrmFO and 203 GidA sequences analyzed (for details, see Figure 5). Moreover, GidA proteins always have an extension (or extensive insertions) that includes an additional characteristic sequence motif at their C-termini.
Third, we demonstrated that B.subtilis TrmFO clearly methylates uridine-54 in the T-psi loop of tRNAs (see above), while E.coli GidA and S.cerevisiae MTO1 (a mitochondrial homolog of bacterial GidA) were shown to be involved in a completely different reaction, namely the multistep formation of the hypermodified uridines at position 34 of anticodon of a few selected tRNAs (44–48). Moreover, in agreement with the fact that the enzymatic activity of TrmFO and GidA is not overlapping, we found that the absence of U-54 methylating activity in the T-psi loop of tRNA of B.subtilis mutant strain BFS2838 (gidermR) does not affect the level of conversion of U-34 into cmnm5s2U-34 in the anticodon loop of B.subtilis tRNAs. This was demonstrated by testing the capability of cmnm5s2U-34 residues, present in naturally occurring tRNAs of both the B.subtilis wild-type strain 168 and the mutant strain BFS2838 (gidermR), to become fully modified in vitro into mnm5s2U-34 upon incubation with purified recombinant MnmC protein of E.coli. This protein is a bifunctional enzyme that is absent from B.subtilis. MnmC removes carboxymethyl group of cmnm5s2U-34 to produce nm5s2U-34 and further methylates it into mnm5s2U-34 (as in naturally occurring E.coli tRNAs) using S-AdoMet as a methyl donor (30). The autoradiographs in Figure 6 show that the formation of mnm5s2U-34 occurs equally well in the tRNAs of both the wild-type B.subtilis strain and the gidermR mutant, thus clearly indicating that absence of TrmFO activity does not interfere with the GidA-dependent formation of cmnm5s2U-34. Conversely, we also verified that purified recombinant B.subtilis GidA protein does not catalyze in vitro a U-54 methylation reaction under the experimental conditions used for m5U-54 formation catalyzed by CH2H4folate-dependent TrmFO (data not shown). Whether GidA proteins, similar to TrmFO enzymes, also act as methylases is currently unclear.
Figure 6 tRNAs from the gidermR mutant are substrates for the MnmC protein. (A and B) correspond to autoradiograms of a TLC analysis of nuclease P1-hydrolysate of bulk tRNAs, isolated from B.subtilis strain 168 (A) or from BFS2838 (gidermR) mutant (B), previously incubated with purified recombinant E.coli MnmC and -labeled AdoMet (see Materials and Methods) (30). Radiolabeled compounds were detected and quantified after 3 days exposure using a PhosphoImager. Results indicated that both bulk tRNAs are equally well methylated by MnmC showing that in both the cases, cmnm5U-34 was prevalent in the tRNAs.
DISCUSSION
The RNA methyltransferases (MTases) add methyl groups to the base or the ribose 2'-hydroxyl of ribonucleotides during the complex process of RNA maturation. The great majority of these MTases use S-AdoMet as methyl donor . However, at least in the case of m5U-54 formation in tRNA of certain organisms, N5, N10-methylenetetrahydrofolate, together with associated oxydo-reduction coenzyme FADH2, has been shown to serve the same purpose . This activity was first detected and the corresponding enzymes purified from S.faecalis almost three decades ago, but the gene encoding this folate-dependent activity had still not been identified. Here, we predicted, using prior experimental knowledge and phylogenetic distribution analyses, that Gid proteins, previously of unknown function (42,43), could correspond to such a folate-dependent tRNA methyltransferase. We have experimentally confirmed this prediction by showing that B.subtilis Gid protein (here renamed TrmFO, FO for the folate) is necessary and sufficient for ribothymidine-54 formation in the T-psi loop of tRNA both in vivo and in vitro.
Based on bioinformatics analyses, one surprising aspect of this work is that the phylogenetic distribution of the folate-dependent pathway appears much wider than originally anticipated. Nevertheless, it appears to be restricted to methylation of uridine-54 in tRNA, not m5U in rRNA as in the case of the S-AdoMet-dependent pathway (12,13). Strikingly, the folate-dependent TrmFO proteins (COG1206) and S-AdoMet-dependent TrmA/Trm2p enzymes (COG2265) acting on tRNA appear to have mutually exclusive phylogenetic distributions (Table 1). Note that the lack of a trmA ortholog in a given organism is difficult to ascertain as paralogous genes that participate in S-AdoMet-dependent methylation of rRNA are also present. For instance, the trmA/TRM2 gene is found in enterobacteriacae (including E.coli and pseudomonaceae) as well as in all Eukarya so far sequenced. In Archaea, the trmA/TRM2 homologs are only found in the Pyrococcus genus (13,37). These organisms do not contain a gene coding for a trmFO ortholog. In contrast, orthologs of trmFO are found in most Gram-positive bacteria (firmicutes and actinobacteria) and in several other bacterial groups (e.g. alpha and delta-proteobacteria, cyanobacteria, Table 1). Strikingly, a subset of bacteria, for instance most Mycoplasma species, seemingly lack either trmFO or trmA genes, suggesting that the uridine at position 54 in their tRNAs may not be methylated. Indeed, in tRNAs of M.capricolum, Mycoplasma mycoides and Mycobacterium smegmatis, for which the primary sequences (including modified nucleotides) are known (2,49), ribothymine-54 is indeed absent, and their bulk ribothymidine-less tRNAs can be used successfully as substrates for U-54 methylation in E.coli extracts (50). Interestingly, M.mycoides has two putative trmFO alleles whose functional role is unclear and is worth investigation. In contrast, thermophilic G.stearothermophilus displays an S-AdoMet-dependent activity for m5U modification in vitro (34,35), but no trmA gene or trmFO has been found in still uncompletely sequenced genome of this Gram-positive bacterium. This observation raises the possibility that one of the S-AdoMet-dependent rRNA MTase paralogs, which we have detected in non-annotated genome sequence of this species, could act as a tRNA methylase. This idea is further supported by an experimental observation indicating that a Pyrococcus abyssi protein highly similar to E.coli rRNA:m5U MTase RumA (13) is actually a site-specific methylase for U-54 in tRNA (J. Urbonavicius, S. Auxilien, K. Trachana and H. Grosjean, unpublished data). We also expect that in many bacteria containing TrmFO, the methylation of U-54 in their tRNAs depends on folate metabolism, while the formation of m5U in their rRNA is dependent on S-AdoMet, as it is in M.lysodeikticus (51). To construct a more comprehensive evolutionary history for this large family of m5U-forming enzymes, as well as other 5-methylpyrimidine MTases, such as those forming m5C in RNA and in DNA (52,53), experimental identification of the exact nucleotide target(s) within an RNA for each of these MTases is needed.
Table 1 Non-exhaustive distribution of the trmAa and trmFO (gid) coding for putative tRNA:m5U-54 methyltransferases in bacteria
In this work, we also considered the possible evolutionary relationship between ribothymidylate synthase TrmFO and the thymidylate synthase ThyX family of enzymes that catalyze a very similar reaction (19). Although both ThyX and TrmFO proteins use flavin (FADH2) nucleotide as cofactor, our studies have indicated that they do not belong to the same family of flavoproteins. In particular, TrmFO proteins lack the characteristic conserved residues required for catalysis, substrate and/or cofactor binding of ThyX proteins (54). In addition, the novel FAD binding fold found in ThyX proteins does not show significant similarity to the classical Rossman-fold predicted for TrmFO proteins (21,38). Thus, in the light of this information, a common origin for TrmFO and ThyX proteins appears unlikely. Therefore, direct comparison of the reaction mechanisms between TrmFO and ThyX proteins cannot be done. Our work suggests, for the first time, that the use of CH2H4folate and FAD in the post-transcriptional methylation of polynucleotides (pre-tRNA) or of a mononucleotide (dUMP) during DNA precursor synthesis has been established independently at least twice during evolution.
ACKNOWLEDGEMENTS
G. Bj?rk, S. Seror, L. Droogmans and R. Moser are acknowledged in the text. K. Trachana is acknowledged for her help in cell extract preparation. All HPLC analysis were performed by K. Jacobsson at the University of Ume?, Sweden (grants from Swedish Cancer Foundation—Project 680—and Swedish Science Research Council—Project BU2930—to G. Bj?rk). The authors thank S. Douthwaite (University of South Denmark, Odense, Denmark), S. Auxilien, B. Golinelli-Pimpaneau (CNRS, Gif-sur-Yvette), B. Holland (University of Paris XI) and V. de Crécy-Lagard (University of Florida, Gainesville, FL, USA) for useful advices and critical reading of the manuscript. H.G. benefits from a research grant from the CNRS (GEOMEX program). H.M. is supported by research funds from INSERM (BIOAVENIR program), CNRS (Program Microbiologie Fondamental) and Fondation Bettencourt-Schueller. J.U. and S.S. are FEBS Postdoctoral Fellow and INSERM Young Researcher, respectively. Funding to pay the Open Access publication charges for this article was provided by CNRS, Programme de Microbiologie Fondamentale to H.M.
REFERENCES
Limbach, P.A., Crain, P.F., McCloskey, J.A. (1994) Summary: the modified nucleosides of RNA Nucleic Acids Res., 22, 2183–2196 .
Sprinzl, M. and Vassilenko, K.S. (2005) Compilation of tRNA sequences and sequences of tRNA genes Nucleic Acids Res., 33, D139–D140 .
Shigi, N., Suzuki, T., Tamakoshi, M., Oshima, T., Watanabe, K. (2002) Conserved bases in the TPsiC loop of tRNA are determinants for thermophile-specific 2-thiouridylation at position 54 J. Biol. Chem., 277, 39128–39135 .
Fleissner, E. and Borek, E. (1962) A new enzyme of RNA synthesis: RNA methylase Proc. Natl Acad. Sci. USA, 48, 1199–1203 .
Svensson, I., Boman, H.G., Eriksson, K.G., Kjellin, K. (1963) Studies on microbial RNA. I. Transfer of methyl groups from methionine to soluble RNA from Escherichia Coli J. Mol. Biol., 16, 254–271 .
Bjork, G.R. (1975) Transductional mapping of gene trmA responsible for the production of 5-methyluridine in transfer ribonucleic acid of Escherichia coli J. Bacteriol., 124, 92–98 .
Ny, T. and Bjork, G.R. (1980) Cloning and restriction mapping of the trmA gene coding for transfer ribonucleic acid (5-methyluridine)-methyltransferase in Escherichia coli K-12 J. Bacteriol., 142, 371–379 .
Kealey, J.T., Gu, X., Santi, D.V. (1994) Enzymatic mechanism of tRNA (m5U54)methyltransferase Biochimie, 76, 1133–1142 .
Fauman, E.B., Blumenthal, R.M., Cheng, X. (1999) Structure and evolution of AdoMet-dependent methyltransferases In Cheng, X. and Blumenthal, R.M. (Eds.). S-Adenosylmethionine-Dependent Methyltransferases: Structures and Functions, Singapore World Scientific pp. 1–38 .
Bujnicki, J.M., Droogmans, L., Grosjean, H., Purushothaman, S.K., Lapeyre, B. (2004) Bioinformatics-guided identification and experimental characterization of novel RNA methyltransferases In Bujnicki, J.M. (Ed.). Nucleic Acids and Molecular Biology: Practical Bioinformatics series, Berlin, Heidelberg Springer-Verlag Vol. 15, pp. 139–168 .
Tatusov, R.L., Fedorova, N.D., Jackson, J.D., Jacobs, A.R., Kiryutin, B., Koonin, E.V., Krylov, D.M., Mazumder, R., Mekhedov, S.L., Nikolskaya, A.N., et al. (2003) The COG database: an updated version includes eukaryotes BMC Bioinformatics, 4, 41 .
Agarwalla, S., Kealey, J.T., Santi, D.V., Stroud, R.M. (2002) Characterization of the 23 S ribosomal RNA m5U1939 methyltransferase from Escherichia coli J. Biol. Chem., 277, 8835–8840 .
Madsen, C.T., Mengel-Jorgensen, J., Kirpekar, F., Douthwaite, S. (2003) Identifying the methyltransferases for m(5)U747 and m(5)U1939 in 23S rRNA using MALDI mass spectrometry Nucleic Acids Res., 31, 4738–4746 .
Delk, A.S., Nagle, D.P., Jr, Rabinowitz, J.C. (1980) Methylenetetrahydrofolate-dependent biosynthesis of ribothymidine in transfer RNA of Streptococcus faecalis. Evidence for reduction of the 1-carbon unit by FADH2 J. Biol. Chem., 255, 4387–4390 .
Samuel, C.E. and Rabinowitz, J.C. (1974) Initiation of protein synthesis by folate-sufficient and folate-deficient Streptococcus faecalis R: partial purification and properties of methionyl-transfer ribonucleic acid synthetase and methionyl-transfer ribonucleic acid formyltransferase J. Bacteriol., 118, 21–31 .
Arnold, H.H. and Kersten, H. (1975) Inhibition of the tetrahydrofolate-dependent biosynthesis of ribothymidine in tRNAs of B.subtilis and M.lysodeikticus by trimethoprim FEBS Lett., 53, 258–261 .
Delk, A.S., Romeo, J.M., Nagle, D.P., Jr, Rabinowitz, J.C. (1976) Biosynthesis of ribothymidine in the transfer RNA of Streptococcus faecalis and Bacillus subtilis. A methylation of RNA involving 5,10-methylenetetrahydrofolate J. Biol. Chem., 251, 7649–7656 .
Delk, A.S., Nagle, D.P., Jr, Rabinowitz, J.C., Straub, K.M. (1979) The methylenetetrahydrofolate-mediated biosynthesis of ribothymidine in the transfer-RNA of Streptococcus faecalis: incorporation of hydrogen from solvent into the methyl moiety Biochem. Biophys. Res. Commun., 86, 244–251 .
Myllykallio, H., Lipowski, G., Leduc, D., Filee, J., Forterre, P., Liebl, U. (2002) An alternative flavin-dependent mechanism for thymidylate synthesis Science, 297, 105–107 .
Giladi, M., Bitan-Banin, G., Mevarech, M., Ortenberg, R. (2002) Genetic evidence for a novel thymidylate synthase in the halophilic archaeon Halobacterium salinarum and in Campylobacter jejuni FEMS Microbiol. Lett., 216, 105–109 .
Mathews, II, Deacon, A.M., Canaves, J.M., McMullan, D., Lesley, S.A., Agarwalla, S., Kuhn, P. (2003) Functional analysis of substrate and cofactor complex structures of a thymidylate synthase-complementing protein Structure (Camb.), 11, 677–690 .
Finer-Moore, J.S., Santi, D.V., Stroud, R.M. (2003) Lessons and conclusions from dissecting the mechanism of a bisubstrate enzyme: thymidylate synthase mutagenesis, function, and structure Biochemistry, 42, 248–256 .
Agrawal, N., Lesley, S.A., Kuhn, P., Kohen, A. (2004) Mechanistic studies of a flavin-dependent thymidylate synthase Biochemistry, 43, 10295–10301 .
Graziani, S., Xia, Y., Gurnon, J.R., Van Etten, J.L., Leduc, D., Skouloubris, S., Myllykallio, H., Liebl, U. (2004) Functional analysis of FAD-dependent thymidylate synthase ThyX from Paramecium bursaria Chlorella virus-1 J. Biol. Chem., 279, 54340–54347 .
Gattis, S.G. and Palfey, B.A. (2005) Direct observation of the participation of flavin in product formation by thyX-encoded thymidylate synthase J. Am. Chem. Soc., 127, 832–833 .
Griffin, J., Roshick, C., Iliffe-Lee, E., McClarty, G. (2005) Catalytic mechanism of Chlamydia trachomatis flavin-dependent thymidylate synthase J. Biol. Chem., 280, 5456–5467 .
Perret, V., Garcia, A., Puglisi, J., Grosjean, H., Ebel, J.P., Florentz, C., Giege, R. (1990) Conformation in solution of yeast tRNA(Asp) transcripts deprived of modified nucleotides Biochimie, 72, 735–743 .
Grosjean, H., Droogmans, L., Giege, R., Uhlenbeck, O.C. (1990) Guanosine modifications in runoff transcripts of synthetic transfer RNA-Phe genes microinjected into Xenopus oocytes Biochim. Biophys. Acta, 1050, 267–273 .
Grosjean, H., Keith, G., Droogmans, L. (2004) Detection and quantification of modified nucleotides in RNA using thin-layer chromatography Methods Mol. Biol., 265, 357–391 .
Bujnicki, J.M., Oudjama, Y., Roovers, M., Owczarek, S., Caillet, J., Droogmans, L. (2004) Identification of a bifunctional enzyme MnmC involved in the biosynthesis of a hypermodified uridine in the wobble position of tRNA RNA, 10, 1236–1242 .
von Ehrenstein, G. (1967) Isolation of sRNA from intact Escherichia coli cells Meth. Enzymol., 12, 588–596 .
Gehrke, C.W. and Kuo, K.C. (1989) Ribonucleoside analysis by reversed-phase high-performance liquid chromatography J. Chromatogr., 471, 3–36 .
Delk, A.S., Nagle, D.P., Jr, Rabinowitz, J.C. (1979) Purification of methylenetetrahydrofolate-dependent methyltransferase catalyzing biosynthesis of ribothymidine in transfer RNA of Streptococcus faecalis In Kisliuk, R.L. and Brown, G.M. (Eds.). Chemistry and Biology of Pteridines, NY Elsevier/North Holland Publishing pp. 389–394 .
Agris, P.F., Koh, H., Soll, D. (1973) The effect of growth temperatures on the in vivo ribose methylation of Bacillus stearothermophilus transfer RNA Arch. Biochem. Biophys., 154, 277–282 .
Schmidt, W., Arnold, H.H., Kersten, H. (1977) Tetrahydrofolate-dependent biosynthesis of ribothymidine in transfer ribonucleic acids of Gram-positive bacteria J. Bacteriol., 129, 15–21 .
Nordlund, M.E., Johansson, J.O., von Pawel-Rammingen, U., Bystrom, A.S. (2000) Identification of the TRM2 gene encoding the tRNA(m5U54)methyltransferase of Saccharomyces cerevisiae RNA, 6, 844–860 .
Constantinesco, F., Motorin, Y., Grosjean, H. (1999) Transfer RNA modification enzymes from Pyrococcus furiosus: detection of the enzymatic activities in vitro Nucleic Acids Res., 27, 1308–1315 .
Dym, O. and Eisenberg, D. (2001) Sequence-structure analysis of FAD-containing proteins Protein Sci., 10, 1712–1728 .
Altschul, S.F., Madden, T.L., Schaffer, A.A., Zhang, J., Zhang, Z., Miller, W., Lipman, D.J. (1997) Gapped BLAST and PSI-BLAST: a new generation of protein database search programs Nucleic Acids Res., 25, 3389–3402 .
Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F., Higgins, D.G. (1997) The CLUSTAL_X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools Nucleic Acids Res., 25, 4876–4882 .
Bjork, G.R. and Isaksson, L.A. (1970) Isolation of mutants of Escherichia coli lacking 5-methyluracil in transfer ribonucleic acid or 1-methylguanine in ribosomal RNA J. Mol. Biol., 51, 83–100 .
White, D.J., Merod, R., Thomasson, B., Hartzell, P.L. (2001) GidA is an FAD-binding protein involved in development of Myxococcus xanthus Mol. Microbiol., 42, 503–517 .
Iwasaki, W., Miyatake, H., Ebihara, A., Miki, K. (2004) Crystallization and preliminary X-ray crystallographic studies of the small form of glucose-inhibited division protein A from Thermus thermophilus HB8 Acta Crystallogr. D Biol. Crystallogr., 60, 515–517 .
Colby, G., Wu, M., Tzagoloff, A. (1998) MTO1 codes for a mitochondrial protein required for respiration in paromomycin-resistant mutants of Saccharomyces cerevisiae J. Biol. Chem., 273, 27945–27952 .
Bregeon, D., Colot, V., Radman, M., Taddei, F. (2001) Translational misreading: a tRNA modification counteracts a +2 ribosomal frameshift Genes Dev., 15, 2295–2306 .
Li, R., Li, X., Yan, Q., Qin Mo, J., Guan, M.X. (2003) Identification and characterization of mouse MTO1 gene related to mitochondrial tRNA modification Biochim. Biophys. Acta, 1629, 53–59 .
Scrima, A., Vetter, I.R., Armengod, M.E., Wittinghofer, A. (2005) The structure of the TrmE GTP-binding protein and its implications for tRNA modification EMBO J., 24, 23–33 .
Umeda, N., Suzuki, T., Yukawa, M., Ohya, Y., Shindo, H., Watanabe, K. (2005) Mitochondria-specific RNA-modifying enzymes responsible for the biosynthesis of the wobble base in mitochondrial tRNAs. Implications for the molecular pathogenesis of human mitochondrial diseases J. Biol. Chem., 280, 1613–1624 .
Andachi, Y., Yamao, F., Muto, A., Osawa, S. (1989) Codon recognition patterns as deduced from sequences of the complete set of transfer RNA species in Mycoplasma capricolum. Resemblance to mitochondria J. Mol. Biol., 209, 37–54 .
Johnson, L., Hayashi, H., Soll, D. (1970) Isolation and properties of a transfer ribonucleic acid deficient in ribothymidine Biochemistry, 9, 2823–2831 .
Schmidt, W., Arnold, H.H., Kersten, H. (1975) Biosynthetic pathway of ribothymidine in B.subtilis and M.lysodeikticus involving different coenzymes for transfer RNA and ribosomal RNA Nucleic Acids Res., 2, 1043–1051 .
Anantharaman, V., Koonin, E.V., Aravind, L. (2002) Comparative genomics and evolution of proteins involved in RNA metabolism Nucleic Acids Res., 30, 1427–1464 .
Bujnicki, J.M., Feder, M., Ayres, C.L., Redman, K.L. (2004) Sequence-structure-function studies of tRNA:m5C methyltransferase Trm4p and its relationship to DNA:m5C and RNA:m5U methyltransferases Nucleic Acids Res., 32, 2453–2463 .
Leduc, D., Graziani, S., Lipowski, G., Marchand, C., Le Marechal, P., Liebl, U., Myllykallio, H. (2004) Functional evidence for active site location of tetrameric thymidylate synthase X at the interphase of three monomers Proc. Natl Acad. Sci. USA, 101, 7252–7257 .
Hall, R.H. The modified nucleosides in nucleic acids, (1971) NY and London Columbia University Press .(Jaunius Urbonaviius, Stéphane Skouloubri)
*To whom correspondence should be addressed. Tel: +33 1 69823468; Fax: +33 1 69823129; Email: henri.grosjean@lebs.cnrs-gif.fr
ABSTRACT
Formation of 5-methyluridine (ribothymidine) at position 54 of the T-psi loop of tRNA is catalyzed by site-specific tRNA methyltransferases (tRNA:m5U-54 MTase). In all Eukarya and many Gram-negative Bacteria, the methyl donor for this reaction is S-adenosyl-L-methionine (S-AdoMet), while in several Gram-positive Bacteria, the source of carbon is N5, N10-methylenetetrahydrofolate (CH2H4folate). We have identified the gene for Bacillus subtilis tRNA:m5U-54 MTase. The encoded recombinant protein contains tightly bound flavin and is active in Escherichia coli mutant lacking m5U-54 in tRNAs and in vitro using T7 tRNA transcript as substrate. This gene is currently annotated gid in Genome Data Banks and it is here renamed trmFO. TrmFO (Gid) orthologs have also been identified in many other bacterial genomes and comparison of their amino acid sequences reveals that they are phylogenetically distinct from either ThyA or ThyX class of thymidylate synthases, which catalyze folate-dependent formation of deoxyribothymine monophosphate, the universal DNA precursor.
INTRODUCTION
Transfer RNAs in all living organisms contain a number of nucleosides that are post-transcriptionally modified on the base and/or the 2'-hydroxyl group of the ribose (1). One such common modified nucleoside is 5-methyluridine (m5U, also designated T for ribothymidine). This C5-methylated uridine is invariably found at position 54, in the so-called T-psi loop of tRNA of almost all Bacteria and Eukarya (2). In thermophilic Bacteria, such as Thermus thermophilus, it is further hypermodified to a 2-thio-derivative , while in certain Eukarya, a 2'-O-methyl-derivative is occasionally found .
Site-specific methylation of U-54 in Escherichia coli tRNA is catalyzed by tRNA:m5U-54 methyltransferase (EC.2.1.1.35). This enzyme, initially designated RUMT for RNA uridine methyltransferase, was the first RNA modification enzyme discovered that acts at the polynucleotide level (4,5). This enzyme is also called TrmA (tRNA methyltransferase A), and a gene trmA encoding this enzyme was first identified in E.coli (6,7). From the standpoint of mechanism and specificity, the tRNA:m5U-54 methyltransferase of E.coli is one of the best characterized RNA modification enzymes . In the majority of RNA methyltransferases studied so far , RUMT uses S-adenosylmethionine (S-AdoMet) as the methyl donor. Automated bioinformatic approaches have included all trmA and TRM2 homologs in the same cluster of orthologous genes . This cluster contains a superfamily of S-AdoMet-dependent RNA:m5U MTases that are specific not only for uridine at position 54 of tRNA, but also paralogs that function in uridine methylation in other RNAs .
Earlier studies have indicated that not all bacterial tRNA:m5U-54 MTases use S-AdoMet as methyl donor. For example, in Enterococcus faecalis (formerly Streptococcus faecalis) and Bacillus subtilis, it was reported that the carbon donor of the methyl group is N5, N10-methylenetetrahydrofolate (CH2H4folate) . The first indication for this came from an observation that bulk tRNAs isolated from folate-deprived E.faecalis cells lacked m5U-54 in their T-psi loop (15). Moreover, in B.subtilis and Micrococcus lysodeikticus, trimethoprim, a specific inhibitor of bacterial dihydrofolate reductase, inhibits formation of m5U-54 in vivo (16), indicating that in these Gram-positive bacteria, the carbon source used in tRNA methylation derives from the folate pool. The results of these in vivo studies were later confirmed by demonstrating that in vitro activity of purified tRNA: m5U-54 MTase of E.faecalis not only requires CH2H4folate but also reduced flavin adenine nucleotide (FADH2) (14,17,18), thus forming a distinct class of tRNA:m5U-54 MTases (EC.2.1.1.74). Strikingly, this observation is reminiscent of the enzymatic mechanism that has been described for the alternative flavin-dependent ThyX class of thymidylate synthases (EC.2.1.1.148) (19–21), but differs from the reaction catalyzed by a canonical thymidylate synthase ThyA, which uses CH2H4folate both as a carbon source and as a reductant . Moreover, in the case of ThyX catalysis, it has been recently demonstrated that a hydride from NAD(P)H is transferred, via a FAD cofactor to reduce the methylene group, to a methyl residue (23–26). The gene encoding the folate-dependent tRNA:m5U-54 MTase has not yet been identified. It is, therefore, not known whether these analogous folate-dependent methylation reactions, involved in RNA or DNA metabolism, are catalyzed by distantly related enzymes, possibly originating from the early RNA World, or, on the contrary, represent independent catalytic mechanisms.
Benefiting from large-scale microbial sequencing and structural genomics projects, we predicted that bacterial Gid proteins would correspond to a novel class of bacterial site-specific tRNA:m5U-54 MTases. This prediction was confirmed through genetic studies and biochemical analyses of tRNA molecules isolated from B.subtilis wild-type and mutants strains. In vitro characterization of the purified recombinant B.subtilis tRNA:m5U-54 MTase indicates that this protein alone is sufficient for tRNA methylation reaction. Our studies further indicate that despite the fact that thymidylate synthase ThyX and Gid proteins catalyze a similar methylation reaction, they lack detectable sequence, and probably structural similarity. Our analyses suggest that the enzymes methylating nucleotides in tRNA and DNA precursor using CH2H4folate and NAD(P)H/FAD as carbon donor and reductant, respectively, have independent evolutionary origins.
MATERIALS AND METHODS
Strains
B.subtilis strain BFS2838 carrying inactivated gidermR gene was kindly provided by S. Seror . E.coli strain GRB113 (metA, trmA5, zij-90::Tn10), encoding an inactivated TrmA protein was a kind gift from G. R. Bj?rk, Ume? University, Sweden. E.coli Sure? strain (e14–(McrA–) mcrCB-hsdSMR-mrr)171 endA1 supE44 thi-1 gyrA96 relA1 lac recB recJ sbcC umuC::Tn5 (Kanr) uvrC was purchased from Stratagene.
Construction of an N-terminal His6-tagged BsuGidA and BsuGid-overexpressing plasmids and purification of the corresponding recombinant proteins
The gidA (GIDA_BACSUB; P25812 ) and gid genes (GID_BACSU; P39815 ; renamed trmFO in this work) were amplified by PCR from B.subtilis strain 168 DNA, using Pfu DNA polymerase (Promega) and the following primers (sequence in small characters correspond to genome sequence): gidAfw (CGGGATCCatggggtatgaagcaggccaatac) and gidArev (TCCCCCGGGctactcggctatcttcgcaatgcg) or gidfw (CGGGATCCatgaaccaacaaacagtgaatgta) and gidrev (TCCCCCGGGctatattgttttcgaaattgtttg). The resulting 1893 or 1314 bp PCR products were then digested with BamHI and SmaI, respectively, and cloned into pQE80L to generate pQE80L-BsuGidA or pQE80L-BsuGid. To purify recombinant GidA and Gid proteins, pQE80L-BsuGidA was transformed into E.coli Sure? strain, and pQE80L-BsuGid was transformed into Sure? or GRB113 (trmA5) strain. Resulting strains were grown at 37°C in 500 ml of Luria–Bertani (LB) medium (Invitrogen) containing 100 μl/ml ampicillin until OD600 = 0.6. After induction of Gid or GidA protein expression by isopropyl ?-D-thiogalactopyranoside (IPTG) (VWR International, final concentration = 1 mM), the cultures were further grown at 37°C for 3 h. After harvesting the cells by centrifugation, the pellet were flash-frozen in liquid N2 and stored at –80°C. Frozen cells were thawn on ice and resuspended in 5 ml of lysis buffer (50 mM sodium phosphate, pH 7.6, 300 mM NaCl, 10% glycerol and 20 mM imidazole) containing 5 μl Protein Inhibitor Cocktail (PIC, Sigma) and 1.5 μl ?-mercaptoethanol. Cells were broken by 2 freeze (liquid N2)/thaw (37°C) cycles and ultrasonication. The lysate was centrifuged for 15 min at 10 000 g at 4°C. Supernatant was loaded onto 2 ml of Ni-NTA resin and washed with 25 ml of lysis buffer. Gid or GidA proteins were eluted with 10 ml elution buffer (same as lysis buffer, but containing 250 mM imidazole). Yellow fractions, containing the Gid or GidA protein, were pooled (to 3 ml of total volume) and dialyzed against 500 ml of 30 mM HEPES buffer, pH 7.5, containing 200 mM NaCl and 10% glycerol. Protein was aliquoted, flash-frozen in N2 and stored at –80°C. To measure any cofactor release from Gid, 5 μg of protein was diluted in 100 μl of distilled water and incubated for 5 min at 90°C. The sample was centrifuged at 10 000 g for 15 min. Absorption and fluorescence spectra of the obtained supernatant were measured.
Preparation of cell-free extracts
Cell-free extracts of B.subtilis strains 168 (wild-type) and BFS2838 (gidermR) were prepared from an exponentially growing cell culture at 37°C. After centrifugation and washing the cell pellet with lysis buffer (25 mM Tris–HCl buffer, pH 7.5, 10 mM MgCl2, 25 mM KCl and 2 mM DTT), it was resuspended in a 1.5 vol of lysis buffer containing 1% (v/v) of PIC, Sigma. An S10 cell-free extract was obtained after ultrasonication and centrifugation for 15 min at 10 000 g. Further centrifugation of supernatant for 1 h at 4°C resulted in S100 cell-free extracts. Cell-free extracts of E.coli strains pQE80L-BsuGid/GRB113 (trmA5) and pQE80L/GRB113 (trmA5) were prepared similarly as for the protein purification, except that they were grown at 37°C to an OD600 of 0.8 in 10 ml of liquid Luria Broth with 100 μl/ml carbenicillin. After the Gid protein induction, harvesting of cells, resuspension in 500 μl lysis buffer, cell disruption by ultrasonication and centrifugation, the S10 cell-free extract was produced.
Enzymatic activity assays
UTP-labeled yeast tRNAAsp transcript, used for determining the tRNA:m5U-54 MTase activity of TrmFO (Gid) protein, was prepared and purified as described elsewhere (27,28). A total of 50–100 fmol of -labeled tRNAAsp were incubated at 37°C in a 50 μl reaction mixture containing 40 mM N- piperazine-N--Na buffer (HEPES-Na, Sigma) at pH 7.0, 0.25 mM FAD, Fluka, 0.5 mM NADH (reduced nicotinamide adenine dinucleotide, Sigma), 1 mM reduced nicotinamide adenine dinucleotide phosphate (NADPH), 0.25 mM (6R)-N5,N10-CH2H4PteGlu-Na2 (methylenetetrahydrofolate, provided by Dr R. Moser, Merck-Eprova, AG, Switzerland), 5 mM DTT (Promega), 15 U of RNase inhibitor (Fermentas) and 10–25 μg of total protein of a B.subtilis or E.coli cell-free extract. At the end of the incubation period, modified tRNA was extracted and digested with nuclease P1 (Roche), the modified nucleotides were separated by 2D thin-layer chromatography (2D-TLC) and data were analyzed as described previously (29). Methylating activity of purified recombinant BsuGid (TrmFO) (1 μg per test) and BsuGidA protein (1 μg per test) were tested using the same experimental conditions as above. Activity of the MnmC enzyme on bulk tRNAs from B.subtilis strains 168 or BFS2838 (gidermR) was tested as follows: five microgram of purified recombinant MnmC protein (provided by Dr L. Droogmans, University of Brussels, Belgium) was added to 300 μl of a reaction mixture containing 50 mM Tris–HCl, pH 8.0, 20 mM NH4Cl, 62 μM -AdoMet (53 Ci/mol, Amersham) and 100 μg of bulk B.subtilis tRNAs. After 1 h incubation at 37°C, tRNA was recovered, digested by nuclease P1, and the resulting radiolabeled nucleotides were analyzed by 2D-TLC as described previously (30).
Isolation of tRNA and chromatographic analysis of tRNA hydrolysates
Bulk tRNAs of B.subtilis strains 168 and BFS2838 (gidermR) were obtained essentially as described previously (31), except that the tRNA deacylation step was omitted and a monoQ column (5 ml from Biorad) was used instead of DEAE-cellulose. For bulk tRNAs from E.coli GRB113 (trmA5), transformed by pQE80L-BsuGid or pQE80L (control experiment), cell cultures were first grown at 37°C in 200 ml of liquid Luria Broth in the presence of 100 μl/ml carbenicillin. At OD600 = 0.6, IPTG was added to the final concentration of 1 mM, and the cells were grown for additional 3 h at 37°C before to be collected in the cold by centrifugation and purified as above. Obtained purified bulk tRNAs were completely degraded to nucleosides with P1 nuclease and alkaline phosphatase (Sigma) and the resulting hydrolysates analyzed by high performance liquid chromatography (HPLC) on a Supelcosil LC18 column (Supelco) with Waters HPLC instrument, as described previously (32).
RESULTS
Comparative genomics identifies a candidate gene encoding a new family of flavin-dependent methyltransferases
An enzyme of two identical subunits of 58 kDa that catalyzes the site-specific formation of 5-methyluridine in position 54 (m5U-54) of tRNA using CH2H4folate as a source of one-carbon unit and a combination of coenzymes NAD(P)H/FAD as reductant, has been purified from E.faecalis (33). We attempted an identification of the gene encoding this activity (described under EC 2.1.1.74 ) based on the facts that a folate-dependent pathway for tRNA methylation exists in some Gram-positive Bacteria species , whereas an S-AdoMet-dependent enzyme is used instead in Eukarya , in gamma-proteobacteria and in a few Archaea .
Our primary searches used an updated version of COG database (http://www.ncbi.nlm.nih.gov/COG/), which currently consists of 4873 gene families (11). Using the phylogenetic distribution analysis tool of this database, we obtained a list of 155 COGs (3% of total number of families) that are present in B.subtilis and Bacillus halodurans, but absent in Archaea, Eukarya and gamma-proteobacteria (data not shown). We have no data for E.faecalis, M.lysodeikticus and G.stearothermophilus, as they are not included in the current data release. Next, among the 155 candidates, we searched for the presence of a characteristic ‘GXGXXG’ motif that is part of the conserved Rossman-fold found in a large number of FAD binding proteins . As a result, one COG family (COG1206) emerged as an evident protein family encoding a putative tRNA:m5U-54 MTase. These COG1206 proteins, also designated as Gid proteins (in reference to the B.subtilis protein) are: (i) currently annotated as ‘NADPH(FAD)-utilizing enzymes possibly involved in translation’, (ii) contain a readily identifiable FAD binding motif ‘GXGXXG’ (in fact G-X-G-L-A-G--E-X-A, see details below) and (iii) their molecular weight is 50 kDa, in close agreement with the 58 kDa determined on SDS–PAGE gels for the -subunit of the E.faecalis folate-dependent tRNA:m5U-54 MTase (33).
COG1206 proteins have a wide phylogenetic distribution
Systematic screening of >200 fully sequenced genomes (http://www.ncbi.nih.gov/genomes/lproks.cgi), using pattern-hit initiated BLAST algorithm (39) and B.subtilis Gid (GID_BACSU; P39815 ) as a query, identified 80 bacterial species containing a gid gene, whereas no hits were found in archaeal nor eukaryal genomes. This phylogenetic distribution of gid genes (Figure 1) is much wider than initially anticipated. In addition to the expected Gram-positive bacteria (Bacillales and Lactobacillales), a gene for Gid-like protein is also found in alpha-proteobacteria, delta-proteobacteria and cyanobacteria. Phylogenetic analyses of a subset of Gid orthologs, using neighbor-joining trees performed with ClustalX program (40), indicate that their phylogeny is congruent with species phylogeny, suggesting a relatively ancient bacterial origin for Gid proteins (Figure 1).
Figure 1 A phylogenetic tree based on a subset of Gid homologs retrieved by pattern-hit initiated BLAST algorithm. Clustal X was used for sequence alignments and phylogenetic trees were constructed using the neighbor-joining methods. GidA sequences from B.subtilis, T.thermophilus and Deinococcus radiodurans were used as an outgroup. Branch points with closed circles indicate a bootstrap support >90%. The shown topology was also supported by quartet puzzling with maximum likelihood analysis performed using Tree-Puzzle 5.1 program implemented at www.pasteur.fr (data not shown).
The Gid protein of B.subtilis is involved in m5U-54 formation in tRNA
To determine whether Gid proteins are involved in the biosynthesis of m5U-54, the presence of this methylated nucleoside was analyzed in B.subtilis tRNA isolated from a BFS2838 (gidermR) strain, lacking functional Gid protein (kindly provided by S. Seror, University of Paris XI). No obvious phenotype has been described for this B.subtilis strain (see http://locus.jouy.inra.fr/cgi-bin/dev/chiapell/strain_pheno_old.pl?STRAIN=BFS2838). Bulk tRNAs from the mutant strain and the corresponding wild-type strain B.subtilis 168 were extracted, and their nucleoside contents were analyzed by HPLC as described in Materials and Methods. Results in Figure 2A and B clearly demonstrate that m5U is absent in the tRNA from the gidermR mutant, whereas tRNA of the wild-type strain 168 contains the m5U modification. The small residual peak eluting at the same position as m5U in Figure 2B was identified as inosine through its characteristic UV absorbance spectrum. The maximum wavelength for inosine is at 250 nm (Figure 2D), compared with 267 nm for 5-methyluridine (Figure 2C).
Figure 2 B.subtilis BFS2838 (gidermR) strain lacks m5U modification in the tRNA. Bulk tRNAs from wild-type B.subtilis 168 (A) or the gidermR mutant BFS2838 (B) were isolated, completely digested to nucleosides by nuclease P1 and alkaline phosphatase and analyzed by HPLC (see Materials and Methods). Alternatively, T7 polymerase transcripts of yeast wild-type tRNAAsp, uniformly labeled with UTP were incubated for 1 h at 37°C with an S100 cell extract from 168 or BFS2838 . After incubation, bulk tRNAs were completely digested to monophosphate nucleosides by nuclease P1 and analyzed by 2D-TLC. Radiolabeled compounds were detected and quantified using PhosphoImager detector. (C) The spectrum analysis of the HPLC fraction corresponding to m5U (A). (D) The spectrum of the small peak detected in (B), corresponding to inosine. Nature of the modified nucleosides in chromatography peaks was determined by comparison with the standards (55).
The absence of C5-methylation activity for U-54 in gidermR mutant strain was further confirmed by testing the corresponding methylation activity in cell extracts. Thus, -radiolabeled T7-transcript of a synthetic yeast tRNAAsp gene was used as substrate, and incubations were performed in the presence of CH2H4folate, NADH/NADPH and FAD as indicated in Materials and Methods. After incubation, the tRNA was digested into 5'-monophosphate nucleosides and the hydrolysate was analyzed by 2D-TLC. The radiolabeled spots, corresponding to -labeled UMP-derivatives were detected by autoradiography. As shown in Figure 2A and B (insets), while the wild-type cell extract was able to catalyze the formation of m5U-54 in vitro, the extract from the gidermR mutant strain did not catalyze such a methylation reaction, thus indicating that the gid gene product is involved in the production of m5U-54 in tRNA.
The Gid protein of B.subtilis is sufficient for methylation of U-54 in E.coli tRNA in vivo
To investigate whether the B.subtilis Gid protein alone can substitute for the S-AdoMet-dependent E.coli TrmA protein for the formation of m5U-54 in vivo, we cloned the B.subtilis gid into an E.coli expression vector pQE80L, under the control of an IPTG inducible promoter. The resulting plasmid, pQE80L-BsuGid was transformed into an E.coli strain GRB113, carrying trmA5 mutation. This E.coli strain grows normally in LB medium but completely lacks S-AdoMet-dependent tRNA:m5U-54 MTase activity (41). After gid expression for 3 h, the cells were collected by centrifugation and divided into two parts. Bulk tRNA was purified from one part, while the remaining cell pellet was used to prepare an S10 cell extract (see Materials and Methods). HPLC analysis of P1/alkaline phosphatase-treated bulk tRNA hydrolysate demonstrated the presence of m5U nucleoside in the E.coli trmA5 strain transformed by pQE80L-BsuGid (Figure 3A), while in the control E.coli mutant strain, transformed by unmodified expression vector, no m5U was detectable (Figure 3B). As described above with B.subtilis bulk tRNAs (Figure 2), we verified that the very small peak migrating at the position expected for m5U in the HPLC analysis of tRNA hydrolysate of the control strain corresponds to inosine (see UV-spectrum in Figure 3D, compare with Figure 3C for m5U). In parallel, the S10 cell extract was incubated together with appropriate cofactors and labeled yeast tRNAAsp transcript. The P1-hydrolyzate of the resulting modified tRNA was then analyzed by 2D-TLC. Results in Figure 3A and B (insets) indicate that m5U-54 in tRNAAsp is formed only when a cell-free extracts from the E.coli trmA5 strain transformed with pQE80L-BsuGid, confirming that the B.subtilis Gid protein efficiently modified tRNAs under the physiological conditions of E.coli cells.
Figure 3 Recombinant BsuGid protein catalyzes the formation of the m5U-54 modification in tRNA in vivo. HPLC analysis performed with bulk tRNAs purified from E.coli GRB113 (trmA5) transformed with pQE80L-BsuGid (A), or with the ‘empty plasmid’ pQE80L (B). Insets in (A) and (B) correspond to autoradiograms of tRNA hydrolysates resulting from in vitro enzymatic activity tests performed with cell-free extracts of E.coli GRB113 (trmA5) strain transformed with pQE80L-Bsu Gid or with pQE80L . As in Figure 2, (C) shows the spectrum analysis of the HPLC fraction corresponding to m5U (A). (D) The spectrum of the small peak detected in (B), which corresponds to inosine.
Purified recombinant BsuGid protein catalyzes the in vitro formation of m5U-54 in tRNA
The B.subtilis Gid protein was tagged with six histidine residues at the N-terminus, and purified to near homogeneity through affinity chromatography, either from E.coli Sure? strain (Figure 4A) or from E.coli trmA5 strain, both transformed with pQE80L-BsuGid. The purified protein is yellow and elutes from an S-200 gel filtration column at 85 kDa (data not shown), suggesting that the functional form of the enzyme may be a homodimer. Heating of the protein at 90°C releases yellow cofactor that has absorption (data not shown) and fluorescence spectra (Figure 4B) characteristic for oxidized flavins. This cofactor is likely FAD that was present in 0.8 mol per 1 mol of Gid from Myxococcus xanthus . Qualitative experiments indicated that the purified protein catalyzes the site-specific formation of m5U-54 in -labeled yeast tRNAAsp transcript (Figure 4C). Specific activity of the purified recombinant protein is low; nevertheless, this data reinforce observations obtained above by means of genetics. Some activity was observed without the addition of a carbon donor in the reaction mixture, suggesting that either a small amount of a carbon donor co-purifies with the enzyme or, alternatively, the purified enzyme contain tightly bound methylene or methyl intermediates. It is of note that the enzymatic test described here is highly sensitive, detecting even femtomolar amount of methylated uridine in the -radiolabeled substrate and has not been systematically optimized during this work.
Figure 4 (A) Electrophoretic analysis of purified recombinant BsuGid protein. An SDS–PAGE analysis was performed using 11% gels, stained with Coomassie blue. Lane 1, soluble proteins from sonicated total cell-free extract of E.coli pQE80L-BsuGid/SURE; lane 2, S10 fraction from E.coli pQE80L-BsuGid/SURE; lane 3, molecular weight markers; lane 4, proteins eluted from the immobilized metal ion adsorption chromatographic column. (B) Fluorescence spectrum of a cofactor released by heat denaturation from purified BsuGid protein. The observed emission maximum (after excitation at 450 nm) at 520 nm is typical for flavin nucleotides. (C) Time course of m5U-54 formation catalyzed by BsuGid. The molar ratio of m5U over total U in yeast tRNAAsp was evaluated over time at 37°C in the presence (closed circles) or absence (open circles) of CH2H4folate.
Taking together all the above information, we now propose to rename the Gid protein as TrmFO (FO for the folate) and the corresponding gene trmFO, in order to differentiate them from the conventional S-AdoMet-dependent TrmA enzyme and trmA gene.
TrmFO (Gid) and GidA proteins are two evolutionarily related families of proteins with distinct functions
TrmFO proteins of 50 kDa (designated Gid in Genome Data Banks) show 40% sequence similarity with another protein family referred to as GidA proteins (in reference to E.coli protein of 70 kDa) (Figure 5). The readily detectable sequence homology, together with the currently used name ‘small GidA’ for Gid proteins , has created confusion regarding the putative cellular functions of TrmFO. Our studies (see also below) have now revealed that in reality, paralogous TrmFO and GidA proteins are two distinct families of proteins that probably evolved from a common ancestor but acquired different, non-overlapping cellular functions during evolution.
Figure 5 Comparison of the amino acid sequences of Gid (TrmFO) and GidA proteins. GidA proteins are systematically longer than Gid proteins, having an additional C-terminal domain. Both proteins apparently bind FAD cofactors (42,47) (data not shown). The two proteins can be discriminated by a sequence motif partially overlapping with the FAD binding motif (‘motif 1’). An additional GidA-specific motif located at the C-terminus of the protein also distinguishes the two paralogs (‘motif 2’).
First, GidA proteins belong to a different cluster of orthologous genes (COG0445) and, in contrast to TrmFO proteins (belonging to COG1206), they are present in mitochondria of Eukarya and in Bacteria, with the exception of high GC% Gram-positive bacteria, such as Mycobacterium and Corynebacterium species (http://string.embl.de/).
Second, consensus motifs implicated in flavin binding are slightly different in the 64 TrmFO and 203 GidA sequences analyzed (for details, see Figure 5). Moreover, GidA proteins always have an extension (or extensive insertions) that includes an additional characteristic sequence motif at their C-termini.
Third, we demonstrated that B.subtilis TrmFO clearly methylates uridine-54 in the T-psi loop of tRNAs (see above), while E.coli GidA and S.cerevisiae MTO1 (a mitochondrial homolog of bacterial GidA) were shown to be involved in a completely different reaction, namely the multistep formation of the hypermodified uridines at position 34 of anticodon of a few selected tRNAs (44–48). Moreover, in agreement with the fact that the enzymatic activity of TrmFO and GidA is not overlapping, we found that the absence of U-54 methylating activity in the T-psi loop of tRNA of B.subtilis mutant strain BFS2838 (gidermR) does not affect the level of conversion of U-34 into cmnm5s2U-34 in the anticodon loop of B.subtilis tRNAs. This was demonstrated by testing the capability of cmnm5s2U-34 residues, present in naturally occurring tRNAs of both the B.subtilis wild-type strain 168 and the mutant strain BFS2838 (gidermR), to become fully modified in vitro into mnm5s2U-34 upon incubation with purified recombinant MnmC protein of E.coli. This protein is a bifunctional enzyme that is absent from B.subtilis. MnmC removes carboxymethyl group of cmnm5s2U-34 to produce nm5s2U-34 and further methylates it into mnm5s2U-34 (as in naturally occurring E.coli tRNAs) using S-AdoMet as a methyl donor (30). The autoradiographs in Figure 6 show that the formation of mnm5s2U-34 occurs equally well in the tRNAs of both the wild-type B.subtilis strain and the gidermR mutant, thus clearly indicating that absence of TrmFO activity does not interfere with the GidA-dependent formation of cmnm5s2U-34. Conversely, we also verified that purified recombinant B.subtilis GidA protein does not catalyze in vitro a U-54 methylation reaction under the experimental conditions used for m5U-54 formation catalyzed by CH2H4folate-dependent TrmFO (data not shown). Whether GidA proteins, similar to TrmFO enzymes, also act as methylases is currently unclear.
Figure 6 tRNAs from the gidermR mutant are substrates for the MnmC protein. (A and B) correspond to autoradiograms of a TLC analysis of nuclease P1-hydrolysate of bulk tRNAs, isolated from B.subtilis strain 168 (A) or from BFS2838 (gidermR) mutant (B), previously incubated with purified recombinant E.coli MnmC and -labeled AdoMet (see Materials and Methods) (30). Radiolabeled compounds were detected and quantified after 3 days exposure using a PhosphoImager. Results indicated that both bulk tRNAs are equally well methylated by MnmC showing that in both the cases, cmnm5U-34 was prevalent in the tRNAs.
DISCUSSION
The RNA methyltransferases (MTases) add methyl groups to the base or the ribose 2'-hydroxyl of ribonucleotides during the complex process of RNA maturation. The great majority of these MTases use S-AdoMet as methyl donor . However, at least in the case of m5U-54 formation in tRNA of certain organisms, N5, N10-methylenetetrahydrofolate, together with associated oxydo-reduction coenzyme FADH2, has been shown to serve the same purpose . This activity was first detected and the corresponding enzymes purified from S.faecalis almost three decades ago, but the gene encoding this folate-dependent activity had still not been identified. Here, we predicted, using prior experimental knowledge and phylogenetic distribution analyses, that Gid proteins, previously of unknown function (42,43), could correspond to such a folate-dependent tRNA methyltransferase. We have experimentally confirmed this prediction by showing that B.subtilis Gid protein (here renamed TrmFO, FO for the folate) is necessary and sufficient for ribothymidine-54 formation in the T-psi loop of tRNA both in vivo and in vitro.
Based on bioinformatics analyses, one surprising aspect of this work is that the phylogenetic distribution of the folate-dependent pathway appears much wider than originally anticipated. Nevertheless, it appears to be restricted to methylation of uridine-54 in tRNA, not m5U in rRNA as in the case of the S-AdoMet-dependent pathway (12,13). Strikingly, the folate-dependent TrmFO proteins (COG1206) and S-AdoMet-dependent TrmA/Trm2p enzymes (COG2265) acting on tRNA appear to have mutually exclusive phylogenetic distributions (Table 1). Note that the lack of a trmA ortholog in a given organism is difficult to ascertain as paralogous genes that participate in S-AdoMet-dependent methylation of rRNA are also present. For instance, the trmA/TRM2 gene is found in enterobacteriacae (including E.coli and pseudomonaceae) as well as in all Eukarya so far sequenced. In Archaea, the trmA/TRM2 homologs are only found in the Pyrococcus genus (13,37). These organisms do not contain a gene coding for a trmFO ortholog. In contrast, orthologs of trmFO are found in most Gram-positive bacteria (firmicutes and actinobacteria) and in several other bacterial groups (e.g. alpha and delta-proteobacteria, cyanobacteria, Table 1). Strikingly, a subset of bacteria, for instance most Mycoplasma species, seemingly lack either trmFO or trmA genes, suggesting that the uridine at position 54 in their tRNAs may not be methylated. Indeed, in tRNAs of M.capricolum, Mycoplasma mycoides and Mycobacterium smegmatis, for which the primary sequences (including modified nucleotides) are known (2,49), ribothymine-54 is indeed absent, and their bulk ribothymidine-less tRNAs can be used successfully as substrates for U-54 methylation in E.coli extracts (50). Interestingly, M.mycoides has two putative trmFO alleles whose functional role is unclear and is worth investigation. In contrast, thermophilic G.stearothermophilus displays an S-AdoMet-dependent activity for m5U modification in vitro (34,35), but no trmA gene or trmFO has been found in still uncompletely sequenced genome of this Gram-positive bacterium. This observation raises the possibility that one of the S-AdoMet-dependent rRNA MTase paralogs, which we have detected in non-annotated genome sequence of this species, could act as a tRNA methylase. This idea is further supported by an experimental observation indicating that a Pyrococcus abyssi protein highly similar to E.coli rRNA:m5U MTase RumA (13) is actually a site-specific methylase for U-54 in tRNA (J. Urbonavicius, S. Auxilien, K. Trachana and H. Grosjean, unpublished data). We also expect that in many bacteria containing TrmFO, the methylation of U-54 in their tRNAs depends on folate metabolism, while the formation of m5U in their rRNA is dependent on S-AdoMet, as it is in M.lysodeikticus (51). To construct a more comprehensive evolutionary history for this large family of m5U-forming enzymes, as well as other 5-methylpyrimidine MTases, such as those forming m5C in RNA and in DNA (52,53), experimental identification of the exact nucleotide target(s) within an RNA for each of these MTases is needed.
Table 1 Non-exhaustive distribution of the trmAa and trmFO (gid) coding for putative tRNA:m5U-54 methyltransferases in bacteria
In this work, we also considered the possible evolutionary relationship between ribothymidylate synthase TrmFO and the thymidylate synthase ThyX family of enzymes that catalyze a very similar reaction (19). Although both ThyX and TrmFO proteins use flavin (FADH2) nucleotide as cofactor, our studies have indicated that they do not belong to the same family of flavoproteins. In particular, TrmFO proteins lack the characteristic conserved residues required for catalysis, substrate and/or cofactor binding of ThyX proteins (54). In addition, the novel FAD binding fold found in ThyX proteins does not show significant similarity to the classical Rossman-fold predicted for TrmFO proteins (21,38). Thus, in the light of this information, a common origin for TrmFO and ThyX proteins appears unlikely. Therefore, direct comparison of the reaction mechanisms between TrmFO and ThyX proteins cannot be done. Our work suggests, for the first time, that the use of CH2H4folate and FAD in the post-transcriptional methylation of polynucleotides (pre-tRNA) or of a mononucleotide (dUMP) during DNA precursor synthesis has been established independently at least twice during evolution.
ACKNOWLEDGEMENTS
G. Bj?rk, S. Seror, L. Droogmans and R. Moser are acknowledged in the text. K. Trachana is acknowledged for her help in cell extract preparation. All HPLC analysis were performed by K. Jacobsson at the University of Ume?, Sweden (grants from Swedish Cancer Foundation—Project 680—and Swedish Science Research Council—Project BU2930—to G. Bj?rk). The authors thank S. Douthwaite (University of South Denmark, Odense, Denmark), S. Auxilien, B. Golinelli-Pimpaneau (CNRS, Gif-sur-Yvette), B. Holland (University of Paris XI) and V. de Crécy-Lagard (University of Florida, Gainesville, FL, USA) for useful advices and critical reading of the manuscript. H.G. benefits from a research grant from the CNRS (GEOMEX program). H.M. is supported by research funds from INSERM (BIOAVENIR program), CNRS (Program Microbiologie Fondamental) and Fondation Bettencourt-Schueller. J.U. and S.S. are FEBS Postdoctoral Fellow and INSERM Young Researcher, respectively. Funding to pay the Open Access publication charges for this article was provided by CNRS, Programme de Microbiologie Fondamentale to H.M.
REFERENCES
Limbach, P.A., Crain, P.F., McCloskey, J.A. (1994) Summary: the modified nucleosides of RNA Nucleic Acids Res., 22, 2183–2196 .
Sprinzl, M. and Vassilenko, K.S. (2005) Compilation of tRNA sequences and sequences of tRNA genes Nucleic Acids Res., 33, D139–D140 .
Shigi, N., Suzuki, T., Tamakoshi, M., Oshima, T., Watanabe, K. (2002) Conserved bases in the TPsiC loop of tRNA are determinants for thermophile-specific 2-thiouridylation at position 54 J. Biol. Chem., 277, 39128–39135 .
Fleissner, E. and Borek, E. (1962) A new enzyme of RNA synthesis: RNA methylase Proc. Natl Acad. Sci. USA, 48, 1199–1203 .
Svensson, I., Boman, H.G., Eriksson, K.G., Kjellin, K. (1963) Studies on microbial RNA. I. Transfer of methyl groups from methionine to soluble RNA from Escherichia Coli J. Mol. Biol., 16, 254–271 .
Bjork, G.R. (1975) Transductional mapping of gene trmA responsible for the production of 5-methyluridine in transfer ribonucleic acid of Escherichia coli J. Bacteriol., 124, 92–98 .
Ny, T. and Bjork, G.R. (1980) Cloning and restriction mapping of the trmA gene coding for transfer ribonucleic acid (5-methyluridine)-methyltransferase in Escherichia coli K-12 J. Bacteriol., 142, 371–379 .
Kealey, J.T., Gu, X., Santi, D.V. (1994) Enzymatic mechanism of tRNA (m5U54)methyltransferase Biochimie, 76, 1133–1142 .
Fauman, E.B., Blumenthal, R.M., Cheng, X. (1999) Structure and evolution of AdoMet-dependent methyltransferases In Cheng, X. and Blumenthal, R.M. (Eds.). S-Adenosylmethionine-Dependent Methyltransferases: Structures and Functions, Singapore World Scientific pp. 1–38 .
Bujnicki, J.M., Droogmans, L., Grosjean, H., Purushothaman, S.K., Lapeyre, B. (2004) Bioinformatics-guided identification and experimental characterization of novel RNA methyltransferases In Bujnicki, J.M. (Ed.). Nucleic Acids and Molecular Biology: Practical Bioinformatics series, Berlin, Heidelberg Springer-Verlag Vol. 15, pp. 139–168 .
Tatusov, R.L., Fedorova, N.D., Jackson, J.D., Jacobs, A.R., Kiryutin, B., Koonin, E.V., Krylov, D.M., Mazumder, R., Mekhedov, S.L., Nikolskaya, A.N., et al. (2003) The COG database: an updated version includes eukaryotes BMC Bioinformatics, 4, 41 .
Agarwalla, S., Kealey, J.T., Santi, D.V., Stroud, R.M. (2002) Characterization of the 23 S ribosomal RNA m5U1939 methyltransferase from Escherichia coli J. Biol. Chem., 277, 8835–8840 .
Madsen, C.T., Mengel-Jorgensen, J., Kirpekar, F., Douthwaite, S. (2003) Identifying the methyltransferases for m(5)U747 and m(5)U1939 in 23S rRNA using MALDI mass spectrometry Nucleic Acids Res., 31, 4738–4746 .
Delk, A.S., Nagle, D.P., Jr, Rabinowitz, J.C. (1980) Methylenetetrahydrofolate-dependent biosynthesis of ribothymidine in transfer RNA of Streptococcus faecalis. Evidence for reduction of the 1-carbon unit by FADH2 J. Biol. Chem., 255, 4387–4390 .
Samuel, C.E. and Rabinowitz, J.C. (1974) Initiation of protein synthesis by folate-sufficient and folate-deficient Streptococcus faecalis R: partial purification and properties of methionyl-transfer ribonucleic acid synthetase and methionyl-transfer ribonucleic acid formyltransferase J. Bacteriol., 118, 21–31 .
Arnold, H.H. and Kersten, H. (1975) Inhibition of the tetrahydrofolate-dependent biosynthesis of ribothymidine in tRNAs of B.subtilis and M.lysodeikticus by trimethoprim FEBS Lett., 53, 258–261 .
Delk, A.S., Romeo, J.M., Nagle, D.P., Jr, Rabinowitz, J.C. (1976) Biosynthesis of ribothymidine in the transfer RNA of Streptococcus faecalis and Bacillus subtilis. A methylation of RNA involving 5,10-methylenetetrahydrofolate J. Biol. Chem., 251, 7649–7656 .
Delk, A.S., Nagle, D.P., Jr, Rabinowitz, J.C., Straub, K.M. (1979) The methylenetetrahydrofolate-mediated biosynthesis of ribothymidine in the transfer-RNA of Streptococcus faecalis: incorporation of hydrogen from solvent into the methyl moiety Biochem. Biophys. Res. Commun., 86, 244–251 .
Myllykallio, H., Lipowski, G., Leduc, D., Filee, J., Forterre, P., Liebl, U. (2002) An alternative flavin-dependent mechanism for thymidylate synthesis Science, 297, 105–107 .
Giladi, M., Bitan-Banin, G., Mevarech, M., Ortenberg, R. (2002) Genetic evidence for a novel thymidylate synthase in the halophilic archaeon Halobacterium salinarum and in Campylobacter jejuni FEMS Microbiol. Lett., 216, 105–109 .
Mathews, II, Deacon, A.M., Canaves, J.M., McMullan, D., Lesley, S.A., Agarwalla, S., Kuhn, P. (2003) Functional analysis of substrate and cofactor complex structures of a thymidylate synthase-complementing protein Structure (Camb.), 11, 677–690 .
Finer-Moore, J.S., Santi, D.V., Stroud, R.M. (2003) Lessons and conclusions from dissecting the mechanism of a bisubstrate enzyme: thymidylate synthase mutagenesis, function, and structure Biochemistry, 42, 248–256 .
Agrawal, N., Lesley, S.A., Kuhn, P., Kohen, A. (2004) Mechanistic studies of a flavin-dependent thymidylate synthase Biochemistry, 43, 10295–10301 .
Graziani, S., Xia, Y., Gurnon, J.R., Van Etten, J.L., Leduc, D., Skouloubris, S., Myllykallio, H., Liebl, U. (2004) Functional analysis of FAD-dependent thymidylate synthase ThyX from Paramecium bursaria Chlorella virus-1 J. Biol. Chem., 279, 54340–54347 .
Gattis, S.G. and Palfey, B.A. (2005) Direct observation of the participation of flavin in product formation by thyX-encoded thymidylate synthase J. Am. Chem. Soc., 127, 832–833 .
Griffin, J., Roshick, C., Iliffe-Lee, E., McClarty, G. (2005) Catalytic mechanism of Chlamydia trachomatis flavin-dependent thymidylate synthase J. Biol. Chem., 280, 5456–5467 .
Perret, V., Garcia, A., Puglisi, J., Grosjean, H., Ebel, J.P., Florentz, C., Giege, R. (1990) Conformation in solution of yeast tRNA(Asp) transcripts deprived of modified nucleotides Biochimie, 72, 735–743 .
Grosjean, H., Droogmans, L., Giege, R., Uhlenbeck, O.C. (1990) Guanosine modifications in runoff transcripts of synthetic transfer RNA-Phe genes microinjected into Xenopus oocytes Biochim. Biophys. Acta, 1050, 267–273 .
Grosjean, H., Keith, G., Droogmans, L. (2004) Detection and quantification of modified nucleotides in RNA using thin-layer chromatography Methods Mol. Biol., 265, 357–391 .
Bujnicki, J.M., Oudjama, Y., Roovers, M., Owczarek, S., Caillet, J., Droogmans, L. (2004) Identification of a bifunctional enzyme MnmC involved in the biosynthesis of a hypermodified uridine in the wobble position of tRNA RNA, 10, 1236–1242 .
von Ehrenstein, G. (1967) Isolation of sRNA from intact Escherichia coli cells Meth. Enzymol., 12, 588–596 .
Gehrke, C.W. and Kuo, K.C. (1989) Ribonucleoside analysis by reversed-phase high-performance liquid chromatography J. Chromatogr., 471, 3–36 .
Delk, A.S., Nagle, D.P., Jr, Rabinowitz, J.C. (1979) Purification of methylenetetrahydrofolate-dependent methyltransferase catalyzing biosynthesis of ribothymidine in transfer RNA of Streptococcus faecalis In Kisliuk, R.L. and Brown, G.M. (Eds.). Chemistry and Biology of Pteridines, NY Elsevier/North Holland Publishing pp. 389–394 .
Agris, P.F., Koh, H., Soll, D. (1973) The effect of growth temperatures on the in vivo ribose methylation of Bacillus stearothermophilus transfer RNA Arch. Biochem. Biophys., 154, 277–282 .
Schmidt, W., Arnold, H.H., Kersten, H. (1977) Tetrahydrofolate-dependent biosynthesis of ribothymidine in transfer ribonucleic acids of Gram-positive bacteria J. Bacteriol., 129, 15–21 .
Nordlund, M.E., Johansson, J.O., von Pawel-Rammingen, U., Bystrom, A.S. (2000) Identification of the TRM2 gene encoding the tRNA(m5U54)methyltransferase of Saccharomyces cerevisiae RNA, 6, 844–860 .
Constantinesco, F., Motorin, Y., Grosjean, H. (1999) Transfer RNA modification enzymes from Pyrococcus furiosus: detection of the enzymatic activities in vitro Nucleic Acids Res., 27, 1308–1315 .
Dym, O. and Eisenberg, D. (2001) Sequence-structure analysis of FAD-containing proteins Protein Sci., 10, 1712–1728 .
Altschul, S.F., Madden, T.L., Schaffer, A.A., Zhang, J., Zhang, Z., Miller, W., Lipman, D.J. (1997) Gapped BLAST and PSI-BLAST: a new generation of protein database search programs Nucleic Acids Res., 25, 3389–3402 .
Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F., Higgins, D.G. (1997) The CLUSTAL_X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools Nucleic Acids Res., 25, 4876–4882 .
Bjork, G.R. and Isaksson, L.A. (1970) Isolation of mutants of Escherichia coli lacking 5-methyluracil in transfer ribonucleic acid or 1-methylguanine in ribosomal RNA J. Mol. Biol., 51, 83–100 .
White, D.J., Merod, R., Thomasson, B., Hartzell, P.L. (2001) GidA is an FAD-binding protein involved in development of Myxococcus xanthus Mol. Microbiol., 42, 503–517 .
Iwasaki, W., Miyatake, H., Ebihara, A., Miki, K. (2004) Crystallization and preliminary X-ray crystallographic studies of the small form of glucose-inhibited division protein A from Thermus thermophilus HB8 Acta Crystallogr. D Biol. Crystallogr., 60, 515–517 .
Colby, G., Wu, M., Tzagoloff, A. (1998) MTO1 codes for a mitochondrial protein required for respiration in paromomycin-resistant mutants of Saccharomyces cerevisiae J. Biol. Chem., 273, 27945–27952 .
Bregeon, D., Colot, V., Radman, M., Taddei, F. (2001) Translational misreading: a tRNA modification counteracts a +2 ribosomal frameshift Genes Dev., 15, 2295–2306 .
Li, R., Li, X., Yan, Q., Qin Mo, J., Guan, M.X. (2003) Identification and characterization of mouse MTO1 gene related to mitochondrial tRNA modification Biochim. Biophys. Acta, 1629, 53–59 .
Scrima, A., Vetter, I.R., Armengod, M.E., Wittinghofer, A. (2005) The structure of the TrmE GTP-binding protein and its implications for tRNA modification EMBO J., 24, 23–33 .
Umeda, N., Suzuki, T., Yukawa, M., Ohya, Y., Shindo, H., Watanabe, K. (2005) Mitochondria-specific RNA-modifying enzymes responsible for the biosynthesis of the wobble base in mitochondrial tRNAs. Implications for the molecular pathogenesis of human mitochondrial diseases J. Biol. Chem., 280, 1613–1624 .
Andachi, Y., Yamao, F., Muto, A., Osawa, S. (1989) Codon recognition patterns as deduced from sequences of the complete set of transfer RNA species in Mycoplasma capricolum. Resemblance to mitochondria J. Mol. Biol., 209, 37–54 .
Johnson, L., Hayashi, H., Soll, D. (1970) Isolation and properties of a transfer ribonucleic acid deficient in ribothymidine Biochemistry, 9, 2823–2831 .
Schmidt, W., Arnold, H.H., Kersten, H. (1975) Biosynthetic pathway of ribothymidine in B.subtilis and M.lysodeikticus involving different coenzymes for transfer RNA and ribosomal RNA Nucleic Acids Res., 2, 1043–1051 .
Anantharaman, V., Koonin, E.V., Aravind, L. (2002) Comparative genomics and evolution of proteins involved in RNA metabolism Nucleic Acids Res., 30, 1427–1464 .
Bujnicki, J.M., Feder, M., Ayres, C.L., Redman, K.L. (2004) Sequence-structure-function studies of tRNA:m5C methyltransferase Trm4p and its relationship to DNA:m5C and RNA:m5U methyltransferases Nucleic Acids Res., 32, 2453–2463 .
Leduc, D., Graziani, S., Lipowski, G., Marchand, C., Le Marechal, P., Liebl, U., Myllykallio, H. (2004) Functional evidence for active site location of tetrameric thymidylate synthase X at the interphase of three monomers Proc. Natl Acad. Sci. USA, 101, 7252–7257 .
Hall, R.H. The modified nucleosides in nucleic acids, (1971) NY and London Columbia University Press .(Jaunius Urbonaviius, Stéphane Skouloubri)