Transfer RNA Methyltransferases from Thermoplasma acidophilum, a Thermoacidophilic Archaeon

We investigated tRNA methyltransferase activities in crude cell extracts from the thermoacidophilic archaeon Thermoplasma acidophilum. We analyzed the modified nucleosides in native initiator and elongator tRNAMet, predicted the candidate genes for the tRNA methyltransferases on the basis of the tRNAMet and tRNALeu sequences, and characterized Trm5, Trm1 and Trm56 by purifying recombinant proteins. We found that the Ta0997, Ta0931, and Ta0836 genes of T. acidophilum encode Trm1, Trm56 and Trm5, respectively. Initiator tRNAMet from T. acidophilum strain HO-62 contained G+, m1I, and m22G, which were not reported previously in this tRNA, and the m2G26 and m22G26 were formed by Trm1. In the case of elongator tRNAMet, our analysis showed that the previously unidentified G modification at position 26 was a mixture of m2G and m22G, and that they were also generated by Trm1. Furthermore, purified Trm1 and Trm56 could methylate the precursor of elongator tRNAMet, which has an intron at the canonical position. However, the speed of methyl-transfer by Trm56 to the precursor RNA was considerably slower than that to the mature transcript, which suggests that Trm56 acts mainly on the transcript after the intron has been removed. Moreover, cellular arrangements of the tRNA methyltransferases in T. acidophilum are discussed.


Introduction
Thermoplasma acidophilum is a thermoacidophilic archaeon that grows optimally at 59 °C and pH 1.9 [1]. The characteristic property of this archaeon is that the cells are very irregular in shape due to the lack of a cell wall [2,3]. Despite this, the cytoplasmic membrane tolerates an acidic environment at high temperatures. Consequently, components of the membrane have been studied in detail [4]. Furthermore, lipoylation of proteins [5], biosynthesis of lipids [6], cell surface glycoproteins [7] and a channel protein [8] have been also studied in T. acidophilum. Given that a prokaryotic histone-like DNA binding protein was discovered first from T. acidophilum [9,10], the bacterium has been used as a model system to investigate DNA replication, DNA repair, and transcriptional initiation in archaea [11][12][13][14][15][16]. Furthermore, the energy metabolism of T. acidophilum has been studied in detail because it can grow under an extreme microaerophilic environment [17,18]. Genome sequencing elucidated that the T. acidophilum genome encodes only approximately 1500 open reading frames [19]. Consequently, large protein complexes such as the proteasome and chaperonin are composed of a relatively limited number of protein subunits, and thus they have been studied and compared with their more complicated counterparts from eukaryotes [20][21][22][23].
Although T. acidophilum has been investigated from various viewpoints as described above, there is little knowledge about tRNA modifications, with the exception of some early studies [24][25][26][27] and our more recent work [28]. In 1981 and 1982, the sequences of the initiator ( [25] and Figure 1A) and elongator tRNA Met ( [24] and Figure 1B) were determined. A novel modification at position 15 (N in Figure 1B), which was named later as archaeosine at position 15 (G + 15) [29], and the typical archaeal tRNA modification of 2'-O-methylcytidine at position 56 (Cm56) [30] were reported. In 1991, Edmonds et al. [27] reported that a mixture of tRNAs from T. acidophilum contains N 7 -methylguanine (m 7 G). In general, the m 7 G modification is found at position 46 in class I tRNAs from eubacteria and eukaryotes [31][32][33][34][35]; class I tRNAs are defined as tRNAs with a variable region of regular size. To verify the location of the m 7 G modification in the tRNA, we analyzed tRNA modifications in T. acidophilum [28]. Unexpectedly, we found that the m 7 G modification was present at a novel position, nucleotide 49 in class II tRNA Leu ( Figure 1C); class II tRNAs have a long variable region. Furthermore, we found several distinct modifications in this tRNA Leu ( Figure 1C): 4-thiouridine at position 9 (s 4 U9) [36][37][38][39], G + 13 [29,40], and 5-carbamoylmethyluridine at position 34 (ncm 5 U34) [41,42]. The modifications s 4 U9 and ncm 5 U34 have been not found in other archaeal tRNAs and G + 13 has not been reported in any other tRNA [31,32]. In the current study, we tested the tRNA methyltransferase activities in crude cell extract from T. acidophilum, analyzed the methylated nucleosides in purified tRNAs and characterized the tRNA methyltransferases by expressing recombinant proteins in Escherichia coli. The regions to which the 3'-biotinylated DNA probes hybridize are illustrated. The abbreviations are as follows: pseudouridine, ψ; N 2 -methylguanosine, m 2 G; 2'-O-methylcytidine, Cm; 1-methyladenosine, m 1 A; 4-thiouridine, s 4 U; unknown modification, N; unidentified G modification, G*; N 6 -thereonylcarbamoyladenosine, t 6 A; archaeosine, G + ; N 2 ,N 2 -dimethylguanosine, m 2 2G; 5-carbamoylmethyluridine, ncm 5 U; 1-methylguanosine, m 1 G; 7-methylguanosine, m 7 G. The precursor of elongator tRNA Met contains an intron at the canonical site (D). Triangles show the cleavage sites of tRNA splicing endonuclease. Figure 1A-C show the cloverleaf structures of initiator tRNA Met , elongator tRNA Met and tRNA Leu UAG, respectively. In the current study, we utilized these tRNA sequences as a basis to characterize tRNA methyltransferases in T. acidophilum. As shown in Table 1, various methylated nucleosides are present in these tRNAs. We predicted the candidate genes for the enzymes responsible for the modifications by BLAST searches [28] and analyzed the corresponding recombinant proteins. For example, given that the G26 in elongator tRNA Met was reported to be a modified G (G*) ( [24] and Figure 1B) we investigated whether recombinant Trm1 methylated G26 in the elongator tRNA Met transcript. The results of the current study are summarized in Table 1. Furthermore, it should be mentioned that the sequence of the initiator tRNA Met that is encoded in the genome of T. acidophilum strain HO-62 differs from that reported in the earlier study ( [25] and Figure 1A): the nucleotide at position 57 is A instead of G in strain HO-62. In archaeal tRNAs, A57 is often modified to 1-methylinosine (m 1 I57) via 1-methyladenosine (m 1 A57) by TrmI and deamination [43][44][45]. Some possible explanations for this discrepancy are elaborated in the Discussion section. Moreover, in the case of the precursor of elongator tRNA Met , a standard intron is inserted at the canonical position between nucleotides 38 and 39 ( Figure 1D). Therefore, it is possible that the presence of the intron might affect the methylations by tRNA methyltransferases.  3 We could not detect m 1 A58 formation activity in the recombinant protein. 1 During the course of the current study, it has been reported that Sulfolobus acidocaldarius TrmJ generates the Cm32 modification in tRNA. The candidate gene in T. acidophilum was predicted by a BLAST search; 2 The sequence of the initiator tRNA Met that is encoded in the genome of the T. acidophilum strain HO-62 differs from that reported in the earlier study [25] (see Figures 1A and 6A): the nucleotide at position 57 is A instead of G in strain HO-62. Furthermore, the tRNA gene in the genome of strain HO-62 contains additional nucleotides, A20b and C22; 3 The Ta0852 gene product was expressed in Escherichia coli as a soluble protein.

Methylated Nucleosides in T. acidophilum tRNAs
However, we could not detect any ability to form m 1 A58; ?, the corresponding enzyme is unknown.

Transfer RNA Methyltransferase Activities in the Crude Cell Extract
Next, we tested tRNA methyltransferase activities in crude extract from T. acidophilum cells. The supernatant fraction from centrifugation at 30,000× g (S-30) was prepared and then the tRNA Leu UAG transcript was subjected to methylation by the S-30 extract with 14 C-S-adenosyl-L-methionine (AdoMet) as the methyl group donor. The methylated tRNA was digested completely with nuclease P1 and then the resultant 14 C-methylated nucleotides were analyzed by two-dimensional thin-layer chromatography (2D-TLC). As shown in Figure 2A, four 14 C-methylated nucleotides (pm 1 G, pm 2 G, pm 2 2G and pm 6 A) could be detected. On the basis of the sequence of tRNA Leu UAG ( Figure 1C) and the candidate enzymes (Table 1), pm 1 G, pm 2 G and pm 2 2G, and pm 6 A were expected to be derived from the activities of Trm5, Trm1, and TrmI, respectively: pm 6 A could be converted from pm 1 A non-enzymatically [46]. However, unexpectedly, pCm and pm 7 G were not detected. In general, the formation of pCm by Trm56 is one of the most common tRNA methyltransferase activities found in crude extract from archaeal cells. For example, it was reported that Trm56 activity in relation to the formation of pCm56 is clearly detected in cell extract from Pyrococcus furiosus [47]. Analysis of the T. acidophilum proteome revealed that various proteins form several large (more than 300 kDa) protein complexes and that some protein complexes might interact with the membrane [48]. Consequently, Trm56 and an unknown tRNA (m 7 G49) methyltransferase might be included in protein complexes and precipitated by centrifugation at 30,000× g. We tested several buffer conditions such as variations in pH, components, detergents, and salt concentrations (data not shown). However, to date, we have not detected enzyme activities responsible for the formation of pCm and pm 7 G in crude extract from T. acidophilum cells, though there are other possibilities; while we used aluminum oxide to prepare the extract in this experiment, other methods for preparation of cell extract should be tested. In addition, tRNA methyltransferases in the crude extract have different affinities for 14 C-AdoMet. The concentration of 14 C-AdoMet in the experiment was 19.5 µM. In this case, tRNA methyltransferases, which have relatively high affinity for AdoMet, might preferentially consume the 14 C-AdoMet. When the supernatant fraction from centrifugation at 100,000× g (S-100) was used as the cell extract instead of the S-30 fraction, the findings were even more marked: only the formation of pm 1 G was detectable ( Figure 2B). Given that tRNA methyltransferases have a general affinity for RNA, the enzymes often bind to ribosomes and are precipitated by centrifugation at 100,000× g. In fact, the majority of TrmI from Thermus thermophilus [49] is precipitated by centrifugation at 100,000× g [50]. However, our findings with the extract from T. acidophilum are unprecedented. In the current study, we characterized tRNA methyltransferases by analyzing purified recombinant proteins. However, these enzymes might interact with other proteins and form large protein complexes in living cells.  14 C-methylated nucleotides were monitored by autoradiography. The solvent systems were as follows: first dimension, isobutyric acid, ammonia, water, 66

Modified Nucleosides in Purified tRNAs
To analyze the modified nucleosides in tRNAs, we purified initiator and elongator tRNA Met by the solid-phase DNA probe method [51], in which tetraalkylammonium salts were used in the hybridization buffer. Tetraalkylammonium salts destabilize the tRNA structure and enhance the formation of DNA-RNA hybrids [51], and thus we were able recently to purify single tRNA species from thermophiles such as Aquifex aeolicus [52], T. thermophilus [50,53], Aerophyrum pernix [54], and T. acidophilum [28] using this approach. In the case of initiator tRNA Met , because the sequence from G15 to U36 was distinct, we designed the DNA probe to this region. In the case of elongator tRNA Met , T. acidophilum two species: the sequence of one was previously determined as shown in Figure 1B [24]. These two tRNA Met species differ in three nucleotides in the D-loop and anticodon arm. Therefore, we designed the DNA probe as shown in Figure 1B. As shown in the insets in Figure 3A,B, initiator and elongator tRNA Met were purified successfully: the 14 C-Met-charging activities were checked by the S-100 fraction (data not shown). The purified tRNAs were digested with snake venom phosphodiesterase, RNase A and bacterial alkaline phosphatase, and then the resultant nucleosides were analyzed by HPLC using a C18 column ( Figure 3A,B). Snake venom phosphodiesterase can cleave the phosphodiester bond adjacent to 2'-O-methylated nucleotide. As shown in Figure 3A, m 1 A, Cm, m 1 I, m 2 G, m 2 2G and m 6 A were detected as methylated nucleosides in the initiator tRNA Met sample. The presence of m 1 A, Cm, m 2 G and m 6 A is consistent with the published RNA sequence ( Figure 1A). However, the presence of m 1 I suggests that A57 in this tRNA is modified to m 1 A57 by TrmI [44,45] and that deamination then generates m 1 I57 as in the case of Haloferax volcanii [43]. Furthermore, a peak for m 2 2G was detected, which suggests that some proportion of the initiator tRNA Met contains m 2 2G26. Moreover, G + was clearly detected, which suggests that G15 is modified to G + 15 in initiator tRNA Met . In the elongator tRNA Met sample, m 1 A, Cm, m 2 G, m 2 2G and m 6 A were detected. The modifications m 2 G and m 2 2G have not been reported at any position in elongator tRNA Met , although the uncharacterized G26 modification represents a possible location (G*26 in Figure 1B and Table 1). Consequently, the modified G26 was expected to be a mixture of m 2 G26 and m 2 2G26. Furthermore, N 6 -threonylcarbamoyladenosine (t 6 A) was also detected, which is consistent with the RNA sequence ( Figure 1B). After these pilot experiments, we established expression systems in E. coli for the candidate genes shown in Table 1. As mentioned in Table 1, we could not obtain soluble protein from the Ta0679 gene. Furthermore, the Ta0852 gene product did not show TrmI activity. Consequently, these gene products were not analyzed further in the current study.  Figure 1A,B. The modified nucleosides in the initiator (A) and elongator (B) tRNA Met were analyzed by reverse phase column chromatography.

Formation of m 1 G37 in tRNA Leu UAG Transcript by Ta0836 Gene Product
The Ta0836 gene product was expressed in E. coli and purified as shown in Figure 4A. The expected amino acid sequence of the Ta0836 gene product shares a high degree of homology (82%) with that of the identified archaeal Trm5 (Mj0883 of Methanocaldococcus jannaschii) [55,56]. When the purified protein was incubated with the tRNA Leu UAG transcript and 14 C-AdoMet, the 14 C-methyl group was clearly incorporated into the transcript (see Figure 4C lane 1). Analysis of the modified nucleotides by 2D-TLC revealed that the 14 C-methylated nucleotide was pm 1 G ( Figure 4B). Furthermore, when the G37 in tRNA Leu UAG transcript was replaced by A, no methyl group incorporation was observed ( Figure 4C lane 2), which indicates that the methylation site is G37. From these results, we concluded that the Ta0836 gene product is the T. acidophilum Trm5 protein. Given that the m 1 G modification was previously found only at position 37 in tRNA Leu UAG in T. acidophilum ( [28] and Figure 1C), the m 1 G modification activity in the S-100 is probably derived from Trm5. The tRNA Leu UAG transcript was methylated by the Ta0836 gene product and then the generated methylated nucleotide was analyzed by 2D-TLC; (C) The methyl group acceptance activities of the wild-type tRNA Leu UAG transcript (lane 1) and the mutant tRNA Leu UAG transcript (lane 2), in which G37 was replaced by A, were investigated. The transcripts were individually incubated with the Ta0836 gene product and 14 C-AdoMet, and then separated by 10% PAGE (7 M urea). The gel was stained with methylene blue (left panel) and the autoradiogram of the same gel was taken (right panel).

Ta0997 Gene Product Is a Single Site-Specific Trm1
The m 2 2G modification was observed only at position 26 in tRNA Leu UAG ( [28] and Figure 1C). Trm1 transfers two methyl groups to the 2-amino group in the target guanine and m 2 G is formed as an intermediate [57,58]. Consequently, the m 2 G and m 2 2G modifications in tRNA Leu UAG transcript by the S-30 ( Figure 2A) were expected to be derived from Trm1 activity. Trm1 enzymes can be divided into two types on the basis of their specificity for the target guanosine(s). One is a single-site-specific Trm1, which modifies only G26 and is found in eukaryotes and archaea [57][58][59][60]. The second is a multi-site-specific Trm1, which modifies both G26 and G27 and is found in the hyperthermophilic eubacterium, A. aeolicus [52]. In addition to the Ta0997 gene product ( Figure 5A lane 1), we prepared two types of Trm1 enzyme from Thermococcus kodakarensis ( Figure 5A lane 2) and A. aeolicus ( Figure 5A lane 3) as controls. The Ta0997 gene product methylated the tRNA Leu UAG transcript (data not shown) and 14 C-nucleotide analysis revealed that the modified nucleotide was pm 2 2G ( Figure 5B). These results showed that the Ta0997 gene product is the T. acidophilum Trm1 protein. To distinguish the site specificity, yeast tRNA Phe and A. aeolicus tRNA Tyr transcripts were prepared ( Figure 5C). These tRNA transcripts were used previously to assess the site specificity of A. aeolicus Trm1 [52]. Yeast tRNA Phe contains the sequence G26C27, whereas A. aeolicus tRNA Tyr contains the sequence A26G27. In addition, the A. aeolicus tRNA Tyr A26G, G27A mutant transcript has the sequence G26A27. As shown in Figure 5D, T. acidophilum Trm1 methylated yeast tRNA Phe and A. aeolicus tRNA Tyr A26G, G27A transcripts, which contain G26. In contrast, T. acidophilum Trm1 did not methylate the wild-type tRNA Tyr transcript ( Figure 5D center), which contains A26. Thus, these results demonstrate that T. acidophilum Trm1 is a single-site-specific Trm1, which methylates only G26. Similar to T. acidophilum Trm1, T. kodakarensis Trm1 methylated only the yeast tRNA Phe and A. aeolicus tRNA Tyr A26G, G27A transcripts ( Figure 5E). In contrast, A. aeolicus Trm1 methylated all the transcripts ( Figure 5F), which indicates that A. aeolicus Trm1 has multi-site specificity.

T. acidophilum Trm1 Can Modify G26 in Initiator tRNA Met Transcript to m 2 2G26 via m 2 G26
The G26 modification in initiator tRNA Met was reported to be m 2 G ( [25] and Table 1). Consequently, we investigated whether T. acidophilum Trm1 can modify G26 in the initiator tRNA Met transcript to m 2 2G. Given that the 5'-end of initiator tRNA Met is an A, T7 RNA polymerase did not synthesize the transcript efficiently. Consequently, the initiator tRNA Met transcript was synthesized with a 5'-leader sequence ( Figure 6A) by T7 RNA polymerase ( Figure 6B, lane 1) and then the 5'-leader sequence was removed with E. coli RNase P ( [61] and Figure 6B, lane 2). The initiator tRNA Met transcript was then purified by 10% polyacrylamide gel electrophoresis in the presence of 7 M urea (PAGE (7 M urea)), ( Figure 6B lane 3). Trm1 from T. acidophilum efficiently methylated the initiator tRNA Met transcript (data not shown), and the methylated nucleotide was pm 2 2G ( Figure 6C). Furthermore, the modified nucleoside analysis showed that native initiator tRNA Met contained m 2 2G ( Figure 3A). Taking these results together, we conclude that initiator tRNA Met from T. acidophilum contained the m 2 2G26 modification in addition to m 2 G26 and that TrmI activity was responsible for these modifications.

T. acidophilum Trm1 Can Methylate the Precursor of Elongator tRNA Met with an Intron
The G26 modification of elongator tRNA Met was uncharacterized ( [24], and Table 1). Furthermore, the precursor of elongator tRNA Met contains an intron at the canonical position between nucleotides 38 and 39 ( Figure 1D). To verify whether Trm1 could methylate the elongator tRNA Met transcript and its precursor, we analyzed the methyl group acceptance activities of these RNAs ( Figure 7A). Both the elongator tRNA Met transcript and its precursor were methylated efficiently by T. acidophilum Trm1. The analysis of modified nucleotides by 2D-TLC revealed that methylated nucleotides were pm 2 G and pm 2 2G ( Figure 7B). The analysis of modified nucleosides revealed that native elongator tRNA Met contained m 2 G and m 2 2G ( Figure 3B). Furthermore, m 2 G and m 2 2G modifications were not reported in the published sequence of elongator tRNA Met although the modification at G26 was uncharacterized [24]. Taking these results together, we conclude that the uncharacterized G26 modification in the elongator tRNA Met is a mixture of m 2 G and m 2 2G, which is formed by Trm1.

Ta0931 Gene Product Is Trm56
To analyze whether the Ta0931 gene product was Trm56, the recombinant protein was purified as shown in Figure 8A. The purified Ta0931 methylated the tRNA Leu UAG transcript (data not shown) and the methylated nucleotide was identified as pCm ( Figure 8B). The Cm modification is only found at position 56 in native tRNA Leu UAG ( Figure 1C). These results showed that the Ta0931 gene product is Trm56. Neither the S-30 nor the S-100 fraction contained activity that was responsible for introducing the Cm modification into the tRNA Leu UAG transcript ( Figure 2); however, the genome does encode Trm56. This discrepancy is addressed in the Discussion section. Finally, we investigated the influence of the presence of intron on Trm56 activity. Trm56 methylated both the elongator tRNA Met transcript and its precursor ( Figure 8C). However, the methyl group acceptance activity of the precursor was considerably lower than that of the mature transcript ( Figure 8D). It should be mentioned that the incubation in Figure 8C was performed for 12 h to show the methylation of the precursor tRNA. These results suggest that the methylation by Trm56 occurs mainly after the removal of the intron. Although the mechanism by which Trm56 recognizes tRNA has not been reported thus far, the results of the current study suggest two possibilities. The first is that the presence of the intron results in steric hindrance that prevents Trm56 binding to the substrate tRNA. The second is that Trm56 directly recognizes the anticodon loop in the tRNA. To clarify the mechanism, further study is required.

Discussion
In the current study, we investigated tRNA methyltransferase activities in crude extract from T. acidophilum cells, analyzed the modified nucleosides in native initiator and elongator tRNA Met , and characterized three tRNA methyltransferases (Trm5, Trm1 and Trm56) by purified recombinant proteins. We utilized the sequences of three tRNAs from T. acidophilum (initiator tRNA Met , elongator tRNA Met and tRNA Leu UAG), which were reported previously in earlier [24,25] and our recent [28] studies, as a basis to predict the candidate genes for the enzymes responsible for the modifications. As summarized in Table 1, our experiments revealed that the genes Ta0997, Ta0931, and Ta0836 encode Trm1, Trm56 and Trm5, respectively, from T. acidophilum. In archaeal tRNA modifications, there are some reports that different enzymes modify the same modification at the same position in tRNA: for example, Cm56 modification is formed by two systems, Trm56 or C/D sRNP [30]. Therefore, to understand tRNA modification systems precisely, the construction of gene disruptant mutant strains is desirable. However, there is no gene disruption method for T. acidophilum. Therefore, in this study, we could not utilize this approach. Consequently, the other gene product except for Trm1, Trm5 and Trm56 may bring the same modification(s) at the same position(s) in tRNA. Unexpectedly, we could not detect Trm56 activity in the crude cell extracts (S-30 and S-100 fractions). Analysis of the T. acidophilum proteome revealed that various proteins form several large (more than 300 kDa) protein complexes and that some of these protein complexes appear to interact with the membrane [48]. Therefore, T. acidophilum Trm56 might be part of a large complex with other proteins in living cells. Similar to the Trm56 activity, we could not detect tRNA (m 7 G49) methyltransferase activity in the crude cell extract. To clarify the intracellular localization of these enzymes, further study will be required. Although the recognition of tRNA by Trm56 might be affected by the presence of other proteins in the putative complex, the purified enzyme at least can act on both the mature elongator tRNA Met transcript and its precursor, which contains an intron at the canonical site. The methyl group was transferred to the mature transcript much more rapidly than to the precursor tRNA with the intron. This result suggests that the methylation by Trm56 occurs mainly after the removal of the intron. The activity of Trm5 was clearly detected in the S-100 fraction and the purified recombinant Trm5 methylated the G37 nucleotide in the tRNA Leu UAG transcript. Among the tRNA methyltransferases from T. acidophilum, only Trm5 seemed to act as a free enzyme, i.e., was not included in a protein complex. It has been reported that Trm5 recognizes the tertiary interaction between the D-and T-arms [55,62]. Consequently, Trm5 might act mainly during and/or after the three-dimensional core of the tRNA has been reinforced structurally by the introduction of other modifications.
The sequence of the initiator tRNA Met that is encoded in the genome of T. acidophilum strain HO-62 [3] differs from that reported in the earlier study [25]. The sequences of the two initiator tRNA Met differ in the D-arm and at position 57: A20b and C22 are inserted in strain HO-62 and this strain also contains A57 instead of G57. These differences might be derived from the different origins of the strains: the strain HO-62 was isolated from Hakone, Japan [3]. There is another possibility as follows. The m 1 I 57 modification in archaeal tRNA and the initiator tRNA Met gene from T. acidophilum were not reported in 1982. The authors used the Kuchino's post-labelling method for tRNA sequencing [25], in which tRNA is partially cleaved by formamide and then the nucleotide at the 5'-end of each fragment is analyzed by 2D-TLC [46,63]. However, the mobility of pm 1 I on 2D-TLC closely resembles that of pm 2 G [64]. Furthermore, in general, formamide cleavage of tRNAs from thermophiles is very difficult due to their structural rigidity. Therefore, it might be difficult technically to distinguish the pm 1 I and pm 2 G on 2D-TLC. In the current study, we detected the m 1 A (m 6 A) formation into the tRNA Leu UAG transcript by the S-30 fraction. However we could not identify the trmI gene, which encodes archaeal tRNA (m 1 A57/m 1 A58) methyltransferase. Consequently, we could not verify whether T. acidophilum TrmI can methylate the A57 in the initiator tRNA Met , although our analysis of modified nucleosides revealed that m 1 I is contained in the initiator tRNA Met . To determine the position of m 1 I modification, the RNA sequence of initiator tRNA Met is required. We are now checking the plasmid vector for TrmI expression and the other gene products. From the results of the current study, we were able to add the following information with respect to initiator tRNA Met from T. acidophilum: (1) initiator tRNA Met from T. acidophilum strain HO-62 contains the modifications G + , m 1 I, and m 2 2G; (2) the m 2 2G26 modification exists in addition to m 2 G26; (3) m 2 G26 and m 2 2G26 are formed by Trm1.
In the current study, we have demonstrated that archaeal Trm1 can methylate the tRNA Leu UAG transcript, which has a long variable region. As far as we know, this is the first time that archaeal Trm1 has been shown to act on class II tRNAs. In the case of class I tRNAs, archaeal Trm1 was reported to recognize the D-stem and the size of the variable region [58]. Therefore, archaeal Trm1 might be able to recognize the large variable region in the class II tRNAs. Furthermore, we showed that Trm1 efficiently methylated both the mature elongator tRNA Met transcript and the precursor with an intron. These results agree well with those of a previous study [58], namely, that archaeal Trm1 does not recognize the anticodon loop. Moreover, from the results of the current study, we were able to add the following information in relation to elongator tRNA Met from T. acidophilum: the unidentified G modification at position 26 is a mixture of m 2 G and m 2 2G, which is formed by Trm1.
During the course of the current study, it has been reported that Sulfolobus acidocaldarius TrmJ is responsible for the Cm32 modification in tRNA [65]. The Cm32 modifications in initiator and elongator tRNA Met are probably formed by this new enzyme as shown in Table 1.

Preparation of S-30 and S-100 Fractions, and Detection of tRNA Methyltransferase Activities
Wet cells (0.3 g) were suspended in 2 mL of buffer A (50 mM Tris-HCl (pH 7.6), 5 mM MgCl2, 6 mM 2-mercaptoethanol, and 50 mM KCl). The cells were ground in a mortar with 0.15 g aluminum oxide and then the suspension was centrifuged at 8000× g for 20 min. The supernatant fraction was centrifuged further at 30,000× g for 2 h. The resultant supernatant fraction was used as the S-30 fraction. The S-100 fraction was the supernatant fraction by that was obtained after centrifugation at 100,000× g for 2 h. Transfer RNA methyltransferase activities in the S-30 and S-100 fractions were analyzed as follows: 30 µg of protein from the S-30 or S-100 fraction, 0.2 A260 units tRNA Leu UAG transcript and 0.78 nmol [methyl-14 C]-AdoMet were incubated in 40 µL of buffer A at 55 °C for 1 h. The RNA was extracted with phenol-chloroform and then recovered by ethanol precipitation. The RNA pellet was dissolved in 3 µL of 50 mM sodium acetate (pH 5.0), and digested with 2.5 units of nuclease P1 (Wako Pure Chemicals, Osaka, Japan). The sample was separated using 2D-TLC as described previously [64]. The 14 C-methylated nucleotides were monitored with a BAS 2000 Bio-imaging Analyzer (Fuji Photo Film, Tokyo, Japan).

Purification of Initiator and Elongator tRNA Met by the Solid-Phase DNA Probe Method
Initiator and elongator tRNA Met were purified by the solid-phase DNA probe method as described in our previous reports [51,52]. The sequences of the 3'-biotinylated DNA oligomers were as follows: for initiator tRNA Met , 5'-ATG AGC CCA TTG GGA TTT CCT GA-biotin 3'; for elongator tRNA Met , 5'-ATG AGT CCG GTG CTC CTC CAG-biotin 3'. The complementary regions are illustrated in Figure 1A,B. The isolated tRNAs were further purified by 10% PAGE (7 M urea).

Nucleoside Analysis
Nucleoside analysis was performed as described in our previous reports [50,53]. The standard marker of G + was kindly provided by Prof. Takashi Yokogawa (Gifu University, Gifu, Japan).

Selection of Candidate Genes
We searched for the candidate genes in the T. acidophilum HO-62 genome by performing a BLAST search using the amino acid sequences of H. volcanii Trm1 and Trm56, and M. jannaschii Trm5. The identification of the other candidate genes was reported in our previous paper [28].

Expression of Gene Products
The underlined regions show restriction enzyme sites (Nde I and Bam HI). The PCR products were individually inserted individually into the multiple cloning linker of expression vector pET-30a (Novagen, Cambridge, MA, USA). The gene products were expressed in the E. coli BL21 (DE3) Rosetta 2 strain (Novagen) in accordance with the manufacturer's instructions.

Purification of Trm5
Briefly, Trm5 was purified by heat treatment at 50 °C for 30 min, followed by successive rounds of column chromatography through HiTrap Q-Sepharose, HiTrap Heparin-Sepharose, and Toyopearl CM-650M (Tosoh, Tokyo, Japan). The final eluted sample was dialyzed against buffer B (50 mM Tris-HCl (pH 7.6), 50 mM KCl, 6 mM 2-mercaptoethanol and 5% glycerol) and concentrated with a Vivaspin 15R centrifugal filter device (Sartorius Japan, Tokyo, Japan). Glycerol was added to the sample to a final concentration of 50% v/v and the sample stored at −30 °C.

Purification of Trm56
Briefly, Trm56 was purified by heat treatment at 50 °C for 30 min, followed by successive rounds of column chromatography through HiTrap Q-Sepharose, HiTrap Heparin-Sepharose, and HiLoad 16/600 Superdex 200 pg. The final eluted protein was dialyzed against buffer B, and concentrated with a Vivaspin 15R centrifugal filter device. Glycerol was added to the purified protein to a final concentration of 50% v/v and the samples stored at −30 °C.

Measurement of tRNA Methyltransferase Activities
The transcripts were prepared by using T7 RNA polymerase and purified by Q-Sepharose column chromatography and 10% PAGE (7 M urea). The standard assay for the purified enzymes was to measure the incorporation of 14 C-methyl groups from [methyl-14 C]-AdoMet into the appropriate tRNA transcript. For the reaction, 66 nM enzyme, 4.25 µM transcript, and 17.3 µM [methyl-14 C]-AdoMet were incubated in 40 µL of buffer A at 50 °C for 5 min. An aliquot (35 µL) of the reaction was then used for the filter assay. To visualize the methyl-transfer reaction, we used 10% PAGE (7 M urea) and autoradiography. Briefly, tRNA (0.1 A260 units) was incubated with 66 nM enzyme and 17.3 µM [methyl-14 C]-AdoMet at 50 °C for 5 min in 40 µL of buffer A, and then loaded onto a 10% polyacrylamide gel that contained 7 M urea. The gel was stained with methylene blue or toluidine blue, and then dried. The incorporation of 14 C-methyl groups into the tRNA was monitored with a Typhoon FLA 7000 laser scanner (GE Healthcare). The 2D-TLC was performed as follows. An aliquot of 1 µg of purified protein, 0.2 A260 units of tRNA transcript and 0.69 nmol of [methyl-14 C]-AdoMet were incubated in 40 µL of buffer A at 55°C for 1 h. The RNA was extracted with phenol-chloroform and then recovered by ethanol precipitation. The RNA pellet was digested with 1.5 units of nuclease P1. The sample was separated using 2D-TLC as described previously [64]. The 14 C-methylated nucleotides were monitored with a Typhoon FLA 7000 laser scanner (GE Healthcare).

Preparation of E. coli RNase P and Removal of 5'-Leader Sequence
The plasmid vectors for the C5 protein and M1 RNA of E. coli RNase P were a gift from Prof. Takashi Yokogawa (Gifu University). The C5 protein was purified as described in the reference [61]. The M1 RNA was synthesized with T7 RNA polymerase. Active RNase P was generated by mixing C5 protein and M1 RNA in accordance with the method described in the reference [61]. The 5'-leader sequence of E. coli tRNA Met f was used. The 5'-leader sequence was cleaved by RNase P in accordance with the reference [61].