A thermostable tRNA (guanosine-2')-methyltransferase from Thermus thermophilus HB27 and the effect of ribose methylation on the conformational stability of tRNA.

An S-adenosylmethionine-=dependent tRNA (guanosine-2'-)-methyltransferase (EC 2.1.1.34) was purified to the homogeneous state (2,400-fold) from a cell-free extract of an extreme thermophile, Thermus thermophilus HB27. The enzyme was highly resistant to heat as reported for other enzymes from thermophilic organism. The enzyme is monomeric and its molecular weight was estimated to be about 20,000. The Km values for S-adenosylmethionine and for Escherichia coli tRNAPhe were determined to be 0.47 microM and 10 nM, respectively, while the Ki for a competitive inhibitor S-adenosylhomocysteine, was 1.67 microM. When yeast tRNAPhe was methylated with the purified Gm-methyltransferase, a stoichiometric amount of methyl group was incorporated into the invariant guanosine at position 18 in the D-loop. Yeast tRNAPhe and E. coli tRNAMet, which were quantitatively methylated with the enzyme, were very similar to the native tRNAs with regard to amino acid acceptor activity and melting temperature, but were more resistant to RNase T1 and RNase A digestions than the corresponding native tRNAs.

It is widely accepted that the modification of tRNA is a post-transcriptional event, and that the reaction is catalyzed with highly site-specific enzymes (1). In tRNA of extreme thermophues, Thermus thermophilus HB8 and T. thermophilus HB27, there exists a unique set of modified nucleosides, 2-thioribothymidine, 1-methyladenosine, and 2°C)-methylguanosine, and among these nucleosides, s2T is most responsible for the thermostability of the thermophile tRNAs (2)(3)(4)(5)(6). In the case of T. thermophilus HB27, it was found that the contents of Gm, m'A, and S2T increased significantly by raising the growth temperature of the bacterial cells (6). This suggests certain functional roles of Gm and m'A in the thermophile tRNA at high temperatures.
In order to elucidate the mechanisms of biosynthesis of * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "uduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
these modified nucleosides and their roles in the function of tRNA, especially at high temperatures, we studied the enzymes concerned with the formation of these nucleosides. As reported previously (6), S-adenosyl-methionine-dependent tRNA (guanosine-2'-)-methyLtransferase (EC 2.1.1.34) and tRNA (adenine-l-bmethyltransferase (EC 2.1.1.36) were isolated from the cell-free extract of T. thermophilus HB27 and it was found that Gm-methykransferase catalyzed the specific methylation of the invariant G18 in the D-loop of tRNA (6) (the numbering of residues conforms to the proposed rule on the basis of the numbering of yeast tRNAPhe (7)). The ribose methylation of G18 in the D-loop is one of the most invariant modifications found in both prokaryotic and eukaryotic tRNAs (8). However, there has been little study on the enzymes responsible for the ribose methylation. Since the enzymes from thermophiles are known to be generally more resistant to heat and protein denaturing reagents than the corresponding enzymes from mesophilic organisms (9, lo), it is worthwhile to use the extreme thermophiles as the enzyme source.
In this paper, we report on the purification and properties, as well as the site specificity, of S-adenosylmethionine-dependent tRNA (guanosine-2'-)-methyltransferase from T. thermophilus HB27. ln addition, the role of Gm18 in the Dloop in maintaining the ordered structure of tRNA is discussed.

Purification of Gm-methyltransferase from the cells of T. thermophilus HB27
All operations were performed at 4 "C. Methyltransferase activity was assayed using yeast tRNAPh' as the acceptor under standard assay conditions at 65 "C.  (Fig. 4 in Miniprint). The Gm-methyltransferase was tightly bound to the column, and attempts to elute the enzyme were unsuccessful under conditions such as high salt concentrations (2 M KC1 or 2 M CaCIZ) or low pH (pH 4.0). Even 2.25 mM AdoHcy failed to elute the enzyme activity from the column. Considering that the enzymes from T, thermophilus are generally resistant to protein denaturants (9, IO), the elution of the enzyme was carried out using the buffer containing 6 M urea. The purified enzyme still retained 93% of its original activity after incubation in 6 M urea solution a t 4 "C for 4 h. Table I summarizes the typical purification of the Gmmethyltransferase. The recovery of the Gm-methyltransferase activity was calculated on the basis of methyltransferase activity in each fraction using yeast tRNAP"" as the substrate.
Since the partially purified m' A-methyltransferase failed to methylate tRNAs containing m'A a t position 58, such as T. thermophilus HB8 tRNA""' and yeast tRNAPh', only the Gm-methyltransferase activity can be detected in the extract of T. thermophilus HB27 when yeast tRNAPhr is used as a substrate.
Purity and Molecular Weight of the Gm-methyltransferase-When the purified preparation of Gm-methyltransferase was subjected to polyacrylamide gel electrophoresis at pH 4.3, a single protein band was observed. The purified enzyme also gave a single band on sodium dodecyl sulfate-polyacrylamide gel electrophoresis (Fig. 5 in Miniprint). From this electrophoresis, the molecular weight of the enzyme was estimated to be 21,000, by comparing the relative mobility of the enzyme with those of standard proteins ( Fig. 6 in Miniprint).
For determination of the molecular weight of the enzyme under nondenaturing conditions, the purified enzyme was subjected to analytical gel filtration on a column of Sephacryl S-200 superfine ( Fig. 7 in Miniprint). The molecular weight of the enzyme was calculated to be 20,000.
Thermal Stability of Gm-methyltransferase-This enzyme was resistant to heat, as with other enzymes obtained from T.
thermophilus (2). In the presence of 10% glycerol at pH 7.5, the enzyme retained 90% of its activity even after a 20-min incubation a t 80 "C; the activity decreased rapidly at 90 "C ( Fig. 8 in Miniprint).
Enzymatic Properties of Gm-methyltransferase-The pH optimum for the enzyme was found to be pH 7.2-7.5. As shown in Table 11, magnesium ion and spermine stimulated the enzyme activity 3 and 8 times, respectively, a t their optimal concentrations, being about 5 mM in each case.
The results of kinetic analysis are also summarized in Table  11. Double reciprocal plots of initial rate data at

Certain enzymatic properties of Gm-methyltransferase
In Experiment 1, the assay was carried out in a standard assay mixture containing a specified compound in place of 5 mM MgC12 at 65 "C. Experiment 2 for determination of kinetic parameters was carried out at 37 "C, as described under "Materials and Methods." Kinetic Darameters were calculated on the basis of a general equation for bireactant sequential mechanism (23). Stoichiometry of Methylating Reaction toward tRNA-Yeast tRNAph' was rapidly methylated and approximately 1 mol of methyl group was incorporated into 1 mol of ( Fig. 9 in Miniprint). Even on adding again the same amounts of enzyme and S-aden~syl[methyl-'~C]methionine initially added to the reaction mixture after 90 min of incubation, only a small increase could be observed in the incorporation. This suggests that the methylation is limited to a single guanosine residue in the tRNA. E. coli tRNAPh' was a less efficient substrate than yeast tRNAPh' (Fig. 9 in Miniprint). The methyl group was hardly incorporated into T. thermophilus HB8 tRNAph', since the tRNAPh' had the Gm18 in situ.  Fig. 2 shows partial digestion patterns on 205 polyacrylamide gel electrophoresis in the presence of 7 M urea. In the absence of Mi", extensive digestion occurred at 4 "C in both the native and methylated tRNA"'"'. On the other hand, in the presence of 10 mM MgCI?, tRNA was considerably resistant to RNase digestion. It appears that G18, G19, and G20 are the most susceptihle sites for RNase T I digestion, as reported hy Wrede et nl. (16). G3, G4, and G30 are also cleaved under the present conditions; however, m'G10 and G15 are scarcely attacked by RNase T I .
For methylated tIINA'""', the band corresponding to G18 is absent not only on the electrophoretogram of the partial digestion products with RNase T I but also on alkaline digestion. This confirms that the Gm residue is, in fact, located at position 18 of the methvlated yeast tHNA"'"'. since at the position where P'-O-methvl residue is present, no hydrolysis should occur either with RNase TI nor with alkali. Except for Gm18, the initial cleavage sites o f the methylated tRNA"'"' with RNase TI was guanosine residues GI9 and G20 in the D-  Effect of Ribose Metl1ylation of GI8 on the Conformntional Stubility of tRNA-By using the purified Gm-methyltransferase, yeast tRNA"'"' and E. coli tRNA1"" were methylated quantitatively on the GI8 residue. In order to examine the role of Gm in the D-loop in stabilizing the conformation of tRNA, these modified tIiNAs were characterized with respect to (i) amino acid acceptor activity, (ii) melting temperature and (iii) susceptibility to RNases. The results are summarized in Table 111. Meth.vlated tIiNAs accepted almost the same amount of amino acid as did unmodified tRNAs, showing that the tRNA samples retained fully the biochemical activity after the modification.
Comparisons of the melting profiles of native yeast tRNA"'"' and yeast tRNA::!:: in the presence and absence of 10 mM MgCI, were macle ( Fig. I : ! in Miniprint). Both tRNAs show the same hyperchromicity at 260 nm. It is intriguing that no significant difference could be observed in the melting profiles between native and modified tHNAs in either the presence or ahsence of Mg". Similar results were obtained with the comparison of melting profiles of E . coli tRNAl""', tRNA:?:,;' and tRNA?:,:&:,, although the melting temperatures of these s in the presence of Mg:' are 5.5-6 "C lower than that of T. thern~ophi1u.s tRNAfh'"' I. Fig. 3 shows the time course of increase in absorbance at 260 nm for native and methylated yeast tRNA"'"'(tRNA1~~:'). caused by RNase TI and RNase A digestions in the presence of 10 nlM M i 2 . " . T h e initial rate of increase in absorbance at 260 nm of t R N A !~~~~ by digestion with RNase T I was much lower than that of native tRNA"'"'. In the case of RNase A digestion. the difference in the initial rates can also be seen between native and modified tRNA"'"'s, although this difference is smaller than that of RNase TI digestion. This is reasonable, since methylation occurs at G18, which is just the cleavage site of RNase T I but not that of RNase A. When excess of RNase T I or RNase A was added to the reaction mixtures, both native and methylated t1INA"'"'s showed the same hyperchromicitv change. This precludes the possibility that the methylated tRNA takes on a special conformation  The melting profiles of tRNAs in 0.01 M Tris-HC1 (pH 7.5), 0.01 M MgC12,O.Z M KC1 were recorded by a Gilford spectrophotometer model * Susceptibilities of tRNAs toward RNases were expressed as the initial increase of absorbance at 260 nm accompanied by RNases-2400-S equipped with autothermoprogrammer model 2527. The rate of temperature increase was set at 0.5 "C/min. digestion, as described in Fig. 3. For details, see the legend to Fig. 3.
The number of methyl groups/mol of tRNA was estimated from the radioactivity of I4C incorporated into tRNA assuming that 1.0 units of tRNA is equal to 1.66 nmol.
The modified tRNA was extracted with 88% (v/v) phenol. The recovered tRNA32 was further methylated with the purified Gmmethyltransferase and purified by gel electrophoresis on a 10% polyacrylamide gel.
... . not cleaved with RNases or that the tRNA has already nicks in the structure, making it insensitive to RNase digestion. (C50 and s2T54 in the thermophile tRNAfnet) (3). The tRNA%,kIA had nearly the same melting temperature as that of either native tRNAfMet or tRNA%, showing that Gm18 and m'A58 are not responsible for the thermostability of tRNA. However, the susceptibility of the toward either RNase TI or RNase A digestion was nearly the same as that of T. therrnophilus tRNA""' ', and quite different from that of native E. coli tRNA"", indicating that both the Gm and m'A residues are related to the conformational stability of tRNA in regard to RNase digestion,

DISCUSSION
In this paper, we describe the purification of tRNA (gua-nosine-2")-methyltransferase from T. tlzermophilus HB27 to 2400-fold by four purification steps. An affinity chromatography of AdoHcy-Sepharose 4B was used for the final purification. In this chromatography, the usual elution conditions using high salts, low pH, or even AdoHcy (2.25 mM) were of no effect. Thus, the elution buffer containing 6 M urea was used, because it is known that the enzyme from T. thermophilus are generally resistant to protein denaturants (10,11). In fact, the Gm-methyltransferase retained most of its activity after treatment with 6 M urea. Affinity chromatography was very effective for the purification (the enzyme was 6-fold purified by this step) and in spite of nonspecific elution condition, the enzyme preparation after this step appeared to be homogeneous, judging from polyacrylamide gel electrophoresis (see Fig. 5 in Miniprint).
The molecular weight of the enzyme was estimated to be 21,000 by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and 20,000 by gel filtration on Sephacryl S-200 column, suggesting that the enzyme is a monomeric protein in solution. The Gm-methyltransferase activity was enhanced from three to several times with either Mg2+ or spermine and inhibited with AdoHcy, the reaction product. These features are similar to those for other tRNA methyltransferases so far reported, such as m'A-methyltransferase, tRNA (N'-guanine)-and tRNA (guanine-1-)-methyltransferases from rat liver or other organisms (1,26,27).
In the present study, we confirmed that the methylation site in tRNA with this enzyme is GI8 in the D-loop by using the purified enzyme preparation and yeast as a substrate (see Fig. 1). This conclusion had been suggested by a previous experiment using the partially purified enzyme and E. coli tRNA"" (6). Thus, it can be concluded that Gmmethyltransferase from T. therrnophilus HB27 catalyzes the specific methylation of G18 in the "invariant" GG sequence of the D-loop in tRNA.
By using this purified Gm-methyltransferase, yeast tRNAPhe and E. coli tRNA"" could be stoichiometrically methylated (Table 111). E. coli t R N A F could be further methylated with partially purified m'A-methyltransferase resulting in double-methylated tRNA"" The charging activity of the methylated tRNAs thus obtained were almost the same as that of unmodified tRNAs, showing that in vitro modification a t G18 and A58 does not affect the conformation of tRNA so much.
It is intriguing that not only the melting temperatures (Table 111) but also the melting profiles of yeast tRNA::

7392
Thermophile tRNA (Guanosine-2'-)-methyltransferase (Fig . 12) and E. coli tRNA>% and tRNAgfl,:LlA hardly differed from those of the corresponding native tRNAs. This clearly precludes the possibility that in addition to s'T54, either Gm18 or m'A58 is also responsible for the thermostability of the thermophile tRNAs as has been speculated (3, 4).
Agris et al. (28) have already reported that the 2'-0-methylnucleosides have little effect on the melting profile of tRNA, using unfractionated tRNAs from Bacillus stearothermophi-Zus, a moderately thermophilic bacteria; the tRNAs extracted from the cells grown at 70 "C contained 2"O-methylnucleosides three times as much as those from the cells grown a t 50 "C, nevertheless both tRNAs showed nearly the same melting profiles. The present study, which is consistent with their observation, further clarified the fact through use of purified tRNA and the site-specific tRNA-methyltransferase.
Agris et al. proposed that Gm18 serves to suppress the susceptibility to RNase TI digestion, since the Gm residue cannot be cleaved with the RNase (28,29). Wrede et al. (16) reported that the initial cleavage sites of yeast tRNAph' with RNase T, in the presence of Mg'+, are G18, G19, and G20 in the D-loop. In addition, G18 and G19 are involved in the tertiary hydrogen-bonding between the D-loop and the T-loop (30,31). The digestion rate shown in Fig. 3 may be interpreted as follows. Once these sites (G18, G19, and G20) are attacked by RNase TI in the native tRNA"h', the tRNA conformation is loosened to some extent so that other G residues such as G3, G4 (Fig. 2), G15, and G57 (Wrede et al. (16)) become to be easily attacked by the enzyme, resulting in a rapid increase in the absorbance shown by Curve I in Fig. 3A. On the other hand, in tRNAE2, one of the most susceptible residues, G18, is closed to RNase TI by the methylation of its ribose moiety, so that a conformational change due to the RNase TI attack, which is necessary for the cleavage of the other G residues, is suppressed considerably, resulting in a slow increase in the absorbance (Curve I1 in Fig. 3A).
In the case of the RNase A digestion, there exists no such time lag in the absorbance increase (Fig. 3B), as observed for RNase T, digestion. However, significant differences in susceptibilities toward RNase A is observed between native tRNA''h' and t RNA: : .
A similar tendency was observed for E. coli tRNA"", tRNAE:', and tRNAE,k,IA, but with less efficiency. In this case, it is evident that the methylation of G18 decreases the susceptibility to 62 and 74% for the RNase TI and RNase A digestions, respectively. Further methylation of A58 again decreases the susceptibility toward both RNases to exactly the same level as that of T. thermophilus tRNA"" I .
The susceptibility of certain residues toward RNases may depend on the local fluctuation surrounding the residues, which may be suppressed to some extent by the presence of the methylated nucleosides. Such a fluctuation may be independent of the melting temperature of the polynucleotide chain as a whole, as Englander et al. (32) proposed on the basis of the hydrogen-exchange study.