Transfer RNA(5-methylaminomethyl-2-thiouridine)-methyltransferase from Escherichia coli K-12 has two enzymatic activities.

The tRNA(5-methylaminomethyl-2-thiouridine)-methyltransferase, which is involved in the biosynthesis of the modified nucleoside 5-methylaminomethyl-2-thiouridine (mnm5s2U) present in the wobble position of some tRNAs, was purified close to homogeneity (95% purity). The molecular mass of the enzyme is 79,000 daltons. The enzyme activity has a pH optimum of 8.0-8.5, is inhibited by magnesium ions, and stimulated by ammonium ions. Two different intermediates in the biosynthesis of mnm5s2U34 are present in tRNA from the mutants trmC1 and trmC2. Unexpectedly, the product present in tRNA from trmC1 cells was identified by mass spectrometric and chromatographic analyses as 5-carboxymethylaminomethyl-2-thiouridine (cmnm5s2U), i.e. a more complex derivative than the final product mnm5s2U. The product present in tRNA from trmC2 cells was identified as 5-aminomethyl-2-thiouridine (nm5s2U). In the presence of S-adenosylmethionine the most purified enzyme fraction converts both cmnm5s2U34 and nm5s2U34 into mnm5s2U34. In the absence of S-adenosylmethionine, however, cmnm5s2U34 is converted into nm5s2U by this enzyme fraction. We conclude that the purified polypeptide has two enzymatic activities; one actually demodifies cmnm5s2U to nm5s2U and the other catalyzes the transfer of a methyl group from S-adenosylmethionine to nm5s2U, thus forming mnm5s2U. The sequential order of the biosynthesis of mnm5s2U34 is suggested to be: (Formula: see text). The molecular activity of the methyltransferase activity (nm5s2U34----mnm5s2U34) is 74 min-1, and the steady state concentration of the enzyme is only 78 molecules/genome equivalent in cells growing at a specific growth rate of 1.0/h.

Transfer RNA(5-Methylaminomethyl-2-thiouridine)-Methyltransferase from Escherichia coli K-12 Has Two Enzymatic Activities* (Received for publication, December 31,1986) Tord G. Hagervall, Charles G. EdmondsS, James A. McCloskeyS, and Glenn R. Bjork The tRNA(5-methylaminomethyl-2-thiouridine)methyltransferase, which is involved in the biosynthesis of the modified nucleoside 5-methylaminomethyl-2-thiouridine (mnm6s2U) present in the wobble position of some tRNAs, was purified close to homogeneity (95% purity). The molecular mass of the enzyme is 79,000 daltons. The enzyme activity has a pH optimum of 8.0-8.5, is inhibited by magnesium ions, and stimulated by ammonium ions. Two different intermediates in the biosynthesis of mnm6s2U34 are present in tRNA from the mutants trmC1 and trmC2. Unexpectedly, the product present in tRNA from trmC1 cells was identified by mass spectrometric and chromatographic analyses as 5-carboxymethylaminomethyl-2-thiouridine (cmnm6s2U), i.e. a more complex derivative than the final product mnm6s2U. The product present in tRNA from trmC2 cells was identified as 5-aminomethyl-2-thiouridine (nm5s2U). In the presence of S-adenosylmethionine the most purified enzyme fraction converts both cmnm6s2U34 and nm6s2U34 into mnm6s2U34. In the absence of S-adenosylmethionine, however, cmnm6s2U34 is converted into nm6s2U by this enzyme fraction. We conclude that the purified polypeptide has two enzymatic activities; one actually demodifies cmnmas2U to nm5s2U and the other catalyzes the transfer of a methyl group from Sadenosylmethionine to nm6s2U, thus forming mnm6s2U. The sequential order of the biosynthesis of mnm6s2U34 is suggested to be: U34 4 s2U34 4 + cmnm6s2U34-nm6s2U-mnm6s2U34. The trmC 1 trmC2 AdoMet t molecular activity of the methyltransferase activity (nm6s2U34 + mnm6s2U34) is 74 min", and the steady state concentration of the enzyme is only 78 molecules/ genome equivalent in cells growing at a specific growth rate of l.O/h.
Transfer RNA from eubacteria contains about 10% modified nucleosides. The biosynthesis of these modified nucleosides constitutes an integral step in the maturation process of tRNA. Most modification reactions occur at the polynucleotide level, i.e. after the primary transcript has been synthe- This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. operating in the maturation process. The biosynthesis of modified nucleosides is catalyzed by highly specific enzymes. At least one enzyme for each type of modified nucleoside is required, and it has been calculated that at least 40 different modifying enzymes are needed in eubacteria to modify tRNA.
A coding capacity of 1% of the total genome is therefore needed to make the necessary tRNA-modifying enzymes. The presence of some of these modified nucleosides has been shown to influence the decoding capacity of the tRNA (reviewed by Bjork et al., 1987).
Some of the modified nucleosides are simple derivatives of the ordinary nucleosides like 5-methyluridine (m'U)' and 1methylguanosine (m'G). However, some modified nucleosides, like 5-methylaminomethyl-2-thiouridine (mnm5s2U), are complex and it is likely that more than one enzyme is involved in the biosynthesis of such modified nucleosides. In addition modification reactions generally occur in a sequential order. The mnm5s2U is present in the wobble position of tRNAreading codons of the type NA& where N can be C, A, or G. This nucleoside has been shown to be important in the decoding step (Hagervall and Bjork, 1984b;Elsevier et al., 1984;Sullivan et al., 1985, reviewed in Bjork et al., 1987. As a first step to elucidate the biosynthetic pathway of mnm5szU, two mutants defective in the biosynthesis of this modified nucleoside were isolated (Marinus et al., 1975;Bjork and Kjellin-Striby, 1978). Transfer RNA from both these mutants accepts methyl groups in uitro using S-adenosylmethionine as methyl group donor. The product formed in both cases was shown to be mnm's2U. Thus, it was suggested that these mutants were defective in the same gene (trmC) and the mutations were denoted trmCl and trmC2, respectively. However, recently we have shown that tRNA from these mutants contain two different derivatives of mnm6s2U, suggesting lesions in two different genes (Hagervall and Bjork, 1984b). We have further shown that a plasmid harboring a 5.2-kb chromosomal fragment including the trmC+ gene complemented both the trmCl and trmC2 mutations (Hagervall and Bjork, 1984a). The 5.2-kb chromosomal fragment has a coding capacity for at least two polypeptides consistent with the concept that the trmCl and trmC2 mutations affect two different loci. However, we have subcloned this area of the chromosome 'The abbreviations used are: m5U, 5-methyluridine; m'G, 1methylguanosine; mnm6s2U, modified 5-methylaminomethyl-2thiouridine; cmnm6s2U, 5-carboxymethylaminomethyl-2-thiouridine; nm5s2U, 5-aminomethyl-2-thiouridine; DTT, dithiothreitob SDS, sodium dodecyl sulfate; AdoMet, S-adenosylmethionine; LC/MS, directly combined liquid chromatography-mass spectrometry; B, base fragment of nucleoside; BH:, protonated base; MH' , protonated molecule; HPLC, high pressure liquid chromatography; kb, kilobase pairs. 8488 tRNA(mnm5s2U)Methyltransferase from E. coli 8489 and found that a smaller plasmid carrying only a 2.4-kb chromosomal fragment is able to complement both mutations. tRNA(mnm6s2U)methyltransferase has a molecular weight of 80,000 (Taya and Nishimura, 1973;1977). Therefore, our results suggest that only one polypeptide of M, = 80,000 with two enzymatic activities, giving the potential s f using tRNA from either a trmCl or trmC2 mutant as substrate, is coded for by the cloned DNA. However, such genetic results do not exclude the possibility of two closely linked genes encoding two polypeptides with molecular weights of about 80,000 and 20,000, respectively. One way to resolve the question whether trmC1 and trmC2 are allelic or not, would be to purify both potential polypeptides using tRNA from both mutants as substrates. This paper demonstrates that tRNA from trmCl cells contains 5-carboxymethylaminomethyl-2-thiouridine (cmnm5s2U) and the tRNA from trmC2 cells contains 5aminomethyl-2-thiouridine (nm5s2U). Unexpectedly, the enzyme is able to demodify cmnm5s2U to nm5s2U and also to act as methyltransferase to catalyze the formation of mnm5s2U from nm5s2U using S-adenosylmethionine as methyl donor.

RESULTS
Purification of tRNA(mnm5s2U)Methyltransferase-As source of enzyme for purification, cells from E. coli strain 1100 harboring plasmid pTH32 were used. Construction of this plasmid is shown in Fig. 1 (Miniprint). Due to the presence of plasmid pTH32 both enzymatic activities, as defined by using tRNA from trmCl and trmC2 mutants as substrates and measuring methyl group incorporation, are overproduced about 17-fold. The cells were disrupted using a French press and by sonication. Ribosomes were removed by high speed centrifugation in the presence of 0.6 M (NH&304, and the recovery was about 50-60% for both activities, indicating that the enzyme(s) has a tendency to bind to ribosomes even at high salt concentration. Following precipitation with (NH&SO, (50% saturation), the enzyme was subjected to molecular sieve chromatography using an Ultrogel AcA34 column ( Fig. 2 A , Miniprint). The majority of the RNA was removed in this step, but the two enzymatic activities were not separated from each other. Both activities eluted as two chromatographically different compounds; one did not penetrate the gel and the other was retarded (Fig. 2 A ) . The fractions containing the compound not penetrating the gel were pooled and rechromatographed on an identical column. The elution profile was the same as earlier (data not shown). Therefore, the enzyme probably forms an unstable complex, which separates from the unaggregated form using such a column. The compound which was retarded on the column was subjected to ion-exchange chromatography on a DEAE-Sephacel column. The two enzymatic activities co-chromatographed as one component at about 0.22 M NaCl with a recovery of 85% for the trmCl activity and 80% for the trmC2 activity. The purification in both cases was about 4 times ( Fig. 2B, Table I). The enzyme(s) was further purified using chromatography on heparin-Sepharose CL-GB (Fig. 2C). The two enzyme activities eluted very late in the gradient (0.43 M Table I,  NaC1). but again as a single component. This step was very efficient and a 105-and 50-fold purification of the trmC2 and trmC1 activities, respectively, was obtained ( Table I).

Portions of this paper (including "Experimental Procedures,"
Enzyme Purity and Molecular Weight-After heparin-Sepharose CL-GB chromatography, both enzyme activities migrated as one single major peptide (M, = 79,000) in sodium dodecyl sulfate-polyacrylamide electrophoresis ( Fig. 3 lane g), which coincides with one of the polypeptides expressed from pTH32 in minicells (data not shown). Some additional polypeptides were present but, judging from the intensities of these bands, the enzyme is at least 95% pure. Furthermore, all impurities appeared to have an M, above 40,000 (see "Discussion"). In conclusion, these results suggest that the two enzymatic activities are expressed by one polypeptide of M, = 79,000.
Optimul Assay Conditiom-The purified tRNA(mnm5s2U)methyltransferase had the highest specific activities using either substrates in the pH range of 8-8.5. Both enzymatic activities were stimulated by NH4C1, with an optimum for both around 20 mM, and they were inhibited when the concentration was increased to 80 mM ( Fig. 4). The enzymatic activities were strongly inhibited by M?', where a concentration of 5 mM M&l2 reduced the trmCl activity to 10% and the trmC2 activity to 25%, respectively (Fig. 4). The trmCl activity was inhibited by addition of 100 mM NaCl while a concentration of 200 mM was needed to significantly reduce the trmC2 activity. The enzymatic activities were also inhibited by p-chloromercuribenzoate (2 mM) (data not shown). Presence of putrescin (6 mM) or spermidine (2 mM) had no significant effect on the activities.
Structures of Precursors S1 and S2 Present in tRNA from tmC2 and trmCl Cells, Respectiuely-To understand on a molecular level what the nature of the two activities possessed by the trmC' polypeptide is, it was necessary to identify the precursors (compounds S1 and S2) to mnm5s2U present in tRNA from trmC2 and trmCl cells, respectively. Since the thiolated nucleoside S1 is present in general methyl deficient tRNA and its conversion to mnm5szU requires AdoMet, this compound is suggested to be nm5s2U (see below). Buck et al. (1982) have shown that tRNA from Salmonella typhimurium contains cmnm5s2U. It migrates in their HPLC system as our compound Furthermore, cmnm5s2U migrates in thinlayer chromatography using the solvent system described by Rogg et al. (1975) in a similar position as S2 (cf. Yamada et al., 1981 and Fig. 5). These results suggest that compounds S1 and S2 might be nm5s2U and cmnm5s2U, respectively.
TO firmly establish the identities of intermediates S1 and S2, enzymatic hydrolysates of trmC2 and trmCl tRNA were examined by directly combined HPLC-mass spectroscopy. In addition, tRNA from E. coli strain THIO was analyzed as a control for presence of the fully modified product mnm5s2U. All three tRNAs were separately analyzed using decylcyano and octadecyl HPLC columns (Systems A and B) which possess differing selectivity for the three nucleosides of interest. The m/z values and structure assignments for ions monitored are shown in Fig. 6. Authentic cmnm5s2U and mnm5s2U were used to establish the suitability of ions chosen for monitoring, as established from their thermospray mass spectra recorded using HPLC System B. The authentic nucleosides were also used as chromatographic markers in separate    FIG. 5. Schematic illustration of the distribution of nucleosides found in tRNA from t r m C . trmCl, a n d trmC2 cells. The illustration indicates the positions of some nucleosides including the different thiolated nucleosides SI, S2, and mnm5ss2U after two-dimensional chromatography and their relative distribution is shown in Table 11. In this paper we show that compound S2 is cmnmsss2U and compound S1 is n m Y U (see Fig. 7). molecular weight (M, = 289, MH' = m/z 290) and the characteristic BH: ion (Edmonds et al., 1985), m/z 158.
Chromatograms in the form of selected ion recordings from the six experiments are shown in Fig. 7 . Nucleoside cmnmsss2U is identified in trmCl tRNA in chromatographic Systems A and B, at the expected elution times (arrows, Fig. 7 , A and  B ) . The unusual breadth of HPLC peaks shown in Fig. 7 , A and B was also observed in chromatograms under the same conditions using authentic cmnmSs'U. As shown in Fig. 7  The chromatographic base line in Fig. 7A does not permit detection of nm5s2U in trmC1 tRNA, but this point is adequately addressed by the System B measurement shown in Fig. 7B. The overall data in Fig. 7 imply the absence of nm5s2U in trrnCl and wild-type tRNAs. These data further support the identity of compound S1 as nm5s2U. The fully modified nucleoside mnm5s2U is clearly observed in the wildtype strain THlO (Fig. 7, E and F) and is essentially absent in the other hydrolysates. Under conditions prevalent for chromatographic System A, the molecular species (MH') for nm5s2U and mnm5s2U are observed but not in the case of System B. In both cases for which authentic nucleosides are available, the elution positions are appropriately shifted when tested in the second chromatographic system in which the relative elution positions have been shifted. Analysis and Specificity of the Enzyme Actiuities-Transfer RNA from trmC2 and trmCl mutant strains which contain nm5s2U (compound Sl) and cmnm5s2U (compound S2), respectively ( Fig. 5 and Table II), were used to analyze the two enzymatic activities. In addition, tRNA from methioninestarved E. coli W6 (relA, metBI ) cells, which is generally methyl-deficient, was used as substrate. This tRNA contains most probably nm5s2U instead of mnm5s2U in position 34 (Table 11). These tRNAs were methylated using I4C-labeled AdoMet as methyl group donor and were digested to nucleosides which were separated by two dimensional thin-layer chromatography. The chromatograms were autoradiographed to visualize radioactive nucleosides. The results of these analyses were identical for the various tRNA preparations since the majority of the radioactivity (95%) was found in a compound comigrating with the nucleoside marker mnm5s2U. One additional minor radioactive compound was also found, indicating that the pure enzyme fraction contains another methyltransferase activity. Alternatively, this compound is an intermediate or secondary product of the normal biosynthesis of mnm5s2U (data not shown). Since the pure enzyme is able to convert cmnm5s2U and nm5s2U into mnm5s2U the 79,000 TABLE I1 Relatiue distribution of nucleosides mnm2s2U, nm5szU, and cmnm5s2u In uiuo 35S04-labeled tRNA from trmC', trmC1, and trmC2 cells were used as substrates. Following incubation under indicated conditions, the tRNA was digested to nucleosides and the distribution of 35S-labeled compounds was determined as shown in Fig. 5. Compound S1 in Fig. 5 is nm5s2U and compound S2 is cmnm5szU. Transfer RNA from strain W6 in uiuo labeled under methionine starvation in the presence of 35S04. Thus, this %-labeled tRNA is generally methyldeficient.

Addition
Relative distribution of tRNA nucleosides from strain
Reaction Order-Since cmnm5s2U and nm5s2U are both thiolated precursors to mnm5s2U, it is possible to use 35Slabeled tRNA as a substrate and monitor the conversion of nm5s2U and cmnm5s2U to mnm5s2U. Thus, the different enzymatic activities might be assayed for, even in the absence of the methyl group donor AdoMet. Strains TH49 (trmC2) and TH69 (trmCl ) were labeled with 35s04, and tRNA was prepared. The tRNA was incubated together with purified enzyme, with or without AdoMet and in the absence or presence of NHZ. The conditions employed allowed the modification reaction to go to completion. The tRNA was next digested to nucleosides and analyzed by thin-layer chromatography. In all cases where AdoMet was added both cmnm5s2U and nm5s2U were converted into mnm5s2U and presence of NH: had no effect (Table 11). However, in the absence of AdoMet, nm5s2U was not converted into mnm5s2U, while cmnm5s2U was converted into nm5s2U. Therefore, the reaction nm5s2U + mnm5s2U must involve the addition of a methyl group which is consistent with the observation that methyl-deficient tRNA contains nm5s2U ( Table 11). The conversion of cmnm5s2U to nm5s2U does not require AdoMet. However, the reaction is enzymatic since it requires the presence of the purified enzyme (Table 11). Thus, the sequential order of the enzymatic reaction is: cmnm5s2U + nm5s2U + mnm5s2U where the first step is AdoMet independent, while the second requires the presence of AdoMet. Other intermediates cannot be excluded.
cmnm5szU May Be Present in tRNA from Wild-type Cells-A low level of what appears to be cmnm5s2U is also found in wild-type cells (Hagervall and Bjork, 1984b; Table 11). To determine if the radioactive compound S2 found in wild-type tRNA in fact is identical to cmnm5s2U found in tRNA from a trmC1 mutant, 35S-labeled tRNA from wild-type and trmCl mutant cells was digested to nucleosides and separated on HPLC according to Buck et al. (1983). Fractions were collected and those containing compound S2 were further analyzed in five different chromatographic systems. No difference in mobility was observed between compounds S2 from trmC+ cells and cmnm5s2U from trmC1 cells in these five chromatographic systems. Furthermore, in the chromatographic analysis according to Rogg et al. (1975) and in the HPLC chromatography used, the two compounds migrated similarly (Table 111). Recently, a mutant (trmE) has been isolated and  * From strain TH69 (trmCI 1. Distribution of mnm5s2U, s2U, and compound 52 (cmnm5s2U)  shown to contain only s2U in position 34 in tRNA (Elseviers et al., 1984). Therefore, the trmE gene should code for an enzyme involved prior to, or directly involved in, the biosynthesis of cmnm5s2U. If so, the trmE mutant should have a reduced level or a complete absence of cmnm5s2U. Transfer RNA from this mutant was labeled with 35S04, degraded to nucleosides and analyzed by two-dimensional thin-layer chromatography. Table IV shows that tRNA from the trmE mutant contains a significantly reduced level of cmnm5s2U. Thus, we therefore conclude that the compound S2 from trmC+ cells might be identical to cmnm5s2U present in trmC1 cells.

DISCUSSION
This paper describes a 400-fold purification and the characterization of the tRNA (mnm5s2U)methyltransferase from E. coli. If the enrichment of the enzyme, due to the presence of a trmC+ plasmid is included, the enzyme has been purified 7,100-fold. The enzyme has an apparent M , of 79,000 on sodium dodecyl sulfate-gel chromatography and was judged to be at least 95% pure (Fig. 3, lanes g and f ) . The enzymatic activities have a pH-optimum of 8.0-8.5 and are stimulated by NH: but severely inhibited by Mg2+ (Fig. 4). This enzyme has been purified before using methyl-deficient tRNAG'" as substrate (Taya and Nishimura, 1973;1977). They reported that the enzyme had an M, of 80,000 and a methyltransferase activity since it transfers methyl groups from AdoMet to tRNAGL" from general methyl-deficient cells. By using tRNA from the trmCl and trmC2 mutants as substrates, we have demonstrated that the 79,000 polypeptide has two different enzymatic activities (Table 11). We also demonstrate that transfer RNA from these mutants contains cmnm5s2U and nm5s2U34, respectively (Fig. 7). Both mutations are complemented by plasmid pTH32, which carries a chromosomal insert of 2.4 kb (data not shown). A polypeptide of 79,000 requires a coding capacity of 2 kb of DNA. Thus, about 0.4 kb of additional coding capacity, corresponding to a 17,000 polypeptide is present on plasmid pTH32. It was therefore not possible from the genetic data to rule out that the two mutations affected two separate but closely located genes, one coding for the 79,000 polypeptide and the other for a 17,000 polypeptide. Here we show that the most purified enzyme fraction is able to convert both cmnm5s2U34 and nm5s2U34 to mnm5s2U in the presence of AdoMet. Furthermore, the few impurities observed have an M, well above 17,000. Thus, it is likely that the 79,000 polypeptide harbors two enzymatic activities. The tRNA(mnm5s2U)methyltransferase is unusually large, since most tRNA-modifying enzymes have a molecular weight of 60,000 or less (Bjork, 1984). This large size may be explained by the presence of two enzymatic activities in the same polypeptide. However, our results do not rule out the existence of overlapping genes contributing to two large polypeptides but which are synthesized in vastly different amounts. We find this unlikely since the 79,000 polypeptide is in itself produced in very small amounts in the cell (see below). At present we favor the hypothesis that the 79,000 polypeptide harbors two separate enzymatic activities, but the final answer has to await the complete DNA sequence of this region of the chromosome.
The methylation reaction was inhibited by magnesium ions irrespective of which substrate was used. Most tRNA-modifying enzymes are stimulated or do not respond to this ion. The conversion of nm5s2U34 to mnm5s2U34 is unique in its sensitivity toward Mg2+ ions. The conversion of cmnm5s2U34 to nm5s2U34 does not require AdoMet. Therefore, the similarities in the response to different conditions by trmCl and trmC2 tRNA measured as methyl group incorporation might only reflect the response in the methylation reaction, i.e. the conversion of nm5s2U34 to mnm5s2U34. Taya and Nishimura (1973;1977) also noted the unusual inhibition by M$+ in the methylating reaction. The specific activity of tRNA-(mnm5s2U)methyltransferase is 0.13 milliunits/mg (trmC2) in E. coli K-12 strain 1100, grown in glucose minimal medium.
Assuming that no factors present in crude extract influence the specific activity of the enzyme, the fraction of total protein which is tRNA(mnm5s2U)methyltransferase is 1.4 x This corresponds to about 78 molecules/genome equivalent calculated as Pedersen et al. (1978), which is comparable with what is reported for other methyltransferases, i.e. 80 molecules/genome for tRNA(m'G)methyltransferase (Hjalmarsson et al., 1983) and 100 molecules/genome for tRNA-(m5U)methyltransferase.4 However, the amount is still low compared to other enzymes acting on tRNA; e.g. the aminoacyl-tRNA-ligases are present in about 500-1000 molecules/ genome (Pedersen et al., 1978). The methylating activity of the tRNA(mnmSs2U)methyltransferase was calculated to 14 min" for the trmCl activity and 74 min-l for the trmC2 activity. This indicates that the step carried out by the trmCl activity is the rate-limiting step in uiuo. However, there is a specific loss of the trmCl activity in the last purification step, i.e. the ratios between the trmC2 and trmC1 activities are 2.6-2.8 in all steps except the last where the trmC2/trmCl ratio is increased to 5.5. Perhaps a low protein concentration is specifically harmful to the trmCl activity. Another possibility is that a cofactor stimulating the trmCl activity is lost in the last purification step. The analysis of enzymatic digests of trmC1 and trmC2 tRNA by thermospray LC/MS provides strong evidence that the precursors present are nm5s2U and cmnm6s2U in tRNA from trmCl and trmC.2 cells, respectively (Fig. 7). Thus, unexpectedly we find that one of the precursors (cmnm5s2U) to mnm5s2U is more complex than the final product in this pathway. These conclusions are based on the use of a mass spectrometer as a mass-specific detector, in conjunction with high performance chromatography, which provides a highly selective means for detection of modified nucleosides in enzymatic digests of tRNA (Edmonds et d., 1985). Because the mass spectrometer can be tuned to ions representing the molecular weight of the nucleoside or corresponding base, the detection of minor components from unfractionated tRNA can in general be made without the necessity of chromatographic separation from higher constituents as would be the case using UV detection. Use of HPLC columns of differing chromatographic selectivity (Systems A and B) provides an additional dimension of selectivity for the identification of eluants from digests of unfractionated tRNA. Table I1 shows that strain TH48 (trmC+) contains a significant amount of cmnm5s2U in its tRNA from cells grown in rich medium containing 35S04. The presence of cmnmss2U was shown by thin-layer chromatography using several chromatographic systems (Table 111). Furthermore, cmnm5s2U is also present in strain DEVl (trmE+) cells, but its level in tRNA from strain DEV16 (trmE) cells is significantly reduced ' T. Ny, unpublished results. (Table IV). A minor compound suggested to be cmnm5s2U was also found upon nucleotide analysis of purified tRNA:' " which is consistent with this idea (Sullivan et al., 1985). tRNA from S. typhimurium contains cmnm6s2U, and although this compound was not observed in tRNA from a trmC+ E. coli strain, its presence in small amounts was not excluded (Buck et aL, 1982). However, cmnm5s2U in strain THlO (trmC') is absent or below 10% of the level of mnmss2U (Fig. 7). Therefore, the level of cmnm5s2U in the tRNA may be influenced by the genetic background, the composition of the growth medium or the growth phase of the cells. If so, cmnm5s2U is an intermediate in the biosynthesis of mnm5s2U in all tRNA chains which contains this nucleoside in the wobble position. Alternatively, one or more tRNA species could normally contain cmnm5s2U and will not be a substrate for the trmCl activity of the trmC+ polypeptide. One such candidate would be tRNAG'" where the structure of the wobble nucleoside is uncertain (Sprinzl et al., 1985). Our results suggest then that the regulation of this particular tRNA species, which potentially only contains cmnm5s2U is influenced by the genetic or physiological conditions. These two alternatives can be experimentally distinguished by identifying the wobble nucleoside in tRNAG'" from wild-type cells and the mutants trmC1, trmC2, and trmE.
The side chain at the 5 position of mnm5szU contains two carbon atoms. Preliminary results from [ methyl-"C]-~-methionine labeling in uiuo experiments suggest that only one of the two carbon atoms originates from AdoMet.6 Taya and Nishimura (1977) also proposed that formyltetrahydrofolate might be donor for the first carbon atom in the side chain. Generally methyl-deficient transfer RNA from methioninestarved cell contains the undermodified derivative nm5s2U in addition to cmnm6s2U, which suggests that the nm5s2U -+ mnm5s2U conversion requires AdoMet as methyl group donor. Table I1 shows that this conversion occurs in uitro in the presence of AdoMet. We therefore propose that the first carbon atom in the side chain at the 5 position of mnm6s2U originates from a source other than methionine and would not be affected in methioninebstarved cells. However, other alternatives certainly exist. Thus, the sequence of events leading to mnm5s2U seems to be complex. The trmE mutant contains s2U, indicating that one step preceding the formation of cmnm5s2U is catalyzed by the trmE gene product and that the thiolation reaction does not require modification in position 5. The asuE gene is suggested to be involved in the thiolation step since tRNA from an asuE mutant contains mnm5U (Sullivan et al., 1985). Therefore, the stepwise for-trmC1 trmC2 -nm5s2U34 7 mnm5s2U34 (Fig. 8). However, the Adohlet thiolation reaction may occur all through the pathway (see above). The conversion of cmnm5s2U to nm5s2U occurs without any addition to the enzyme and does not require any external energy. The actual mechanism of this reaction awaits further studies.