INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

All tRNA species from the three domains, Archaea, Bacteria, and Eucarya, contain modified nucleosides, which are derivatives of the four nucleosides, adenosine, guanosine, cytidine, and uridine. At present, more than 79 different modified nucleosides from the tRNA of various organisms have been characterized (23). Some of these are present in tRNA from only one domain, but a few are present in the same subset of and at the same position in the tRNAs from all three domains (3). One such conserved group of modified nucleosides is the threonylated adenosine (t⁶A) derivatives. These modified adenosines are present adjacent to and 3' of the anticodon (position 37) in the subset of tRNAs that reads codons starting with A. The universal presence of t⁶A derivatives suggests that these kinds of modifications may have been present in the tRNA of the progenitor, unless a convergent evolution has occurred. This conservation also suggests that the functions of these modified nucleosides may be principally the same in all organisms.

In Escherichia coli, the t⁶A37 derivative N⁶-methyl-N⁶- threonylcarbamoyladenosine (m⁶t⁶A37) is present in only two tRNA species, the tRNA_GGU^Thr species, with the same anticodon (20). Threonine is the precursor in the synthesis of t⁶A (10, 32), and in vitro threonylation requires carbonate and ATP (15, 21). Here we show that the methyl group of m⁶t⁶A37 originates from methionine. So far, no mutant deficient in any t⁶A37 derivative has been characterized. As a first step to elucidate the syntheses of these groups of modified nucleosides and their roles in vivo, we have isolated and characterized a mutant deficient in the synthesis of m⁶t⁶A37. We show that the tsaA gene most likely encodes the tRNA(m⁶t⁶A37)methyltransferase that transfers a methyl group from S-adenosylmethionine (AdoMet) to the two tRNA_GGU^Thr species containing the t⁶A moiety. The tsaA gene was localized to the 4.6 min site on the E. coli chromosome. We also show that the methyl group of m⁶t⁶A37 slightly improves the translational efficiency of the two tRNA_GGU^Thr species.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Bacteria, growth conditions, and genetic procedure. Bacterial strains used were all derivatives of E. coli K-12 and are listed in Table 1. Cultures were grown in either Luria-Bertani (LB) medium (1) or nutrient broth (0.8%; Difco Laboratories, Detroit, Mich.) supplemented with 0.5% NaCl. The minimal medium was made from the basal medium (medium E) described by Vogel and Bonner (40) supplemented with 0.2% glucose, thiamine (1 µg/ml), and required amino acids (50 µg of the L isomer per ml). In some experiments the MOPS (morpholinepropanesulfonic acid)-glucose minimal medium (28) and M9 medium (25) were used. Kanamycin and carbenicillin were used at 50 µg/ml. The Hfr mapping procedure was adapted from Singer et al. (39). P1 transduction was done as described by Miller (25).

Analysis of modified nucleosides in tRNA. Cells were grown in 45 ml of LB medium, harvested at a cell density of about 4 × 10⁸ cells/ml by centrifugation, and washed with 0.9% NaCl. tRNA was prepared by lithium chloride fractionation (9) and degraded to nucleosides by nuclease P1 and alkaline phosphatase (17). The hydrolysate was then analyzed by high-performance liquid chromatography (HPLC) according to the method of Gehrke and Kuo (17). For the analysis of methylated nucleosides by thin-layer chromatography, cells were grown in rich MOPS medium (27) in the presence of L-[methyl-¹⁴C]methionine (0.074 µg/ml; 20 or 55 µCi/µM). The methyl-¹⁴C-labelled tRNA was degraded to nucleosides as described previously (2) and analyzed by thin-layer chromatography as described by Rogg et al. (34). The various radioactive compounds were located by autoradiography, and the radioactivity in each compound was determined by scintillation counting (2).

Determination of tRNA methyltransferase activity in vitro. Cells were grown in 50 ml of LB medium at 37°C and harvested at a density of 3 × 10⁸ cells/ml. The cells were pelleted, washed twice with 0.9% NaCl and once with buffer A (25 mM Tris-HCl [pH 7.4], 10 mM magnesium acetate, 0.1 mM dithiothreitol, 1 mM EDTA, 10% [vol/vol] ethylene glycol), and resuspended in 0.5 ml of buffer A. Cells were disrupted by sonication three times for 5 s at 20% power on a VCX400 sonicator (Sonics & Materials Inc., Danbury, Conn.). Cell debris was removed by centrifugation at 15,000 rpm with a Beckman JA-20 rotor for 10 min at 4°C, and the supernatant was transferred to a new tube. Ribosomes were removed by ultracentrifugation at 300,000 × g. The obtained supernatant was used as an enzyme source. The reaction mixture contained 100 µg of bulk tRNA from strain GRB1109 (tsaA1) or strain GRB1108 (tsaA⁺), 80 µl of enzyme extract, 10 µl of [methyl-¹⁴C]AdoMet (60 µCi/µM), 20 mM Tris-HCl (pH 7.5), 10 mM MgCl₂, 10 mM NH₄Cl, and 0.1 mM dithiothreitol in a final volume of 1 ml. The mixture was incubated for 3 h at 37°C. The reaction was stopped by adding 1 ml of phenol saturated with water followed by vigorous shaking for 10 min at room temperature. RNA in the aqueous phase was precipitated by ethanol. The precipitate was washed twice with 80% ethanol, dried, and digested either to nucleotides or to nucleosides. The distribution of methylated nucleotides or nucleosides was determined by two-dimensional thin-layer chromatography as described by Nishimura (30) and Rogg et al. (34), respectively. The radioactive compounds were detected by a PhosphorImager (Molecular Dynamics) (data not shown). Meanwhile, the rest of the nucleoside sample was assayed by HPLC and the radioactivity in the eluate was monitored by a flow scintillation analyzer (Radiomatic FLO-ONE beta; Packard Instrument Co., Meriden, Conn.) (Fig. 2C).

Determination of sensitivity to various amino acid analogs. Strains GRB1108 (tsaA⁺) and GRB1109 (tsaA1) were tested for sensitivity to amino acid analogs as described by Ericson and Björk (16) and Cortese et al. (11). All the 27 analogs specified in reference 16 were obtained from Sigma Chemical Co., St. Louis, Mo.

Determination of growth and polypeptide chain elongation rates. Growth rates at 37°C in rich MOPS medium and three MOPS minimal media with three different carbon sources (glucose, glycerol, and acetate) were determined as described by Björk and Neidhardt (5). Polypeptide chain elongation rates were determined as described by Ericson and Björk (16) and Schleif et al. (36).

Determination of ribosomal A-site selection and P-site frameshifting. The rate of aminoacyl-tRNA selection to the A site and the P-site frameshifting ability were determined as described earlier (13, 14, 18).

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Strain GRB105 lacks three modified nucleosides in its tRNA. During the analysis of various aro strains of E. coli for the presence of two modified nucleosides (cmo⁵U and its methylester, mcmo⁵U) (2), we noticed that one such strain (RE113 [aroE]), in addition to cmo⁵U and mcmo⁵U deficiencies (due to the aroE mutation) also lacks an unknown methylated nucleoside denoted no. 5 in Fig. 1 of reference 2. A spontanous Str^r derivative of strain RE113 was isolated for future Hfr mapping. This strain (GRB105 [aroE strL]) was grown in rich MOPS medium containing L-[methyl-¹⁴C]methionine. tRNA was prepared, digested to nucleosides, and analyzed by two-dimensional thin-layer chromatography (34). Like strain RE113, strain GRB105 also lacked compound no. 5 (2). Further analyses suggested that compound no. 5 may be m⁶t⁶A. First, m⁶t⁶A migrates similarly to our compound no. 5 (34). Furthermore, compound no. 5 becomes radioactively labelled when wild-type cells are grown in the presence of either [¹⁴C]HCO₃ or [¹⁴C]threonine, conditions that are known to label m⁶t⁶A (15).

The tsaA gene is located at 4.6 min on the E. coli chromosome. We used strain GRB105 (aroE stcA1 trmG1 tsaA1 strL) in Hfr crosses with strains (39) that have Tn10 at defined locations to localize the tsaA1 mutation to the region between 96 and 7 min (between cysA [96.5 min] and zag-3198::Tn10 [6.8 min]). We scored the allelic states of genes tsaA, stcA, and trmG from recombinants by HPLC analysis of the modification patterns of tRNA. The stcA and trmG genes are located in the regions from 28 to 36 min and from 10 to 28 min, respectively, (33a) and were not further characterized.

View larger version (14K): [in this window] [in a new window]   FIG. 1.   Mapping of the tsaA gene by P1 transduction. A physical map of the donor and recipient is shown. Strain CAG18447 (proAB81::Tn10) was used as the donor, and strain GRB1037 (tsaA1 zae-3095::Tn10kan) was used as the recipient. Tcr transductants were selected. tRNA was prepared from 100 transductants, and the TsaA phenotype was determined by HPLC analysis of tRNA for the presence or absence of m6t6A. The cotransduction frequencies between proAB81::Tn10 and tsaA1 and between proAB81::Tn10 and zae-3095::Tn10kan are indicated (see text for the order of the genes). The cotransduction frequencies between zae-3095::Tn10kan and tsaA1 were obtained from another transduction in which strain GRB1037 (tsaA1 zae-3095::Tn10kan) was used as the donor and strain GRB1450 (thrA::lacZ fusion) was used as the recipient. One hundred Kmr transductants were selected, and the TsaA phenotype was scored as derepressed in thrA::lacZ transcription, where the level of transcription was measured as -galactosidase activity.

The tsaA gene probably encodes tRNA(m⁶t⁶A37)methyltransferase. To further characterize the tsaA1-related nucleoside deficiency, tRNA was prepared from the two congenic strains GRB1108 (tsaA⁺) and GRB1109 (tsaA1) and digested to nucleosides. Figure 2A and B show that the only difference in the tRNA modification patterns of the wild-type tsaA⁺ and the mutant tsaA1 observed was due to the compound that we now identify as m⁶t⁶A. We identify this compound as m⁶t⁶A by the following criteria. (i) Its migration in the chromatography system devised by Gehrke and Kuo (17) is the same as that observed for m⁶t⁶A in their analysis. (ii) The UV spectrum is the same as that published for m⁶t⁶A. (iii) By mass spectroscopy, the molecular weight of the unknown compound was found to be 427 as expected for the protonated form of m⁶t⁶A. Only two tRNA species, tRNA_GGU^Thr1 and tRNA_GGU^Thr3 in E. coli, have m⁶t⁶A37 (20). If only the methyl group were lacking in the mutant tRNA_GGU^Thr, we would expect an increase in the level of t⁶A in bulk tRNA. Indeed, the level of t⁶A in the mutant increased (the ratio of t⁶A to m²A, measured as the absorbancy at 254 nm, increased from 0.71 in the wild type to 0.96 in the tsaA1 mutant). Together, these data suggest that tsaA1 prevents the methylation of t⁶A.

View larger version (24K): [in this window] [in a new window]   FIG. 2.   HPLC of nucleosides from tRNA from strain GRB1108 (tsaA+) (A) and strain GRB1109 (tsaA1) deficient in m6t6A (B). tRNA was degraded to nucleosides and a part (retention time, 0 to 80 min) of the chromatogram is shown. No difference in the patterns of modified nucleosides for the wild type and the mutant than that shown in the figure was observed. The identification of the compound migrating at min 67.538 as m6t6A was based on the following criteria: (i) the relative retention time was the same as that found for m6t6A with the HPLC system devised by Gehrke and Kuo (17) and used in the experiment described and also the same as that found by Buck et al. (9), as revised by Limbach et al. (23); (ii) the spectrum of this compound was the same as that reported for m6t6A (17); and (iii) the molecular weight was that of protonated m6t6A (427), as determined by the electrospray method. (C) HPLC of nucleosides from in vitro-methylated tRNA from strain GRB1109 (tsaA1). The radioactivity was monitored by a Radiomatic FLO-ONE scintillation counter, and the absorbance at 254 nm was measured by a photodiode array detector (Waters AB). The retention time of the major radioactive compound, which was not observed with tRNA from GRB1108 (tsaA+) as the substrate (data not shown), corresponded to that of m6t6A, as determined by measuring the UV absorbance at 254 nm. Although the retention time for m6t6A in panel C was several minutes shorter than in panel A, the retention time of m6t6A relative to m2A was the same for m6t6A in panel A (0.956) as in panel C (0.957). The small radioactive compound migrating at 31 min coincided with m5U, and a similar amount of radioactive m5U was also observed with tRNA from strain GRB1108 (tsaA+).

Lack of m⁶t⁶A37 does not affect the growth rate, the polypeptide step time, or the response to 27 amino acid analogs. Several tRNA modification-deficient mutants grow slower than the wild type and translate mRNA at a reduced rate (4). Therefore, we investigated whether the lack of m⁶t⁶A37 also reduced the growth rate and the polypeptide chain elongation rate. Steady-state cultures of the congenic pairs GRB1108 (tsaA⁺) and GRB1109 (tsaA1) were prepared in MOPS-glucose, MOPS-glycerol, MOPS-acetate, and rich MOPS media at 37°C. We did not observe any differences between the growth rates of the wild-type control and the tsaA1 mutant in any of these media (data not shown). Furthermore, we also determined the time by which ribosomes translate the lacZ mRNA (16, 36). No difference between the tsaA1 mutant and the wild-type control was observed. It has been shown earlier that a deficiency in 38, 39, 40 (11), ms²io⁶A37 (16), or m¹G37 (22) results in an altered response to various amino acid analogs, suggesting that expression of the corresponding biosynthetic enzymes may be altered (see reference 4 for a recent review). We therefore tested 27 different amino acid analogs (specified in reference 16), but no effect was observed with any of the tested analogs. Note, however, that no threonine analog was tested.

Lack of the methyl group of m⁶t⁶A37 does not induce +1 frameshifting at an ACC-A or ACC-U site and does not influence the Thr-tRNA selection to the ribosomal A sites at ACC and ACU codons. There is precedent for the possibility that a deficiency in a modified nucleoside, such as m¹G37 or ms²i⁶A37, may cause the tRNA to make +1 frameshift errors (6, 18, 33, 38). In Salmonella typhimurium, a +1 frameshift suppressor (sufJ105) was isolated and shown to be a derivative of tRNA_GGU^Thr3 (8). The altered tRNA_GGU^Thr3 promotes +1 frameshifting at sites in the mRNA where any of the four-base sequences ACC-A, ACC-U, and ACC-C is encountered (7). (The ACC codon is in the zero frame, and a +1 frameshift moves the reading frame one step to the right, e.g., to the CC-A triplet.) Since m⁶t⁶A37 is present in the wild-type form of this tRNA, we wanted to determine whether the lack of m⁶t⁶A37 induced +1 frameshifting at a similar site. A plasmid (pTHF71) that harbored a hybrid lacZ gene was constructed such that the ribosomes must shift from the zero frame to the +1 frame to obtain a functional -galactosidase enzyme (18). An ACC-A sequence, which may be a potential +1 frameshifting site for this tRNA, was included within the frameshifting window. This plasmid was then introduced into tsaA⁺ (GRB1319) and tsaA1 (GRB1320) strains. However, no significant difference in the -galactosidase specific activities between the tsaA⁺ and tsaA1 strains was observed (data not shown). Furthermore, the tsaA1 mutation did not increase the frameshifting at a prfB (release factor 2 [RF2]) programmed-frameshift site that was modified to have ACC at the slip site (13). These results thus indicate that the lack of the methyl group of m⁶t⁶A does not induce a +1 frameshift at ACC-A and ACC-U sites.

The presence of m⁶t⁶A37 improves the reading of the ACC present in the leader mRNA of the thr operon. The threonine (thrABC) operon of E. coli consists of three genes whose expression is regulated by an attenuator located upstream of the first gene, thrA (Fig. 3). The leader mRNA contains a 21-codon open reading frame in which there are four Ile codons and eight Thr codons (24). Seven of the Thr codons are ACC, which is read by m⁶t⁶A37-containing tRNA_GGU^Thr1 and tRNA_GGU^Thr3. The rate with which the ribosome traverses these control Ile and Thr codons determines the level of transcription termination of the leader mRNA (24). We introduced the tsaA1 mutation into a strain containing a transcription fusion in the thrA gene such that the activity of -galactosidase reflects the level of transcription of the thr operon (35). The data indicate that both in rich (LB) medium and in MOPS-glucose minimal medium the m⁶t⁶A37 deficiency of tRNA_GGU^Thr resulted in a twofold derepression, suggesting that the undermodified tRNA is less efficient than the fully modified counterpart in decoding the ACC codon. The measured -galactosidase activities are as follows: for LB medium, the activity of the wild type was 9.3 ± 1.8 Miller units and that of the tsaA1 mutant was 16 ± 1.0 Miller units; for MOPS-glucose minimal medium, the activity of the wild type was 177 ± 17 Miller units and that of the tsaA1 mutant was 363 ± 22 Miller units. Strains were grown at 37°C to about 3 × 10⁸ cells/ml, and -galactosidase activity was determined as described in reference 25.

View larger version (22K): [in this window] [in a new window]   FIG. 3.   Features of the regulatory region of the thrABC operon (modified from reference 24). The structure shown is that when the terminator (region 3-4) is formed and when region 1B and region 2 are paired. Region 1A can also pair with region 2, resulting in the pairing of regions 3 and 4 and the formation of the terminator. Only when the ribosome is stalled, especially in region 1B, does derepression of the operon occur.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

This paper describes the first mutant found to be defective in the synthesis of a t⁶A derivative; these derivatives are present in the same subset of the tRNAs from all organisms. One such derivative is m⁶t⁶A (Fig. 4), which is present only in two tRNA species in E. coli, tRNA_GGU^Thr1 and tRNA_GGU^Thr3. Gene tsaA was localized to min 4.6 of the chromosomal map of E. coli (Fig. 1) and most likely encodes tRNA(m⁶t⁶A37)methyltransferase, which catalyzes the addition of the methyl group to the t⁶A in tRNA. Although lack of this methyl group does not influence the growth rate, the average polypeptide chain elongation rate, and the selection of the Thr-tRNA_GGU^Thr-GTP-EF-Tu ternary complex to the cognate codons ACC and ACU, we observed a twofold derepression of the thr operon (see data in Results). These results suggest that the m⁶t⁶A37-modified nucleoside improves the efficiency of tRNA_GGU^Thr1 and tRNA_GGU^Thr3, probably in a step after the aminoacyl-tRNA selection step.

View larger version (18K): [in this window] [in a new window]   FIG. 4.   Structure of m6t6A (N6-methyl-N6-threonylcarbamoyladenosine).

Position 37, which is next to and 3' of the anticodon, is hypermodified for tRNAs reading codons starting with U (isopentenyl-adenosine derivatives [i⁶A, ms²i⁶A, and ms²io⁶A] or Y base) and A (t⁶A derivative). It was therefore suggested that the intrinsically weak interaction of the A-U/U-A base pair in the first position of the codon has to be stabilized by a hypermodification at position 37 (19, 29). Their universal occurrence suggests that the t⁶A derivatives play some essential role in the performance of the tRNA, perhaps in the stabilization of the codon-anticodon interaction by stacking onto the base-paired complex. Bulk tRNA isolated from E. coli starved for threonine and, therefore, deficient for the t⁶A modifications, does not function normally in in vitro protein synthesis (26). However, we find that the tsaA1 mutation has no effect on the average polypeptide chain elongation rate in vivo. There may be several reasons for the apparent difference between the results of these experiments. First, our mutation does not prevent threonylation, but only the methylation of t⁶A, and it is possible that the threonyl group provides much of the stabilizing effect. This is consistent with the fact that most of the tRNAs with t⁶A37 are also not methylated. It will be of interest to isolate and study mutants that fail to threonylate tRNA. Second, the assay may not be sensitive enough to detect minor effects at the relatively small set of codons (codons ACC and ACU) affected by the absence of the 6-methyl group.

To amplify any possible small reduction in the efficiency due to tsaA1, we examined the expression of the thr leader, which contains seven consecutive ACC codons. Here, tsaA1 caused a twofold derepression. This effect may be due either to the summation of a small signal from each codon or to some context effect specific to the cluster. In any case, these results show that the 6-methyl group of m⁶t⁶A37 does improve the efficiency of reading the cognate codon ACC. It is possible that this observed derepression could be caused by decreased aminoacylation of tRNA_GGU^Thr1 or tRNA_GGU^Thr3. However, the native tRNA_GGU^Thr3 and the completely unmodified form show similar aminoacylation kinetics (37), suggesting that m⁶t⁶A37 does not influence this reaction. Therefore, the derepression of the thr operon-lacZ fusion was likely to be caused by a less-efficient decoding of the ACC codons in the thr leader. Whether this also applies to the other cognate codon (ACU) read by these tRNA species awaits further analysis.

It is useful to compare the effects of tsaA1 to those of the hisT mutation, which prevents the conversion of uridine in the 3' side of the anticodon arm of primary transcripts to pseudouridine (). The hisT and tsaA1 mutations have comparable effects on the deattenuation of thr leader constructs. Lynn et al. (24) replaced thr regulatory codon ACC (Thr) with the CAU (His) codon in thr leader mRNA. The CAU codon is translated by tRNA^His, which normally contains  at positions 38 and 39. The hisT mutation derepresses this thr operon allele two- to threefold, i.e., to an extent similar to that observed by us for the tsaA1 mutation. Therefore, the reduction in the translational efficiency of tRNA_GGU^Thr by the m⁶t⁶A37 deficiency that was observed may be quantitatively similar to that caused by the absence of  in the anticodon region of tRNA^His. However, unlike the tsaA1 mutation, the hisT mutation strongly reduces growth rate and polypeptide chain elongation rate (31). These substantial differences between the effects of these two mutations on global protein synthesis may be related to large differences in the numbers of affected codons. The tsaA1 mutation affects only two tRNAs and a correspondingly small set of codons. In contrast, nearly all of the tRNAs in E. coli contain  at one or more of positions 38, 39, and 40, and so the hisT mutation affects most tRNAs and codons.