INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

The central aerobic metabolic pathways of Salmonella typhimurium, the citric acid cycle (CAC), which oxidizes acetyl units to CO₂, and the Embden-Meyerhof-Parnas (EMP) pathway (glycolysis), which converts glucose-6-phosphate to pyruvate in conjunction with the pentose phosphate cycle, which oxidizes glucose-6-phosphate to CO₂, are responsible for generating energy and for supplying precursors for biosynthesis (Fig. 1). The choice of different central pathways is regulated in response to several environmental factors. When S. typhimurium cells grow on glucose (Fig. 1), most energy is generated by the Embden-Meyerhof-Parnas pathway, whereas the CAC mainly provides precursor metabolites such as oxaloacetate, 2-oxoglutarate, and succinyl coenzyme A (succinyl-CoA). During growth on acetate (Fig. 1), the CAC generates both energy and precursor metabolites. Under certain growth conditions specific anaplerotic pathways are induced to direct the synthesis of necessary precursors. The genes for enzymes needed for growth on "poor" substrates such as CAC intermediates and acetate are regulated by the phosphotransferase:sugar transport system (PTS) through activation by the cyclic AMP-cyclic AMP receptor protein (cAMP-CRP) complex (reviewed in references 45, 49, and 50). Among the inducible anaplerotic enzymes are phosphoenolpyruvate synthase (PPS), phosphoenolpyruvate carboxykinase (PEPCK), and the enzymes of the glyoxylate shunt. These enzymes are also regulated by the product of the cra gene (previously designated fruR) of the so-called fructose regulon (50). Special regulatory molecules such as acetyl phosphate are also involved in regulation of central metabolic pathways when cells are grown on acetate (reviewed in references 24 and 34). Thus, the central aerobic metabolic pathways are affected by several regulons, which adjust their activity in a strict dependence on the environment of the cell.

View larger version (19K): [in this window] [in a new window]   FIG. 1.   Biosynthetic pathways with some of the anaplerotic pathways, including the enzymes in those steps, depicted as dashed arrows with the enzyme names in italics. Some of the arrows depicting the pathways represent more than one biosynthetic step. Underlined carbon sources are those on which miaE mutants have a longer lag time than wild-type S. typhimurium, while the miaE mutants do not grow at all on carbon sources marked in boldface.

tRNA from all organisms contains modified nucleosides, which are derivatives of the four canonical nucleosides. About 90 different modified nucleosides have been characterized so far (32). The synthesis of all these occurs after the primary transcript of the tRNA has been made (for reviews on modified nucleosides in tRNA, see references 3 and 4). Modification of tRNA is influenced by the cellular metabolism; e.g., starvation for methionine leads to methyl-deficient tRNA (33), and iron starvation in S. typhimurium results in lack of the 2-methylthio group in those tRNAs that normally have 2-methylthio-N-6-(cis-hydroxy)isopentenyl adenosine (ms²io⁶A) (8, 47, 60). The modification level of tRNA can also affect the expression of biosynthetic enzymes (reviewed in references 3, 4, and 40). The effects, so far found, of tRNA modification on metabolism have all concerned metabolic pathways such as amino acid biosynthesis and iron uptake. In this paper we show that the ms²io⁶A modification of tRNA influences central metabolic pathways.

The synthesis of ms²io⁶A in S. typhimurium occurs as shown in Fig. 2. Judged from the chemical reactions involved, at least four enzyme activities are required for the synthesis of ms²io⁶A. The corresponding genetic loci are designated miaA, miaB, miaC, and miaE. The last step, hydroxylation, is dependent on the presence of molecular oxygen (O₂). The tRNA (ms²io⁶A37)hydroxylase is present under anaerobic conditions, but the hydroxylation reaction does not occur under these conditions (6), and it has been suggested that ms²io⁶A may function as a regulator of aerobiosis.

View larger version (15K): [in this window] [in a new window]   FIG. 2.   Suggested biosynthetic pathway for ms2io6A in S. typhimurium. Known cofactors and substrates are indicated. The gene designations correspond either to identified genetic loci or postulated, but not yet identified, functions. DMAPP, dimethylallyl diphosphate (2-isopentenylpyrophosphate); SAM, S-adenosyl methionine.

To further study the synthesis and function of the hydroxyl group of ms²io⁶A, we have previously isolated miaE::MudJ mutants of S. typhimurium LT2, which are unable to hydroxylate 2-methylthio-N-6-isopentenyl adenosine (ms²i⁶A). Such mutants can not grow on the CAC intermediates succinate, fumarate, or malate (41). Here, we show that the ability of S. typhimurium to grow on these dicarboxylic acids is strictly correlated with the presence of the hydroxyl group on the isopentenyl moiety of ms²io⁶A in position 37 (ms²io⁶A37).

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Strains and growth conditions. Strains used are listed in Table 1. Cultures were grown either in Luria broth (LB) according to the work of Bertani (2) or in Difco nutrient broth (0.8%; Difco Laboratories, Detroit, Mich.) supplemented with 0.5% (wt/vol) NaCl, adenine, tryptophan, tyrosine, phenylalanine, p-hydroxybenzoate, dihydroxybenzoate, and p-aminobenzoate (concentrations as stated in the work of Davis et al. [13]). As rich solid medium, TYS-agar (10 g of Trypticase Peptone, 5 g of yeast extract, 5 g of NaCl, and 15 g of agar per liter) was used. As defined liquid medium, morpholine propanesulfonic acid (MOPS) medium (37) supplemented with 0.2% of the relevant carbon source was used. Solid minimal medium was medium E (57) with no citrate added (NCE plates), supplemented with agar (1.5%) and 0.2 to 0.4% (wt/vol) of the relevant carbon source. Antibiotics were used in concentrations of 50 µg/ml for both kanamycin and carbenicillin. Anaerobic growth on agar plates was achieved by incubation in a Generbox (bioMérieux, Marcy l'Etoile, France).

HPLC analysis of tRNA nucleosides. tRNA was prepared as described by Buck et al. (7). RNA samples were digested to nucleosides by the method of Gehrke et al. (20) by using nuclease P1 (Boehringer Mannheim) and bacterial alkaline phosphatase (Sigma, St. Louis, Mo.). After centrifugation, appropriate amounts (usually 100 µg for stable RNA and 50 µg for tRNA preparations) were applied on a Supelcosil LC-18S column on a Waters high-performance liquid chromatography (HPLC) system. The gradient profile developed by Gehrke and Kuo was used (19).

HPLC analysis of ubiquinone. Ubiquinone was extracted from cells grown to mid-logarithmic phase according to the method of DiSpirito et al. (14) with the modification suggested for Escherichia coli. Before the extraction procedure, the wet weight was determined and ubiquinone-10 (Sigma) was added as an internal standard for the efficiency of extraction. The ubiquinone content was analyzed by HPLC as described by Sherman et al. (54). The number of isoprene units of the S. typhimurium ubiquinone was determined by comparison to known markers ranging from ubiquinone-7 to ubiquinone-10 (Sigma).

Uptake of [¹⁴C]malate. Cells were grown to a density of 2 × 10⁸ in MOPS-lactate medium at 37°C, harvested, washed, and resuspended in MOPS buffer without a carbon source and containing 100 µg of chloramphenicol/ml. To 6 × 10⁸ cells, 44.4 nmol of nonradioactive malate and 1.26 nmol of [¹⁴C]malate (48 µCi/µmol) were added, bringing the final malate concentration to 7.6 µM. Samples were taken at 0, 5, 10, and 15 min after the addition of [¹⁴C]malate and were immediately filtrated. The filters were washed twice with 10 ml of MOPS medium without a carbon source and containing chloramphenicol; then they were dried, and the radioactivity was determined in an LKB Wallac 1214 Rack- Beta Excel scintillation counter. For measurement of malate uptake in the presence of competitive inhibitors, either succinate or citrate was added to the cell suspension to achieve a concentration of 0.8 mM before addition of the labeling mix.

Preparation of extracts for enzyme assays. For extracts of soluble proteins, cells were grown in 200 ml of LB medium and harvested in late-logarithmic phase. The cells were washed in MOPS-HCl (pH 7.4), pelleted, resuspended in 2 ml of MOPS-HCl (pH 7.4), and lysed by sonication (6 times, 20 s each time, on a Vibra Cell [Sonic and Materials, Inc., Danbury, Conn.] at a setting of 2.5). The sonicated extract was centrifuged in a Beckman SW50:1 rotor at 100,000 × g for 1 h at 4°C. The upper two-thirds of the supernatant was collected and used for assays of soluble enzymes. Membrane preparations for assays of succinate dehydrogenase, succinate oxidase, glycerol-3-phosphate oxidase, and NADH oxidase activities were made from bacteria grown in LB plus 0.2% (wt/vol) glycerol to a density of 4 × 10⁸ cells per ml. The membranes were prepared according to the method of Fridén et al. (18). Protein concentration determinations were made with the BCA (bicinchoninic acid) Protein Assay Reagent kit (Pierce, Rockford, Ill.) as recommended by the manufacturer.

Enzyme assays. All enzymes were assayed at 30°C. 2-Oxoglutarate dehydrogenase complex activity was measured according to the method of Carlsson and Hederstedt (10). Succinyl-CoA synthetase was assayed by the method of Bridger et al. (5). Succinate dehydrogenase was assayed with artificial electron acceptors according to the method of Hägerhäll et al. (23). The fumarase assay was performed by the method of Racker (46), and malate dehydrogenase was assayed according to the work of Ochoa (38). Succinate and glycerol-3-phosphate oxidase activities were measured as oxygen consumption by using a Clark oxygen electrode as described elsewhere (28). The assay buffer was 20 mM MOPS (pH 7.4) plus 10 mM potassium phosphate (pH 7.4). The oxygen concentration of the buffer at 30°C was taken to be 0.237 µmol/ml (28). Substrates were added to a final concentration of 10 mM. The enzymatic origin of the oxidase activities was confirmed by inhibition with 1 mM KCN. -Galactosidase activity was determined according to the method of Miller (35), except that cells were lysed by the addition of 20 µl of chloroform and 50 µl of 0.1% sodium dodecyl sulfate per 500-µl sample.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Growth characterization of the miaE::MudJ mutants. The miaE::MudJ mutants were tested for aerobic growth on various carbon sources on minimal-medium plates (Table 2). The mutants appeared to grow slower than the wild type, LT2, on citrate and acetate, and they did not grow at all on the CAC intermediates malate, fumarate, and succinate (41). However, no difference was observed for growth on glucose, glycerol, or lactate. Determination of the kinetics of growth of cells shifted into a medium containing a particular CAC intermediate gave clearer insight into the growth behavior of the miaE::MudJ strains. Figure 3 shows the results of shifts from MOPS-glucose medium to MOPS-acetate, -citrate, -malate, and -succinate media. Wild-type cells resumed growth after a short lag period. In contrast, mutant cells immediately stopped growing when they were shifted to a medium containing either malate or succinate and did not resume growth at all. When shifted to citrate or acetate minimal medium, the miaE mutant resumed growth after a longer lag period compared to wild-type cells but then grew at the same growth rate as the wild type at steady state (Fig. 3; Table 3). If glucose was added to the mutant cells 4 h after they were shifted to MOPS-malate or MOPS-succinate medium, the cells immediately resumed growth at a rate typical for growth in MOPS-glucose medium (data not shown). We used the amino acid and base pool plates of Davis et al. (13) and aspartic acid and glutamic acid together to see if the inability of the miaE mutant to grow on succinate was due to starvation for any amino acid. The growth deficiency of the miaE mutant could not be relieved by the addition of any amino acids (data not shown). Thus, not only is the miaE::MudJ mutant unable to grow on some CAC intermediates, but upon a shift from glucose to any of these compounds, growth stops immediately and is never resumed.

View larger version (30K): [in this window] [in a new window]   FIG. 3.   Growth of S. typhimurium LT2 () and the miaE::MudJ mutant GT3099 (). At time zero, exponentially growing cells were shifted from MOPS-glucose to MOPS-citrate (A), MOPS-acetate (B), MOPS-malate (C), or MOPS-succinate (D).

The growth phenotype of miaE::MudJ mutants can be complemented by the wild-type miaE gene. We have previously demonstrated that the growth phenotype is 100% linked to the miaE::MudJ insertion (41), which shows that this phenotype is not caused by a nearby mutation. However, it was important to establish that the growth defect is due to a defective miaE gene and not to some polar effects induced by the MudJ insertion. The expression of the downstream open reading frame 17.8 (ORF17.8) gene may be altered by the MudJ insertion if a strong promoter is created at the insertion point, although ORF17.8 is transcribed in the opposite direction from the miaE operon. Also, a gene product encoded by the short intercistronic region between the miaE gene and ORF17.8 could cause the growth deficiency. Therefore, several plasmids covering either the whole miaE operon or parts of it were used to complement the growth deficiency of the miaE mutant. Only plasmids carrying the entire miaE gene could restore growth on malate, fumarate, and succinate (data not shown). Plasmid pUST120, which carries the miaE gene, has an inactive ORF15.6 gene, and includes only 43 nucleotides downstream of the miaE gene (41), complemented the growth deficiency. These results show that the observed growth deficiency on CAC intermediates is caused by the inactivated miaE gene.

The inability of miaE strains to grow on succinate, malate, and fumarate is correlated with the presence of the cis-hydroxypentenyl modification in position 37 of the tRNA. The growth deficiency of the miaE::MudJ mutant may be independent of the deficiency in ms²io⁶A modification. Another tRNA modification enzyme, the trmA gene product (42), has been shown to have two functions in E. coli, and this could also be the case for the MiaE enzyme. Therefore, we constructed strains carrying both the miaE::MudJ insertion mutation and either, or both, of the miaA and miaB mutations, which affect earlier steps in the ms²io⁶A synthetic pathway. If the miaE mutation causes the growth deficiency on CAC intermediates through a second function rather than through the hydroxylation deficiency of the tRNA, we would expect that the miaE mutants would retain their growth phenotype in the presence of the miaA or miaB mutation. If, however, the growth deficiency is caused by the lack of the hydroxyl group of ms²io⁶A, we would expect the miaA and the miaB mutations to be epistatic to the phenotype of the miaE mutant, since miaA and miaB single mutants can grow well on malate. The strains were tested for growth on malate, and the modification states of their tRNAs were confirmed by HPLC analysis. Note that the miaB mutant has a small amount of io⁶A, since apparently the MiaE hydroxylase is able, albeit poorly, to use as a substrate a tRNA lacking the ms² group, provided that the i⁶ group is present (Table 4). As can be seen in Table 4, only miaA is epistatic to miaE. The miaA mutation suppresses the growth deficiency of the miaE mutant, whereas the miaB mutation does not. Thus, when the A in position 37 is lacking the isopentenyl side chain (as in the miaA mutant), the miaE gene product is not necessary for growth on succinate, fumarate, and malate, but when the isopentenyl side chain is present (as in wild-type cells and in the miaB mutant), growth occurs only if this side chain is hydroxylated. Since the miaA miaE double mutant grows on malate (Table 4), the MiaE enzyme does not appear to have an alternative function (besides modifying tRNA) the lack of which causes the inability of the miaE mutant to grow on malate. Thus, it seems as if S. typhimurium in some way senses the hydroxylation status of the isopentenyl side chain of the nucleoside in position 37 of tRNA.

Dicarboxylic acid uptake is normal in the mutant. It was possible that the inability of the miaE mutants to grow on succinate, fumarate, and malate was due to a deficient dicarboxylic acid transport system (dct) (39). We measured [¹⁴C]malate uptake in cells grown in MOPS-lactate but found no difference between the wild type, LT2, and the miaE mutant GT3099 (data not shown). As expected, the uptake of malate was competitively inhibited by the addition of succinate or fumarate but not by the addition of citrate, which is taken up by the tricarboxylic acid transport system (27). Also, the miaE mutant is able to grow on malate in the presence of small amounts of lactate. We conclude that the inability of the miaE mutants to grow on dicarboxylic acids is not due to a deficient uptake of these carbon sources.

Activity of CAC enzymes. The inability of the miaE mutants to grow on malate, fumarate, and succinate might be due to lack of, or reduced levels of, enzymes metabolizing dicarboxylic acids. We measured the activities of the CAC enzymes malate dehydrogenase, fumarase, succinate dehydrogenase, succinate thiokinase, and the 2-oxoglutarate dehydrogenase complex in cells grown in LB. Extracts from wild-type and mutant cells showed no significant difference (data not shown).

Activity of the respiratory chain. A characteristic of E. coli mutants defective in the synthesis of ubiquinone is their ability to grow on glucose but not on succinate or malate (21). Thus, the phenotype of ubi mutants is reminiscent of that of miaE mutants. Therefore, we analyzed S. typhimurium wild-type and miaE::MudJ mutant cells for ubiquinone. Both strains contained ubiquinone-8, in accordance with what has been found earlier for S. typhimurium (12). The amount of ubiquinone in the cells (wet weight) was the same for the wild type and the mutant (data not shown). To exclude a malfunction of the respiratory chain in the miaE mutant, succinate oxidase and glycerol-3-phosphate oxidase activities of isolated membranes from cells grown in LB supplemented with 0.2% glycerol were measured. No differences in activity between wild-type and miaE::MudJ mutant cells could be detected.

The miaE growth phenotype can be suppressed by the addition of cAMP. The growth phenotype of miaE::MudJ mutants was compared to those of other mutants known to be affected in growth on CAC intermediates. Mutants (crr) defective in the IIA^Glc protein of the PTS sugar transport system are unable to grow on malate, fumarate, and succinate. In these mutants adenylate cyclase cannot be activated, and the cAMP necessary to relieve catabolite repression of genes needed for gluconeogenesis cannot be synthesized (reviewed in references 43, 45, and 48). miaE::MudJ mutants, as well as crr mutants, can grow on malate, fumarate, and succinate in the presence of 10 mM cAMP (Table 2). However, much higher concentrations of cAMP are needed to supplement the growth of the miaE mutant (Table 5). The inability of mutants (ptsI) defective in enzyme I of the PTS system to grow on glycerol can be suppressed by a mutation in the crr gene (51); the miaE mutation, however, cannot suppress the ptsI phenotype (Table 2). Furthermore, the growth abilities of an miaE::MudJ crr::Tn10 double mutant are different from those of either of the single mutants (Table 2).

Isolation and genetic mapping of mutations causing suppression of miaE::MudJ. Strains that were able to grow on succinate, although they still retained the miaE::MudJ insertion and were deficient in ms²io⁶A hydroxylation, could be isolated at a frequency of 10⁶. We designate these suppressor mutations ses (for suppressors of miaE-mediated inability to grow on succinate). All ses mutations also allowed for growth of the miaE mutants on malate. Tn10dTc transposon (59) insertions cotransducible with ses-1 were isolated. Preliminary mapping using the MudP and MudQ bank (Mud-P22 recombinant phage inserted at specific positions in the S. typhimurium chromosome [61]) as described by Benson and Goldman (1) indicated that the ses-1 mutation is located between 38 and 42 min on the S. typhimurium chromosome (data not shown). None of the other suppressors were cotransducible with the transposons next to ses-1. Thus, mutations at more than one locus can suppress the inability of miaE::MudJ mutants to grow on malate and succinate.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

S. typhimurium miaE-defective mutants cannot grow aerobically on succinate, fumarate, or malate (reference 41 and this work). This growth defect is correlated with the absence of the hydroxyl group of the isopentenyl moiety of those tRNAs that normally have ms²io⁶A37. In this work we have searched for an explanation to this unusual connection between central carbon metabolism and tRNA modification.

Mutants defective in the miaE gene were found to take up dicarboxylic acid from the growth medium, to contain functional CAC enzymes, and to respire like a wild-type strain. Growth on CAC intermediates is principally different from growth on glucose in that it requires gluconeogenesis. In cells grown in the presence of glucose, sufficient energy is obtained from the glycolysis pathway and the enzymes of the CAC are present in small amounts (Fig. 1). The finding that small amounts of compounds normally synthesized through gluconeogenesis, e.g., ribose, can support the growth of the miaE mutants indicates that gluconeogenesis is affected in these strains. Among the enzymes induced by growth on CAC intermediates are fructose-1,6-phosphatase, PPS, the malic enzymes (MAEs), and PEPCK (Fig. 1). Since either MAEs together with PPS or PEPCK can synthesize PEP from CAC intermediates, the defect in the miaE mutants must be in at least two enzymes, either in PEPCK and PPS or in PEPCK and MAEs. Since the miaE mutants grow well on lactate and pyruvate, the PPS appears functional. When E. coli cells grow aerobically on acetate, the flux from oxaloacetate (OAA) to PEP is about 3.6 mmol · g¹ · h¹ (24). In a cell growing on citrate, this flux is increased due to the increased growth rate and, more importantly, because the acetyl-CoA needed for fatty acid biosynthesis and other acetyl-CoA-requiring reactions must be generated from carbon coming through the CAC. During growth on succinate, fumarate, and malate, the strain on the enzymes catalyzing the flow from OAA to PEP is even higher, since the acetyl-CoA needed for biosynthesis in the first half of the CAC (to 2-oxoglutarate) is also synthesized through this pathway. The flux from OAA to PEP in cells growing on malate thus has to be severalfold higher than when the cells grow on acetate (calculated from reference 24). These differences in the requirements of gluconeogenesis during growth on acetate, citrate, or malate may explain the differences in phenotype seen for these carbon sources.

MiaE is most likely an enzyme that catalyzes the hydroxylation of the isopentenyl moiety of ms²i⁶A of tRNA (Fig. 2) (41). We found that if A37 is unmodified, as in an miaA-defective mutant, S. typhimurium can grow on malate independently of whether the miaE gene is functional or not. In contrast, MiaE is required for growth on malate if the isopentenyl part is present, in the form of i⁶A37 or ms²i⁶A37. Thus, it seems as if S. typhimurium has a mechanism to recognize the isopentenyl chain of i⁶A and sense whether it is hydroxylated. The presence of the hydroxyl group will then stimulate the activity or synthesis of some enzyme(s) needed for growth on succinate, fumarate, and malate. This suggestion is based on the assumption that the MiaA enzyme does not add an isopentenyl group to some other substrate in addition to tRNA. We find this unlikely, since the MiaA enzyme requires both the sequence A36-A37-A38 in the anticodon loop of the tRNA and a 5-bp anticodon stem for activity (22, 31, 36, 55). The structural requirements of the substrate for the MiaE enzyme are not known, but hydroxylation of i⁶A is more efficient when the ms² group is present (16, 41).

An miaA null mutation suppresses the phenotype of an miaE mutant. We interpret this as a result of the absence of the signaling molecule, the isopentenyl side chain of ms²i⁶A. There is, however, a possibility that the suppression is caused by a derepression of some key enzyme, which in turn metabolically suppresses the growth defect induced by the miaE mutation. Tsui et al. (56) have shown that a strain containing the miaA:: insertion has an increased oxidation rate of some amino acids and CAC intermediates and that it has a changed utilization of some primary carbon sources. Jones et al. (26) have shown that a mutation in the miaA gene derepresses phosphogluconate dehydrogenase, which is encoded by the gnd gene and is part of the pentose phosphate cycle. This derepression is specific, since the miaA1 mutation does not derepress the synthesis of glucose 6-phosphate dehydrogenase, which is encoded by the zwf gene and is upstream of the gnd gene product in the synthesis of ribose phosphate. Although derepressed synthesis of an enzyme in a complex metabolic pathway does not necessarily result in an increased pool of a downstream metabolite, we tested whether the miaA1-induced suppression required an intact gnd gene product. We found that this is not the case; i.e., the miaA1-mediated suppression of the miaE-induced growth defect did not require a gnd⁺ allele (data not shown). Thus, although we cannot completely rule out a mechanism in which the miaA1 mutation changes a pool of any of the glycolytic intermediates, we have so far no indications that this is the case.

The most apparent way the cell may sense the presence of the hydroxyl group of ms²io⁶A37 would be through translation. However, we have not been able to demonstrate any defect in translation in an miaE mutant (40). Also, E. coli lacks the miaE gene; consequently, it lacks ms²io⁶A in its tRNA and has ms²i⁶A instead (9, 41). Still, E. coli grows well on CAC intermediates (E. coli cannot grow on citrate due to the absence of a functional uptake system [29, 30]). In this respect, i.e., in lacking the hydroxylation group and still being able to grow on succinate, E. coli resembles the Salmonella miaE mutant strain carrying extragenic suppressor mutations (see below). Thus, we do not believe that the cis-hydroxyl group of ms²io⁶A exerts its effect in translation. However, the possibility that ms²io⁶A is essential for proper function of a certain tRNA species at certain sites in the translation of gene products necessary for growth on succinate, fumarate, or malate cannot be ruled out.

Screenings for extragenic suppressors have shown that there are several cellular components that can be genetically altered to compensate for the absence of the miaE gene, i.e., to allow growth on dicarboxylic acids. The suppressor-containing strains retain the ms²i⁶A hydroxylation deficiency. Preliminary results suggest that one of the suppressor mutations is not in the gene for any gluconeogenetic enzyme, but in a gene that might encode an uncharacterized transcription factor (data not shown). One way this suppressor mutation might act is by causing derepression of the synthesis of a gluconeogenic enzyme or by increasing the availability of a cofactor required for the activity of a key enzyme.

If the cis-hydroxyl group of ms²io⁶A37 does not exert its effect through translation, then our finding that the presence of the hydroxyl group on the tRNA regulates the ability of S. typhimurium to grow on succinate, fumarate, and malate represents a completely new mode of metabolic regulation. Few examples are known where tRNA participates in metabolism in ways other than through translation. A tRNA^Glu is required for -aminolevulinic acid synthesis in plants, algae, and many bacteria (25, 53, 58). Cross-linking by near-UV light of 4-thiouridine in position 8 (s⁴U8) and C13 in tRNA is believed to act as a photosensor that triggers a protective response in E. coli (17). Perhaps the hydroxylation of the i⁶A residue is required for some gluconeogenetic enzyme reactions or as part of a regulatory complex influencing gluconeogenesis.