One-Carbon Metabolism
Chapter
36
ROWENA G. MATTHEWS
One-carbon units at various oxidation states can be formed by the reduction of carbon dioxide as shown in Fig. 1. In this chapter, I will discuss transfers of one-carbon units at the oxidation state of formic acid (the formyl group), formaldehyde (the methylene group), and methanol (the methyl group). Although carboxylations and decarboxylations are formally one-carbon transfers, they occur by entirely different mechanisms than the aforementioned one-carbon transfers and are discussed in other chapters. Methane, which plays important roles in one-carbon cycles in archaebacteria, is neither formed nor metabolized in enteric bacteria.
Table 1 lists reactions in Escherichia coli that involve one-carbon transfers of formyl, methylene, or methyl groups. Two cofactors, S-adenosylmethionine (AdoMet) (Fig. 2, 1) and tetrahydrofolate (H4folate) (Fig. 2, 2), play important roles in these one-carbon transfers. AdoMet serves as a donor of methyl groups, whereas H4folate can transfer formyl, methylene, or methyl groups, and these groups can be interconverted while bound to H4folate. The structures of 10-formyltetrahydrofolate (10-formyl-H4folate) (3), 5,10-methylenetetrahydrofolate (CH2- H4folate) (4), and 5-methyltetrahydrofolate (CH3-H4folate) (5), as well as the enzymatic reactions required for their interconversion, are shown in Fig. 3. The biosynthesis of dihydrofolate and its reduction to H4folate are discussed in chapter 41. As noted in that chapter, the intracellular forms of folic acid typicially contain from five to eight glutamyl derivatives on the p-aminobenzoyl substituent at C-6, and these polyglutamate side chains increase the retention of folates and enhance their binding to and reaction with folate-dependent enzymes.
Table 1Metabolic roles for one-carbon transfer reactionsa |
As indicated in Table 1, the biosynthesis of methionine and histidine involves steps in which a one-carbon unit bound to H4folate must be provided. A deficiency of one-carbon units is therefore expected to reduce biosynthesis of these amino acids and slow the rate of protein translation in the cell. In the presence of trimethoprim, which inhibits the reduction of H2folate to H4folate and depletes cellular stores of H4folate derivatives, E. coli strains grow very slowly and fail to formylate their Met-tRNAfMet (34). Treatment with trimethoprim also induces a marked stringent response, as assessed by elevated levels of ppGpp in nucleotide assays (72, 85) and by alterations in RNA synthesis. Studies of the specific auxotrophic requirements created by trimethoprim revealed that methionine, glycine, and a purine base (e.g., hypoxanthine or inosine) are required for continued protein synthesis (2, 3). In the presence of these supplements, cell growth is unbalanced and proceeds in the absence of thymidylate, resulting in cell death. Initiation of protein synthesis is particularly sensitive to one-carbon deficiency, because translation of all proteins in E. coli and Salmonella typhimurium (official designation, Salmonella enterica serovar Typhimurium) is initiated with formylmethionine, and two one-carbon units, one for the production of methionine and a second for its formylation, are required to produce this amino acid.
Methionine Biosynthesis.
In methionine biosynthesis, the one-carbon unit is introduced in the terminal step of the pathway. This reaction is catalyzed by two homocysteine-methyltetrahydrofolate methyltransferases, the metE and metH gene products:
CH3-H4folate + l-homocysteine → H4folate + l-methionine
MetE is produced during growth in minimal media and shows an absolute specificity for CH3-H4folate polyglutamates. MetH is a cobalamin-dependent enzyme and is active only in media supplemented with vitamin B12. In nature, this enzyme is probably functional only during growth in the gut, where intestinal flora synthesize cobalamin and supplement dietary sources of the vitamin. The enzyme also requires activation by AdoMet (45) and a reducing system. Two flavoproteins that provide electrons derived from NADPH to methionine synthase during reductive activation have been isolated from E. coli (24). These proteins have been identified as flavodoxin, the fldA gene product (64), and an NADPH-flavodoxin (ferredoxin) oxidoreductase, the fpr gene product (4).
MetE has been shown to be one of the proteins induced during a shift from anaerobic to aerobic growth conditions (84). This observation is consistent with the assumption that the MetH protein is primarily active in nature during growth in the intestinal tract, where anaerobic conditions prevail and cobalamin is available. MetE is also strongly induced following the shift of a groEL mutant strain to a nonpermissive temperature (35), where it is the major soluble protein synthesized. Cystathionine lyase, which catalyzes another step in methionine biosynthesis, is also overproduced under these conditions. The physiological significance of these observations is not known.
Histidine Biosynthesis.
Histidine is synthesized from ATP and shares its requirements for one-carbon units with the pathway for purine biosynthesis, which will be discussed below. During the complex biosynthetic reaction sequence (reviewed in chapter 29), the intermediate phosphoribulosylformimino-5-aminoimidazole carboxamide ribonucleotide is amidated and cleaved to form 5-aminoimidazole-4-carboxamide ribonucleotide (AICAR) and imidazole glycerol phosphate. In this cleavage, one of the one-carbon units incorporated into ATP is retained in the imidazole ring, and the other is liberated in the formation of AICAR and can be returned to the purine biosynthetic pathway.
Glycine and Serine Biosynthesis and Interconversion.
During growth on glucose, serine is generated from 3-phosphoglyceric acid (71). However, during growth on fructose, lactose, glycerol, or acetate (23, 69, 71), serine can be derived from glycine, which in turn is synthesized from threonine (see below). Thus, during growth on these carbon sources, one-carbon units bound to H4folate are required for the conversion of glycine to serine catalyzed by serine hydroxymethyltransferase:
Glycine + CH2-H4folate → serine + H4folate
Initiation of Translation: Formylation of Initiator Met-tRNA.
The formylation of methionine for the initiation of translation occurs with methionine bound to the initiator tRNAfMet. The reaction catalyzed by Met-tRNAfMet formyltransferase is
Met-tRNAfMet + 10-formyl-H4folate → fMet-tRNAfMet + H4folate
The enzyme, first characterized by Dickerman et al. (13), has more recently been purified to homogeneity (38). The encoding gene, fmt, has been cloned, its sequence has been determined, and the gene has been shown to lie at 72.4 min on the E. coli chromosome (32). Guillon and coworkers (32) disrupted the fmt gene and showed that the resulting mutant strain exhibited severely decreased growth rates at 37°C in a rich medium (0.28 doublings per hour, in contrast to 2.3 for the control strain) and was unable to grow at 42°C. These authors also demonstrated that during growth of the mutant strain at 37°C, acylated (formylated) Met-tRNAfMet could not be detected (<3%). Thus, formylation of Met-tRNAfMet is important for protein translation but not essential at 37°C. The formyl group seems important for IF2-promoted selection of fMet-tRNAfMet in the 30S initiation complex (33).
One-carbon units in the form of 10-formyl-H4folate are required for the biosynthesis of purines. The conversion of CH2-H4folate to 10-formyl-H4folate requires the action of the bifunctional methylenetetrahydrofolate dehydrogenase/methenyltetrahydrofolate cyclohydrolase (Fig. 3). This bifunctional protein is the product of the folD gene, located at 12 min (570 kb) on the genetic map (14).
There are two steps in de novo purine biosynthesis that involve the incorporation of a one-carbon unit into the nascent purine. These are the reactions catalyzed by the two glycineamide ribonucleotide (GAR) transformylases and AICAR transformylase. The two GAR transformylases differ in that the formyl group is derived from 10-formyl-H4folate in the reaction catalyzed by one enzyme and from formate in the reaction catalyzed by the other. These three reactions, which are discussed in chapter 34, are as follows:
10-Formyl-H4folate + GAR → H4folate + formyl-GAR
Formate + ATP + GAR → ADP + Pi + formyl-GAR
10-Formyl-H4folate + AICAR → H4folate + formyl-AICAR
Many early textbooks and reviews indicate that the substrate for GAR transformylase is methenyl-H4folate rather than 10-formyl-H4folate. This error resulted from early observations that 10-formyl -(6-R,S)-H4folate, synthesized chemically and racemic at C-6 of the pterin ring, was not a substrate for the enzyme, whereas methenyl-H4folate was. The confusion was resolved when it was found that the unnatural stereoisomer of 10-formyl-H4folate is a potent inhibitor of GAR transformylase (10, 83) and that partially purified preparations of GAR transformylase from chicken liver or cellular extracts from S. typhimurium or E. coli are contaminated with methenyl-H4folate cyclohydrolase activity, which converts methenyl-H4folate to 10-formyl-H4folate but reacts only with methenyl-H4folate bearing the natural configuration at C-6.
The existence of two GAR transformylase enzymes, only one of which uses 10-formyl-H4folate as the formyl donor, explains some very puzzling observations in E. coli. Treatment of strains of E. coli with sulfanilamide, which inhibits folate biosynthesis and leads to a depletion of all reduced folate cofactors, was shown to result in the excretion of a breakdown product of AICAR, which had accumulated in the cells (82, 88). Since GAR transformylase catalyzes the first 10-formyl-H4folate-requiring step in purine biosynthesis, this observation was quite surprising. Investigation of the sources of one-carbon units introduced into the purine ring established that serine and glycine contribute all of the C-2 carbons of the purine ring (the reaction catalyzed by AICAR transformylase) but only 50% of the C-8 carbons (the reaction catalyzed by GAR transformylase), and formate is utilized extensively for the C-8 carbon (11). As will be discussed below, serine and glycine serve as precursors for one-carbon groups bound to H4folate, but these labeling studies indicated that formate is not derived from serine and glycine.
One-carbon units are required for the synthesis not only of the purines but also of the unique nucleotide component of DNA, thymidylate. The conversion of deoxyuridylate (dUMP) to thymidylate involves a methylation at C-6 of the uracil ring with a one-carbon unit donated by CH2-H4folate:
dUMP + CH2-H4folate → dTMP + H2folate
In this complex reaction, catalyzed by thymidylate synthase, the thyA gene product, H4folate supplies the reducing equivalents necessary to convert the methylene group to a methyl group. This is the only reaction in the cell resulting in the net oxidation of H4folate, and it contributes to the cellular requirement for H2folate reductase. Many chemotherapeutic drugs directed at enteric bacteria inhibit dihydrofolate reductase, because the resulting thymidylate deficiency is lethal to the organisms (2, 3). Additional information on the thyA gene and on thymidylate synthase can be found in chapter 35.
DNA Methylases in E. coli and S. typhimurium.
Two DNA methylases that are responsible for the bulk of DNA methylation in that organism have been identified in E. coli (reviewed in chapter 53). The dam (DNA adenine methylation) methylase methylates hemimethylated or unmethylated DNA duplexes on adenine residues in -GATC- sequences. AdoMet serves as the methyl donor, and the methylation product is 6-methyladenine. The dcm (DNA cytosine methylation) methylase methylates cytosine residues in -CCA/TGG- sequences to produce 5-methylcytosine on the second cytosine in the sequence.
DnaA and Methylated GATC Sites Are Required for Initiation of Replication.
One-carbon units may play important roles in the control of the initiation of DNA replication. Assembly of the chromosomal DNA replication complex occurs at the origin of replication, oriC. The origin contains an abundance of GATC sites (94). These sequences are sites of methylation by the dam methyltransferase of E. coli, and immediately after replication, the GATC sites on the newly synthesized DNA strand will be unmethylated. The DnaA protein, which plays a major role in initiating assembly of the replication complex, binds to multiple sites in the region of the origin and appears to induce a nucleoprotein structure that may involve wrapping of the origin DNA around a core of DnaA protein (25). The dnaA gene is located adjacent to oriC, and the expression of this gene has been shown to be regulated by the methylation status of a GATC site in the promoter region. When DNA in the origin of replication is hemimethylated, expression of the dnaA gene is repressed, and initiation of a new round of replication is prevented. Once the dam methylase has methylated the newly synthesized strand, expression of dnaA is increased, and DnaA binding to the chromosomal origin initiates a new round of DNA replication. Recent studies (6) have shown that oriC is protected from the dam methyltransferase for a significant period of time after passage of the chromosomal replication fork and that the kinetics of remethylation of oriC and of the GATC site upstream of dnaA are the same. These studies suggest a mechanism for timing the replication of DNA with respect to the cell cycle. Immediately after replication of oriC, this region is sequestered from remethylation, and the synthesis of DnaA is reduced. Once the region is released from sequestration, methylation of GATC sites in the region by the dam methylase is required before the synthesis of DnaA is induced and a new round of replication can be initiated. Deficiencies of AdoMet leading to decreased rates of methylation of GATC sites might then be expected to slow the rate of DNA replication, although this does not appear to have been tested experimentally. As will be noted below, the regulation of AdoMet synthesis is significantly affected by medium composition, and the possibility exists that this regulation relates the timing of the cell cycle and DNA replication to the availability of nutrients in the medium.
Hemimethylation of GATC sites on newly replicated double-stranded DNA is also essential for identification of the daughter strand in mismatch repair (54). The mutH, mutL, and mutS genes encode components of a mismatch repair complex that scans double-stranded DNA for mismatches. Once such a mismatch is found, MutH protein interacts with adjacent GATC sites in the parent strand and cleaves the unmethylated strand (92). The complex then excises about 2 kb of DNA, and the region is resynthesized with the parent strand as the template. Thus, mismatch repair must occur in the time between synthesis of the daughter strand of the DNA and methylation of the daughter strand by the dam methylase. Strains with reduced dam methylase activity are hypermutable (46). We might also expect that strains with acute deficiency of AdoMet would be hypermutable, although this has not been demonstrated. The requirement for methylation of both strands of the newly replicated DNA before a new round of replication is initiated (see above) may minimize the effects of lowered cellular levels of AdoMet.
Pantothenate Biosynthesis.
Pantothenate is a precursor of coenzyme A, and its synthesis from α-ketoisovaleric acid (an intermediate in branched-chain amino acid biosynthesis) is described in chapter 44. The first step in pantothenate biosynthesis involves the addition of a hydroxymethyl group to a-ketoisovalerate to form ketopantoic acid (67). This reaction, which is analogous to the synthesis of serine from glycine catalyzed by serine hydroxymethyltransferase, uses CH2-H4folate as the one-carbon donor.
Thiamine Biosynthesis.
There appear to be several pathways for the biosynthesis of the pyrimidine ring of thiamine. The classical pathway diverges from the pathway for de novo purine biosynthesis following the formation of 5-aminoimidazole ribonucleotide (AIR), which already contains a one-carbon unit derived from either formate or 10-formyl-H4folate (59, 60). Labeling studies suggest that the conversion of the five-membered imidazole ring of AIR to the six-membered pyrimidine ring of thiamine involves incorporation of three additional carbon atoms from the ribose ring of AIR (21).
If AIR were produced only via the pathway for de novo purine biosynthesis, then strains carrying mutations in the steps of purine biosynthesis leading up to the formation of AIR should all be thiamine auxotrophs. However, studies of Downs and Roth (17) demonstrated that strains with null mutations of the first gene of purine-thiamine synthesis (purF) can grow without thiamine either in the presence of exogenous pantothenate or if they carry a mutation in the panR locus that leads to overproduction of pantothenate. They postulated the existence of an additional path to AIR that does not involve PurF but requires the activities of the other enzymes involved in the biosynthesis of AIR (PurD, PurN, and PurI). Downs (15) also suggests the existence of yet another route to the pyrimidine of thiamine, termed alternative pyrimidine biosynthesis, that incorporates label derived from aspartate into the pyrimidine ring. This pathway does not require any of the purine biosynthetic enzymes and functions in their absence during growth under anaerobic conditions. The first mutations of genes in this pathway have now been isolated and are being characterized (16).
The reader may be wondering why this chapter contains such a detailed description of thiamine biosynthesis, when there does not appear to be a direct one-carbon transfer reaction involved in the conversion of AIR to the pyrimidine of thiamine or in the conversion of pantothenate or aspartate to pyrimidine. One of the sequelae of one-carbon deficiency is adenine sensitivity, which results in thiamine auxotrophy. When folate biosynthesis is limited, as occurs during treatment with sulfanilamide drugs, AICAR accumulates as mentioned above, owing to the inability to catalyze its formylation to formyl-AICAR. Under such conditions, strains exhibit acute adenine sensitivity (9). This sensitivity to adenine can be relieved by thiamine or its pyrimidine moiety, by panthothenate, by histidine, or by methionine and lysine. Mutations in the AICAR transformylase gene also lead to an adenine-sensitive phenotype (29). In view of the routes to the pyrimidine of thiamine discussed immediately above, it is reasonable to assume that adenine sensitivity in folate-deficient strains results from inhibition of the synthesis of AIR by the combined action of AICAR and adenine and can be alleviated either by supplying the pyrimidine of thiamine or by supplying a precursor for the PurF-independent pathway to pyrimidine. Methionine/lysine auxotrophy has been characterized as deriving from a functional coenzyme A deficiency (41, 68), resulting from decreased conversion of α-ketoisovalerate to pantothenate, and providing these amino acids may reduce the coenzyme A requirements of the cell and spare one-carbon units for formylation of AICAR. Histidine biosynthesis involves the formation of AICAR (see above), and the administration of exogenous histidine blocks the accumulation of AICAR from this source.
It is not yet clear why accumulation of AICAR, whether caused by one-carbon deficiency, decreased AICAR transformylase activity, or histidine production, should exacerbate adenine sensitivity. Adenine is known to inhibit the early steps of purine biosynthesis, both by feedback inhibition of phosphoribosyl pyrophosphate amidotransferase (PurF) and by repression of the purine biosynthetic genes (see chapter 34). However, thiamine deficiency does not normally result when adenine is present in the medium, suggesting that sufficient flux through the pathway occurs to meet the needs of thiamine biosynthesis. AICAR, also called ZMP, is converted to a trinucleotide, ZTP, by pyrophosphorylation, and it has been suggested that this compound serves as an alarmone in S. typhimurium, signalling a deficiency of formyl-H4folate (5). It is conceivable that a combination of ZTP and adenine shuts down purine biosynthesis more completely than adenine alone and thus starves the cell for thiamine unless the additional pathways of thiamine pyrimidine biosynthesis are activated.
Posttranslational Modification of Proteins and Nucleic Acids.
A number of proteins are methylated using AdoMet as the methyl donor. Methylations are known to occur on lysine, carboxylate groups, and other amino acid residues. Methylation/demethylation of glutamyl residues of proteins plays a particularly important role in bacterial chemotaxis. The methyl-accepting chemotaxis proteins are regulated by CheR (an AdoMet-dependent methyltransferase) and CheB (a protein methyl esterase) to regulate the sensitivity of the chemotactic mechanism to exogenous nutrients that bind the methyl-accepting chemotaxis proteins (see chapter 73).
Nucleic acids are also extensively modified in reactions that require one-carbon units, particularly by methylation or hydroxymethylation. A wide variety of AdoMet-dependent methyltransferases exist to catalyze specific methylations on sites in DNA (including the dam methyltransferase), RNA, and tRNA (reviewed in chapters 53 and 57). In general, these reactions impose only a very minor drain on the cellular one-carbon pool. The exception is the requirement for hydroxymethylation of bacteriophage DNA following viral infection by T-even bacteriophages. The DNA of T-even bacteriophages contains 5-hydroxymethylcytosine in place of cytosine, and the phage genome contains a gene coding for dCMP hydroxymethylase (22). This enzyme catalyzes the transfer of the methylene group of CH2-H4folate to C-5 of dCMP and the subsequent hydration of this methylene group to produce the hydroxymethyl substituent:
CH2-H4folate + dCMP + H2O + H+ → H4folate + CH2OH-dCMP
During phage infection, this reaction may impose a highly significant requirement for cellular one-carbon units.
The ultimate source of one-carbon units is of course the carbon source (or sources) in the medium. For cells grown on glucose as the carbon source, the major source of cellular one-carbon units is the β carbon of serine. Serine is synthesized from 3-phosphoglyceraldehyde, an intermediate in the glycolytic pathway. The activity of serine hydroxymethyltransferase then converts serine to glycine and a one-carbon unit:
Serine + H4folate → glycine + CH2-H4folate
Thus, during growth on glucose, the regulation of serine hydroxymethyltransferase is critical for regulating the flow of carbon into pathways that use H4folate-bound one-carbon units for biosynthetic pathways. As might be expected, the regulation of serine hydroxymethyltransferase is complex and occurs both to modulate enzyme activity and to modulate expression. This regulation is discussed extensively in chapter 30, but a brief overview of regulation pertinent to control of the synthesis of H4folate-bound one-carbon units will be presented here.
Two major systems are responsible for the transcriptional regulation of serine hydroxymethyltransferase expression. Transcription of the glyA gene is repressed by PurR, and this regulates serine hydroxymethyltransferase expression in response to the availability of purine trinucleotides in the cell (87). In addition, glyA transcription is positively regulated by MetR in response to the level of homocysteine, a precursor of methionine. Thus, MetR regulates serine hydroxymethyltransferase expression in response to demands for methionine biosynthesis. As might be expected, these systems modulate glyA expression but do not switch expression on or off. Thus, the total level of glyA transcription is independently responsive to the need for one-carbon units in many different metabolic pathways. The details of regulation by PurR and MetR are presented below.
PurR is a 38-kDa repressor protein (73) that binds specifically to 16-bp operator sequences upstream of its target genes (7). Repression by PurR requires the binding of a corepressor, hypoxanthine or guanine, to the protein (36, 52). PurR mediates inosine-dependent repression of glyA (87), which results in an ∼40% decrease in expression in a wild-type background.
MetR is a positive transcriptional regulator of methionine biosynthesis, and homocysteine serves as a coregulator (see chapter 30). The glyA gene is positively controlled by MetR, and homocysteine acts as a coactivator (66). The level of serine hydroxymethyltransferase activity is 1.6-fold higher in a metR + strain than in an isogenic metR strain (87). Furthermore, addition of methionine to the medium has no effect on serine hydroxymethyltransferase activity in a metR background but leads to a 50% reduction in activity in a metR + background.
The earlier studies of Dev and Harvey (12) had shown a correlation between the ratio of homocysteine to AdoMet and the rate of serine hydroxymethyltransferase synthesis, and these authors suggested that homocysteine acts as an inducer and AdoMet acts as a corepressor of enzyme synthesis. The identification of MetR as the protein responsible for the effect of homocysteine on glyA transcription also provides an explanation for the effect of AdoMet on serine hydroxymethyltransferase expression. Transcription of metR is regulated by the MetJ repressor protein, which uses AdoMet as a corepressor (89). It is probable that AdoMet has only an indirect effect on glyA expression, exerted via its effect on metR transcription.
Many of the H4folate derivatives that contain bound one-carbon units are interconvertible by reversible reactions. Thus, CH2-H4folate and H4folate can be interconverted by the action of serine hydroxymethyltransferase, and CH2-H4folate and 10-formyl-H4folate can be interconverted by the action of methylenetetrahydrofolate dehydrogenase/cyclohydrolase in a reversible reaction:
CH2-H4folate + NADP+ + H2O →10-formyl-H4folate + NADPH + H+
Thus, the needs for cofactors for the biosynthesis of purines and thymidylate can be balanced by interconversion of the different cofactors. In mammalian cells, methylenetetrahydrofolate reductase catalyzes an irreversible reduction of CH2-H4folate to CH3-H4folate (31) that commits one-carbon units to the pathway of regeneration of the methyl group of methionine in support of AdoMet-dependent methylation reactions:
CH2-H4folate + NADPH + H+ → CH3-H4folate + NADP+
The mammalian enzyme is allosterically regulated by the AdoMet/adenosylhomocysteine (AdoHcy) ratio in the cell, in a classic example of feedback regulation (40). Since reduced pyridine nucleotides are the ultimate source of reducing equivalents for this reaction in E. coli (see below), this reduction must also be irreversible and should represent a potential point of control of the flux of one-carbon units in the cell.
The metF genes from E. coli (77) and S. typhimurium (86) have been sequenced and shown to code for a protein with a subunit molecular mass of 33 kDa. The MetF protein was partially purified by Katzen and Buchanan (39), who showed that it was a flavoprotein that contained an flavin adenine dinucleotide prosthetic group. They assayed the enzyme by measuring the oxidation of CH3-H4folate to CH2-H4folate in the presence of menadione. More recently, a strain overexpressing metF has been constructed by transformation with a plasmid bearing the metF gene (18). By employing a maxicell system, the metF gene product was identified as an ∼32.5-kDa protein. Studies from my laboratory (C. Sheppard and R. Matthews, unpublished data) have established that the homogeneous purified enzyme is isolated with the flavin noncovalently bound (i.e., as holoenzyme) and it is directly reduced by NADH. In addition to catalyzing the oxidation of CH3-H4folate in the presence of menadione, the enzyme can also catalyze the reduction of menadione in the presence of NADH. Significant homologies were identified between the deduced amino acid sequence of MetF and the sequences of the human and yeast methylenetetrahydrofolate reductase proteins (30).
Expression of the metF gene was shown to be repressed by methionine (39), and this repression was later shown to be mediated by the MetJ repressor protein and its corepressor, AdoMet (76). The metF gene is also regulated by the MetR protein (8). MetR appears to antagonize MetJ-mediated methionine repression of the metF promoter. metF expression is repressed by vitamin B12 and the metH gene product (vitamin B12-dependent methionine synthase) (53, 55). It remains to be established whether vitamin B12-mediated repression of metF requires a functional MetR protein.
Dev and Harvey (11) studied the sources of one-carbon units in wild-type strains. When grown aerobically in glucose minimal medium containing [14C]formate, these strains incorporated label from formate into C-8 of guanine and adenine but not into C-2 of the purines or into methionine, histidine, serine, glycine, or thymine. Thus, the one-carbon units bound to H4folate were unlabeled, as expected if they were derived from serine and ultimately from glucose, but formate was incorporated into C-8 of purines.
These studies were puzzling from two points of view. First, E. coli lacks a formyltetrahydrofolate synthetase, the mammalian enzyme used to convert formate to 10-formyl-H4folate according to the stoichiometry shown below:
HCOOH + ATP + H4folate → 10-formyl-H4folate + ADP + H2PO4 –
The only enzyme then known to catalyze the formylation of GAR, PurN, uses 10-formyl-H4folate as the source of the one-carbon unit. Second, while it is known that formate is generated by the action of pyruvate formate-lyase during anaerobic growth on glucose, this enzyme is induced only under anaerobic growth conditions, and it is not known how formate is generated in aerobically grown cells.
In 1993, Nygaard and Smith described the isolation and cloning of an E. coli gene that complemented an E. coli mutant strain lacking any GAR transformylase activity and requiring purines for growth (63). This gene, designated purT, restored purine prototrophy to the mutant strain, as did the previously identified 10-formyl-H4folate-dependent GAR transformylase (the purN gene product). The PurT gene product required formate as the one-carbon donor. The PurT enzyme has been characterized by Marolewski et al. (47). The protein is a 42-kDa monomer and catalyzes the production of formyl-GAR from formate, ATP, and GAR. Catalysis is thought to involve a formyl phosphate intermediate. Aerobic growth of purN mutants is inhibited by glycine or by threonine, and the inhibitory effect of threonine is potentiated by leucine (63). Inhibition by glycine or by threonine and leucine is relieved by the addition of formate to the medium, suggesting that glycine reduces the concentration of formate in aerobically grown cells.
Very recent studies from Zalkin’s laboratory (56) have identified a gene, purU, that can also lead to purine auxotrophy in a purN background. A purU purN double mutant is auxotrophic for either purine or formate under aerobic growth conditions, suggesting that the purU gene product is responsible for providing formate during aerobic growth. However, the enzyme has not yet been characterized fully, and the reaction that it catalyzes remains undefined.
Labeling studies indicate that the source of one-carbon units in the cell is highly dependent on the carbon source. During growth on acetate, glycine is not derived from exogenous serine, but rather serine is derived from exogenous glycine. During growth on fructose, endogenous serine and glycine can be derived from either exogenous serine or exogenous glycine (71). These studies suggest that during growth on acetate or fructose, the concentration of the glycolytic precursor of serine, 3-phosphoglyceric acid, may be lower than it is during growth on serine. Fraser and Newman (23) subsequently demonstrated that a pseudorevertant of a glyA strain of E. coli lacking detectable serine hydroxymethyltransferase activity was not auxotrophic for glycine if threonine was provided in the medium. Pseudorevertants are extragenically suppressed mutants that retain their original mutation. l-Leucine was shown to exert a sparing effect on the amount of glycine required by the pseudorevertant strain to reach a given cell density and to stimulate the incorporation of label from threonine into purine. Newman and coworkers (61) subsequently demonstrated the activity of l-threonine dehydrogenase in extracts of E. coli and showed that the enzyme was induced by l-leucine. The metabolic pathway leading to the synthesis of glycine and serine from threonine was further defined by the studies of Ravnikar and Somerville (69), who termed the reaction sequence the threonine utilization cycle. They showed that strains that were unable to synthesize serine from glucose could grow in the absence of glycine on lactate or glycerol, provided that the medium was suppplemented with leucine, arginine, lysine, threonine, and methionine. These strains showed long lag phases when inoculated into similar media containing glucose. Addition of cyclic AMP (cAMP) to the glucose-containing medium did not shorten the lag period, although it is unclear whether this results from inducer exclusion or from regulation independent of the cAMP receptor protein-cAMP complex.
The pathway for conversion of threonine to glycine and serine is shown in Fig. 4. The net reaction for serine biosynthesis is
2 glycine + NAD+ → serine + CO2 + NH4 + + NADH
Thus, during growth on acetate, lactose, fructose, or glycerol, both the α and β carbons of serine may be derived from the α carbon of exogenous glycine.
Threonine dehydrogenase expression is very low in cells grown on glucose minimal medium, and its induction upon addition of leucine to the medium has recently been shown to be mediated by the leucine-responsive regulatory protein (Lrp), which negatively regulates the tdh operon (44, 70). Lrp also positively regulates the serA gene, which encodes the enzyme catalyzing the first step in serine biosynthesis from glucose (44, 70). Thus, Lrp appears to play a major role in controlling the source of cellular one-carbon units, exerting reciprocal effects on biosynthesis of serine and glycine from glucose (positively regulated by Lrp) and from threonine (negatively regulated by Lrp). Lrp is also a positive regulator of the gcv operon encoding the enzymes required for glycine cleavage (42).
AdoMet (Fig. 2, 1) is an activated methyl donor by virtue of the positive charge on the sulfur, which labilizes the bonds to each of the substituent groups. It serves as the methyl donor in a wide variety of biosynthetic reactions, including the methylation of DNA, RNA, and protein residues, the formation of cyclopropane fatty acids, and the biosynthesis of biotin. As shown in Fig. 5, reactions in which a methyl group is transferred from AdoMet to another molecule result in the production of AdoHcy (3). This metabolite can be hydrolyzed to form adenosine and homocysteine:
AdoHcy + H2O → homocysteine + adenosine + H+
Homocysteine can then be remethylated by the action of methionine synthase, using CH3-H4folate as the source of the methyl group:
Homocysteine + CH3-H4folate → methionine + H4folate
Methionine can be converted to AdoMet by the action of adenosylmethionine synthetase:
Methionine + ATP → AdoMet + PPi + Pi
Thus, the ultimate source of the methyl group of AdoMet is a one-carbon unit bound to H4folate, and during growth on glucose, this will be derived from serine.
Regulation of the flux of one-carbon units into AdoMet may be achieved by regulation of expression of MetF, which commits one-carbon units to methylation of homocysteine, by regulation of the expression of the two isozymes of methionine synthase (see below), which are regulated primarily in response to the availability of methionine in the medium, and also by regulation of the expression of adenosylmethionine synthetase. There also appear to be two isozymes of adenosylmethionine synthetase (80). One of these isozymes, the metK gene product, is the major isozyme during growth on glucose minimal medium, whereas the other, the metX gene product, is proposed to be induced during growth on Luria broth or after entry into stationary phase in minimal medium. Little is known about the regulation of metX, but metK expression is regulated by the MetJ repressor and AdoMet as a corepressor (78).
The magnitude of the cellular requirement for H4folate-bound one-carbon units is closely linked to the nutritional content of the medium. For cells growing on glucose minimal medium, it has been estimated that ∼15% of the carbon in glucose is used to provide serine, glycine, and one-carbon units (71). The detailed analysis of metabolic processes provided by Neidhardt et al. (57) can be used to estimate the approximate requirements for serine, glycine, and one-carbon units during growth on glucose minimal morpholinepropanesulfonic acid (MOPS) medium and on glucose-rich MOPS medium (supplemented with amino acids, nitrogen bases, and vitamins). During growth on glucose minimal MOPS medium, the cell has the following requirements to produce 1 g of cell mass (or 550 mg of protein mass): 90 μmol of His (requires one ATP and produces one molecule of AICAR per molecule); 146 μmol of Met (requires one one-carbon unit per molecule and one cysteine, derived from serine); 614 μmol of Ser (for serine residues in proteins, and for Met, Cys, Trp, phospholipids, and lipopolysaccharides); and 582 μmol of Gly. In addition, the production of 1 g of cells requires a total of 508 μmol of purine (165 μmol of ATP, 203 μmol of GTP, 24.7 μmol of dATP, and 25.4 μmol of dGTP, plus 90 μmol of ATP for histidine biosynthesis). Synthesis of 1 μmol of purine requires 1 μmol of glycine and 1.5 μmol of one-carbon units (CH2-H4folate equivalents, assuming that 50% of C-8 is supplied by formate). As a by-product of histidine synthesis, 90 μmol of AICAR is formed, which supplies the equivalent of 90 μmol of glycine and 90 μmol of one-carbon units toward purine biosynthesis. An additional 24.7 μmol of one-carbon units is required for the synthesis of 24.7 μmol of dTTP, and 14 μmol of one-carbon units is required for the formylation of fMet-tRNAfMet. Given the stoichiometry of the reaction catalyzed by serine hydroxymethyltransferase, which converts a molecule of serine into a molecule of glycine and a one-carbon unit, 1,565 μmol of Ser is required to meet the cell’s needs for 1,000 μmol of glycine, 614 μmol of serine, and 902 μmol of one-carbon units. This represents 10% of the carbon derived from the glucose consumption needed to produce 1 g of cell mass (7,869 μmol of glucose). It should be noted also that because of the imbalance between cellular requirements for glycine and for one-carbon units, which are produced in equimolar amounts by the action of serine hydroxymethyltransferase on serine, the glycine cleavage system is needed to convert 49 μmol of one-carbon units to glycine.
In contrast, during growth on a defined glucose-rich MOPS medium, the amino acids and nitrogen bases necessary for protein and nucleic acid biosynthesis are supplied, and the cell’s needs for serine, glycine, and one-carbon units are greatly reduced. The cell must still synthesize dTTP and must formylate Met-tRNAfMet; therefore, 39 μmol of one-carbon units is still required, as is 122 μmol of serine for phospholipid and lipopolysaccharide biosynthesis. Note that growth on glucose minimal MOPS leads to requirements for both one-carbon units and glycine, whereas growth on glucose-rich MOPS requires only one-carbon units, with glycine being supplied in the medium. The glycine cleavage system is still required but now must convert only 39 μmol of glycine to one-carbon units.
From these detailed calculations, the following principles emerge. (i) The expression of enzymes required for serine biosynthesis and generation of glycine and one-carbon units from serine must be regulated. (ii) The different needs for glycine and one-carbon units must be balanced, presumably by the action of the glycine cleavage system (the gcv operon). The enzymes in this operon catalyze the conversion of glycine to a one-carbon unit:
Glycine + H4folate + NADP+ → CO2 + NH4 + + NADPH + CH2-H4folate
How are shifts in demand for one-carbon units, glycine, and serine accommodated in E. coli? As we have seen, during growth on glucose, the key enzymes in the generation of these metabolites will be the enzymes required for serine biosynthesis, particularly 3-phosphoglyceraldehyde dehydrogenase (SerA), which catalyzes the committing step in serine biosynthesis during growth on glucose. In addition, serine hydroxymethyltransferase (GlyA) must be regulated, to balance requirements for glycine and one-carbon units with requirements for serine. Finally, the glycine cleavage enzymes (the gcv gene products) must be regulated to catalyze the conversion of glycine to one-carbon units. I have already discussed the regulation of serine hydroxymethyltransferase by PurR and MetR, which report needs for de novo biosynthesis of purines and methionine, respectively. The other major regulatory player is Lrp, which positively regulates the expression of both serA (44, 70) and the gcv operon (42). Recent studies have shown that the level of Lrp expression is decreased during growth of strains on rich media (20, 42), where the demand for one-carbon units will be greatly reduced. As we have seen, the requirement for the glycine cleavage system appears to be somewhat greater during growth on glucose minimal medium than during growth in a rich medium. This finding is consistent with the positive regulation of the gcv operon by Lrp (42). Decreased levels of Lrp in the cells should result in decreased biosynthesis of serine and in reduced levels of the glycine cleavage enzymes needed to interconvert glycine and one-carbon units.
In a review in 1988, Gold (27) suggested that global regulation of translational initiation might be achieved by controlling the availability of one-carbon units. He hypothesized that depletion of the H4folate pool would block initiation by eliminating formylation of the Met-tRNAfMet and that methionine plays a special role in metabolism as the major one-carbon donor. Since initiation uses two one-carbon donors, methionine and H4folate, to allow complete construction of an initiation complex, this process should be particularly sensitive to the availability of one-carbon units in the cell. Six years later, we may ask whether there is any evidence that the availability of one-carbon units is regulated to control the initiation of translation or of other processes, such as initiation of DNA replication, that depend on these one-carbon units.
If prototrophic strains of E. coli that are growing exponentially in a glucose minimal medium at 37°C are shifted to temperatures between 40 and 47°C, they exhibit methionine auxotrophy (74). This methionine auxotrophy was traced to decreased activity of the first enzyme in the methionine biosynthetic pathway, homoserine transsuccinylase, the metA gene product. In a medium supplemented with all amino acids except methionine, the effect of a temperature shift is even more dramatic (49). In a metK lrp strain that overproduces all of the enzymes of methionine biosynthesis, no methionine auxotrophy is seen following a temperature shift (49).
Heat shock induces other physiological changes in E. coli that affect the generation of one-carbon units. The rate of catabolism of serine to pyruvate catalyzed by l-serine deaminase is significantly elevated after a temperature shift (50, 62), thus decreasing the availability of serine for production of CH2-H4folate. During growth at 42°C, E. coli W3110 uses threonine rather than serine as the precursor from which glycine is made(26). Gage and Neidhardt (26) have also demonstrated that the fMet-tRNAfMet/Met-tRNAfMet ratio in prototrophic strain W3110 falls 10-fold following a shift from 28 to 42°C, providing strong support for Gold’s hypothesis that the initiation of translation may be regulated by the availability of one-carbon units (27). These observations may underlie the observation of Patterson and Gillespie (65) that a shift in temperature from 30 to 44°C results in transient inhibition of the initiation of protein synthesis.
If heat shock results in "pulling the plug in the bathtub drain," catabolizing serine and forcing the cell to use the threonine utilization cycle for the generation of glycine, and limiting the availability of one-carbon units for formylation of initiator Met-tRNAfMet, we would like to know how this reprogramming of one-carbon metabolism is accomplished. It is tempting to hypothesize that Lrp, known to be a global regulator of one-carbon metabolism (see above and chapter 94), might be involved in this reprogramming. Lrp negatively regulates both serine catabolism, by its effect on sdaA transcription, and synthesis of glycine from threonine, by its regulation of the tdh operon. Thus, if Lrp levels were decreased following a temperature shift, or if the effect of Lrp as a repressor were diminished, then the metabolic changes seen following a temperature shift might be explained. Furthermore, since Lrp positively regulates the gcv operon responsible for glycine cleavage to form one-carbon units, the diminished availability of one-carbon units for formylation of Met-tRNAfMet would be explained. However, there is no direct evidence that the level or activity of Lrp is altered by heat shock, and initial attempts to measure the level of formylation of Met-tRNAfMet in an lrp strain failed to demonstrate a reduced fMet-tRNAfMet/Met-tRNAfMet ratio during growth at 37°C (D. Gage and F. C. Neidhardt, personal communication). It should be noted that the expression of another of the proteins negatively regulated by Lrp, the lysU gene product lysyl-tRNA synthetase form II (43), is also increased following a temperature shift (58). Further investigation of the possible relationship between Lrp and the reprogramming of one-carbon metabolism associated with heat shock should be a high priority.
Strains with reduced adenosylmethionine synthetase activity associated with mutations in the metK gene rapidly accumulate secondary mutations in lrp (44). These double-mutant strains exhibit abnormal induction of the heat shock proteins following a temperature shift (49). LysU is constitutively expressed in these lrp metK strains, and two other heat shock proteins, C14.7 and G13.5, are not induced following a temperature shift. C14.7 and G13.5 are also known to be induced in response to heterologous protein expression in E. coli; the genes encoding these proteins, ibpA and ibpB, have been cloned, and their sequences have been determined (1). The abnormal induction of IbpA and IbpB following a heat shock is not seen in an lrp strain (19). LysU, IbpA, and IbpB are the three heat shock proteins whose expression is not induced at 28°C by overproduction of the heat shock sigma factor RpoH (90). Since these heat shock polypeptides are not induced on a shift to 42°C in rpoH mutants, it is thought that their synthesis at elevated temperature requires the rpoH gene product, but it presumably also requires a metabolic signal that is missing when rpoH is induced at 28°C with isopropylthiogalactopyranoside (IPTG). It is possible that metK lrp strains fail to produce the metabolic signal needed for alteration of the levels of these proteins following a temperature shift and that the normal metabolic signal is in fact an altered flux of one-carbon metabolites.
Lrp is an unusual DNA binding protein in that its binding to the promoter regions of some of its target genes is highly sensitive to the methylation status of GATC sequences in the binding region (91; see chapter 94). This opens the possibility of a complex interplay between the availability of AdoMet for DNA methylation and the binding of Lrp to these target sequences and might even underlie the rapid accumulation of lrp mutations in metK strains.
Although a role of Lrp in the reprogramming of one-carbon metabolism following a heat shock remains speculative, the facts cited above suggest a linkage between Lrp expression, one-carbon metabolism, and heat shock.
When the cellular concentration of 10-formyl-H4folate is decreased, the formylation of AICAR in de novo purine biosynthesis becomes the rate-limiting step in the pathway (see above). AICAR (also called ZMP) was shown be converted to a trinucleotide derivative, ZTP, in human blood cells (93). Bochner and Ames (5) used two-dimensional thin-layer chromatography to demonstrate that ZMP and ZTP are formed when S. typhimurium is subjected to treatment with psicofuranine (an inhibitor of xanthosine 5'-monophosphate amidotransferase that blocks the formation of guanosine nucleotides), trimethoprim (an inhibitor of dihydrofolate reductase), and sodium sulfathiazole (an inhibitor of folate biosynthesis) or when a strain auxotrophic for p-aminobenzoate (a precursor in folate biosynthesis) is starved for p-aminobenzoate. They postulated that ZTP serves as an alarmone for one-carbon folate deficiency and accumulates rapidly when 10-formyl-H4folate levels are depleted in the cell. They demonstrated that a purF strain unable to make AICAR during purine biosynthesis is hypersensitive to trimethoprim and/or sulfa drugs when ZMP and ZTP production is repressed by histidine.
The accumulation of ZTP has also been observed when E. coli cells are treated with psicofuranine, trimethoprim, or sodium sulfathiazole or when auxotrophic strains of E. coli are starved for p-aminobenzoic acid (72). However, accumulation of ZTP and ZMP did not correlate well with folate stress in E. coli, as measured by the determination of the folate/protein ratios of extracts of treated cells. These studies failed to provide evidence in E. coli for a folate stress regulon controlled by ZTP.
In summary, the availability of one-carbon units has been shown to limit the rate of initiation of protein translation during heat shock. However, the factors leading to reprogramming of one-carbon metabolism in heat shock have not been elucidated, although it is possible that the Lrp plays a role. ZTP, which is formed by treatments that lead to the accumulation of AICAR during purine biosynthesis, does not appear to be induced by one-carbon deficiency per se, and there is little evidence that this putative alarmone plays a role in regulation of one-carbon metabolism in E. coli.
As pointed out above, there is every reason to postulate that one-carbon metabolism is regulated in response to the nutritional content of the medium. The factors responsible for such regulation have not been identified, although again Lrp is likely to play a major role. It is hoped that this review will encourage further research on this important aspect of cellular regulation.
Research from my laboratory described in this review has been funded by NIH grant R37 GM24908 and NSF grant MCG-9203447.
References
1. Allen, S. P., J. O. Polazzi, J. K. Gierse, and A. M. Easton. 1992. Two novel heat shock genes encoding proteins produced in response to heterologous protein expression in Escherichia coli. J. Bacteriol. 174:6938–6947.
2. Amyes, S. G. B., and J. T. Smith. 1974. Trimethoprim action and its analogy with thymine starvation. Antimicrob. Agents Chemother. 5:169–178.
3. Angehrn, P., and R. Then. 1973. Nature of trimethoprim-induced death in Escherichia coli. Arzneim. Forsch. 23:421–430.
4. Bianchi, V., P. Reichard, R. Eliasson, E. Pontis, M. Krook, H. Jornvall, and E. Haggard-Ljungquist. 1993. Escherichia coli ferredoxin NADP+ reductase: activation of E. coli anaerobic ribonucleotide reduction, cloning of the gene (fpr), and overexpression of the protein. J. Bacteriol. 175:1590–1595.
5. Bochner, B. R., and B. N. Ames. 1982. ZTP (5-amino 4-imidazole carboxamide riboside 5'-triphosphate): a proposed alarmone for 10-formyl-tetrahydrofolate deficiency. Cell 29:929–937.
6. Campbell, J. L., and N. Kleckner. 1990. E. coli oriC and the dnaA gene promoter are sequestered from dam methyltransferase following the passage of the chromosomal replication fork. Cell 62:967–979.
7. Choi, K. Y., and H. Zalkin. 1992. Structural characterization and corepressor binding of the Escherichia coli purine repressor. J. Bacteriol. 174:6207–6214.
8. Cowan, J. M., M. L. Urbanowski, M. Talmi, and G. V. Stauffer. 1993. Regulation of the Salmonella typhimurium metF gene by the MetR protein. J. Bacteriol. 175:5862–5866.
9. Dalal, F. R., R. E. Gots, and J. S. Gots. 1966. Mechanism of adenine inhibition in adenine-sensitive mutants of Salmonella typhimurium. J. Bacteriol. 91:507–513.
10. Dev, I. K., and R. J. Harvey. 1978. N10-formyltetrahydrofolate is the formyl donor for glycinamide ribotide transformylase in Escherichia coli. J. Biol. Chem. 253:4242–4244.
11. Dev, I. K., and R. J. Harvey. 1982. Sources of one-carbon units in the folate pathway of Escherichia coli. J. Biol. Chem. 257:1980–1986.
12. Dev, I. K., and R. J. Harvey. 1984. Regulation of synthesis of serine hydroxymethyltransferase in chemostat cultures of Escherichia coli. J. Biol. Chem. 259:8402–8406.
13. Dickerman, H. W., E. Steers Jr., B. G. Redfield, and H. Weissbach. 1967. Methionyl soluble ribonucleic acid transformylase. I. Purification and partial characterization. J. Biol. Chem. 242:1522–1525.
14. Dimri, G. P., G. Ferro-Luzzi Ames, L. D’Ari, and J. C. Rabinowitz. 1991. Physical map location of the Escherichia coli gene encoding the bifunctional enzyme 5,10-methylene-tetrahydrofolate dehydrogenase/5,10-methenyl-tetrahydrofolate cyclohydrolase. J. Bacteriol. 173:5251.
15. Downs, D. 1992. Evidence for a new, oxygen-regulated biosynthetic pathway for the pyrimidine moiety of thiamine in Salmonella typhimurium. J. Bacteriol. 174:1515–1521.
16. Downs, D. M., and L. Petersen. 1994. apbA, a new genetic locus involved in thiamine biosynthesis in Salmonella typhimurium. J. Bacteriol. 176:4858–4864.
17. Downs, D. M., and J. R. Roth. 1991. Synthesis of thiamine in Salmonella typhimurium independent of the purF function. J. Bacteriol. 173:6597–6604.
18. Emmett, M. R., and J. R. Johnson. 1986. Control of metF gene expression in maxicell preparations of Escherichia coli K-12: reversible action of the metJ protein and effect of vitamin B12. J. Bacteriol. 168:1491–1494.
19. Ernsting, B. R., M. R. Atkinson, A. J. Ninfa, and R. G. Matthews. 1992. Characterization of the regulon controlled by the leucine-responsive regulatory protein in Escherichia coli. J. Bacteriol. 174:1109–1118.
20. Ernsting, B. R., J. W. Denninger, R. M. Blumenthal, and R. G. Matthews. 1993. Regulation of the gltBDF operon in Escherichia coli: how is a leucine-insensitive operon regulated by the leucine-responsive regulatory protein? J. Bacteriol. 175:7160–7169.
21. Estramareix, B., and M. Therisod. 1984. Biosynthesis of thiamine: 5-aminoimidazole ribotide as the precursor of all the carbon atoms of the pyrimidine moiety. J. Am. Chem. Soc. 106:3857–3860.
22. Flaks, J. G., and S. S. Cohen. 1957. The enzymic synthesis of 5-hydroxymethyldeoxycytidylic acid. Biochim. Biophys. Acta 25:667–668.
23. Fraser, J., and E. B. Newman. 1975. Derivation of glycine from threonine in Escherichia coli K-12 mutants. J. Bacteriol. 122:810–817.
24. Fujii, K., and F. M. Huennekens. 1974. Activation of methionine synthetase by a reduced triphosphopyridine nucleotide-dependent flavoprotein system. J. Biol. Chem. 249:6745–6753.
25. Fuller, R. S., B. E. Funnell, and A. Kornberg. 1984. The dnaA protein complex with the E. coli chromosomal replication origin (oriC) and other DNA sites. Cell 38:889–900.
26. Gage, D. J., and F. C. Neidhardt. 1992. One-carbon metabolism and the modulation of the heat shock response in Escherichia coli, abstr. H-199, p. 217. In Abstracts of the 92nd General Meeting of the American Society for Microbiology 1992. American Society for Microbiology, Washington, D.C.
27. Gold, L. 1988. Posttranscriptional regulatory mechanisms in Escherichia coli. Annu. Rev. Biochem. 57:199–233.
28. Gonzalez, J. C., R. V. Banerjee, S. Huang, J. S. Sumner, and R. G. Matthews. 1992. Comparison of cobalamin-independent and cobalamin-dependent methionine synthases from Escherichia coli: two solutions to the same chemical problem. Biochemistry 31:6045–6056.
29. Gots, J. S. 1971. Regulation of purine and pyrimidine metabolism, p. 225–254. In H. J. Vogel (ed.), Metabolic Regulation, vol. V. Academic Press, New York.
30. Goyette, P., J. Sumner, R. Milos, A. M. V. Duncan, D. S. Rosenblatt, R. G. Matthews, and R. Rozen. 1994. Isolation and chromosomal assignment of a cDNA for human methylenetetrahydrofolate reductase (MTHFR), with identification of mutations in MTHFR deficiency. Nat. Genet. 7:195–200.
31. Green, J. M., D. P. Ballou, and R. G. Matthews. 1988. Examination of the role of methylenetetrahydrofolate reductase in the incorporation of methyltetrahydrofolate into cellular metabolism. FASEB J. 2:42–47.
32. Guillon, J.-M., Y. Mechulam, J.-M. Schnitter, S. Blanquet, and G. Fayat. 1992. Disruption of the gene for Met-tRNAfMet formyltransferase severely impairs growth of Escherichia coli. J. Bacteriol. 174:4294–4301.
33. Hartz, D., D. S. McPheeters, and L. Gold. 1989. Selection of the initiator tRNA by E. coli initiation factors. Genes Dev. 3:1899–1912.
34. Harvey, R. J. 1973. Growth and initiation of protein synthesis in the presence of trimethoprim. J. Bacteriol. 114:309–322.
35. Horwich, A., K. B. Low, W. A. Fenton, I. N. Hirshfield, and K. Furtak. 1993. Folding in vivo of bacterial cytoplasmic proteins: role of GroEL. Cell 74:909–917.
36. Houlberg, U., and K. F. Kensen. 1983. Role of hypoxanthine and guanine in regulation of Salmonella typhimurium pur gene expression. J. Bacteriol. 153:837–845.
37. Jones, C. E., J. M. Brook, D. Buck, C. Abell, and A. G. Smith. 1993. Cloning and sequencing of the Escherichia coli panB gene, which encodes ketopantoate hydroxymethyltransferase, and overexpression of the enzyme. J. Bacteriol. 175:2125–2130.
38. Kahn, D., M. Fromant, G. Gayat, P. Dessen, and S. Blanquet. 1980. Methionyl-transfer-RNA transformylase from Escherichia coli. Purification and characterisation. Eur. J. Biochem. 105:489–497.
39. Katzen, H. M., and J. M. Buchanan. 1965. Enzymatic synthesis of the methyl group of methionine. VIII. Repression-derepression, purification and properties of 5,10-methylenetetrahydrofolate reductase from Escherichia coli. J. Biol. Chem. 240:825–835.
40. Kutzbach, C., and E. L. R. Stokstad. 1971. Mammalian methylenetetrahydrofolate reductase: partial purification, properties, and inhibition by S-adenosylmethionine. Biochim. Biophys. Acta 250:459–477.
41. Langley, D., and J. R. Guest. 1974. Biochemical and genetic characteristics of deletion and other mutant strains of Salmonella typhimurium LT2 lacking α-keto acid dehydrogenase complex activities. J. Gen. Microbiol. 82:319–335.
42. Lin, R., R. D’Ari, and E. B. Newman. 1992. λ placMu insertions in genes of the leucine regulon: extension of the regulon to genes not regulated by leucine. J. Bacteriol. 174:1948–1955.
43. Lin, R., B. Ernsting, I. N. Hirshfield, R. G. Matthews, F. C. Neidhardt, R. L. Clark, and E. B. Newman. 1992. The lrp gene product regulates expression of lysU in Escherichia coli K-12. J. Bacteriol. 174:2779–2784.
44. Lin, R. T., R. D’Ari, and E. B. Newman. 1990. The leucine regulon of Escherichia coli K-12: a mutation in rblA alters expression of l-leucine-dependent metabolic operons. J. Bacteriol. 172:4529–4535.
45. Mangum, J. H., and K. G. Scrimgeour. 1962. Cofactor requirements and intermediates in methionine biosynthesis. Fed. Proc. 21:242.
46. Marinus, M. G., and N. R. Morris. 1974. Biological function for 6-methyladenine residues in the DNA of Escherichia coli K-12. J. Mol. Biol. 85:309–322.
47. Marolewski, A., J. M. Smith, and S. J. Benkovic. 1994. Cloning and characterization of a new purine biosynthetic enzyme: a non-folate glycinamide ribonucleotide transformylase from E. coli. Biochemistry 33:2531–2537.
48. Matthews, R. G. 1984. Methionine biosynthesis, p. 497–553. In R. L. Blakley and S. J. Benkovic (ed.), Folates and Pterins, vol. 1. Chemistry and Biochemistry of Folates. John Wiley & Sons, New York.
49. Matthews, R. G., and F. C. Neidhardt. 1988. Abnormal induction of heat shock proteins in an Escherichia coli mutant deficient in adenosylmethionine synthetase activity. J. Bacteriol. 170:1582–1588.
50. Matthews, R. G., and F. C. Neidhardt. 1989. Elevated serine catabolism is associated with the heat shock response in Escherichia coli. J. Bacteriol. 171:2619–2625.
51. Meinnel, T., and S. Blanquest. 1993. Evidence that peptide deformylase and methionyl-tRNAfMet formyltransferase are encoded within the same operon in Escherichia coli. J. Bacteriol. 175:7737–7740.
52. Meng, L. M., and P. Nygaard. 1990. Identification of hypoxanthine and guanine as the corepressors for the purine regulon genes of Escherichia coli. Mol. Microbiol. 4:2187–2191.
53. Milner, L., C. Whitfield, and H. Weisbach. 1969. Effect of l-methionine and vitamin B12 on methionine biosynthesis in Escherichia coli. Arch. Biochem. Biophys. 133:413–419.
54. Modrich, P. 1987. DNA mismatch correction. Annu. Rev. Biochem. 56:435–466.
55. Mulligan, J. T., W. Margolin, J. H. Krueger, and G. C. Walker. 1982. Mutations affecting regulation of methionine biosynthetic genes isolated by use of met-lac fusions. J. Bacteriol. 151:609–619.
56. Nagy, P. L., G. M. McCorkle, and H. Zalkin. 1994. purU, a source of formate for purT-dependent phosphoribosyl-N-formylglycinamide synthesis. J. Bacteriol. 175:7066–7073.
57. Neidhardt, F. C., J. L. Ingraham, and M. Schaechter. 1990. Physiology of the Bacterial Cell: a Molecular Approach. Sinauer Associates, Sunderland, Mass.
58. Neidhardt, F. C., R. A. VanBogelen, and E. T. Lau. 1983. Molecular cloning and expression of a gene that controls the high-temperature regulon of Escherichia coli. J. Bacteriol. 153:597–603.
59. Newell, P. C., and R. G. Tucker. 1968. Precursors of the pyrimidine moiety of thiamine. Biochem. J. 106:271–277.
60. Newell, P. C., and R. G. Tucker. 1968. Biosynthesis of the pyrimidine moiety of thiamine. Biochem. J. 106:279–287.
61. Newman, E. B., V. Kapoor, and R. Potter. 1976. Role of l-threonine dehydrogenase in the catabolism of threonine and synthesis of glycine by Escherichia coli. J. Bacteriol. 126:1245–1249.
62. Newman, E. G., D. Ahmad, and C. Walker. 1982. l-Serine deaminase activity is induced by exposure of Escherichia coli K-12 to DNA-damaging agents. J. Bacteriol. 152:702–705.
63. Nygaard, P., and J. M. Smith. 1993. Evidence for a novel glycinamide ribonucleotide transformylase in Escherichia coli. J. Bacteriol. 175:3591–3597.
64. Osborne, C., L.-M. Chen, and R. G. Matthews. 1991. Isolation, cloning, mapping and nucleotide sequencing of the gene encoding flavodoxin in Escherichia coli. J. Bacteriol. 173:1729–1737.
65. Patterson, D., and D. Gillespie. 1972. Effect of elevated temperatures on protein synthesis in Escherichia coli. J. Bacteriol. 112:1177–1183.
66. Plamann, M. D., and G. V. Stauffer. 1989. Regulation of the Escherichia coli glyA gene by the metR gene product and homocysteine. J. Bacteriol. 171:4958–4962.
67. Powers, S. G., and E. E. Snell. 1976. Ketopantoate hydroxymethyltransferase II. Physical, catalytic and regulatory properties. J. Biol. Chem. 251:3786–3793.
68. Primerano, D. A., and R. O. Burns. 1982. Metabolic basis for the isoleucine, pantothenate or methionine requirement of ilvG strains of Salmonella typhimurium. J. Bacteriol. 150:1202–1211.
69. Ravnikar, P. D., and R. L. Somerville. 1987. Genetic characterization of a highly efficient alternate pathway of serine biosynthesis in Escherichia coli. J. Bacteriol. 169:2611–2617.
70. Rex, J. H., B. D. Aronson, and R. L. Somerville. 1991. The tdh and serA operons of Escherichia coli: mutational analysis of the regulatory elements of leucine-responsive genes. J. Bacteriol. 173:5944–5953.
71. Roberts, R. B., P. H. Abelson, D. B. Cowie, E. T. Bolton, and R. J. Britten. 1957. Studies of Biosynthesis in Escherichia coli. Publication 607. Carnegie Institution of Washington, Washington, D.C.
72. Rohlman, C. E., and R. G. Matthews. 1990. Role of purine biosynthetic intermediates in response to folate stress in Escherichia coli. J. Bacteriol. 172:7200–7210.
73. Rolfes, R. J., and H. Zalkin. 1988. Escherichia coli gene purR encoding a repressor protein for purine nucleotide synthesis. J. Biol. Chem. 263:19653–19661.
74. Ron, E. Z., and B. D. Davis. 1971. Growth rate of Escherichia coli at elevated temperatures: limitation by methionine. J. Bacteriol. 107:391–396.
75. Rowe, P. B. 1984. Folates in the biosynthesis and degradation of purines, p. 329–344. In R. L. Blakley and S. J. Benkovic (ed.), Folates and Pterins, vol. 1. Chemistry and Biochemistry of Folates. John Wiley & Sons, New York.
76. Saint-Girons, I., J. Belfaiza, Y. Guillou, D. Perrin, N. Guiso, O. Barzu, and G. N. Cohen. 1986. Interaction of the Escherichia coli methionine repressor with the metF operator and with its corepressor, S-adenosylmethionine. J. Biol. Chem. 261:10936–10940.
77. Saint-Girons, I., N. Duchange, M. M. Zakin, I. Park, D. Margarita, P. Ferrara, and G. N. Cohen. 1983. The nucleotide sequence metF, the E. coli structural gene for 5,10-methylene tetrahydrofolate reductase and of its control region. Nucleic Acids Res. 11:6723–6732.
78. Saint-Girons, I., C. Parsot, M. M. Zakin, O. Barzu, and G. N. Cohen. 1988. Methionine biosynthesis in enterobacteriaceae: biochemical, regulatory and evolutionary aspects. Crit. Rev. Biochem. 23:1–42.
79. Santi, D. V., and P. V. Danenberg. 1984. Folates in pyrimidine nucleotide biosynthesis, p. 345–398. In R. L. Blakley and S. J. Benkovic (ed.), Folates and Pterins, vol. 1. Chemistry and Biochemistry of Folates. John Wiley & Sons, New York.
80. Satishchandran, C., J. C. Taylor, and G. D. Markham. 1993. Isozymes of S-adenosylmethionine synthetase are encoded by tandemly duplicated genes in Escherichia coli. Mol. Microbiol. 9:835–846.
81. Schirch, L. 1984. Folates in serine and glycine metabolism, p. 399–431. In R. L. Blakley and S. J. Benkovic (ed.), Folates and Pterins, vol. 1. Chemistry and Biochemistry of Folates. John Wiley & Sons, New York.
82. Shive, W., W. W. Ackermann, M. Gordon, M. E. Getzendaner, and R. E. Eakin. 1947. 5(4)-Amino-4(5)-imidazole carboxamide, a precursor of purines. J. Am. Chem. Soc. 69:725–726.
83. Smith, G. K., W. T. Mueller, P. A. Benkovic, and S. J. Benkovic. 1981. On the cofactor specificity of glycineamide ribonucleotide and 5-aminoimidazole-4-carboxamide ribonucleotide transformylase from chicken liver. Biochemistry 20:1241–1245.
84. Smith, M. W., and F. C. Neidhardt. 1983. Proteins induced by aerobiosis in Escherichia coli. J. Bacteriol. 154:344–350.
85. Smith, R. J., and J. E. M. Midgley. 1973. The effect of trimethoprim on macromolecular synthesis in Escherichia coli: regulation of RNA synthesis with "magic spot" nucleotides. Biochem. J. 136:249–257.
86. Stauffer, G. V., and L. T. Stauffer. 1988. Cloning and nucleotide sequence of the Salmonella typhimurium LT2 metF gene and its homology with the corresponding sequence of Escherichia coli. Mol. Gen. Genet. 212:246–2151.
87. Steiert, J. G., R. J. Rolfes, H. Zalkin, and G. V. Stauffer. 1990. Regulation of Escherichia coli glyA gene by the purR gene product. J. Bacteriol. 172:3799–3830.
88. Stetton, M. R., and C. L. Fox, Jr. 1945. An amine formed by bacteria during sulfonamide bacteriostasis. J. Biol. Chem. 161:333–349.
89. Urbanowski, M. L., and G. V. Stauffer. 1987. Regulation of the metR gene of Salmonella typhimurium. J. Bacteriol. 169:5841–5844.
90. VanBogelen, R. A., M. A. Acton, and F. C. Neidhardt. 1987. Induction of the heat shock regulon does not produce thermotolerance in Escherichia coli. Genes Dev. 1:525–531.
91. van der Woude, M. W., B. A. Braaten, and D. A. Low. 1992. Evidence for global regulatory control of pilus expression in Escherichia coli by Lrp and DNA methylation: model building based on analysis of pap. Mol. Microbiol. 6:2429–2435.
92. Welsh, K. M., A.-L. Lu, S. Clark, and P. Modrich. 1987. Isolation and characterization of the Escherichia coli mutH gene product. J. Biol. Chem. 262:15624–15629.
93. Zimmerman, T. P., and R. D. Deeprose. 1978. Metabolism of 5-amino-1-β-d-ribofuranosyl-imidazole-4-carboxamide and related five-membered heterocycles to 5'-triphosphates in human blood and L5178Y cells. Biochem. Pharmacol. 27:709–716.
94. Zyskind, J. W., and D. W. Smith. 1986. The bacterial origin of replication, oriC. Cell 46:489–490.
Chapter36
