1. Regulation of Methionine Biosynthesis by Gene Expression
The metA gene codes for the first step in the methionine biosynthetic pathway, starting from homoserine (see Methionine Regulon). It has two starting sites for transcription located 74 nucleotides apart. Both promoters are used in vivo, but only one is regulated by intracellular methionine levels. The metB, metC, and metF genes, coding respectively for cystathionine synthase, cystathionase, and 5,10-methylene tetrahydrofolate reductase, each have a single promoter. The metA, metB, metC, metE, and metF genes are all repressed when excess methionine is added to the minimal growth medium, albeit to different degrees. The level of the enzyme varies 300-fold for homoserine transsuccinylase, 40-fold for cystathionine synthetase, 6- to 12-fold for cystathionase, 20-fold for the methylene tetrahydrofolate reductase, and 60-fold for the vitamin B12-independent transmethylase. Thus there is no coordinate repression, reflecting the fact that the met genes do not form a single operon, but are scattered through the bacterial chromosome.
There is a notable exception: metB and metL, coding for cystathionine synthetase and aspartate kinase II-homoserine dehydrogenase II, constitute an operon and belong to a gene cluster, metJBLF, located around 89 min. on the chromosome (see Methionine Regulon). Methionine also regulates the synthesis of aspartokinase II-homoserine dehydrogenase II (metL) and S-adenosylmethionine (AdoMet) synthetase (metK). Although methionine affects the level of the B12-dependent transmethylase, coded by metH, this effect appears to be the indirect result of repression of the synthesis of an activator protein (MetR, see below). Expression of the metE gene is repressed by vitamin B^.
The E. coli metJ gene is 312 bp long and codes for the Met repressor, a molecule of 104 amino acid residues that exists as a homodimer (Fig. 1). metJ is transcribed divergently from metB. Consequently a complex 276-bp regulatory region is found between them. There is a single promoter for metB, whereas metJ is transcribed from three separate points. Only two of the three metJ promoters are regulated by methionine. The repressor regulates its own synthesis.
Figure 1. Sequence of the E.coli methionine repressor. The * sign indicates the amino acid residues making contacts with DNA, inferred from the structure of the repressor/DNA complex. The underlined sections indicate the b-strand and two of the a-helices of the protein (1).
The effect of the repressor on the synthesis of the methionine biosynthetic enzymes has been assessed by measuring Beta-galactosidase synthesis in a cell-free system where the synthesis of this enzyme is under the control of metF (metF-lacZ fusion). Methionine had to be present in these experiments as a necessary building block for protein synthesis. Expression of b-galactosidase was progressively repressed by increasing concentrations of the aporepressor. Complete repression was attained at 600 nM. At a constant concentration of aporepressor, AdoMet enhances the repression in a concentration-dependent manner. Half-maximum inhibition is reached at 10 mM AdoMet and complete repression at 100 mM. Similar in vitro assay systems demonstrated that the aporepressor and AdoMet have the same effect on the expression of the metB and metL genes and on that of the metJ gene itself.
Equilibrium Dialysis demonstrates that the methionine repressor binds two molecules of S-AdoMet per dimer. The corresponding Scatchard Plot is linear, showing that there is no cooperativity. Methionine does not bind to the aporepressor. Because the repressor binding sites should be similar for all of the genes regulated by methionine, comparison of the four 5′ regions of the metC, metB, metA, and metF genes reveals a repetitive unit (R) eight nucleotides long, named the "Met box." In the alignment presented in Figure 2 of 128 positions, 89 matches, and 21 transitions are found when the repetitive units are compared to the consensus sequence R. This consensus sequence is a perfect palindrome, AGACGTCT, which is present in an altered form two to five times in the Met boxes. The metB and metJ genes are transcribed divergently and share the same Met box.
Figure 2. Comparison of the upstream regions of the metC, metB, metF, and metA genes. The sequences 5′ to the structural metC, metB, metF, and metA genes are presented discontinuously and have been aligned in order to focus on the presence of the underlying repetitive palindromic repetitive unit. Nucleotides matching the consensus sequence presented in line R are in boldface. Numbers indicate positions relative to the adenine of the respective start codon taken as +1. The -10 promoter sequences are overlined and arrowheads indicate the transcription start signals. In the case of metB, the overlined hexamer is the -35 box: the two underlined promoter sequences represent the -35 and -10 boxes of the first promoter of metJ (2).
The metE and metR genes also possess these repetitive units (3 for metE and 4 for metR). The two exceptions are metK and metH which, although regulated by methionine, do not possess the consensus sequence. The differences in the number and/or sequences of the Met boxes may be related to the different extents of repression elicited by the Met repressor. A filter-binding assay using nitrocellulose and radioactive oligodeoxynucleotides has shown that DNA fragments containing two consecutive consensus Met boxes are tightly bound by repressor in the presence of saturating AdoMet, whereas nucleotide sequences containing only one Met box are not bound. Binding is cooperative with respect to repressor concentration. The repressor protects the operator against nuclease digestion (see Footprinting), and the results strongly indicate the binding of an array of repressor dimers centered on the 16-bp operator site but extending into the neighboring DNA. The same was found for a fragment containing the metF regulatory region, where five boxes are protected. These observations are consistent with the following in vivo data: the level of repression of defined mutants of the metC operator, the smallest one known (having two Met boxes), is increased when the sequence becomes closer to the consensus sequence. Similar results were obtained with the five Met boxes of metF. Furthermore, the arrangement of all five boxes in tandem repeats is important for effective repression, since some point insertions within the operators lower repression 100-fold.
2. Three-Dimensional
Structure of the Methionine Repressor of its Binary Complex with AdoMet and of its Ternary Complex with AdoMet and DNA As in the case of the Trp repressor, (see TRP Operon) the two subunits of the Met repressor are strongly intertwined to form the dimer. AdoMet binds to the dimer at two independent but symmetrical sites, one on each monomer. The purine nucleus of AdoMet inserts into a hydrophobic pocket, normally occupied in the aporepressor by the side-chain of Phe65, whereas the methionine moiety lies at the protein surface. This explains why methionine itself does not bind to the aporepressor, whereas S-adenosylhomocysteine binds with about half the affinity of AdoMet. The positively charged trivalent sulfur atom lies at the C-terminus of helix B. In contrast to the Trp repressor, there is no major conformational change associated with corepressor binding. The apo and holo structures are indistinguishable, except for the absence or presence of AdoMet and small conformational changes at the N-terminus and Phe65. AdoMet does not contact DNA in the repressor-operator complex (see below), and how it regulates binding remains unclear. One possibility is a long-range electrostatic effect based on the positive charge on the sulfur atom.
Each monomer of the Met repressor is composed of three a-helices and one b-strand, accounting for about 50% of the sequence (see Fig. 1). In the protein dimer, the b-strands pair to form an antiparallel Beta-sheet, whereas the a-helical regions pack against the sheet and against each other to stabilize the dimer. The molecule lacks a helix-turn-helix motif which makes it different from many other repressors and from the cyclic AMP receptor protein.
By X-ray crystallography it was found that a 19-bp oligonucleotide containing two adjacent 8-bp Met boxes binds two dimeric Met repressor molecules. One dimer binds to each half-site, each of which contains a twofold axis of symmetry that coincides with the twofold axis of the protein b-sheet. The a-helix 1 on one dimer interacts with the same a-helix 1 on the adjacent dimer to form a tetrameric protein structure. Sequence specificity is achieved by inserting the double-stranded, antiparallel, protein b-sheet into the major groove of B-form DNA. Direct hydrogen-bonding occurs between amino acid side-chains on the exposed face of the sheet and the base pairs. Lys23 from each b-strand contacts a guanine base, and the neighboring Thr-25 on each strand contacts an adenine. Residues from the N-terminus of a-helix 2 make backbone contacts. The repressor also recognizes sequence-dependent distortion or flexibility of the operator phosphate backbone and confers specificity even for inaccessible base pairs.
3. The metR Gene and its Product
Surprisingly it was found that some mutants with normal metF, metE and metH genes express the two transmethylases from the metE and metH genes at very low levels, resulting in a growth requirement for methionine. This auxotrophy was overcome in strains carrying multicopy plasmids with either the metE or the metH gene. This implied that the methionine auxotrophy was caused by independent mutations resulting in an inability to synthesize enough homocysteine transmethylase enzyme to permit growth. It was found that these mutations are linked to metE but lie outside the metE structural gene. Their locus was called metR. The metR gene from E. coli encodes a 35,628-Da polypeptide 317 amino acid residues long. The native protein is a homodimer and is believed to contain a leucine zipper, a motif characteristic of many eukaryotic DNA binding proteins (and of the E. coli Lac repressor).
Expression of metR is repressed by the Met repressor, as the 5′ flanking region of metR contains four Met boxes, and by the product of metR itself. Conversely, the MetR protein plays no role in regulating the metJ gene for the Met repressor. MetR activates the expression of both metE and metH. Homocysteine, the substrate of both transmethylases, is the coactivator for metE but not metH. This positive regulatory mechanism may apply when vitamin B12 is absent, ie, when there is no B12- dependent transmethylase activity (catalyzed by the metH product). The resulting methionine starvation would then cause derepression of the biosynthetic enzymes and accumulation of homocysteine. The latter could then act as a signal to activate the synthesis of the vitamin B- independent transmethylase, coded by metE.
In S. typhimurium, the MetR protein activates metA, the gene for homoserine succinyltransferase, but homocysteine inhibits this activation. This result explains why expression of the metE gene is repressed by vitamin B^. It is primarily caused by a loss of MetR-mediated activation through depletion of the coactivator homocysteine, rather than a direct repression by the metH-B holoenzyme.
