Direct Sulfhydrylation for Methionine Biosynthesis in Leptospira meyeri

Abstract

A gene library of the Leptospira meyeri serovar semaranga strain Veldrat S.173 DNA has been constructed in a mobilizable cosmid with inserts of up to 40 kb. It was demonstrated that a Leptospira DNA fragment carrying metY complemented Escherichia coli strains carrying mutations in metB. The latter gene encodes cystathionine γ-synthase, an enzyme which catalyzes the second step of the methionine biosynthetic pathway. The metY gene is 1,304 bp long and encodes a 443-amino-acid protein with a molecular mass of 45 kDa as determined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The deduced amino acid sequence of the Leptospira metY product has a high degree of similarity to those of O-acetylhomoserine sulfhydrylases from Aspergillus nidulans and Saccharomyces cerevisiae. A lower degree of sequence similarity was also found with bacterial cystathionine γ-synthase. The L. meyeri metY gene was overexpressed under the control of the T7 promoter. MetY exhibits an O-acetylhomoserine sulfhydrylase activity. Genetic, enzymatic, and physiological studies reveal that the transsulfuration pathway via cystathionine does not exist in L. meyeri, in contrast to the situation found for fungi and some bacteria. Our results indicate, therefore, that the L. meyeri MetY enzyme is able to perform direct sulfhydrylation for methionine biosynthesis by using O-acetylhomoserine as a substrate.

The biosynthetic pathways of sulfur amino acids are well documented. Two alternative methionine biosynthetic pathways exist in microorganisms (Fig. 1). One, called the transsulfuration pathway, involves cystathionine formation, and the other bypasses cystathionine via direct sulfhydrylation of O-acylhomoserine to homocysteine (29).

FIG. 1.

Biosynthetic pathways of sulfur amino acids in E. coli (A) and S. cerevisiae (B). Enzyme steps: 1, O-succinylhomoserine transferase (metA); 1′, O-acetylhomoserine transferase; 2, cystathionine γ-synthase (metB); 3, cystathionine β-lyase (metC); 4, O-acetylhomoserine sulfhydrylase (Met17 [or Met25]); 5, cystathionine β-synthase; 6, cystathionine γ-lyase. Genes shown in parentheses in this legend are the corresponding E. coli genes; for step 4 the S. cerevisiae gene is indicated.

In enteric bacteria, the sulfur atom is incorporated first into a serine ester (O-acetylserine) to yield cysteine (16). Sulfur is then transferred from cysteine to homocysteine via transsulfuration. In Escherichia coli, it requires the sequential action of cystathionine γ-synthase (EC 4.2.99.9), the product of the metB gene (7), and cystathionine β-lyase (EC 4.4.1.8), the metC gene product (1), with the intermediary formation of cystathionine (Fig. 1A, steps 2 and 3).

The direct sulfhydrylation pathway has been reported to be the main pathway for homocysteine biosynthesis in Saccharomyces cerevisiae (5) and bacteria such as Brevibacterium flavum and Pseudomonas aeruginosa (10, 23). In S. cerevisiae, which is the best-studied example, the direct synthesis of homocysteine is catalyzed by an O-acetylhomoserine sulfhydrylase, the Met25 (or Met17) product (Fig. 1B, step 4) (5, 34). The resulting homocysteine is used as a direct precursor for methionine and is converted to cysteine via the reverse transsulfuration pathway (Fig. 1B, steps 5 and 6).

In addition, it should be kept in mind that the ester of homoserine used for homocysteine biosynthesis differs depending on the organisms: enteric bacteria use O-succinylhomoserine, while fungi and most gram-positive bacteria use O-acetylhomoserine (Fig. 1, steps 1 and 1′) (for a review, see reference 33).

Little is presently known about the regulation of the metabolite flux of the methionine pathway. However, it has been reported that the control at the enzymatic level in bacteria and S. cerevisiae occurred at an early step of the methionine biosynthetic pathway. The first enzyme of the methionine biosynthetic pathway in E. coli, O-succinylhomoserine transferase, is feedback inhibited by methionine and S-adenosylmethionine (20), while the activity of O-acetylhomoserine transferase from S. cerevisiae is inhibited only by S-adenosylmethionine (6). In previous work, we demonstrated that O-acetylhomoserine transferase activity in Leptospira meyeri is not regulated by methionine and/or S-adenosylmethionine (2).

Our goal was to investigate the evolution of sulfur metabolism in L. meyeri. We report here the construction of a representative cosmid L. meyeri DNA library and the cloning of a biosynthetic gene, metY, which complements E. coli metB mutants. Analysis of the inferred L. meyeri MetY amino acid sequence, growth impairment of E. coli mutants carrying metY, and results of enzymatic assays allow us to propose a direct sulfhydrylation pathway catalyzed by an O-acetylhomoserine sulfhydrylase.

MATERIALS AND METHODS

Bacterial strains, plasmids, and media.

L. meyeri serovar semaranga strain Veldrat S.173 (National Reference Center, Paris, France), isolated from a rat, was grown in EMJH medium at 30°C (8, 14).

E. coli strains (Table 1) were grown in LB broth or on L agar plates at 37°C except when indicated (27). The antibiotics used and their concentrations were as follows: kanamycin, 25 μg/ml; tetracycline, 8 μg/ml; chloramphenicol, 30 μg/ml; and ampicillin, 100 μg/ml. Minimal medium M9 supplemented with 0.04% glucose and 1 μg of thiamine per ml, plus appropriate amino acids (1 mM), was used to characterize E. coli transformants at 30°C (27).

TABLE 1.

E. coli strains used in this study

E. coli strain	Genotype	Origin or reference
HB101	hsdR hsdM recA supE44 lacZ4 leuB6 proA2 thi-1 Smr	3
WA802	metB1 lac-3 (or lacY1) galK2 galT22 supE44 hsdR Rifr	31
AB1932	metA28 argH1 thi-1 lacY1 lacZ4 galZ2 xyl4 (or -5) tsx6 F Rifr	E. A. Adelberg
EC972	metB185 araD139 Δ(argF-lac)205 trpB202 flB350 ptsF25 relA1 rpsL150 deoG λ−	B. Bachmann
β254	metC::Tetr	26
GT1107	metC::Tetr metB1 lac-3 (or lacY1) galK2 galT22 supE44 hsdR Rifr	Transduction of WA802 by a P1 lysate grown on β254
NK3	cysK cysM	N. M. Kredich
β180	metA::Cmr	26
GT1140	metA::CmrmetB1 lac-3 (or lacY1) galK2 galT22 supE44 hsdR Rifr	Transduction of WA802 by a P1 lysate grown on β180

pill₂₀₀ (Km^r Mob⁺ Tra⁻) (6.9 kb long), derived from pILL575 by deletion of a 3.2-kb HindIII-PstI fragment (18), was used as a cosmid vector to clone L. meyeri DNA. pUC18 (Amp^r), pBR322 (Tet^r Amp^r), and pSU18 (Cm^r) (22, 30, 35) were used for subcloning experiments.

Construction of an L. meyeri genomic DNA library in a cosmid.

Total DNA from L. meyeri was prepared as described previously (11) from 500 ml of culture at 5 × 10⁸ bacteria/ml. Genomic DNA partially cleaved with restriction endonuclease Sau3A and sized on a 10 to 40% sucrose density gradient was ligated into the BamHI-digested and alkaline-phosphatase-treated cosmid vector pill₂₀₀ (1 μg). The ligated mix was packaged into phage lambda particles as described by the supplier (Gigapack III Gold 11 kit; Stratagene) and used to infect E. coli HB101 harboring helper plasmid pRK212.1 (9).

Complementation of E. coli methionine auxotrophs.

Two different strategies were used for complementation of E. coli methionine auxotrophs: (i) direct infection of E. coli methionine auxotrophs with lambda phage particles from the cosmidic library at 30°C and (ii) mobilization of the recombinant cosmids as follows. Individual clones of the L. meyeri gene library (HB101 harboring the IncP helper plasmid plus the hybrid plasmid to be mobilized) were grown in LB broth at 30°C with kanamycin. Aliquots (2 μl) of each clone were transferred to L agar plates spotted with the recipient cells (2 μl of E. coli spontaneous Rif^r AB1932 and WA802 methionine auxotrophs).

In each case, the plates were incubated at 30°C overnight and cells were replica plated on minimal medium containing tetracycline, kanamycin, rifampin, and the appropriate amino acids.

DNA analysis, DNA sequencing, and computer analysis.

Restriction endonucleases purchased from Boehringer Mannheim were used as described in the manufacturer’s instructions. Calf intestinal alkaline phosphatase, T4 DNA ligase (high concentration), and the Wizard DNA cleanup system were purchased from Promega, Madison, Wis. TaqI DNA polymerase was from Cetus. Ligations, agarose gel electrophoresis, and electroporation were performed by standard procedures (27).

Double-stranded plasmid DNA was sequenced by using the Pharmacia T7 sequencing kit, [α-³³P]dATP (111 TBq/mmol; ICN), and synthetic oligonucleotides. Comparisons to protein databases were done by using the BLAST e-mail server.

Overexpression of metY.

A plasmid allowing expression of the metY structural gene under the control of the T7 promoter was constructed by oligonucleotide mutagenesis. The final construct (pETmetY), verified by nucleotide sequencing, contained the whole metY gene (starting at the ATG and continuing to the BamHI site located 44 bp beyond the metY stop codon) inserted into pET20b⁺ (Novagen, Madison, Wis.). The pETmetY plasmid was transformed into E. coli BL21 (DE3) (Novagen), which carries the T7 RNA polymerase gene on λDE3 integrated on the chromosome of BL21. The conditions of overexpression of metY under the control of the T7 promoter were as described previously for overexpression of L. meyeri metX (2).

Enzymatic assays.

O-Acetylhomoserine sulfhydrylase and O-acetylserine sulfhydrylase activities were assayed as described by Ravanel et al. (25). The amount of homocysteine or cysteine formed in a 0.1-ml reaction mixture was determined by the nitroprusside reaction (32) or by the procedure described by Kredich and Becker (17). Since MetY does not contain cysteine, dithiothreitol was omitted from the reaction mixture. The reaction was started by addition of 2 mM sodium sulfide (Na₂S). The reaction mixtures overlaid with 50 μl of paraffin oil were incubated at 30°C for 10 and 30 min. When aliphatic thiols were determined with the nitroprusside test, the incubation was stopped by 3 min of heating at 100°C instead of acid precipitation (which can cause formation of thiolactone from homocysteine at a lower pH). The assays were found to be reproducible.

Cystathionine γ-synthase was assayed in the same reaction mixture as that described above by using a 1 mM final concentration of l-cysteine. Disappearance of l-cysteine was measured by using either DTNB (13) or the ninhydrin reaction (25). Proteins were estimated by the method of Bradford (4).

RESULTS

Cloning of an L. meyeri DNA fragment able to complement E. coli metB and metA mutants.

A cosmid library with 25- to 40-kbp inserts of L. meyeri DNA was obtained. Km^r clones were screened for complementation of the metB1 E. coli mutant (WA802) at 30°C. Seven recombinant cosmids were found in two separate experiments using either mobilization (1,152 clones) or direct infection of WA802 with transducing phage particles (5,760 clones). Several restriction fragments were common to the seven cosmids. Cosmid pb10 containing a 25-kbp insert was kept for further study and was shown to also complement E. coli EC972 carrying the metB185 allele. No major rearrangement had occurred during the cloning experiment since BamHI restriction fragments from cosmid pb10 were the same size as fragments of the genomic DNA of L. meyeri as determined by Southern blotting (data not shown).

The pb10 cosmid carrying L. meyeri metY (the metB complementing activity) was shown to also carry the metX gene able to complement E. coli metA mutants (2). Further subcloning allowed us to locate metY more precisely within the 25-kbp insert of pb10 (Fig. 2A). Plasmid pb12 carrying a 6.8-kbp PstI fragment in pBR322 still complemented metB and metA mutants. Plasmids pb13s8, pb13c8, and pb13c9, generated by cloning an XbaI-PstI insert (2.9 kb) from pb12 into pSU21, pUC18, and pUC19, respectively, complemented only the metB mutant. Such expression of metY in both orientations (pUC18 and pUC19) is evidence for transcription of the metY gene from its own promoter. When a BamHI-XbaI fragment was cloned into pUC18 to yield pcn (Fig. 2A), no complementation was found, indicating that a BamHI site is located within the L. meyeri metY gene.

FIG. 2.

(A) Subcloning of L. meyeri DNA able to complement an E. coli metB mutant (WA802) and/or an E. coli metA mutant (AB1932). Large arrows indicate the orientations of the metX and metY genes. Names of the recombinant plasmids (or cosmid for pb10) are indicated on the left. Restriction site abbreviations: Ac, AccI; Ba, BamHI; Bg, BglII; Cfr, Cfr101; E, EcoRI; H, HindIII; P, PstI; Xb, XbaI. c8, c9, and s8 indicate the vectors pUC18, pUC19, and pSU21, respectively. ND, not done. (B) Cloning of the L. meyeri metX and metY genes under the control of the plac promoter (small arrows).

High similarity of MetY to O-acetylhomoserine sulfhydrylases.

The determination of the sequence of a 2.8-kb PstI-XbaI insert of pb13s8 (Fig. 2A) demonstrated an open reading frame of 1,304 nucleotides encoding a 443-amino-acid protein. Two putative start codons, separated by 21 nucleotides, were found, and the first one was chosen conservatively. The amino acid sequence deduced from metY and analyzed by the BLAST program showed the following sequence identities: 55% with Aspergillus nidulans O-acetylhomoserine sulfhydrylase (28), 50% with yeast O-acetylhomoserine sulfhydrylase (Met17 or Met25), and 40.5% with P. aeruginosa O-succinylhomoserine sulfhydrylase (MetZ) (10, 19). Amino acid sequence alignments of these proteins indicated a high overall similarity (Fig. 3). E. coli cystathionine γ-synthase (MetB) (7) was less similar to MetY (30% identity). In this respect, the deduced MetY amino acid sequence showed a stronger similarity to the sequences of fungal transsulfuration enzymes than to those of the corresponding bacterial enzymes, MetB and MetZ, which, as already mentioned (10, 12), had a gap of about 40 amino acids in their middle portion, unlike the sequences of the yeast enzymes (see Fig. 3 for MetZ).

FIG. 3.

Alignment of L. meyeri serovar semaranga metY (MetYLm) deduced amino acid sequence with the O-acetylhomoserine sulfhydrylases of A. nidulans (OAHSAn; accession no. U19394) and S. cerevisiae (MT17Sc; accession no. P06106) and the O-succinyl homoserine sulfhydrylase of P. aeruginosa (MetZPa; accession no. U10904). Identical residues are represented by bold characters, and similar residues are indicated by dots. A hyphen indicates a gap. The alignment was performed by use of Clustal V. The proposed lysine (K) of the pyridoxal phosphate binding site is indicated by # above the sequences.

All transsulfuration enzymes and enzymes catalyzing the incorporation of reduced sulfur in carbon chains utilize pyridoxal phosphate as a cofactor. From the sequence alignment depicted in Fig. 3, Lys-216 from L. meyeri O-acetylhomoserine sulfhydrylase appears to be strictly conserved in all other enzymes. It suggests that this lysine is the pyridoxal phosphate binding residue, in agreement with the location of the pyridoxal phosphate binding site found experimentally for β-cystathionase and cystathionine γ-synthase (21).

Evidence for direct sulfhydrylation in L. meyeri.

E. coli metB mutants carrying the pb12 plasmid (metX metY) grew very slowly (generation time, about 24 h) and stopped growing at an optical density at 600 nm (OD₆₀₀) of 0.2 (Table 2). This slow growth could be due to a weak expression of the L. meyeri methionine genes in E. coli. To improve expression, L. meyeri metX and metY genes were cloned under the control of the lac promoter. pxc8 (metX) was obtained by cloning the AccI-HindIII fragment from pb12 into pUC18, and pyc9 (metY) was obtained by deletion of the PstI-Cfr101 insert of pb13c9 (Fig. 2B). However, the growth of metA and metB E. coli strains (AB1932 and WA802) harboring pxc8 and pyc9, respectively, was not improved.

TABLE 2.

Growth of E. coli mutants bearing Leptospira metX and metY

Plasmida	Leptospira gene(s)	Relevant E. coli chromosomal mutation(s)	OD600b
pb12	metX metY	metB1	0.25
pxc8	metX	metA	NDc
pyc9	metY	metB1	0.2
pb13s8	metY	metB1	0.3
pxc8 and pb13s8	metX metY	metB1	0.18
pb13c9	metY	metB1 metC	0.83
pb13c9	metY	metB1	0.73
pb12	metX metY	metA metB1	1

^ac8, c9, and s8 indicate the vectors pUC18, pUC19, and pSU18, respectively. 

^bGrowth in minimal medium without methionine was evaluated after 24 h. 

^cND, not done. 

Based on the knowledge that E. coli cystathionine γ-synthase preferentially uses O-succinylhomoserine as a substrate (15) and that the product of the L. meyeri metX gene (O-acetyhomoserine transferase) uses only acetyl coenzyme A (2), two other hypotheses were tested. (i) Overexpression of the L. meyeri metX gene leading to high synthesis of O-acetylhomoserine might compensate for the deficit in E. coli of the substrate required for MetY activity. However, the growth of the E. coli metB1 mutant was not improved after transformation by both plasmids pxc8 and pb13s8. (ii) The cause of impaired growth could be found at the level of homocysteine biosynthesis, which could be either direct or via cystathionine (Fig. 1). The comparison of growth rates of E. coli mutants carrying the L. meyeri metY gene in a metB1 metC double mutant (indication of direct pathway) or in a metB1 metC⁺ mutant (indication of homocysteine synthesis via cystathionine) revealed that the growth rates were the same. This is in keeping with the fact that E. coli cystathionine β-lyase (Fig. 1A, step 3), the second enzyme of the transsulfuration pathway, is not needed within the recombinant E. coli carrying metY.

It is therefore suggested that MetY uses acetylhomoserine much better than it uses succinylhomoserine. Along these lines, much improved growth (an OD₆₀₀ of 1 was reached with a generation time of 8 h) was obtained with E. coli metA metB1 double mutants carrying the pb12 plasmid (metX metY) compared to that from the E. coli metB1 single mutant carrying the pb12 plasmid (metX metY) (generation time, 24 h; growth stopped at an OD₆₀₀ of 0.2) (Table 2). This suggested the proposed roles for reaction steps 1′ and 4 (Fig. 1B) for homocysteine biosynthesis in vivo.

MetY, an enzyme with O-acetylhomoserine sulfhydrylase activity.

MetY was overproduced by cloning the metY gene under the control of a very strong and inducible T7 promoter. The molecular mass of the overexpressed MetY protein as determined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (data not shown) was 45 kDa, in agreement with that calculated from the sequence. Crude cellular extracts of E. coli BL21 (D23) transformed with pET20b⁺ bearing metY exhibited, after induction, two very high activities, O-acetylhomoserine sulfhydrylase (15 μmol of homocysteine/min/mg of protein) and O-acetylserine sulfhydrylase (13 μmol of cysteine/min/mg of protein). A control was made with crude extracts of the same strain transformed with the pET20b⁺ vector alone. No acetylhomoserine sulfhydrylase activity was found, whereas O-acetylserine sulfhydrylase activity was observed at a level similar to that found in the strain bearing metY. However, no cystathionine γ-synthase activity was found. The O-acetylhomoserine sulfhydrylase activity was specific for the acetyl substrate; succinylhomoserine was not used as a substrate.

A question arises as to the regulation of the enzymatic activity by the end products of the pathway. The results indicate that L. meyeri MetY exhibited an O-acetylhomoserine sulfhydrylase activity which is feedback inhibited (33 or 25% residual activity) at very high concentrations (10 mM) of methionine or S-adenosylmethionine.

DISCUSSION

The L. meyeri metY gene was cloned by functional complementation of an E. coli metB mutant. However, the results suggest that cystathionine is not an intermediary metabolite of methionine synthesis in L. meyeri. The data are consistent with the enzymatic activity exhibited in vitro by MetY. Indeed, MetY protein can utilize sulfide for reaction with O-acetylhomoserine to yield homocysteine (Fig. 1B, step 4). In contrast to E. coli MetB protein (Fig. 1A, step 2), MetY protein does not use cysteine as a substrate, indicating that MetY protein is devoid of cystathionine γ-synthase activity. It is clear that complementation of E. coli metB mutants by the L. meyeri metY gene, however poor, became possible by marginal use of E. coli substrates. In fact, two substrates used by MetY protein differ from those used in E. coli: the homoserine derivative is an O-acetyl derivative, and the source of sulfur is not cysteine. A previous study from our laboratory has also shown that the L. meyeri metX product, O-acetylhomoserine transferase, the enzyme catalyzing the step upstream of MetY (2), does not transfer the acetyl group of acetyl coenzyme A to serine, showing its exclusive specificity for homoserine.

In S. cerevisiae, the Met17 (or Met25) enzyme exhibits both O-acetylserine sulfhydrylase and O-acetylhomoserine sulfhydrylase activities in vitro (32). Unfortunately, it was not possible to determine if L. meyeri MetY had an O-acetylserine sulfhydrylase activity in vitro since the latter activity was also present in E. coli extracts (see Results). The synthesis of cysteine by MetY was thus experimentally tested in vivo by functional complementation of the double E. coli cysK cysM mutant (these two genes specify two isoenzymes with O-acetylserine sulfhydrylase activity) (16). However, no complementation was found (data not shown). This could suggest that L. meyeri MetY is indeed similar to S. cerevisiae Met17 (or Met25), which was found ultimately to behave in vivo only as an O-acetylhomoserine sulfhydrylase (5), and that MetY may also have evolved from the same common ancestor of the γ-family of the transsulfuration enzymes (24).

At this point, the key goal is to determine the physiological role of the O-acetylhomoserine sulfhydrylase activity of L. meyeri MetY and whether it represents a major or alternative pathway. The organization of the two genes metX and metY in an operon (2) suggests the participation of both genes in the methionine pathway. Since the complementation of an E. coli metB mutant by the L. meyeri metY gene was effective, we rather expected to isolate an L. meyeri gene encoding cystathionine γ-synthase. However, the L. meyeri inserts from the seven recombinant cosmids able to complement both metA and metB E. coli mutants overlapped, indicating that they originated from the same region of the chromosome. We thus propose that the transsulfuration pathway via cystathionine does not exist in L. meyeri; this is in contrast to the situation found for fungi, which have both operating pathways for methionine biosynthesis (transsulfuration and sulfhydrylation). The proposed pathway for methionine biosynthesis for L. meyeri is shown in Fig. 4. With regard to metabolic regulation, we have reported that MetX, the first enzyme of the L. meyeri pathway, is not feedback inhibited (2). The concentration of methionine and S-adenosylmethionine giving 67 and 75% inhibition of O-acetylhomoserine sulfhydrylase (MetY), respectively, was 10 mM. The methionine or S-adenosylmethionine inhibition of O-acetylhomoserine sulfhydrylase seems not to be physiologically significant. Further studies are needed to examine regulation at the level of repression by methionine or its metabolites.

FIG. 4.

Proposed pathway for methionine biosynthesis in L. meyeri. MetX, O-acetylhomoserine transferase; MetY, O-acetylhomoserine sulfhydrylase.

It was of interest to compare the methionine biosynthetic pathway in a pathogenic species of Leptospira to that in the saprophytic L. meyeri species. A 525-bp DNA fragment from Leptospira interrogans serovar icterohaemorrhagiae strain Verdun was amplified by PCR assay with appropriate oligonucleotides chosen within metY (data not shown). The deduced amino acid sequence of the amplified product (Fig. 3, amino acids 110 to 285) was 86% identical to the amino acid sequence of MetY. Interestingly, the large insertion of 40 amino acids characteristic of MetY protein (Fig. 3, amino acids 234 to 276) and of yeast enzymes was found in the corresponding amino acid sequence of this pathogenic species. These results could suggest that the direct sulfhydrylation pathway for methionine is also operating in a pathogenic species of Leptospira.