An Unusual Mutation Results in the Replacement of Diaminopimelate with Lanthionine in the Peptidoglycan of a Mutant Strain of Mycobacterium smegmatis

Abstract

Mycobacterial peptidoglycan contains l-alanyl-d-iso-glutaminyl-meso-diaminopimelyl-d-alanyl-d-alanine peptides, with the exception of the peptidoglycan of Mycobacterium leprae, in which glycine replaces the l-alanyl residue. The third-position amino acid of the peptides is where peptidoglycan cross-linking occurs, either between the meso-diaminopimelate (DAP) moiety of one peptide and the penultimate d-alanine of another peptide or between two DAP residues. We previously described a collection of spontaneous mutants of DAP-auxotrophic strains of Mycobacterium smegmatis that can grow in the absence of DAP. The mutants are grouped into seven classes, depending on how well they grow without DAP and whether they are sensitive to DAP, temperature, or detergent. Furthermore, the mutants are hypersusceptible to β-lactam antibiotics when grown in the absence of DAP, suggesting that these mutants assemble an abnormal peptidoglycan. In this study, we show that one of these mutants, M. smegmatis strain PM440, utilizes lanthionine, an unusual bacterial metabolite, in place of DAP. We also demonstrate that the abilities of PM440 to grow without DAP and use lanthionine for peptidoglycan biosynthesis result from an unusual mutation in the putative ribosome binding site of the cbs gene, encoding cystathionine β-synthase, an enzyme that is a part of the cysteine biosynthetic pathway.

Mycobacteria and other closely related actinobacterial species have a unique cell envelope that sets them apart from other bacteria (7). This envelope contains an asymmetric bilayer, with an outer layer composed of lipids with short- to medium-chain-length fatty acids and an inner layer composed of long-chain mycolic acids covalently attached to an arabinogalactan polysaccharide. The arabinogalactan is in turn covalently attached to the peptidoglycan via a disaccharide linker, forming the mycolylarabinogalactan-peptidoglycan complex (6). Thus, the peptidoglycan is the anchor for the principal mass of this unique cell envelope.

Bacterial peptidoglycan is comprised of a β-1,4-linked polymer of N-acylated muramic acid and glucosamine carbohydrates with cross-linked peptides of varied compositions (13). In virtually all bacteria, both carbohydrates in the glycan chain are N acetylated; in the mycobacteria, however, the muramic acid moieties are N glycolylated instead (2). The pentapeptide portion of mycobacterial peptidoglycan is one of the most common types found in bacteria, l-alanyl-d-iso-glutaminyl-meso-diaminopimelyl-d-alanyl-d-alanine, with glycine replacing the l-alanyl residue in Mycobacterium leprae, but it differs from the other common types in that the diaminopimelyl residue can be monoamidated (18, 25). The third-position amino acid in peptidoglycan is an important residue, involved with peptide cross-linking and attachment of cell envelope proteins to the peptidoglycan. Cross-linking of the mycobacterial peptidoglycan is direct, with standard linkages occurring between a meso-diaminopimelate (DAP) residue of one peptide and the penultimate Ala residue of another peptide and with unusual linkages occurring between two DAP residues of adjacent peptides (30). In Mycobacterium smegmatis, approximately 75% of the peptidoglycan peptides are cross-linked, with 50% of the cross-links consisting of DAP-Ala linkages and 25% of the cross-links consisting of DAP-DAP linkages (30). The significance of these two types of cross-links in the peptidoglycan is unknown. It has been proposed that the unusual DAP-DAP cross-links may be required to stabilize the cell wall during stationary phase and may have a role in long-term survival under nonreplicating conditions (14). It has also been proposed that the enzymes responsible for the formation of DAP-DAP cross-links are insensitive to inhibition by β-lactam antibiotics and might be new drug targets (14).

We previously constructed DAP-auxotrophic mutants of M. smegmatis and showed that they undergo lysis (DAP-less death) when deprived of DAP (9, 22). We then isolated a set of spontaneous mutants of the DAP auxotrophs that can grow without DAP, whose phenotype we termed “Sud” (suppressor of DAP-less death) (9). The Sud mutants are grouped into seven classes (I to VII), depending on how well they grow without DAP and whether they are sensitive to DAP, temperature, or detergent (9). All of the mutants are hypersusceptible to β-lactam antibiotics when grown in the absence of DAP, suggesting that the peptidoglycan architecture is altered in these mutants (9). We proposed that these Sud mutants utilize an alternative amino acid in place of DAP and that this amino acid affects peptidoglycan biosynthesis and may alter cell wall assembly. In this study, we report that one of the class II Sud mutants, strain PM440, replaces DAP with lanthionine in the peptidoglycan when grown without DAP. We also show that the ability of PM440 to grow without DAP is due to an unusual mutation in the putative ribosome-binding site (RBS) of the cbs gene, encoding cystathionine β-synthase, an enzyme involved with cysteine biosynthesis. Furthermore, we demonstrate that the cbs gene is required for the Sud phenotype for all but two of the seven classes of Sud mutants.

MATERIALS AND METHODS

Bacterial strains, media, and reagents.

The strains used in this study are shown in Table 1. All M. smegmatis mutants in this study were derived from the common laboratory strain mc²155 (26), used for the M. smegmatis genome sequencing project (http://www.tigr.org/tdb/mdb/mdbinprogress.html). Escherichia coli cultures were grown in Luria-Bertani (LB) broth (Difco, BD Biosciences, San Jose, Calif.) or on LB agar. Mycobacterial cultures were grown in LB broth with 0.05% Tween 80, in Middlebrook 7H9 broth (Difco, BD Biosciences), or on Middlebrook 7H11 medium (Difco, BD Biosciences). All Middlebrook media were supplemented with a solution containing 0.05% Tween 80, 0.2% glycerol, 0.5% bovine serum albumin (fraction V; Roche Applied Science, Indianapolis, Ind.), 0.2% dextrose, and 0.85% NaCl. Sucrose was added to media at a concentration of 2% (wt/vol) when required. l-Lysine, l-methionine, and l-threonine were added at 40 μg/ml to all M. smegmatis cultures in Middlebrook media, dl-homoserine was used at 80 μg/ml, and meso-diaminopimelate was added at 200 μg/ml to both LB broth with 0.05% Tween 80 and Middlebrook media. When necessary, ampicillin (50 μg/ml; Sigma-Aldrich Chemical, St. Louis, Mo.), hygromycin (50 μg/ml; Roche Applied Science), kanamycin (10 μg/ml; Sigma-Aldrich Chemical), or apramycin (50 μg/ml; Sigma-Aldrich Chemical) were added to media. M. smegmatis plates were incubated for 3 to 6 days at 37°C unless otherwise indicated. Sephadex G-25 and Superdex peptide columns were from Amersham Biosciences (Piscataway, N.J.). The Hypersil ODS (C₁₈) column and the EZ:faast amino acid analysis kit were from Phenomenex (Torrance, Calif.). High-performance liquid chromatography-grade water and acetonitrile were purchased from Burdick and Jackson (Muskegon, Mich.). All other chemicals were purchased from Sigma-Aldrich Chemical.

TABLE 1.

Strains and plasmids used in this study

Strain or plasmid	Description	Reference or source
Strains
E. coli K-12
HB101	F− Δ(gpt-proA)62 leuB1 glnV44 ara-14 lacY1 hsdS20 rpsL20 xyl-5 mtl-1 recA13	4
DH10B	F−mcrA Δ(mcrBC-hsdRMS-mrr) (φ80dlacZΔM15) ΔlacX74 deoR recA1 endA1 araD139 Δ(ara leu)7697 galU galK λ−rpsL nupG	Gibco-BRL
STBL2	F−mcrA Δ(mcrBC-hsdRMS-mrr) recA1 endA1 gyrA96 thi supE44 relA1 λ− Δ(lac-proAB)	Gibco-BRL
M. smegmatis
PM274	ept-1 ΔlysA4 rpsL6 cbs+	9
PM321	ept-1 ΔlysA4 rpsL6 ask1::aph cbs+ (Met Thr Lys, DAP auxotrophic parent, Kmr)	9
PM440	ept-1 ΔlysA4 rpsL6 ask1::aph cbs1 (class II sud-21 mutant of PM321)	9
PM762	ept-1 ΔlysA4 rpsL6 ask1::aph cbs1 (PM440 cbs allelic-exchange derivative of PM321)	This work
PM777	ept-1 ΔlysA4 rpsL6 ask1::aph cbs1::aacC4 (cbs deletion derivative of PM440, Amr)	This work
Plasmids
pKSI+	Apr, high-copy-number cloning vector	Stratagene
pMV261.aph	Kmr, E. coli-mycobacterial shuttle vector, groELp, replicative	27
pMV261.hyg	Hygr, E. coli-mycobacterial shuttle vector, groELp, replicative	AECOMa
pMV361.hyg	Hygr, E. coli-mycobacterial shuttle vector, groELp int attP, nonreplicative but integration proficient in mycobacteria	AECOM
pYUB412	Apr Hygr, E. coli-mycobacterium PacI-excisable shuttle cosmid vector, int attP, nonreplicative but integration proficient in mycobacteria	23
pYUB415	Apr Hygr, E. coli-mycobacterium PacI-excisable shuttle cosmid vector, replicative	3
pYUB657	Apr Hygr, mycobacterial sacB suicide vector	23
pMP139	pKSI bearing the aacC4 gene, conferring resistance to apramycin (Amr)	10
pMP144	pYUB412::PM440, sud-21 (cbs1) cosmid from PM440 genomic library	This work
pMP165	pYUB415::pMP144, PacI insert of pMP144 repackaged in pYUB415 cosmid vector	This work
pMP167	1,068-bp EcoRI-SpeI fragment from pMP139 containing the aacC4 gene, into same sites of pMV261.aph, replacing the aph gene	This work
pMP194	12-kb EcoRI fragment from pMP165, bearing the sud-21 allele (cbs1) cloned into the EcoRI site of pKSI	This work
pMP195	12-kb EcoRI fragment from pMP165, bearing the sud-21 allele (cbs1) cloned into the EcoRI site of pMP167	This work
pMP196.1	5-kb BamHI fragment of pMP195 cloned into the BamHI site of pMV261.hyg; the 5′ end of the cbs1 gene is downstream of groELp	This work
pMP196.2	Same as pMP196.1 except that the insert is in the opposite orientation	This work
pMP226	3-kb fragment bearing sud-21 (cbs1) of PM440; PCR amplified from pMP194 with primers Pv172 and Pv173 and cloned into pMV261.hyg with HindIII and HpaI; the 5′ end of the cbs1 gene is downstream of groELp	This work
pMP242	pYUB657 bearing a 3-kb PstI fragment containing the cbs1 gene of PM440 with flanking DNA: suicide plasmid for cbs1 allelic exchange in M. smegmatis	This work
pMP244	pYUB657 bearing a 3-kb PstI fragment containing the Δcbs1::aacC4 allele with flanking DNA; suicide plasmid for disruption of the cbs gene of M. smegmatis	This work
pMP263	3-kb fragment bearing the wild-type cbs gene; PCR amplified from PM321 genomic DNA with primers Pv172 and Pv173 and cloned into pMV261.hyg with HindIII and HpaI; the 5′ end of the cbs+ gene is downstream of groELp	This work
pMP310	1.9-kb fragment bearing the 3′ BamHI deletion derivative (cbs′) of the wild-type cbs gene; PCR amplified from PM321 genomic DNA with primers Pv172 and Pv173 and cloned into pMV261.hyg with HindIII and BamHI; the 5′ end of the cbs+ gene is downstream of groELp	This work
pMP311	1.9-kb fragment bearing the cbs1 3′ BamHI deletion derivative (cbs1′); PCR amplified from PM440 genomic DNA with primers Pv172 and Pv173 and cloned into pMV261.hyg with HindIII and BamHI; the 5′ end of the cbs1′ gene is downstream of groELp	This work

^aAECOM, Albert Einstein College of Medicine.

DNA manipulation and plasmid construction.

DNA methods were performed essentially as described previously (1). DNA fragments were isolated using agarose gel electrophoresis and absorption to a silica matrix (GeneClean; Bio101, Vista, Calif.) or with QIAquick spin columns (QIAGEN, Inc., Chatsworth, Calif.). Preparation of genomic and plasmid DNA from M. smegmatis was done as previously described (5, 23). Oligonucleotides for PCR were synthesized by Invitrogen Life Technologies (Carlsbad, Calif.). DNA sequencing was done by ACGT, Inc. (Wheeling, Ill.). All restriction and DNA-modifying enzymes were from New England Biolabs (Beverly, Mass.) or Fermentas (Hanover, Md.).

Plasmids used in this study are listed in Table 1. Detailed descriptions of plasmid construction can be obtained from the corresponding author. Plasmids were constructed in E. coli DH10B or in M. smegmatis and were prepared by an alkaline lysis protocol or with QIAGEN (Valencia, Calif.) columns if the plasmids were used for recombination. Mycobacterial genomic inserts from integrated cosmid library clones of M. smegmatis were recovered as previously described (23).

PCR amplification of the M. smegmatis cbs gene region.

The oligonucleotide primers Pv172 (5′-CGCTAAGCTTCGCGGCACGGATGCTGCGGCCCAG-3′) and Pv173 (5′-TGCCGTTAACGATGCCCTGCAGGATGGCGTAG-3′) were used to amplify the cbs coding region, including 840 bp of upstream DNA and 831 bp of downstream DNA, from M. smegmatis genomic DNA and to introduce a 5′ HindIII recognition site and a 3′ HpaI recognition site. The PCR was performed in a Perkin-Elmer GeneAmp 2400 temperature cycler (Applied Biosystems, Foster City, Calif.) with the GC-rich PCR system (Roche Applied Science) with 3 mM MgCl₂ and 1.2 M GC-rich resolution solution. PCR parameters included an initial melt at 94°C for 5 min, followed by 25 cycles of a melting step at 94°C for 1 min and an annealing step at 55°C for 1 min and then by an extension step at 68°C for 3 min.

Purification and analysis of UDP-acylmuramyl-pentapeptides.

Cells for muropeptide preparation were harvested from mid-exponential-phase cultures (optical density at 600 nm, 0.4 to 0.8) that were quickly chilled. The cells were washed in ice-cold phosphate-buffered saline (pH 7), pelleted, and frozen at −20°C. Ten grams of cell pellet was suspended in ice-cold 50 mM MOPS buffer [3-(N-morpholino) propanesulfonic acid, pH 8] and subjected to probe sonication on ice. The cell lysate was centrifuged at 27,000 × g, and the supernatant was transferred to a new tube to which trichloroacetic acid was added at a concentration of 10%. The mixture was stirred on ice and centrifuged at 15,000 × g and 4°C. The clear supernatant was transferred to a new tube, and the trichloroacetic acid was removed by three extractions with diethyl ether. The resulting solution was dried on a rotary evaporator, reconstituted in water, and loaded on a Sephadex G-25 (116- by 2.5-cm) column equilibrated with 75 mM ammonium acetate (pH 5). The column was calibrated with authentic UDP-acylmuramyl-pentapeptide prepared using recombinant E. coli enzymes (31). Fractions containing UDP-acylmuramyl-pentapeptide were pooled and lyophilized to remove ammonium acetate. These partially purified nucleotide-linked precursors were resuspended in 2 M trifluoroacetic acid and incubated at 60°C for 1 h to remove the UDP moiety. The resulting muropeptides were further purified by size exclusion chromatography on a Superdex peptide 10/300 GL column, equilibrated, and eluted with 30% acetonitrile containing 0.1% trifluoroacetic acid. An aliquot of the muropeptide-containing fraction was applied to a Hypersil ODS C₁₈ column connected to a model 1100 high-performance liquid chromatography system (Agilent Technologies, Pal Alto, Calif.). The eluent was directly introduced into an LCQ Duo electrospray mass spectrometer (Finnigan-Thermoquest, San Jose, Calif.), and the muropeptides were analyzed by mass spectrometry (MS) and tandem mass spectrometry (MS-MS). Amino acid analyses of the muropeptides were performed using an EZ:faast gas chromatography-MS kit.

Cystathionine β-synthase assays.

Cystathionine β-synthase (CBS) assays were performed on whole-cell lysates obtained from mid-exponential-phase cultures (optical densities at 600 nm, 0.4 to 0.8). M. smegmatis cells from 50-ml cultures were pelleted, washed two times in cold reaction buffer (200 mM Tris-HCl [pH 8.0]), and resuspended in 3 ml of reaction buffer with 50% (wt/vol) glycerol DNase (100 U; Roche Applied Science), RNase A (100 μg; Sigma-Aldrich), and protease inhibitor [3 mM 4-(2-aminoethyl) benzenesulfonylfluoride (AEBSF; Calbiochem, San Diego, Calif.)] added. Cells were broken in a French pressure cell (14,000 lb/in², four applications), and cell debris was removed by centrifugation (12,000 × g, 5 min). The whole-cell lysates were stored at −20°C until assayed. Protein concentration was determined with the Bradford assay (Bio-Rad, Hercules, Calif.). The cystathionine β-synthase activity of the lysates was determined essentially as previously described (17), using a Beckman DU 530 spectrophotometer that was fitted with a temperature control module programmed for 37°C and had a 3-min scan time; the absorbance at 390 nm was recorded every 10 s. Assays were done in duplicate on duplicate cultures. The cystathionine β-synthase activity, expressed as the change in absorbance at 390 nm/min/mg of protein, was in the linear range for the amount of protein (0.4 to 1.0 mg) used for each assay.

RESULTS

Identification of lanthionine in the peptidoglycan of strain PM440 grown in the absence of DAP.

Strain PM440 is a spontaneous sud-21 mutant derived from M. smegmatis strain PM321, an ask1::aph mutant auxotrophic for amino acids of the aspartate family pathway (9, 29). The PM321 mutant requires Met, Thr, homoserine, Lys, and DAP supplementation for growth (9). Strain PM440 is a spontaneous class II Sud (suppressor of DAP-less death) mutant isolated as described previously that grows well in the absence of DAP but still has the ask1::aph mutation and requires Met, Thr, Lys, and homoserine for growth (9). We examined the UDP-acylmuramyl-pentapeptide peptidoglycan precursors purified from extracts of the parental DAP-auxotrophic strain PM321 and of the Sud mutant PM440 grown with and without DAP, respectively, by liquid chromatography-MS to determine what amino acid replaced DAP in the peptidoglycan of strain PM440. The positive ion mass spectra of the muropeptides obtained from the parent, PM321, were dominated by two molecular ions, m/z 808.2 and 824.2 (Fig. 1A), which correspond to MurNAc-l-Ala-d-Glu-DAP-d-Ala-d-Ala and MurNGlyc-l-Ala-Glu-DAP-d-Ala-d-Ala, respectively. We confirmed the identities of these two molecular ions by MS-MS. The spectra of muropeptides from the Sud mutant, PM440, were dominated by two molecular ions, m/z 826.2 and 842.2 (Fig. 1B), but ions of m/z 808.2 and 824.2 were not detected. The MS-MS data suggest that the ions having m/z of 826.2 and 842.2 were N acetylated and N glycolylated muramyl peptides where the DAP was replaced by another amino acid with a molecular mass 18 atomic mass units higher than the molecular mass of DAP. Amino acid analyses (gas chromatography-MS) identified Ala, Glu, and DAP in the muropeptides from the wild-type strain (Fig. 2A) and lanthionine in the muropeptides from the Sud mutant PM440 (Fig. 2B). The identity of the lanthionine peak was confirmed by performing the same analysis on the lanthionine-containing lantibiotic nisin (Fig. 2C). In nisin synthesis, lanthionine is produced by posttranslational modification of cysteine residues in the nisin polypeptide (15). Very little has been discovered about the de novo synthesis of lanthionine in bacteria, and we wanted to better understand how mutant PM440 was able to use lanthionine in place of DAP. Therefore, we sought to identify the sud-21 mutation responsible for the mutant phenotype.

FIG. 1.

Positive ion mass spectra of muropeptides from PM321 (A) and the Sud mutant PM440 (B). Muropetides were isolated from nucleotide-linked peptidoglycan precursors isolated from PM321 grown with DAP and from PM440 grown without DAP. Ions with m/z values of 830.3 and 846.3 (A) and 848.2 and 864.2 (B) represent monosodium adducts of the dominant muropeptides.

FIG. 2.

Amino acid analysis of muropeptides isolated from PM321 (A), PM440 (B), and nisin (C). Analyses were done using the EZ:faast amino acid analysis system. The retention times and identities of relevant amino acids are indicated above the peaks; the identity of each amino acid was confirmed by mass spectrometry. Structural diagrams of DAP and lanthionine are inset into panels A and B, respectively. The stereoconfiguration of lanthionine was not determined.

Identification of the sud-21 allele as cbs1 in M. smegmatis PM440.

We identified the sud-21 mutation in strain PM440 by genetic complementation. We constructed a genomic library from PM440 in the integrative mycobacterial cosmid shuttle vector pYUB412 (Table 1), electroporated the library into the parental strain PM321, and selected for transformants that could grow in the absence of DAP. Positive clones were then screened for hygromycin resistance encoded by the cosmid vector. The PM440 genomic DNA from one such Hyg^r Sud clone was recovered from the chromosome of the complemented strain and packaged as cosmid pMP144 (Table 1). We found that this integrating cosmid was unstable in E. coli and thus repackaged the genomic insert within pMP144 into the replicating mycobacterial vector pYUB415 (Table 1), yielding cosmid pMP165. All subsequent manipulations with pMP165 and subcloning were done using M. smegmatis and the mycobacterial replicating vector pMV261.hyg without passage of the DNA in E. coli. One of the unstable cosmid clones derived from pMP144 that no longer conferred the Sud phenotype was missing an ∼12-kb EcoRI DNA fragment present in the original Sud⁺ cosmid. We cloned the 12-kb EcoRI DNA fragment to produce plasmid pMP195 and showed that this plasmid carried the sud-21 allele (Fig. 3). Subcloning of pMP195 revealed that the sud-21 mutation was located on the 5-kb BamHI fragment in pMP196.1 (Fig. 3). We sequenced the DNA at each end of the insert in pMP196.1 and found that the right side of the insert included a truncated cbs gene encoding a protein similar to eukaryotic cystathionine β-synthases but missing one-third of the 3′ end of the gene. We removed the insert within pMP196.1, cloned it in the opposite orientation with respect that of to the groEL promoter of the same vector, and found that the new plasmid, pMP196.2, unlike pMP196.1, did not permit the parental strain PM321 to grow in the absence of DAP (Fig. 3).

FIG. 3.

Complementation of the ask mutant PM321 with various cbs plasmids. The DAP-auxotrophic strain PM321 (ask1::aph) was transformed to have a hygromycin resistance phenotype with various plasmids bearing different cbs alleles. The transformed cells were then tested for their ability to grow in media with (+DAP) and without (−DAP) diaminopimelate acid. The lipU gene is a putative lipase located just upstream of the cbs gene of M. smegmatis. In all of the plasmids except pMP195 and pMP196.2, the DNA inserts with the cbs alleles were oriented such that the cbs genes were transcribed by the groEL promoter of the vector.

We identified the full-length cbs gene on DNA contig 3439 of The Institute for Genomic Research M. smegmatis genome sequencing project (http://www.tigr.org/tdb/mdb/mdbinprogress.html). The M. smegmatis CBS protein is 86% identical to the Rv1077 protein of Mycobacterium tuberculosis H37Rv (8), which is annotated as a putative cystathionine β-synthase (http://genolist.pasteur.fr/TubercuList/), and includes a pyridoxal phosphate-binding domain and two cystathionine β-synthase regulatory domains (data not shown). The amino acid sequence homologies and the fact that CBS is a pyridoxal phosphate-dependent enzyme (17) are consistent with the idea that the gene identified in this study encodes the cystathionine β-synthase of M. smegmatis.

Cystathionine is an intermediate in the methionine biosynthetic pathway in prokaryotes and an intermediate in the synthesis of cysteine in eukaryotes and actinobacteria, including mycobacteria (Fig. 4) (17, 21). In the latter pathway, cystathionine is produced from serine and homocysteine by the enzyme cystathionine β-synthase. Because cystathionine is a precursor to cysteine, which may be a precursor for lanthionine (24), we decided to examine the cbs gene in more detail. We used PCR to amplify the cbs gene from the parental strain PM321 and the Sud mutant PM440, cloned each gene into the expression vector pMV261.hyg, and tested the constructs for the Sud phenotype of PM321. As shown in Fig. 4, only plasmid pMP226 bearing the cbs gene (designated cbs1) from the Sud mutant PM440, and not pMP263 bearing the cbs⁺ allele, permitted PM321 to grow in the absence of DAP. The original plasmid subclone, pMP196.1 (Fig. 3), has a 3′-end-truncated version of the cbs1 gene encompassing the CBS regulatory domains, which are not required for the activity of the eukaryotic enzymes (16, 17) To test whether these domains are dispensable for the mycobacterial enzyme and to confirm that the cbs1 allele is responsible for the Sud phenotype, we took plasmids pMP226 (cbs1) and pMP263 (cbs⁺), truncated each cbs allele at the same BamHI site of the truncation in pMP196.1, and found that pMP311, which has the truncated cbs1 allele, but not pMP310, which has the truncated wild-type allele, conferred the Sud phenotype to PM321 (Fig. 3).

FIG. 4.

Role of cystathionine in amino acid biosynthesis. Cystathionine is an intermediate in the biosynthesis of methionine in prokaryotes (top) and an intermediate in the biosynthesis of cysteine in eukaryotes and actinobacteria (bottom). In cysteine synthesis, cystathionine β-synthase, encoded by the cbs gene, is responsible for the production of cystathionine, while the corresponding enzyme in methionine synthesis is cystathionine γ-synthase, encoded by the metB gene.

The mutation in the cbs1 allele is in the putative ribosome-binding site of the cbs gene.

We sequenced the cbs1 gene using a PCR product from PM440 genomic DNA as the template and compared its sequence to that of the wild-type gene. The only difference between the sequences was a G-to-T mutation in the putative RBS of the cbs1 gene of PM440 (Fig. 5). A consensus sequence for mycobacterial ribosome-binding sites is AGGAGGA, with a 4- to 11-nucleotide spacer between the RBS and the start codon (11). The only sequence in the RBS region upstream of cbs that matches any part of the consensus sequence is GAG, positioned 12 nucleotides upstream of the cbs start codon (Fig. 5). While this is a short sequence, the mycobacterial translation machinery is believed to not be dependent upon the entire consensus RBS sequence (28). Because of this view, we tentatively describe the GAG sequence as the RBS for the cbs open reading frame. There are 29 nucleotides between the stop codon of the upstream gene lipU (a putative lipase) and the start codon of the cbs gene. In addition, there are several pairs of inverted repeats just upstream of the cbs gene, and one of these repeat sequences encompasses the putative RBS (Fig. 5).

FIG. 5.

Point mutation in the cbs1 allele of the sud mutant PM440. Shown is the sense-strand DNA sequence in the region near the start codon of the cbs gene of M. smegmatis. Upstream is the 3′ end of the lipU gene, followed by a 29-nucleotide intergenic region (in lightface) containing the putative RBS, followed by the ATG start codon of the cbs open reading frame. The mutation in the cbs1 allele is a G-to-T change (in boldface italic) in the putative RBS. There are also four sets of inverted repeats (IR1 to IR4) in this region, designated by arrows above the sequence. Note that IR1 includes the mutation site.

We were surprised by the sequence data, and so we performed two allelic-exchange experiments to confirm that the RBS mutation was responsible for the Sud phenotype of PM440. First, we constructed a sacB-based suicide plasmid, pMP242, bearing the cbs1 allele and used it to replace the wild-type cbs gene in PM321 with cbs1. The resultant mutant, strain PM762 (Table 1), was capable of growing without DAP and was hypersusceptible to β-lactam antibiotics like the original cbs1 strain PM440 (data not shown). The mutation in the cbs1 allele removes a DdeI site at the RBS sequence, and thus the replacement of cbs⁺ with cbs1 in PM762 was confirmed by DdeI restriction enzyme digestion of a PCR product amplified from PM762 genomic DNA with cbs-specific primers. For the second experiment, we built another sacB suicide plasmid, pMP244, bearing the mutated cbs1 allele disrupted with the aacC4 gene, conferring apramycin resistance. We used this plasmid to knock out the cbs1 allele in the Sud mutant PM440. The resultant derivative mutant, PM770 (Table 1), like the wild-type strain PM321, could not grow in the absence of DAP (data not shown). These experiments demonstrate that the G-to-T mutation in the putative RBS of the cbs gene of strain PM440 is responsible for the Sud phenotype.

The cbs1 mutation results in increased cystathionine β-synthase activity.

The mutation in the RBS of the cbs1 gene of PM440 is unusual, and so we sought to better understand the mechanism behind the phenotype of this mutation. We hypothesized that the RBS mutation somehow affects the translation of the cbs mRNA. As mentioned above, there are several inverted repeats upstream of the cbs gene (Fig. 5). We surmise that some of these repeats might contribute to secondary structures in the mRNA that might regulate cbs translation. We could not obtain relevant information about the mRNA secondary structure from RNA-folding algorithms, as we do not know the location and size of the mRNA for the cbs gene. However, one of the repeat sequences (IR1) includes the 5′ end of the putative RBS and contains the G-to-T mutation. If the mRNA for this region adopts a secondary structure that includes this repeat, it might potentially interfere with ribosome access to the RBS. In this model, the G-to-T mutation in the cbs1 allele would serve to destabilize that part of the repeat, thus allowing better translation initiation. This could result in increased protein translation and a higher level of CBS enzyme activity. We tested part of this hypothesis by assaying the cystathionine β-synthase activity in whole-cell lysates prepared from strains with various chromosomal cbs mutations, which were grown in media with or without DAP. As shown in Fig. 6, ask1::aph mutant strains bearing the cbs1 mutation had elevated CBS activity relative to that of the parental cbs⁺ strain PM321. Both PM440, the original cbs1 mutant, and PM762, the cbs1 mutant constructed by allelic exchange, had elevated CBS activities in both types of media, although the activity of PM440 grown with DAP was lower than that of cells grown without DAP. Parental strain PM321 had a level of CBS activity that was essentially the same as that of the cbs knockout mutant PM777 (Table 1 and Fig. 6). Interestingly, PM274, the cbs⁺ ask⁺ parent of PM321 (Table 1), had elevated CBS activity in both media (Fig. 6), indicating that the ask1::aph mutation in PM321 has some influence on the expression of CBS activity.

FIG. 6.

CBS activity is elevated in cbs1 mutants. The graph shows the CBS activity of whole-cell extracts of cbs⁺ and cbs1 mutant strains. Duplicate cultures of each strain were assayed in duplicate, and the average values were reported with standard deviations. Strain PM321 is the ask1::aph DAP-auxotrophic parent, PM440 is the original spontaneous cbs1 mutant, PM762 is the cbs1 mutant constructed by allelic exchange with the PM321 parent, and PM777 is the cbs1::aacC4 knockout derived from the cbs1 mutant PM440. Strain PM274 is the ask⁺ cbs⁺ DAP-prototrophic parent of PM321.

Analysis of the cbs gene in other classes of Sud mutants.

We performed cbs allelic-exchange experiments with representatives of each class of Sud mutants and found that for all but two classes, the cbs gene was required for the Sud phenotypes (data not shown). In these experiments, we disrupted the cbs gene of each mutant with the suicide plasmid pMP244 (above data and Table 1) and showed that only the class IV and V Sud mutants lacking CBS could still grow in the absence of DAP. These two classes consist of the only temperature-sensitive Sud mutants in our collection.

DISCUSSION

We previously isolated a set of mutants from a DAP-auxotrophic strain of M. smegmatis that no longer required DAP for growth (9). We proposed that the mutants can replace DAP with a different amino acid that is not used as efficiently as DAP and that although the cells grow, they do not produce a normal peptidoglycan. In this study, we showed that one of the mutants, strain PM440, uses lanthionine in place of DAP, the consequence of an unusual mutation that results in the overexpression of cystathionine β-synthase activity in the mutant.

We believe that the elevated cystathionine β-synthase levels in the cbs1 mutant PM440 likely results in an excessive amount of cysteine that is subsequently converted to lanthionine. Both cystathionine and lanthionine are analogs of DAP that can act as alternative substrates for MurE, the enzyme that adds DAP to peptidoglycan precursors, with lanthionine as the preferred alternative substrate (19). Lanthionine is a normal constituent of the peptidoglycan of Fusobacterium capsulatum (12), but little is known about the de novo biosynthesis of this amino acid. To the best of our knowledge, this is the first description of lanthionine as a metabolite in a mycobacterial species. We surmise that lanthionine is synthesized by the condensation of cysteine and serine, or perhaps directly from cysteine, as previously proposed (24).

The replacement of DAP with isomers of cystathionine and lanthionine in the peptidoglycan of E. coli has been reported. DAP auxotrophs of E. coli can be grown without DAP if the cultures are supplemented with either cystathionine or lanthionine (19). It is also possible to redirect amino acid pathways in E. coli such that a DAP auxotroph can grow in the absence of DAP (24). An E. coli strain engineered to overproduce the cystathionine γ-synthase enzyme (MetB) and to lack the cystathionine β-lyase enzyme (MetC) of the methionine biosynthetic pathway synthesizes an excess of l-cystathionine and l-allo-cystathionine and produces meso-lanthionine, all of which are incorporated into the peptidoglycan in the absence of DAP (24). However, it is not understood how E. coli synthesizes lanthionine. The mycobacterial mutant described here is unique in that it has a single, spontaneous mutation resulting in a total replacement of DAP with lanthionine but not cystathionine. This might indicate that peptidoglycan biosynthesis in mycobacteria has more constraints on peptide structure than that in E. coli.

The Sud mutant PM440 is hypersusceptible to β-lactam antibiotics when grown in the absence of DAP. The use of lanthionine in place of DAP in the mutant might potentially impact several steps in peptidoglycan biosynthesis, resulting in an abnormally assembled peptidoglycan which could influence the susceptibility of the cell to β-lactam antibiotics. One explanation for this phenotype is that the production of lanthionine is inefficient and the intracellular pools of lanthionine are smaller than the pools of DAP when the cells grow in DAP-supplemented media. This might affect the overall peptidoglycan precursor level, which would result in a smaller amount of mature peptidoglycan and hence in increased susceptibility to β-lactam antibiotics. It is also possible that the addition of lanthionine to the precursors is less efficient than the addition of DAP, resulting in a decrease in the steady-state pool of precursors. We think that this is unlikely, given that the addition of amino acids to the precursors is not the rate-limiting step in peptidoglycan biosynthesis. The presence of lanthionine in the precursors would probably not interfere with the translocation of the precursors across the cytoplasmic membrane. We propose that the lanthionine interferes with the cross-linking reactions carried out by the peptidoglycan transpeptidases during assembly of the mature peptidoglycan. This idea is supported by the observations that E. coli DAP auxotrophs using lanthionine or cystathionine in place of DAP had reduced cross-links in the mature peptidoglycan (19). Similarly, E. coli cells engineered to insert lysine into peptidoglycan precursors in place of DAP can incorporate the altered precursors into the mature peptidoglycan, but the lysine residues do not function as acceptors in peptide cross-linking (20).

We hypothesize that lanthionine is not efficiently recognized by the peptidases responsible for peptidoglycan cross-linking. The side chain of lanthionine differs from that of DAP by the replacement of a CH₂ group with a thioester. This substitution results in a longer lanthionine side chain than that of DAP; the greater length might be enough to decrease the overall efficiency of peptidoglycan cross-linking reactions, resulting in a mature peptidoglycan with a reduced number of cross-links. Alternatively, the overall number of cross-links may not decrease; rather, the ratio of the two types of cross-links found in mycobacterial peptidoglycan (DAP-DAP and DAP-Ala) might be skewed. The cbs1 mutant may be able to link lanthionine to d-alanine but may not be able to efficiently link lanthionine to lanthionine. In either case, the net result would be a decrease in the amount of peptidoglycan cross-linking that could manifest itself as an increase in susceptibility to β-lactam antibiotics.

In wild-type peptidoglycan, the l center of DAP is present in the backbone of the peptidoglycan peptide, and the d center is used for cross-linking. We do not know the chirality of the lanthionine in the peptidoglycan of mutant PM440. We assume that the lanthionine is synthesized with both of its chiral centers in the l configuration and that it is epimerized to the meso form by an epimerase, possibly DapF, the same enzyme responsible for the epimerization of l,l-DAP to meso-DAP in the DAP biosynthetic pathway. Such is the case for the E. coli mutant that produces meso-lanthionine and l-allo-cystathionine for the peptidoglycan synthesis described above (24).

We have shown that the cbs gene is required for the majority of the Sud mutants in our collection. We originally thought that these classes represented replacement of DAP by different amino acids. Here we show that all but the class IV and V mutants likely place lanthionine in the peptidoglycan when they are grown without DAP, as disruption of the cbs gene eliminates the Sud phenotype of these mutants. We hypothesize that the cbs1 mutation was the first to occur in each strain and that the different phenotypes of these mutants arose from secondary mutations.

The unique class IV and V Sud mutants that do not have cbs mutations are distinguished from the other classes by their inability to grow without DAP at 37 and 42°C and by their sensitivity to Tween 80 when grown at 30°C in the absence of DAP (9). Furthermore, mutants in class IV grow better than mutants in class V without DAP at 30°C. These mutants may have another amino acid other than lanthionine replacing DAP; however, it is also possible that lanthionine might be produced in these mutants via a non-CBS-dependent mechanism.

The biochemical and genetic characterization of the mutant described in this study might allow us to examine the role of DAP cross-linking in the architecture of mycobacterial peptidoglycan. Isolation of extragenic suppressors of the β-lactam hypersusceptibility phenotype might allow us to identify enzymes involved with peptidoglycan cross-linking enzymes, specifically those involved with the unusual DAP-DAP linkages first described for mycobacteria (30) and now considered to be potential drug targets (14).