ABSTRACT
Certain Strr mutations that confer low-level streptomycin resistance result in the overproduction of antibiotics by Bacillus subtilis. Using comparative genome-sequencing analysis, we successfully identified this novel mutation in B. subtilis as being located in the mthA gene, which encodes S-adenosylhomocysteine/methylthioadenosine nucleosidase, an enzyme involved in the S-adenosylmethionine (SAM)-recycling pathways. Transformation experiments showed that this mthA mutation was responsible for the acquisition of low-level streptomycin resistance and overproduction of bacilysin. The mthA mutant had an elevated level of intracellular SAM, apparently acquired by arresting SAM-recycling pathways. This increase in the SAM level was directly responsible for bacilysin overproduction, as confirmed by forced expression of the metK gene encoding SAM synthetase. The mthA mutation fully exerted its effect on antibiotic overproduction in the genetic background of rel+ but not the rel mutant, as demonstrated using an mthA relA double mutant. Strikingly, the mthA mutation activated, at the transcription level, even the dormant ability to produce another antibiotic, neotrehalosadiamine, at concentrations of 150 to 200 μg/ml, an antibiotic not produced (<1 μg/ml) by the wild-type strain. These findings establish the significance of SAM in initiating bacterial secondary metabolism. They also suggest a feasible methodology to enhance or activate antibiotic production, by introducing either the rsmG mutation to Streptomyces or the mthA mutation to eubacteria, since many eubacteria have mthA homologues.
INTRODUCTION
Streptomycin (Sm) was first shown to be a particularly potent drug against Mycobacterium tuberculosis in 1944 (1), and mutants resistant to Sm were reported as early as 1946 (2). Because of its clinical importance, molecular mechanisms of resistance to Sm have been extensively studied, especially in M. tuberculosis (3–6). These mutants could be classified into two distinct types, depending on whether they exhibit high- or low-level Sm resistance. Type I mutants carry a mutation within rpsL, which encodes the ribosomal protein S12, or within rrn, which encodes 16S rRNA, and exhibit high-level Sm resistance (MIC, >100 μg/ml). In contrast, type II mutants possess the wild-type rpsL gene and exhibit low-level Sm resistance (MIC, 5 to 10 μg/ml). Most mutations within S12 that confer resistance to, or dependence on, Sm are known to lead to a hyperaccurate phenotype (5), which compensates for the effect of the drug without affecting the interaction between the drug and the ribosome. A number of mutations within 16S rRNA, including those within the 530 loop, also result in both Sm resistance and hyperaccuracy (7–9). However, the mechanisms underlying low-level resistance to Sm (i.e., type II mutations) have remained obscure because they were thought to be less clinically important.
Some mutations causing low-level resistance have been characterized recently. Using the comparative genome-sequencing (CGS) technique, we successfully determined that low-level resistance is caused by mutations in rsmG (rRNA small subunit methyltransferase G), which encodes an S-adenosymethionine (SAM)-dependent 16S rRNA methyltransferase (10). Analysis of the 16S rRNA by high-performance liquid chromatography (HPLC) showed that the rsmG mutant lacked a 7-methylguanosine (m7G) modification (11, 12).
Our laboratory has focused on strain improvement for antibiotic overexpression and has developed a new method to activate or enhance antibiotic production in bacteria. Current methods of antibiotic production, ranging from classical random approaches to metabolic engineering, are either costly or labor-intensive. In contrast, our methods are characterized by simplicity, i.e., introduction of a mutation by isolating spontaneously developed drug-resistant mutants. Thus, the method requires no induced mutagenesis, providing a rational approach to elicit bacterial capabilities within industrial applications (13–16). Importantly, certain rpsL or rsmG mutants display a markedly increased ability to produce antibiotics, indicating that bacterial gene expression can be altered dramatically by modulating ribosomal proteins and/or rRNA (17–19).
The mechanisms underlying this remarkable activation have been studied. The rpsL mutant ribosomes carrying an amino acid substitution in S12, which confers a high level of resistance to Sm, are more stable than those of wild-type controls, indicating that this increase in stability could enhance protein synthesis during the late growth phase (20). We later found that increased expression of the translation factor ribosome-recycling factor also contributes to the enhanced protein synthesis observed during the late growth phase in the rpsL K88E mutant (21). This finding suggested that both the greater stability of the 70S ribosomes and the elevated levels of ribosome-recycling factor caused by the rpsL K88E mutation are responsible for the enhanced protein synthesis seen during the late growth phase and that this underlies antibiotic overproduction by the rpsL K88E mutant (21). In contrast, rsmG mutants display markedly enhanced expression of SAM synthetase in Streptomyces (10). SAM synthetase activity was shown to be important in initiating antibiotic production, as demonstrated by the fact that overexpression of metK (encoding SAM synthetase) stimulates antibiotic production in a Streptomyces coelicolor rsmG mutant and by the finding that exogenous addition of SAM to the culture medium induces antibiotic biosynthesis in wild-type cells (10, 22).
Bacillus subtilis strains are reported to produce three ribosomal antibiotics, TasA, subtilosin, and sublancin; three nonribosomal antibiotics, surfactin, bacilysin, and plipastatin; and the novel phospholipid antibiotic bacilysocin (23, 24). The dipeptide bacilysin is one of simplest peptide antibiotics produced by B. subtilis, consisting of an l-alanine at the N terminus and an unusual amino acid, l-anticapsin, at the C terminus (Fig. 1). Although the biosynthesis of bacilysin has been studied extensively (25, 26), little is known about its regulation. B. subtilis may also be able to synthesize the antibiotic neotrehalosadiamine (NTD) (Fig. 1), an aminosugar produced by Bacillus pumilus and Bacillus circulans (27, 28). Although B. subtilis, unlike B. pumilus and B. circulans, normally does not produce this antibiotic, certain rifampin resistance (rpoB) mutations activate the dormant ability to produce NTD, possibly by more efficient transcription from σA-dependent promoters resulting from alteration of its ternary structure, since the promoter for NTD-biosynthetic genes (ntdABC) was recognized by σA (29). Among prokaryotes, B. subtilis provides a feasible system for the study of various biological functions, as denoted by the presence of a transformation system and the availability of genomic information courtesy of the completed genome project.
Structures of bacilysin and neotrehalosadiamine.
We previously reported that certain B. subtilis mutants possess low-level Sm resistance and exhibit a marked increase in antibiotic production (18). However, none of these strains had a mutation in the rsmG gene, indicating that another mechanism, fundamentally different from that involving rsmG mutation and possibly involving activation of secondary metabolism, is responsible for the low-level Sm resistance of B. subtilis. Thus, based on the feasibility of the CGS technique (30), we attempted to identify this novel type of Sm resistance mutation, as well as to clarify the mechanism regulating antibiotic production as a model of bacterial secondary metabolism. This paper describes the identification and analysis of this novel mutation in relation to antibiotic overproduction and activation of dormant antibiotic-biosynthetic genes.
MATERIALS AND METHODS
Bacterial strains and culture conditions.The bacterial strains and plasmids used in this study are listed in Table 1. All strains were derived from B. subtilis strain 168. Spontaneous Sm-resistant mutants were obtained as colonies that grew within 3 days after cells were spread on solid L medium (10 g tryptone, 5 g yeast extract, and 5 g NaCl per liter) containing Sm at a concentration of 10 to 15 μg/ml. We used a relA1 mutation with a slightly leaky phenotype, accumulating 1/10 of the amount of ppGpp accumulated by the wild type upon nutritional downshift (31), rather than the relA-null mutation, because the latter results in a severe growth defect (32).
Bacterial strains and plasmids used in the study
Strain KJ04 (mthA1) was obtained by transformation with congression using B. subtilis strain YO-005 (hisC101) as a recipient (33). The amino acid auxotrophic marker genes, hisC and trpC, were cotransformed at high frequency (approximately 70%). The histidine auxotrophic recipient strain YO-005 was transformed with the genomic DNA of strain ST20 with selection for histidine prototrophy. Several Sm-resistant transformants were selected from 100 His+ Trp− transformants, with DNA sequencing analysis confirming that all Sm-resistant transformants contained the expected mthA1 mutation. One of these transformants was designated KJ04 and used for further study. To obtain KJ05 (mthA::pMutinT3), the DNA fragment containing a partial coding region of the mthA gene was amplified using the primers mthAH-F and mthAB-R (see Table S1 posted at https://drive.google.com/file/d/0Bwj6L7P2wWCwQXZoc2JsX0k0Umc/edit?usp=sharing), digested with HindIII and BamHI, and cloned into the corresponding sites of pMutinT3 (34). The resulting plasmid, pMutinT3-mthA, was used for transformation, with selection for erythromycin resistance (1 μg/ml).
To overexpress the metK gene, its complete coding region was amplified using the primers metK-F and metK-R and cloned into pCR2.1. The cloned fragment was fully sequenced to confirm its correctness and then digested with EcoRI and subcloned into the corresponding site of pUB18, a derivative of pUB110 with an M13mp18 multiple-cloning site, generating pUB18-metK. The resulting plasmid, pUB18-metK, was used to transform B. subtilis 168 with selection for kanamycin resistance (5 μg/ml).
B. subtilis strains were grown at 37°C with vigorous shaking in L medium or NG medium (10 g nutrient broth, 10 g glucose, 2 g NaCl, 5 mg CuSO4 · 5H2O, 7.5 mg FeSO4 · 7H2O, 3.6 mg MnSO4 · 5H2O, 15 mg CaCl2 · 2H2O, and 9 mg ZnSO4 · 7H2O per liter), a medium developed for antibiotic production by B. subtilis (18).
Detection of the mthA gene mutation.The mthA gene was PCR amplified from B. subtilis genomic DNA using the primers mthA-F and mthA-R (see Table S1 posted at https://drive.google.com/file/d/0Bwj6L7P2wWCwQXZoc2JsX0k0Umc/edit?usp=sharing). The PCR products were directly sequenced using an ABI Prism 310 Genetic Analyzer.
Determination of susceptibility to Sm.Cells were grown in L medium for 18 h and diluted 200-fold to approximately 106 cells/ml. Five-microliter aliquots of cell suspension were spotted onto L agar plates containing various concentrations of Sm and incubated at 37°C for 18 h. Susceptibility to Sm or other antibiotics was expressed as the MIC.
Antibiotic production by B. subtilis strains.B. subtilis strains were grown in NG medium at 37°C with vigorous shaking. Antibiotic production was determined by the paper disk-agar diffusion assay using Staphylococcus aureus 209P as a test organism (35). Briefly, 50-μl aliquots of culture supernatant obtained after centrifugation were applied to 8.0-mm paper disks (Advantec), which were placed onto half-strength Mueller-Hinton agar (Difco) plates inoculated with S. aureus 209P and incubated for 15 h at 37°C. One unit of bacilysin was defined as the amount of bacilysin that produced a halo 1 mm wide around the disk. To assess the production of neotrehalosadiamine, strains were grown in S7N medium (29) for 24 h and assayed as described above, except for the inclusion of 1 mM glucosamine in the Mueller-Hinton assay plate to negate the effect of any bacilysin that might be produced.
Determination of SAM level.The level of intracellular SAM in B. subtilis wild-type (168) and mutant strains was determined by reverse-phase high-pressure liquid chromatography, as described previously (22, 36). Cells grown to various growth phases were harvested, transferred to a petri dish containing 10 ml of 1 M formic acid, and allowed to stand at 4°C for 1 h. The formic acid was collected, filtered through a membrane filter (pore size, 0.45 μm), and lyophilized. The lyophilized samples were dissolved in a small amount of water and analyzed with a Capcell-Pak C18 column (4.6 by 250 mm; Shiseido, Tokyo, Japan). To normalize the number of picomoles of SAM per milliliter of culture to the number of picomoles per number of cells, intracellular SAM levels were expressed as pmol/optical-density (OD) unit, where 1 OD unit was defined as the number of cells that would produce an optical density at 650 nm (OD650) of 1 if suspended in 1 ml.
Determination of polyamines.Polyamines were identified as described previously (37, 38). Aliquots of bacterial cultures were pelleted, washed, and extracted with 0.2 M perchloric acid. The polyamines were subsequently dansylated and extracted with toluene for analysis by thin-layer chromatography (TLC) on silica gel G plates (Merck) developed in ethyl acetate-cyclohexane (2:3 [vol/vol]). The spots were visualized under UV light.
Identification of neotrehalosadiamine.Neotrehalosadiamine produced by mthA mutants was identified by 1H (500 MHz) and 13C (125 MHz) nuclear magnetic resonance (NMR) spectra recorded on a JEOL ECP-500 spectrometer in D2O.
RT-qPCR.Strains were grown in NG medium at 37°C until the OD650 reached 0.5 (exponential growth phase), 1.5 (transition between exponential and stationary phases), or 3.0 (stationary phase). Total RNA was prepared as described previously (39), with contaminating DNA removed by incubation of 2 μg total RNA with 2 U of DNase I (Invitrogen) for 15 min at 25°C. The RNAs were reverse transcribed using a High Capacity RNA-to-cDNA kit (ABI) according to the manufacturer's instructions. After termination of the reaction by incubation for 5 min at 95°C, samples were analyzed using a 7300 real-time quantitative PCR (RT-qPCR) system (ABI) and Thunder Bird SYBER qPCR Mix (Toyobo, Osaka, Japan). Amplification of the 16S rRNA gene was used as an internal control (see Table S1 posted at https://drive.google.com/file/d/0Bwj6L7P2wWCwQXZoc2JsX0k0Umc/edit?usp=sharing).
RESULTS
Development of B. subtilis mutants with low-level Sm resistance.Low-level Sm-resistant mutants of B. subtilis developed spontaneously at a high frequency of 10−5 to 10−6. When selected with an Sm concentration of 10 times the MIC, most of these mutants had mutations in the rsmG gene (12). Few rsmG mutations, however, were detected when selected with an Sm concentration at 3 times the MIC. Importantly, many of the mutants with slight resistance to Sm, as selected at 3 times the MIC, overproduced bacilysin, suggesting an as yet unidentified type of Sm resistance mutation.
Identification of mutations conferring low-level Sm resistance.As we have successfully identified novel mutations using the CGS technique, a method (30) that uses microarray-based DNA sequencing to identify single-nucleotide polymorphisms (SNPs) and insertion-deletion sites within the genome, we utilized the method to identify the mutation in the B. subtilis genome conferring low-level Sm resistance. Genomic DNA was obtained from the mutant strain ST20, with low-level Sm resistance (Table 2) and bacilysin overproduction, and from the parental strain, 168. Using the CGS technique, we identified a putative insertion-deletion site within the gene mthA, which encodes the enzyme S-adenosylhomocysteine (SAH)/methylthioadenosine (MTA) nucleosidase. Direct sequencing showed that the mutation consisted of an 11-bp deletion. Strikingly, mthA mutations were found in many of the colonies that developed spontaneously on the plates containing an Sm concentration of 3 times the MIC (listed in Table 2 as KO-1225 to KO-1233). The frequency of appearance of mthA mutations among these colonies was as high as 2% to 5%. The mthA mutations were characterized by the frequent appearance of deletions that resulted in stop codons just downstream of the mutations (Table 2). All of these mthA mutants overproduced bacilysin (about 3-fold compared to the wild-type strain) (data not shown).
Locations of mutations in the B. subtilis mthA gene and resulting amino acid changes in the MthA protein
To confirm the causal relationship between the identified mutation and Sm resistance or bacilysin overproduction, we constructed the strain KJ04 (trpC mthA) by transformation (Table 1). Similar to the original mutant, ST20, the mthA transformant KJ04 showed increased resistance to Sm (Fig. 2) and overproduction of bacilysin (2.5-fold higher than the wild type) (Fig. 3A and C). Bacilysin has been reported to interfere with glucosamine synthesis (40). As expected, antibacterial potency was negated completely by the addition of glucosamine to the assay plate (Fig. 3C), indicating that the observed antibiotic activity was due to bacilysin. Similarly, disruption of mthA using pMutinT3, resulting in the disruptant KJ05 (Table 1), increased resistance to Sm (Fig. 2). Disruption of the mthA gene in Escherichia coli also caused slight resistance to Sm (MIC, 8 μg/ml compared with 4 μg/ml for the wild type BW25113), as determined using L medium and Mueller-Hinton medium (data not shown). These results indicate that the mthA mutation was responsible for both increased Sm resistance and bacilysin overproduction.
Sm susceptibility of the mthA mutant. The B. subtilis strains 168 (wild type [WT]), KJ04 (mthA1 transformant), and KJ05 (mthA disruptant) were grown to stationary phase in L medium at 37°C. Approximately 2 × 108 cells were diluted with distilled water, and 5-μl aliquots of the cell suspension were spotted onto medium containing 0, 10, or 20 μg/ml of Sm, followed by incubation at 37°C for 12 h.
Growth, antibiotic production, and transcriptional analysis of parental (168) and mthA mutant (KJ04) strains. (A) Strains were grown in NG medium at 37°C. Antibiotic production was determined by the paper disk-agar diffusion method and expressed as the size of the inhibition zone around the disk (diameter, 8 mm). Growth (closed symbols) and antibiotic production (open symbols) of strain 168 (circles) and strain KJ04 (triangles) are shown. (B) Transcriptional analysis of the ywfB gene involved in bacilysin biosynthesis. The strains were grown as for panel A, and the levels of expression of the ywfB gene were determined by real-time qPCR. (C) Strains 168 (wild type), KJ04 (mthA), K2-007 (relA), and KO-1234 (relA mthA) were grown for 15 h as for panel A, and bacilysin production was determined by the paper disk-agar diffusion method. (D) Transcriptional analysis of the relA gene involved in ppGpp synthesis. Strains were grown as for panel A, and levels of expression of the relA gene were determined by real-time qPCR. The error bars indicate the standard deviations of the means of three or more samples.
We previously reported that certain low-level Sm-resistant mutants show an increased ability to produce antibiotic(s) (18). The results described above suggested that those antibiotic-overproducing strains also had mutations in mthA. Strikingly, all strains tested (KO-274 to KO-278 in Table 2) had a mutation in mthA, confirming the association between mthA mutations and antibiotic overproduction.
Characterization of the mthA mutant.The mthA transformant KJ04 grew more slowly (Fig. 3A) and formed smaller and browner colonies with mountain form on L agar plates (Fig. 4A) than the parental strain, 168. When cultured in liquid L medium, KJ04 cells were filamentous, with cross walls (Fig. 4B). The high-level Sm resistance (rpsL) mutations often confer resistance to Sm concentrations greater than 1,000 μg/ml (12, 18). In contrast, the mthA mutants were resistant only to Sm concentrations of 50 μg/ml, less than the resistance of rsmG mutants (100 μg/ml), and displayed cross-resistance to kanamycin and paromomycin, but not to the antibiotics (chloramphenicol, erythromycin, lincomycin, and thiostrepton) that act on the 50S subunit of the ribosome (Table 3). The mthA mutants sporulated as well as the wild-type strain, 168 (sporulation frequency, 90% to 95%), as determined microscopically, when cultured in Schaeffer's sporulation medium for 2 days. The MthA protein, encoded by mthA, catalyzes two steps (SAH→SRH and MTA→MTR) in the SAM-recycling pathway of B. subtilis (Fig. 5). The mthA mutation did not cause severe growth defects, even in the chemically defined medium. This growth defect was not repaired by supplementation with methionine, cysteine, or homocysteine (see Fig. S1 in the supplemental material), indicating that the observed growth defect was not due to limitations in these amino acids. Polyamines play an important role in microorganisms, including B. subtilis, especially in macromolecular syntheses and in the modulation of translation accuracy (41, 42). Since SAM-recycling pathways are closely involved in polyamine synthesis (Fig. 5), we measured intracellular concentrations of spermidine, spermine, and putrescine and found that none differed significantly in the mutant and parental strains (see Fig. S2 in the supplemental material). Moreover, the addition of spermidine, putrescine, methionine, or cysteine did not affect the resistance of these strains to Sm (data not shown), indicating that polyamines are not relevant to the changes caused by mthA mutations.
Morphological appearance of parental (168) and mthA mutant (KJ04) strains. (A) Colony morphology observed after 5 days of culture on L agar plates. (B) Microscopic observation (magnification, ×1,000) of cells grown to mid-growth phase (5 h) in L medium.
Antibiotic susceptibilities of B. subtilis wild-type and mutant strains
Outline of SAM-recycling pathways in B. subtilis. The genes and enzymes involved in SAM recycling are as follows: metI, cystathionine γ-synthase; metC and patB, cystathionine β-lyase; metE, methionine synthase; metK, SAM synthetase; speD, SAM decarboxylase; speE, spermidine synthase; mthA, SAH/MTA nucleosidase (mthA is synonymous with mtnA); mtnK, methylthioribose kinase; the mtnA and mtnWXBD gene products, involved in the MTR-to-KMBA recycling pathway; mtnE, aminotransferase; luxS, S-ribosylhomocysteine hydrolase; yrhA (mccA), cystathionine β-synthase; yrhB (mccB), cystathionine γ-lyase and homocysteine γ-lyase; SAM (S-adenosylmethionine); SAH (S-aenosylhomocysteine); KMBA (α-keto-γ-methyl-thiobutyric acid); MTA (methylthioadenosine); MTR (methylthioribose); and SRH (S-ribosylhomocysteine). AI-2 indicates autoinducer-2. The figure was drawn on the basis of work by Hullo et al. (53).
Unlike antibiotic production, mthA mutations did not alter the expression of enzymes, such as α-amylase and proteases, as determined by RT-qPCR analysis of the amyE gene, which encodes α-amylase, and the bpr, epr, mpr, nprB, nprE, and wprA genes, which encode proteases (data not shown). Although high-level Sm-resistant (rpsL) mutants were found to emerge at 200-fold-higher frequency from rsmG mutants than from the wild type (12), similar results were not observed in mthA mutants (data not shown).
Transcriptional analysis of the bacilysin-biosynthetic gene.The biosynthesis of bacilysin is controlled by a polycistronic operon (ywfBCDEFG) and a monocistronic operon (ywfH) (23), both considered structural genes for bacilysin biosynthesis. We therefore analyzed the transcription of the ywfB gene, which may encode the enzyme that synthesizes the anticapsin moiety (23), by real-time quantitative PCR. Although expression of ywfB was detected during the late growth phase (OD650 = 5) in wild-type strain 168, high expression was detected during the early growth phase (OD650 = 0.5) in the mthA mutants, with mutant expression during the late growth phase being much higher than in the wild type (Fig. 3B). Similar results for expression of ywfB were obtained when amplification of the sigA gene was used, instead of the 16S rRNA gene, as an internal control (data not shown), accounting for the increased production of bacilysin by the mthA mutant (Fig. 3A and C). The stringent response, one of the most important adaptation systems in bacteria (43), is closely involved in initiating secondary metabolism (13, 44). This response depends on the transient increase in hyperphosphorylated guanosine nucleotides ([p]ppGpp), which are synthesized from GDP or GTP by the relA gene product (ppGpp synthetase) in response to nutrient limitation (43). Transcription of both ywfBCDEFG and ywfH is enhanced directly or indirectly under conditions that elicit the stringent response (23). Introduction of a relA mutation into the wild-type 168 or mthA mutant strain severely inhibited bacilysin production (Fig. 3C), indicating that the mthA mutation can fully exert its effect on antibiotic overproduction in the genetic background of rel+. Therefore, it was possible that the mthA mutation, like a certain thiostrepton resistance (tsp) mutation in Streptomyces (19), increased relA expression, leading to enhanced ppGpp synthesis and eventual bacilysin overproduction. We found, however, that the transcription of relA was somewhat decreased in the mthA mutant (Fig. 3D).
Increase in SAM pool size is responsible for bacilysin overproduction.In Streptomyces spp., SAM is involved in triggering the onset of secondary metabolism; overexpression of metK, which encodes SAM synthetase, or addition of SAM into the medium causes antibiotic overproduction (22, 45). SAM metabolism was partially blocked by the mthA-null mutation in B. subtilis, because MthA catalyzes two steps (SAH→SRH and MTA→MTR) in the SAM-recycling pathway (Fig. 5). Consequently, the mthA mutation may increase the SAM pool size, eventually activating secondary metabolism. To address this hypothesis, we measured SAM pool size in wild-type and mutant cells grown to various growth phases. As expected, SAM pool sizes were 2-fold larger in the mthA mutant than in the wild type (Fig. 6A). The expression of metK was not increased in the mthA mutant but rather somewhat decreased (Fig. 6B), suggesting that the mutation increases SAM pool size by blocking SAM recycling.
Intracellular SAM level and transcriptional analysis of the metK gene in wild-type (168) and mthA mutant (KJ04) strains. (A) The strains were grown in NG medium at 37°C to an OD650 of 0.5, 1, or 1.5. Intracellular SAM levels were determined by reverse-phase high-performance liquid chromatography. (B) Transcriptional analysis of the metK gene. Strains were grown in NG medium as for panel A. Total RNAs were extracted from the cells and used for real-time qPCR analysis. The transcription level of metK was normalized relative to the amount of 16S rRNA in each RNA sample. The error bars indicate the standard deviations of the means of three or more samples.
To confirm that SAM initiates antibiotic production in B. subtilis, we attempted the forced expression of the metK gene in wild-type cells by utilizing the multicopy vector pUB18, in which the metK gene was integrated. As expected, the SAM pool size was 1.5 to 1.8 times greater in the resulting transformant, KJ06, harboring pUB18-metK, than in the parental strain throughout all growth phases (Fig. 7). KJ06 also showed enhanced production of bacilysin (10 units), similar to the mthA mutant KJ04, but did not show elevated resistance to Sm (data not shown). These results indicate that the increased SAM pool size is responsible for the increased production of bacilysin, but not for low-level resistance to Sm. Since the rsmG-null mutation confers low-level resistance to Sm (12), we hypothesized that the mthA mutation may severely decrease the level of expression of rsmG, making the mthA mutant slightly resistant to Sm. This possibility was ruled out, however, as the level of expression of rsmG was as high in KJ04 as in the parental strain (see Fig. S3 in the supplemental material), although these results do not completely exclude the possibility that mthA mutation inhibited RsmG activity at the posttranslational level.
Effect of metK overexpression on the intracellular SAM level. Strains were grown in NG medium at 37°C to an OD650 of 0.5, 1, or 2. Intracellular SAM levels were determined by reverse-phase high-performance liquid chromatography. The error bars indicate the standard deviations of the means of three or more samples.
mthA mutation activates even dormant genes.B. subutilis 168 can produce the antibiotic NTD, although the genes responsible for its biosynthesis are dormant (i.e., silent) under ordinary culture conditions (29). Although the physiological condition(s) that permits the expression of these genes has not been determined, the S487→L mutation in the rpoB gene, which encodes the β-subunit of RNA polymerase, can activate the NTD-biosynthetic genes, resulting in the production of 500 to 1,000 μg/ml NTD (29). Wild-type B. subtilis produces less than detectable levels (<1 μg/ml). We therefore assessed whether the mthA mutation accompanied by an increase in the SAM level can activate these dormant genes. The wild-type (168) and mthA mutant (KJ04) strains were grown for 24 h in S7N medium, developed for NTD production (29), and NTD production was assayed by the disk-agar diffusion method using Mueller-Hinton assay plates containing 1 mM glucosamine to negate the effects of bacilysin. Strikingly, the mthA mutant produced 150 to 200 μg/ml NTD, whereas none was produced by the wild type (Fig. 8A). Isolation of the antibacterial compound, followed by 1H NMR and 13C NMR analyses, confirmed that it was NTD (see Fig. S4A and B in the supplemental material) (46).
NTD production and transcriptional analysis in parental (168) and mthA mutant (KJ04) strains. (A) Strains were grown in S7N medium at 37°C for 24 h, and NTD production was determined by the paper disk-agar diffusion method. (B) Transcriptional analysis of the ntdA gene. The strains were grown as for panel A for 12, 16, or 20 h, and the level of expression of the ntdA gene was determined by real-time qPCR. The error bars indicate the standard deviations of the means of three or more samples.
NTD biosynthesis is controlled by the polycistronic gene ntdABC for NTD biosynthesis and a monocistronic gene, a positive regulator of ntdABC (29). Transcriptional analysis of the ntdA gene showed that the mthA mutation resulted in the expression of ntdA mRNA during the late growth phase (16 to 20 h) (Fig. 8B), accounting for the burst of NTD production. Thus, modulation of the SAM pool size not only enhanced the production of antibiotics, but activated the expression of dormant genes.
DISCUSSION
We previously reported that antibiotic production by B. subtilis is markedly activated by introducing certain low-level Sm resistance mutations, which differ from rsmG and rpsL mutations (12, 18). In the present study, we used the CGS technique to successfully identify this third mutation, which confers low-level Sm resistance, at least in B. subtilis and E. coli. Our principal findings were as follows: (i) the third mutation conferring low-level Sm resistance was located in the mthA gene, which encodes SAH/MTA nucleosidase; (ii) the mthA mutation increased intracellular SAM by arresting the SAM-recycling pathways; (iii) the increased SAM level directly or indirectly enhanced antibiotic production and activated even dormant genes involved in secondary-metabolite biosynthesis; and (iv) the activation of antibiotic production was apparently achieved at the transcriptional level, not as a methyl donor for antibiotic biosynthesis, by markedly enhancing the expression of biosynthetic genes.
The high frequency of emergence of spontaneous mthA mutants is likely due to the dispensability of the gene, allowing cells to remain viable. Our finding that SAM plays an essential role in triggering the onset of secondary metabolism is not unprecedented. For example, rsmG mutants in S. coelicolor exhibit enhanced expression of the metK gene encoding SAM synthetase, accompanied by increased protein synthesis during the late growth phase, eventually leading to the overproduction of antibiotics (10). Thus, increases in SAM synthetase and protein synthesis activity resulting from the rsmG mutation likely activated secondary metabolism in S. coelicolor (10). In B. subtilis, rsmG mutations did not result in increased antibiotic production, consistent with findings that these mutants did not show increased protein synthesis during the late growth phase or increased SAM synthetase activity (12). Thus, in contrast to Streptomyces, secondary metabolism in B. subtilis is not activated by rsmG mutations. Rather, mutations in mthA, not rsmG, increase intracellular SAM, but by a completely different mechanism. In S. coelicolor, rsmG mutations enhance metK expression, whereas in B. subtilis, mthA mutations block SAM-recycling pathways, with both leading to increased SAM levels (Fig. 9). The causal relationship between antibiotic production and the SAM level in B. subtilis was confirmed by forced expression of metK in the wild-type strain. As reviewed by Parveen and Cornell (42), the importance of SAH/MTA nucleosidase encoded by the mthA gene is currently increasing because a comprehensive analysis of its various roles demonstrates that it is an integral component of the activated methyl cycle, which recycles adenine and methionine through SAM-mediated methylation reactions. Inhibition of this enzyme by certain substrate analogues limits synthesis of autoinducers and hence causes reduction in biofilm formation and may attenuate virulence in certain bacteria.
Schematic showing the signal transduction pathways in B. subtilis and Streptomyces spp., from the mthA or rsmG mutation to the enhancement of antibiotic production or the activation of silent genes. The scheme was based on the work presented here and previous studies in B. subtilis (12, 18, 23, 29, 31) and Streptomyces spp. (13, 44, 48).
Mutations in rsmG have been shown to cause low-level Sm resistance in E. coli, M. tuberculosis, S. coelicolor, and B. subtilis (10–12, 47). Loss of the m7G modification in 16S rRNA results in resistance to Sm, providing a molecular basis for rsmG mutation-induced Sm resistance. In contrast, the mechanism by which mthA mutations cause low-level Sm resistance is presently unclear, especially since an increased SAM level did not cause Sm resistance. As the mthA mutation-mediated Sm resistance in B. subtilis (and in E. coli) was slight, the resistance was likely due to a secondary effect of the mutation. Figure 9 shows a comparison of B. subtilis and Streptomyces signal transduction systems, starting from the mthA and rsmG mutations, respectively, to the ability to enhance antibiotic production or activate silent genes (10, 12, 13, 23, 29, 44, 48). Interestingly, Streptomyces spp. and Zymomonas spp. have no mthA homologue, whereas many bacteria (e.g., E. coli, S. aureus, Salmonella enterica, Thermus thermophilus, Helicobacter pylori, and Listeria monocytogenes) do. Initiation of secondary metabolism in Streptomyces is characterized by involvement of autoinducers, such as A factor (48).
SAM is synthesized from methionine and ATP by an SAM synthetase encoded by the metK gene. SAM is highly reactive and plays a central role in many cellular functions (42, 49, 50). In contrast to Streptomyces cells (22), extracellular SAM cannot be incorporated by E. coli or B. subtilis cells (51, 52). The major steps for SAM recycling in B. subtilis have been recently characterized by Hullo et al. (53). The intracellular SAM level should be tightly regulated in native B. subtilis cells, since a 2-fold increase markedly altered cellular physiology and morphology, as shown in the present study. The filamentous form of the mthA mutant accompanied by elevated SAM (Fig. 4) was striking, due to recent findings that E. coli cells starved for SAM are very long, suggesting that SAM and methylation are important for cross wall formation (54). The results described here, together with previous findings in rsmG mutants, establish the significance of SAM in initiating antibiotic production in bacteria, indicating an intrinsic role for SAM in microbial secondary metabolism (55). In addition, introduction of a multicopy plasmid containing the Streptomyces spectabilis metK gene into Streptomyces lividans was found to induce antibiotic production (45), and the addition of SAM to the culture medium increased antibiotic production by S. coelicolor, Streptomyces griseus, and Streptomyces griseoflavus (22, 56). SAM is the methyl donor for the methylation of cytosine and adenosine bases in DNA, rRNA, and tRNA; of various proteins; and of small molecules important for both lower and higher organisms (42, 57, 58). Thus, SAM synthetase is likely essential for the viability of E. coli (59) and B. subtilis (60), with overexpression in these bacteria causing methionine auxotrophy, perhaps by depleting the intracellular methionine pool, either by consuming it to synthesize SAM or by repressing the methionine-biosynthetic genes. In E. coli, SAM was shown to be a corepressor of the methionine regulation system (49). Although there is no evidence to date for the existence of a methyltransferase involved in the regulation of antibiotic production, SAM-dependent protein methylation may play a role in controlling the activity of the regulatory proteins encoded by these developmental genes. It may be possible that the increased SAM pool size stimulated SAM-dependent methylation of RelA (or YjbM/YwaC), resulting in enhancement of ppGpp synthesis, followed by acceleration of antibiotic production. Alternatively, DNA or RNA methylation may be involved in the expression of these regulatory genes. For example, the extent of methylation of a particular macromolecule (e.g., a particular DNA region) could determine the probability that a second macromolecule (e.g., a repressor or activator) can bind to the first, significantly altering gene expression, similar to findings in E. coli (61). In addition to being a methyl donor, SAM may be directly involved in regulating antibiotic synthesis as a corepressor or inducer, as shown for methionine regulon expression in E. coli (49). It is also noteworthy that SAM is a precursor for certain autoinducers, which control many different processes, including antibiotic production, biofilm formation, and virulence (62–64). Autoinducer 2 (AI-2), a furanosyl borate diester, is produced from SAM in B. subtilis (Fig. 5) (53). Although mthA mutation likely did not enhance AI-2 production in this case, it is possible that the elevated level of SAM caused enhanced production of another, yet unknown autoinducer (if any), eventually accelerating the secondary metabolism.
ppGpp is crucial in triggering the onset of antibiotic production in Streptomyces spp., whereas morphological differentiation is triggered by a reduction in GTP (13, 44, 65). In B. subtilis, both GTP and ppGpp are key factors initiating antibiotic production (23). Bacilysin production in B. subtilis is apparently controlled by CodY, a GTP-binding protein, because codY disruption increased the level of expression of genes involved in bacilysin biosynthesis (23). However, the effect of GTP can be elicited only in rel+ wild-type cells, not in relA mutant-type cells. Neither decoyinine treatment nor codY disruption activated transcription from the ywfBCDEFG operon responsible for the bacilysin biosynthesis in relA mutant cells (23). Thus, ppGpp plays a pivotal role as a positive regulator in antibiotic production by B. subtilis, in accordance with the results from the present work showing that the mthA mutation fully exerted its effect on antibiotic overproduction in rel+ but not relA mutant cells (Fig. 3C). Since ppGpp is essential for transcription of the ywfBCDEFG and ywfH genes via the CodY-mediated regulation system, bacilysin production in B. subtilis is controlled by a dual regulation system composed of ppGpp and GTP (23), unlike antibiotic production in Streptomyces spp. (Fig. 9). Recent advances with respect to ppGpp-RNA polymerase interrelationship have clarified that, unlike E. coli, where ppGpp decreases rRNA promoter activity by directly inhibiting RNA polymerase, in B. subtilis and T. thermophilus (species distantly related to E. coli), ppGpp reduces the available GTP pools, thereby modulating rRNA promoter activity indirectly (66–69). In addition, direct regulation of GTP homeostasis by ppGpp was analyzed in detail in B. subtilis, demonstrating that two GTP biosynthesis enzymes (Gmk and HprT) are major posttranscriptional targets of ppGpp whose activities are strongly inhibited by ppGpp in vitro, while inhibition of IMP dehydrogenase (GuaB) activity by ppGpp is likely at least a minor contributor in B. subtilis (70, 71) and perhaps in T. thermophilus (67). This situation could be prominent, especially in the absence of starvation (with lower ppGpp levels) rather than under starvation conditions (with higher ppGpp levels). Recent work by Belitsky and Sonenshein (72) clarified the CodY-binding motifs using a motif-searching algorithm. CodY regulates transcription in several ways, including (i) negative or positive regulation by binding within or near a promoter site, (ii) negative regulation by interfering with the binding of a positive regulator, and (iii) negative regulation by acting as a roadblock to RNA polymerase. It is unclear at present what the direct critical target(s) of ppGpp is in initiating basilysin biosynthesis in B. subtilis. The current achievements described above may be helpful in clarifying the mechanisms underlying regulation of the secondary metabolism in Bacillus and related bacteria at the molecular level.
Sequencing of the genomes of Streptomyces, fungi, and myxobacteria has shown that, although each strain contains genes encoding the enzymes necessary to synthesize a plethora of potential secondary metabolites, only a fraction are expressed during fermentation (73, 74). Methods of activating dormant antibiotic-biosynthetic gene clusters are therefore of interest in both basic and industrial microbiology (16, 75–77). Since the mthA mutation activated the expression of dormant genes involved in NTD production, alterations in the intracellular SAM level, by introducing an rsmG (in Streptomyces) or an mthA (in eubacteria) mutation, could be a feasible way to activate the dormant genes, as they require neither induced mutagenesis nor gene-engineering technique. The forced expression of the metK gene should also be effective, though labor-intensive. Thus, our method provides a powerful tool for screening novel compounds and for strain improvement to overproduce useful compounds. Clinically, mthA mutations appear to be microbiologically insignificant because, unlike rsmG mutations, these mutations, despite their high frequency, did not trigger the emergence of mutants with high-level Sm resistance.
ACKNOWLEDGMENTS
This work was supported by grants to K.O. from the Ministry of Education, Culture, Sports, and Technology of the Japanese Government (Effective Promotion of Joint Research of Special Coordination Funds) and from the National Agriculture and Food Research Organization (Program for Promotion of Basic and Applied Research for Innovations in Bio-Oriented Industry).
We are grateful to Yusuke Motoi and Yasuko Tanaka for valuable technical assistance throughout the work. We also acknowledge Roche NimbleGen, Inc. (Madison, WI), for supporting the mutation search using the comparative genome-sequencing technique.
S.T. and J.-Y.K. mainly worked with transcriptional analysis and physiological analysis, respectively. Y.H. conducted structural analysis of NTD, and K.O. designed the research work and wrote the article.
FOOTNOTES
- Received 12 December 2013.
- Accepted 23 January 2014.
- Accepted manuscript posted online 7 February 2014.
- Address correspondence to Kozo Ochi, k.ochi.bz{at}it-hiroshima.ac.jp.
↵* Present address: Ji-Yun Kim, Viral Infectious Diseases Unit, RIKEN, Wako, Saitama, Japan.
S.T. and J.-Y.K. contributed equally to this research.
Supplemental material for this article may be found at http://dx.doi.org/10.1128/JB.01441-13.
REFERENCES
- Copyright © 2014, American Society for Microbiology. All Rights Reserved.
