The Global Regulator Spx Functions in the Control of Organosulfur Metabolism in Bacillus subtilis

Abstract

Spx is a global transcriptional regulator of the oxidative stress response in Bacillus subtilis. Its target is RNA polymerase, where it contacts the α subunit C-terminal domain. Recently, evidence was presented that Spx participates in sulfate-dependent control of organosulfur utilization operons, including the ytmI, yxeI, ssu, and yrrT operons. The yrrT operon includes the genes that function in cysteine synthesis from S-adenosylmethionine through intermediates S-adenosylhomocysteine, ribosylhomocysteine, homocysteine, and cystathionine. These operons are also negatively controlled by CymR, the repressor of cysteine biosynthesis operons. All of the operons are repressed in media containing cysteine or sulfate but are derepressed in medium containing the alternative sulfur source, methionine. Spx was found to negatively control the expression of these operons in sulfate medium, in part, by stimulating the expression of the cymR gene. In addition, microarray analysis, monitoring of yrrT-lacZ fusion expression, and in vitro transcription studies indicate that Spx directly activates yrrT operon expression during growth in medium containing methionine as sole sulfur source. These experiments have uncovered additional roles for Spx in the control of gene expression during unperturbed, steady-state growth.

The global regulator Spx of Bacillus subtilis functions in the oxidative stress response by activating transcription of genes that function in thiol homeostasis (33, 34, 49). In this capacity, it is required for the transcriptional activation of the thioredoxin (trxA) and thioredoxin reductase (trxB) genes in response to disulfide stress. It also represses genes that function in a variety of metabolic and developmental pathways (34, 35). The activity and synthesis of Spx are stimulated when cells undergo accelerated disulfide generation resulting from encounters with toxic oxidants. The accumulated Spx interacts with RNA polymerase (RNAP) in part by contacting residues of the alpha subunit C-terminal domain (CTD), while showing no sequence-specific DNA-binding activity (36). An important feature of Spx is the N-terminal CxxC redox disulfide motif that controls Spx activity. Transcriptional activation from the trxA and trxB promoters requires that the two cysteines be in the oxidized, disulfide state. The presence of a reductant that converts the cysteines to the thiol state results in loss of transcription-stimulating activity (33). It is currently not known how the interaction of oxidized Spx protein with RNA polymerase activates transcription at promoters under Spx positive control.

In addition to its role in oxidative stress, recent studies indicate that Spx functions as a regulator of gene expression in cells undergoing steady-state growth. Spx was found to negatively affect the transcription of genes that function in the utilization of organosulfur compounds as alternative sources of sulfur (13). Such operons are repressed in minimal glucose media containing either of the two preferred sulfur sources, sulfate and cysteine. These operons are derepressed when cysteine or sulfate are replaced with methionine or an organosulfonate as sole sulfur source (8, 44, 46). The ytmI, yxeI, and ssu operons are derepressed in sulfate medium when cells bear an spx null mutation (13).

Recent studies have clarified the control of sulfur metabolism in B. subtilis, and these findings likely apply to sulfur control systems in other low-GC content gram-positive bacteria. First, cysteine, the product of sulfur assimilation, can be produced from the aforementioned organosulfur compounds, which undergo oxygenolytic cleavage to yield sulfite (10, 24) (Fig. 1), an intermediate in cysteine sulfur assimilation. Second, sulfite for cysteine production can also be obtained through uptake and reduction of inorganic sulfate. Third, cysteine can be made from methionine via the S-adenosyl homocysteine (SAH) shunt and reverse transsulfuration of homocysteine (18, 40) (Fig. 1). This is believed to involve the activities of the YrrT protein, an S-adenosylmethionine methyl (SAM) transferase; the Mtn protein, an S-adenosyl homocysteine nucleosidase; LuxS, the ribosyl homocysteinase; YrhA, a putative cystathionine β-synthase; and YrhB, a cystathionine γ-lyase. All of these proteins, except LuxS, are encoded by the yrrT operon (14, 40). Like the ytmI, yxeI, and ssu operons, the yrrT operon is repressed by sulfate and cysteine but is derepressed in methionine medium. Another route to cysteine synthesis from methionine involving the enzyme methionine-γ-lyase might exist in B. subtilis (18). This would generate methanethiol, which is a substrate for oxygenolytic cleavage, possibly catalyzed by the enzymes encoded by the ssu, ytmI, or yxeI operon, and would yield sulfite. However, no candidate gene encoding the methionine-γ-lyase has been identified.

FIG. 1.

Diagram of the organosulfur metabolic pathways of B. subtilis (18, 40). Cysteine can be produced from sulfate or from methionine/SAM. Production of sulfite is from sulfate and organosulfonates, and the genes believed to function in the process are shown. The SAH shunt-transsulfuration pathways (indicated) function in converting methionine to cysteine via SAM and SAH. The genes whose products function at each step are indicated, as are the effects of Spx activity; an arrowhead pointing up denotes Spx-dependent stimulation of expression, and an arrowhead pointing down indicates Spx-dependent repression (34). α-kb, α-ketobutyrate; O-ahs, O-acetylhomoserine; pyr, pyruvate; AI-2, autoinducer-2 (48).

The repressor CymR (formerly YrzC) has been shown to function in the negative control of operons whose products participate in the biosynthesis of cysteine from a variety of sources (14) (Fig. 1). Both CymR and Spx exert negative control of the ytmI, yxeI, and ssu operons in the presence of sulfate, while only CymR is required for negative control in cysteine medium (13, 14) (see below).

As previously observed in microarray transcriptome analysis, the Spx-RNAP interaction increased yrrT operon and cysK transcript levels (34) in cultures grown with sulfate as sole sulfur source. It also repressed genes that function in methionine biosynthesis (34) (Fig. 1). Thus, when Spx is overproduced, sulfur metabolism appears to undergo a shift that favors cysteine biosynthesis over methionine biosynthesis. An increase in yrrT operon expression is characteristic of cells undergoing oxidative stress, as this was also observed in cells having depleted levels of the antioxidant thioredoxin (41) and cells treated with the thiol-specific oxidant diamide (26).

In this report, we show that Spx positively controls the cymR operon, which accounts, in part, for the role of Spx in sulfate-dependent negative control of the ytmI and yrrT operons. However, Spx has another role in organosulfur metabolism under steady-state growth. This report provides evidence that Spx controls the production of cysteine from methionine by regulating the transcription of the yrrT operon.

MATERIALS AND METHODS

Bacterial strains and culture conditions.

The B. subtilis strains used in this work are listed in Table 1 and are all derivatives of strain JH642. Escherichia coli DH5α was grown at 37°C in 2× YT (yeast extract, tryptone) liquid or on LB agar medium. B. subtilis cells were grown at 37°C in 2× YT, Difco sporulation medium (20), or TSS minimal medium (15) with some modification (37.4 mM NH₄Cl, 2 mM K₂HPO₄, 49.5 mM Tris pH 7.5, 5 mM MgCl₂, 0.5% glucose, 0.004% FeCl₃, 0.004% citric acid, 0.1% l-glutamate, 0.005% amino acids for auxotrophic requirements) and supplemented with one of the following sulfur sources at a concentration of 1 mM: l-cysteine, l-methionine, or MgSO₄. Isopropyl-β-d-thiogalactopyranoside (IPTG) was added to culture media at a concentration of 1 mM where appropriate. When necessary, the following antibiotics were added (at the concentration shown in parentheses): ampicillin (25 μg/ml), neomycin (5 μg/ml), chloramphenicol (5 μg/ml), erythromycin plus lincomycin (1 and 25 μg/ml, respectively), and spectinomycin (75 μg/ml).

TABLE 1.

Bacillus subtilis strains

Strain	Genotype	Source or reference
MH5636	trpC2 pheA1 rpoCHis10	5
ORB3834	trpC2 pheA1 spx::aphA3	32
ORB4028	trpC2 pheA1 rpoCHis10 spx::aphA3	This study
ORB4794	trpC2 pheA1 ytmI′-lacZ erm ΔytmI	13
ORB4802	trpC2 pheA1 ytmI′-lacZ ΔytmI spx::aphA3	13
ORB4871	trpC2 pheA1 amyE::ytlI′-lacZ spx::aphA3	13
ORB5691	trpC2 pheA1 amyE::ytlI′-lacZ	This study
ORB5724	trpC2 pheA1 cymR::spc	This study
ORB5730	trpC2 pheA1 ytmI′-lacZ erm ΔytmI cymR::spc	This study
ORB5732	trpC2 pheA1 ytmI′-lacZ erm ΔytmI ytlI::aphA3 cymR::spc	This study
ORB5733	trpC2 pheA1 amyE::ytlI′-lacZ cymR::spc	This study
ORB5809	trpC2 pheA1 amyE::ytlI′-lacZ cymR::spc thrC::cymR (pSY11)	This study
ORB6080	trpC2 pheA1 SPβc2 del2::Tn917::cymRTL-lacZ	This study
ORB6081	trpC2 pheA1 SPβc2 del2::Tn917::cymRTL-lacZ (pSY25) spx::aphA3	This study
ORB6082	trpC2 pheA1 amyE::ytlI′-lacZ thrC::Pspank-hy-cymR (pSY27) spx::aphA3	This study
HNU25	trpC2 pheA1 yrrT′-lacZ erm ΔyrrT	This study
HNU26	trpC2 pheA1 yrrT′-lacZ erm ΔyrrT cymR::spc	This study
HNU27	trpC2 pheA1 yrrT′-lacZ erm ΔyrrT spx::aphA3	This study

Plasmids and construction of Bacillus strains.

Plasmids constructed in this study are listed in Table S1 of the supplemental material. Primers used in this study are listed in Table S2 of the supplemental material.

B. subtilis strain ORB5724 (cymR::spc) was constructed as follows. The B. subtilis JH642 chromosomal DNA region covering the complete coding sequence of cymR (nucleotides −303 to +617 relative to the translational start site) was amplified by PCR using oligonucleotides oSY2 and oSY3 with the creation of EcoRI and AseI sites. The amplified DNA fragment was inserted into the EcoRI and NdeI sites of pUC18, producing pSY5. A spectinomycin resistance cassette (Sp^r) was isolated from pDG1727 (7) by digestion with StuI and EcoRV, inserted into pSY5 digested with NdeI, and then filled in using T4 DNA polymerase and deoxynucleotide triphosphates. The ligated recombinant pSY7, which bears the spc cassette in the same orientation with the cymR transcription unit, was introduced into JH642. Growth of the resultant mutant was normal, indicating that there was no adverse effect on the expression of the downstream genes of the cymR operon, one of which is the essential trmU gene (1). Plasmid pSY11 is a derivative of pDG795 that contains the cymR gene and promoter region. The cymR fragment was synthesized by PCR using primers oSY4 and oSY5 (see Table S2 in the supplemental material) and then treated with BamHI and EcoR1. The fragment was then inserted into BamHI-EcoRI-cleaved pDG795 (43).

B. subtilis strain ORB6082 carries the cymR coding sequence under the control of the IPTG-inducible P_spank-hy promoter (6). The P_spank-hy construct was made as follows. Plasmid pSY13 was constructed by inserting a 450-bp PCR fragment (created using primers oSY07 and oSY05) containing the cymR coding sequence and ribosome-binding site into SphI/SalI-cleaved pDR66. Plasmid pSY18 was constructed by obtaining the SphI-SalI fragment bearing the cymR coding sequence and ribosome-binding site from pSY13 and inserting it into SphI/SalI-cleaved pDR111, thus placing the cymR gene under P_spank-hy control. Plasmid pSY20 was created by exchanging the EcoRI-BamHI fragment of pDR66 (containing the Pspac promoter and associated lacI gene) with the homologous EcoRI-BamHI fragment of pSY18, thus placing the P_spank-hy-cymR construct with a linked chloramphenicol resistance (cat) marker. pSY24 was created by inserting a PCR fragment (obtained by using primers oSY12 and oSY05) containing the cymR gene and upstream DNA to the transcription start site into SphI/SalI-cleaved pSY20, thus exchanging the shorter cymR fragment with the complete cymR transcription unit. Plasmid pSY27 was created by obtaining the EcoR1-BamH1 fragment of pSY24, containing the P_spank-hy-cymR construct with associated lacI gene, and inserting it into EcoRI-BamH1-cleaved pDG795, thus creating a P_spank-hy-cymR plasmid that could integrate into the thrC locus. The sequence of cymR was confirmed by sequencing the plasmid with both Pspac-up and Pspac-down primers. Strain ORB4871 was transformed with plasmid pSY27 with selection for erythromycin resistance (Erm^r) and screening for Thr⁻, which created strain ORB6082. ORB5809 was constructed by transforming ORB5733 with plasmid pSY11 with selection for Erm^r and screening for Thr⁻.

A Bacillus subtilis strain carrying a cymR-lacZ translational fusion was constructed as follows. The DNA region containing the 5′-regulatory and the first 16 codons of cymR was amplified by PCR from the chromosome of strain JH642 using primers oSY13 and oSY14. The amplified 350-bp fragment corresponding to position −303 to +49 relative to the cymR translational start site was digested with SalI and ligated with plasmid pTKlac (22) digested with the same enzyme, which placed the cymR coding sequence in frame with the lacZ coding sequence. The resulting plasmid, pSY25, was then introduced into B. subtilis ZB307A by homologous recombination at the SPβ prophage locus (50), and the phage lysate prepared from the transformants was used to transduce ZB278 (50). Strains ORB6080 and ORB6081, carrying the cymR-lacZ translational fusion, were constructed by transducing JH642 and ORB3834 (spx::aphA3), respectively, with the lysate prepared from the ZB278 transductants.

The yrrT′-lacZ fusion strain HNU25 was constructed as follows. The 330-bp DNA containing 5′ promoter and coding sequence of yrrT was amplified by PCR using chromosomal DNA of strain 168 and the primer pair forward oHSY17 and reverse oHSY18. The PCR product was doubly digested with HindIII and BamHI and then ligated to the HindIII- and BamHI-cleaved pMUTIN2 plasmid. The ligated DNA was used for transformation of E. coli DH5α. After the sequence of the cloned DNA was confirmed by DNA sequencing, the resulting plasmid, pSY42, was introduced into the chromosome of B. subtilis JH642 by a single crossover event for the integrational disruption of yrrT. The disruption was confirmed by resistance to erythromycin and PCR analysis of the chromosomal DNA of the erythromycin-resistant recombinant. Bacillus subtilis strain ORB4028 (rpoC-_His10 spx::aphA3) was constructed by transforming strain MH5636 (38) with chromosomal DNA derived from ORB3834 (spx::aphA3).

β-Galactosidase assays.

Cells from a frozen stock or an isolated colony were used to inoculate 2 ml of 2× YT plus antibiotics. The culture was allowed to grow for several hours, after which it was diluted 100-fold into 1 ml of TSS medium supplemented with a 1 mM concentration of different sulfur sources. When required, 1 mM IPTG was added at the initial optical density at 600 nm (OD₆₀₀) of 0.05. Growth of cultures and assay of β-galactosidase activity were carried out as previously described (13). The two samples from a single time point were averaged to provide the β-galactosidase activity in Miller units (31). Experiments were repeated three times, and the data are presented as the mean of three independent experiments ± 1 standard deviation.

Primer extension analysis.

ORB4794 (spx [wild type]) and ORB4802 (spx::aphA3) were grown at 37°C in TSS medium containing either MgSO₄, cysteine, or methionine to an OD₆₀₀ of 0.7. The purification of the total RNA and the primer extension reactions were performed as described previously (34). Primer oSY9 was used to detect the cymR transcript. As a control, the rpsD transcript was detected by primer extension using oligonucleotide primer oSN03-87.

Overproduction and purification of CymR protein.

For production of recombinant CymR protein, the IMPACT system (New England BioLabs), which utilizes the inducible self-cleaving intein tag, was used. The cymR gene was amplified by PCR using JH642 chromosomal DNA and oSY10 and oSY11 primers. The amplified DNA fragment, after being digested with NcoI and SmaI, was inserted into pTYB4 (New England BioLabs) digested with the same enzymes to generate pSY22. The nucleotide sequence of cloned cymR was verified by DNA sequencing. E. coli ER2566 (New England BioLabs) carrying pSY22 was grown in 2× YT at 37°C. When the culture reached an OD₆₀₀ of 0.8, 0.5 mM IPTG was added to induce cymR expression. Incubation was continued overnight at 17°C. The procedure for further purification was described previously (34).

RNAP purification.

RNAP was purified according to the method previously described (27, 38) with minor modifications. After elution from a Ni-nitrilotriacetic acid (QIAGEN) column, the proteins were then passed through a heparin-agarose (Sigma) column with a 100 to 800 mM NaCl gradient. The enzyme was further purified with Bio-Rad High-Q column chromatography with a 100 to 500 mM NaCl gradient prior to concentration. Finally, RNAP was dialyzed against 10 mM Tris-HCl, pH 8.0, 100 mM KCl, 10 mM MgCl₂, and 50% glycerol and stored at −20°C.

In vitro transcription assay.

In vitro transcription experiments were carried out as described elsewhere (33). Linear DNA templates for rpsD and trxB promoters were generated as previously described (33, 35), while cymR and yrrT fragments were derived using primers oSY04 and oDYR05-025 (encoding an 86-base transcript) and oyrrT1 and oyrrT2 (encoding a 79-base transcript), respectively.

Microarray transcriptome analysis.

RNA was purified from JH642 and from ORB5724 cells using the previously reported procedure (34). RNA was used as template for the synthesis of dye-labeled cDNA, which was applied in a hybridization reaction with microarray slides of B. subtilis genomic DNA (2). Microarray slides were constructed from Corning UltraGAPS slides and an oligo set from Sigma-Genosys that represented 4,106 B. subtilis genes according to the Subtilist R16.1 release (http://genolist.pasteur.fr/SubtiList/). Analysis of hybridization data was carried out as previously described (2).

DNase I footprinting.

A DNA fragment (180 bp) containing the yrrT promoter DNA was synthesized by PCR amplification using primers oyrrT-1 and oyrrT-2. Plasmid pJSY3-3 (see Table S1 in the supplemental material) was used for the PCR template. The coding strand primer (oyyrT-1) was treated with T4 polynucleotide kinase and [γ-³²P]ATP before PCR. The PCR product was separated on a nondenaturing polyacrylamide gel and purified with Elutip-d columns (Schleicher & Schuell). Dideoxy sequencing ladders were obtained using the Thermo Sequenase cycle sequencing kit (USB) and the same primer used for the footprinting reactions. DNase I footprinting reactions were performed in 10 mM Tris-HCl pH 8.0, 30 mM KCl, 10 mM MgCl₂, 0.5 mM β-mercaptoethanol. Proteins were incubated with 100,000 cpm-labeled probe at 37°C for 20 min prior to DNase I treatment. The reactions were then precipitated by ethanol and subjected to 8% polyacrylamide-8 M urea gel electrophoresis.

RESULTS

CymR is required for both sulfate- and cysteine-dependent repression of the ytlI operon.

Spx is known to control sulfur metabolism in B. subtilis under steady-state, log-phase growth conditions (13). As reported previously (13), one of the Spx-controlled operons identified was the ytmI operon (ytmIJKLMNO ytnIJribRytnLM) (8, 14), which encodes highly conserved proteins functioning in the transport and utilization of alternative organosulfur sources such as sulfonates (3, 10, 23, 25, 45). A LysR homolog, encoded by a divergently oriented ytlI gene, exerts positive control over the ytmI operon (8), and Spx indirectly regulates ytmI expression by controlling the expression of the ytlI gene in medium containing sulfate. Recently, it was also reported that CymR is a negative regulator of the ytmI operon and ytlI gene expression (8, 14, 42). To investigate the role of CymR as a negative regulator for ytlI expression, we expressed the ytlI′-lacZ fusion (13) in wild-type (ORB5691), spx (ORB4871), and cymR mutant (ORB5733) strains. As shown in Fig. 2A, a much higher level of ytlI′-lacZ expression was observed in the cymR mutant and in the spx mutant grown in medium containing sulfate as a sole source of sulfur. In contrast to the spx mutant, which showed increased ytlI expression only in sulfate medium, the cymR mutant exhibited derepression of ytlI in the presence of either cysteine or sulfate (Fig. 2A), suggesting that there is a mechanism governing CymR production or activity that is responsive to the presence of cysteine or sulfate.

FIG. 2.

Control of ytmI and ytlI by Spx and CymR. A. Expression of ytlI′-lacZ in the wild type (ORB5691) and spx (ORB4871) and cymR (ORB5733) mutant strains. Cells of each fusion strain were grown in TSS medium containing MgSO₄ (SO₄) or cysteine (Cys) as sole sulfur source. Samples were collected at log phase and assayed for β-galactosidase activity, which was expressed in Miller units. B. Expression of ytmI′-lacZ in wild-type (ORB4794), cymR (ORB5730), and cymR ytlI (ORB5732) strains. Mutation of ytlI (ORB5732) eliminates the derepression of ytmI in the cymR mutant. C. Complementation of cymR with a thrC::cymR construct. ytlI-directed β-galactosidase activity of the cymR mutant strain bearing a ytlI-lacZ fusion (strain ORB5733) and the fusion-bearing partial diploid strain, ORB5809, with a thrC::cymR⁺ construct are shown.

The ytmI operon is induced in the presence of methionine, and induction requires the LysR-like regulator YtlI (8, 14). The expression of ytmI′-lacZ and ytlI-lacZ was examined in strains ORB4794, ORB5730 (cymR::spc), and ORB5732 (cymR::spc ytlI::aphA3) grown with sulfate or cysteine as a sole sulfur source (Fig. 2B). CymR has a negative effect on ytmI expression, as shown by the observed derepression of ytmI in the cymR mutant background. An additional mutation in the ytlI gene eliminates the derepression of ytmI-lacZ expression in the cymR mutant strain, consistent with the conclusion that CymR exerts negative control over ytmI by repressing the expression of the ytlI gene encoding its activator.

The cymR mutation is an insertion of a spectinomycin resistance cassette into the coding sequence. Although the resistance gene is oriented such that it can drive transcription of the downstream genes of the operon (Fig. 3A), polarity effects are still possible. The cymR gene is the first gene of an operon (Fig. 3A) that includes the yrvO gene, encoding a putative NifS homolog, and the trmU gene, encoding tRNA (5-methylaminomethyl-2-thiouridylate)-methyltransferase, an essential protein of B. subtilis (1), the product of which is required for thiouridine formation in tRNA. Hence, a complementation experiment was undertaken to determine whether repression of ytlI-lacZ would be restored if an ectopically expressed copy of cymR were introduced into the unlinked thrC locus of the B. subtilis genome. The complementation result in Fig. 2C confirms that CymR is a negative regulator for the ytlI gene. In the cymR mutant, repression of ytlI is observed when a copy of cymR (expressed from its own promoter) is supplied in trans (ORB5809). The expression of cymR (under control of the P_spank-hy promoter) also results in the repression of ytmI-lacZ in the cymR mutant (data not shown), confirming that CymR is responsible for negatively regulating the expression of ytIl and the ytmI operon.

FIG. 3.

A. The diagram top right shows the organization of the cymR operon containing the yrvO and trmU genes, based on primer extension analysis of cymR RNA in wild-type and spx mutant cells. Cultures were grown in TSS minimal medium containing either MgSO₄ (S), cysteine (C), or methionine (M) to an OD₆₀₀ of 0.7, and total RNA was isolated. Primer extension reactions and the sequencing reaction were performed using oligonucleotide oSY09. Primer extension products were resolved by polyacrylamide gel electrophoresis and visualized by phosphorimaging. Four lanes on the left are the gel profiles of sequencing reactions A, G, C, and T. Primer extension was also done with an rpsD-specific primer for a control of constitutive expression. B. Diagram of the nucleotide sequence of the cymR promoter region, showing the −10 sequence (TG promoter) and −35 along with the regions of dyad symmetry upstream of the promoter (dotted arrows). Coding sequences of the cymR and divergently oriented yrvN genes are shown with associated ribosome-binding sites (rbs) and a putative transcriptional start site for cymR (bent arrow).

Expression of cymR is reduced in the spx mutant.

As shown above, ytlI-lacZ expression is derepressed in both spx and cymR mutants in cultures grown with sulfate as sole sulfur source. We sought to determine if cymR expression is affected by Spx, indicating that Spx exerts its control of ytlI through the expression of the repressor gene, cymR. Primer extension analysis of total RNA purified from a wild-type strain and an spx mutant grown in TSS minimal medium containing sulfate, cysteine, or methionine shows that the start site of transcription for the cymR gene is at G located 57 bases upstream from the translation start point of cymR and downstream of the coding sequence belonging to the divergently oriented yrvN gene (Fig. 3A and B). The primer extension result indicates that cymR is constitutively transcribed in the medium supplemented with sulfate, cysteine, or methionine as a sole source of sulfur. The spx mutation appeared to have no effect on the level of primer extension product. The primer extension product of rpsD (ribosomal protein S4 [16]), transcript levels of which are not expected to change under the growth conditions tested, was included as a control.

In order to examine whether expression of cymR is reduced in spx mutant cells in vivo, a translational fusion of cymR-lacZ was constructed and cymR-directed β-galactosidase activity in wild-type (ORB6080) and spx mutant (ORB6081) cells was assayed. Figure 4A and B show that in sulfate- or cysteine-containing medium cymR-lacZ is reduced to 30% of the wild-type level in the spx mutant. The result suggested that Spx was required for optimal expression of cymR and that the reduced production of repressor might account for the increase in ytlI and ytmI expression in the spx mutant.

FIG. 4.

Control of cymR expression by Spx. Expression of a translational cymR-lacZ fusion over time in ORB6080 (cymR-lacZ; diamonds) and ORB6081 (spx::aphA3 cymR-lacZ; squares) grown in TSS medium containing either MgSO₄ (SO₄) (A) or cysteine (Cys) (B) is shown. The x axis represents the time (in hours) of cell growth.

Western analysis was performed on two sets of protein extracts from sulfate- and cysteine-containing cultures of wild-type and spx mutant strains. The results showed that the level of CymR protein was twofold less in the spx mutant than in wild-type cells when cultures were grown in TSS-sulfate medium (Fig. 5A). However, in cysteine medium, only a slight decrease in CymR protein level was observed when an spx mutation was introduced. To determine if the Spx enhances cymR transcription by interacting with RNAP, in vitro transcription was carried out using RNAP holoenzyme from an spx null mutant, with and without added Spx protein. There was no observable effect of Spx on in vitro transcription initiation from the cymR promoter DNA or the negative control promoter of the rpsD gene (Fig. 5B), while Spx was observed to activate transcription from the trxB promoter, as previously reported (33).

FIG. 5.

Spx controls cymR indirectly. A. Western analysis of CymR protein levels in wild-type (WT) and spx mutant cells. Strains ORB5691 (ytlI-lacZ) and ORB4871 (ytl-lacZ spx::aphA3) were grown in TSS medium containing MgSO₄ or cysteine as the sulfur source. Cleared lysates from mid-log-phase cells were obtained after lysis by French press and centrifugation. Lysate from ORB5724 (cymR::aphA) was used as a negative control. B. In vitro runoff transcription reactions of the cymR promoter fragment (5 nM) in the reaction mixtures containing purified RNAP in the presence and absence of Spx. Concentrations of Spx in the reaction mixtures are indicated. A positive control is the trxB promoter fragment (10 nM), which shows activation by Spx. The rpsD template DNA (20 nM) containing the ribosomal protein S4 gene promoter shows no effect by Spx. C. The CymR repressor, supplied with the Pspac promoter in the spx mutant, restores the repression of ytlI. Cells of WT and spx mutant strains (indicated) bearing the ytlI-lacZ fusion were transformed with plasmid pSY20, in which the cymR gene was placed under the control of P_spank-hy. Cultures were grown to mid-log phase in sulfate medium in presence or absence of IPTG. Samples were collected and assayed for β-galactosidase activity.

If Spx were necessary for optimal cymR expression, then releasing cymR expression from Spx control should result in repression of ytlI in an spx mutant. Expression of the cymR gene under P_spank-hy promoter control in the spx mutant (ORB6082) restored repression of ytlI (Fig. 5C) in the presence of IPTG, which indicates that the requirement of Spx in negative control of ytlI is due, at least in part, to its role in stimulating cymR expression.

The amount of CymR produced in the spx mutant in sulfate medium might not be sufficient to repress ytlI expression, while the amount produced by the mutant in cysteine medium is enough to repress the expression. The reduced amount of CymR in the spx mutant results in derepression of ytlI and enhanced expression of the ytmI operon, as shown in Fig. 2A. In cysteine medium, the low expression of cymR-lacZ in the spx mutant evidently is not an indication of the concentration of active CymR, since protein levels are higher in cysteine-grown spx mutant cells than in mutant cells grown in sulfate.

Microarray hybridization analysis uncovers the CymR regulon.

Previous studies showed that Spx exerts negative control of the ytmI, yxeI, and ssu operons in medium containing sulfate as sole sulfur source. As with ytmI, the role of spx in yxeI and ssu expression is likely the enhancement of cymR expression, which leads to repression of the two operons by the CymR protein. We sought to determine if this was the case by performing a genomic microarray hybridization experiment to uncover the genes that are negatively controlled by CymR. A transcriptional profile of the CymR regulon has recently been published by the Martin-Verstraete group (14); in this profile genes functioning in cysteine biosynthesis and those involved in the general stress response were derepressed in the cymR in-frame deletion mutant. To undertake transcriptional profiling of our cymR mutant, we used an insertion of the spc resistance cassette, which bears a constitutively utilized promoter to drive expression of the downstream genes of the operon, while disrupting the cymR gene.

RNA was purified from triplicate, parallel TSS-sulfate cultures of wild-type JH642 cells and cells of the cymR mutant, ORB5724. The RNA was used as template for Cy3- and Cy5-labeled cDNA synthesis. The cDNA was then used for hybridization analysis of a B. subtilis genomic microarray using previously described procedures (2). Table 2 lists the genes whose expression is elevated in the cymR mutant. All of the genes of the ytmI, yxeI, and ssu operons show derepression in the cymR mutant background. It is likely, therefore, that the reduced expression of yxeI and ssu in the spx mutant (13) is due to the reduced expression of cymR in the spx background. Additionally, the yrrT operon genes show derepression in the cymR mutant. Genes of the yrrT operon encode enzymes of the S-adenosylhomocysteine shunt and reverse transsulfuration pathway that yield cysteine from S-adenosylmethionine (Fig. 1 and 6A). Unlike the previously published cymR transcriptome profile (14), we observed little significant induction of general stress operons, although the katA gene, a member of the hydrogen peroxide-induced PerR regulon (7) and encoding the vegetative catalase, is significantly induced in the cymR mutant. It is possible that differences in expression profiles between microarray hybridization analyses reported herein and published previously could be the result of different growth media (TSS glucose-glutamate in our work and phosphate-buffered medium containing glucose and glutamine in the work of Even et al. [14]) used in cultures of wild-type and cymR mutant strains. The use of different cymR alleles could have also affected the results of the microarray analyses.

TABLE 2.

Genes induced by cymR mutation

Gene	Function	Inductiona
ytmL	ytmI operon, sulfonate utilization	6.72
ytmI	ytmI operon, sulfonate utilization	6.51
ytmJ	ytmI operon, sulfonate utilization	6.42
ytmO	ytmI operon, sulfonate utilization	6.42
ssuB	ssu operon, sulfonate utilization	6.35
ytmN	ytmI operon, sulfonate utilization	6.23
ssuC	ssu operon, sulfonate utilization	6.16
ssuA	ssu operon, sulfonate utilization	6.13
ytmK	ytmI operon, sulfonate utilization	6.07
ytnI	ytmI operon, sulfonate utilization	5.93
ytmM	ytmI operon, sulfonate utilization	5.75
ygaN	ssu operon, sulfonate utilization	5.39
ytnJ	ytmI operon, sulfonate utilization	5.09
yezD	Unknown	4.97
ssuD	ssu operon, sulfonate utilization	4.55
ytnM	ytmI operon, sulfonate utilization	4.44
ribR	ytmI operon, sulfonate utilization	4.19
yrhB	Cysteine synthesis, cystathionine gamma synthase	3.67
yrhA	Cysteine synthesis	3.51
yrrT	S-Adenosylmethionine-dependent methyltransferase	3.45
cah	Cephalosporin C deacetylase	3.37
ytlI	LTTR,b activator of ytmI operon	2.81
katA	Catalase	2.68
ytlB	Unknown	2.65
yjnA	Predicted permease	2.53
yhcL	tcyP, cystine transport	2.53
yhzA	Ribosomal protein S14, ytmI operon	2.49
yxeP	yxeI operon, organosulfur utilization	2.4
yxeQ	yxeI operon, organosulfur utilization	2.39
yxeN	yxeI operon, organosulfur utilization	2.37
mtn	Methylthioadenosine nucleosidase	2.31
yxeO	yxeI operon, organosulfur utilization	2.28
yoqM	SPβ phage gene	2.27
ycgN	Similar to 1-pyrroline-5-carboxylate dehydrogenase	2.25
yrhC	yrrT-mtn operon, unknown	2.16
yxeR	yxeI operon, organosulfur utilization	2.15
yxeM	yxeI operon, organosulfur utilization	2.11
yqxI	Unknown, secreted	2.02
tgl	Transglutaminase	2.01
tasA	Translocation-dependent antimicrobial spore protein	2
cysK	Cysteine synthase	1.95
yxeL	yxeI operon, organosulfur utilization	1.94
yxeD	yxeI operon, organosulfur utilization	1.9
ytoI	CBSc domain protein, putative thioesterase	1.85
ycgM	Similar to proline oxidase	1.83
sipW	Signal peptidase I	1.77
ydbM	Similar to butyryl-coenzyme A dehydrogenase	1.75
yxeK	yxeI operon, organosulfur utilization	1.72
yuaB	Unknown	1.7
cssS	Secretion stress sensor kinase	1.68
cotJB	Spore coat	1.67
yqxJ	Unknown	1.63
ywfD	Similar to glucose 1-dehydrogenase	1.63
ywcI	Unknown	1.62
pel	Pectate lyase	1.53
lrpB	Leucine-responsive protein, represses glyA and kinB	1.53
luxS	SAM metabolism, autoinducer-2 production protein	1.53

^aLog₂ ratio of the yrzC mutant transcript level and wild-type transcript level.

^bLTTR, LysR-type transcriptional regulator.

^cCB5, cystathione beta synthase.

FIG. 6.

Spx and CymR control of the yrrT operon. A. Organization of the yrrT operon (see Fig. 1 for details of gene functions.). B. Expression of yrrT′-lacZ in wild-type (WT) and cymR mutant cells. Cells were grown to mid-log phase in TSS medium containing either MgSO₄ (SO₄), cysteine (Cys), or methionine (Met) as the sole sulfur source. Samples were collected for assay of β-galactosidase activity. C. Expression of yrrT′-lacZ in wild-type and spx mutant cells. Cultures were grown and samples were processed as for panel B.

Control of the yrrT operon by CymR and Spx.

In contrast to the ytmI, yxeI, and ssu operons, which are under the negative control of both Spx and CymR, the yrrT operon is positively controlled by Spx and is one of the most highly induced transcription units observed in microarray analysis of Spx-overexpressing cells (34). Thus, yrrT is under CymR negative control (14), while Spx appears to exert positive control. We validated these results using a yrrT operon-lacZ fusion introduced into the yrrT locus. As shown previously (14), yrrT-lacZ is derepressed in the cymR cells (Fig. 6B) and, like ytlI and the ytmI operon, repression in sulfate medium requires Spx (Fig. 6C). In methionine medium, yrrT-lacZ is derepressed in wild-type and cymR mutant cells; however, derepression requires Spx. This is the first indication that Spx can activate transcription during log phase in the presence of an alternative sulfur source (methionine).

Spx activates transcription from the yrrT promoter in vitro.

To determine if Spx can directly stimulate transcription from the yrrT promoter, a fragment containing the intergenic region between the divergently organized yrrT and yrzA coding sequences (Fig. 7A) was obtained by PCR. This region contains the yrrT operon promoter, the sequence for which can be discerned by inspection of the intergenic DNA sequence (Fig. 7A). The TG-type promoter −10 region (9) is located between 66 and 77 bp upstream of the start of the yrrT coding sequence. Promoters with extended −10 TG elements are more prevalent in B. subtilis than in E. coli, and such promoters have a reduced dependence on the RNAP interaction with the −35 element. Using the downstream PCR primer that hybridizes to the first eight codons of yrrT, template DNA was synthesized by PCR for in vitro transcription reactions. A transcript of 82 bases was generated in the in vitro transcription reaction containing the PCR product, as shown in Fig. 7B (the trxB transcript is 88 bases). This transcript corresponds to the predicted start point and promoter sequence indicated (Fig. 7A). Transcription from this promoter is stimulated by Spx, as evident from the elevated amount of transcript synthesized in reactions containing RNAP and Spx protein (Fig. 7B). As shown in previously reported work (33), the addition of the reductant dithiothreitol (DTT) eliminates trxB transcription due to the conversion of the CxxC redox disulfide center of Spx to the thiol state. In contrast, the reaction containing the yrrT promoter DNA, RNAP, and Spx shows stimulation of yrrT transcription by Spx even in the presence of DTT, although the amount of transcript accumulation is less than in the absence of reductant. Consistent with the previous microarray transcriptome analysis (34), yrrT is a member of the Spx regulon and is under the direct positive control of Spx, but the oxidized form of Spx is not essential for the activation, as suggested by the β-galactosidase assay carried out with yrrT-lacZ fusion-bearing cells under steady-state, log-phase growth conditions (or under nonoxidative conditions in vitro).

FIG. 7.

Transcription from the yrrT promoter is activated by Spx in vitro. A. Organization of the yrrT operon. Dotted arrows indicate regions of dyad symmetry. The −10 TG promoter is indicated, as is the −35 sequence. The dashed arrow indicates the complementary sequence of the downstream primer used to construct the transcription promoter fragment by PCR. The bent arrow indicates the approximate location of the transcription start site as deduced from the size of the in vitro transcript. B. Gel profiles of transcription reactions containing promoter fragments trxB or yrrT (10 nM) and RNAP (50 nM) in the presence and absence of Spx and DTT. Concentrations of Spx and DTT are indicated. C. Repression of yrrT transcription by CymR. In vitro runoff transcription reaction mixtures contained RNAP (25 nM) and either trxB promoter (20 nM) or yrrT promoter (10 nM) DNA. Spx and CymR were added to the reaction mixtures at the indicated concentrations. The lower panel shows a plot of band intensities of the in vitro runoff transcription of yrrT shown in the upper panel. Band intensities were quantified using ImageQuant version 5.2, and error bars were derived from two repeats of the experiment. D. DNase I footprinting of the yrrT promoter with CymR. The yrrT promoter prepared by PCR using oyrrT-1 and oyrrT-2 is shown. Reaction mixtures contained CymR and promoter DNA with a radiolabeled coding strand. The concentrations of CymR used were 0.5, 1, 2, and 4 μM. The positions relative to the transcription start site are shown on the left. The −10 and +1 are shown next to the dideoxy sequencing ladders. The protected region is indicated by the solid line on the right side.

We used purified CymR protein to determine if it could repress transcription of yrrT in transcription reactions containing purified RNAP and Spx. As shown in Fig. 7C, added CymR caused repression of yrrT in the presence and absence of Spx but had no effect on Spx-dependent transcription from the trxB promoter. This result is in keeping with the previously reported finding that CymR protein binds to DNA bearing the yrrT regulatory region (14). The direct binding of CymR to an operator element residing downstream of the transcription start site was reported (14). This was confirmed by DNase I footprinting using a fragment of yrrT promoter DNA and purified CymR protein. As shown in Fig. 7D, CymR protected a region between +11 and +28, which corresponds closely with the CymR-binding sequence suggested in the previously reported electrophoretic mobility shift assay analysis (14).

DISCUSSION

Organosulfur metabolism in Bacillus is controlled by a number of regulatory systems that act at the level of transcription initiation and during transcription elongation. An example of the latter is the S-box control mechanism that governs expression of genes that function in methionine synthesis in response to levels of SAM (17). SAM binds directly to the nascent RNA at a folded sequence known as the S-box, within the untranslated leader region encoded by met genes (30). High SAM levels stabilize a terminator that reduces transcription of the downstream coding regions. As for regulation of transcription initiation, LysR-type regulators show operon-specific activity. For example, CysL and YtlI activate transcription of cysIJ (sulfite reductase) and ytmI (sulfonate utilization and cysteine transport), respectively (8, 11, 19).

The work described herein has uncovered multiple roles for Spx in sulfur metabolism. First, Spx functions in sulfate-dependent control of organosulfur utilization operons through stimulation of cymR expression. Secondly, microarray transcriptome analysis reveals that transcript levels of genes required for cysteine synthesis, the yrrT operon and cysK, increase when Spx interacts with RNAP (34). The stimulation of homocysteine production and transsulfuration is characteristic of cells responding to oxidative stress (4, 12, 21, 28, 29, 37, 39, 47). In mammalian systems, cysteine production is accelerated during bacterial infection because of an increased need for glutathione. Oxidative stress stimulates the activity of cystathionine β-synthase, which catalyzes the first step of homocysteine transsulfuration. Thus, by stimulating the SAH shunt and homocysteine transsulfuration, Spx appears to perform a regulatory function that operates in species representing several phylogenetic levels.

The CymR repressor exerts wide control over a number of sulfur metabolism operons and is responsible for their repression when the preferred sulfur sources, sulfate and cysteine, are present (14). Evidence was reported that CymR binds directly to the promoter DNA of genes under its control (14). Optimal levels of CymR protein require the activity of Spx as shown herein by Western blot analysis and by examining the expression of a cymR-lacZ fusion. The contribution of Spx to control of cymR is likely indirect, as Spx has no effect on transcription from the cymR promoter in vitro. That the promoter of cymR is utilized by purified RNAP in vitro suggests that cymR does not require a positive transcriptional control factor for expression. Spx could function indirectly by affecting the expression of a gene encoding the direct transcription factor. Alternatively, through its control of other genes that function in organosulfur metabolism and cysteine biosynthesis, Spx could affect the concentration of a metabolite, perhaps an intermediate of sulfur assimilation or organosulfur utilization, which serves as an effector of cymR expression control.

The third function of Spx in sulfur metabolism is the derepression of certain organosulfur metabolism genes in methionine medium. The yrrT gene is derepressed in a cymR mutant, but this derepression requires Spx (Fig. 6). This is observed under nonstress conditions when methionine is the sole sulfur source. Consistent with this observation is the finding that Spx can stimulate yrrT transcription in vitro in the presence or absence of reductant, unlike the control it exerts in the transcription from the trxA and trxB promoters (33, 34). Spx can interact with RNA polymerase in vitro in the presence of reductant, as observed in reactions in which Spx repressed ComA-dependent transcription from the srf operon promoter (35; Y. Zhang and P. Zuber, unpublished data). From the evidence detailed herein, we can conclude that Spx can activate gene transcription under steady-state growth conditions.

A diagram (Fig. 8) is shown that summarizes the proposed Spx-dependent control that governs the expression of genes functioning in sulfur metabolism. Spx promotes negative control (by stimulating CymR production) of operons that function in cysteine synthesis from alternative sulfur sources but stimulates yrrT operon expression in the absence of cysteine, thus promoting cysteine synthesis through S-adenosyl methionine catabolism (Fig. 1). Microarray analysis (34) also suggests that Spx represses expression of genes required for methionine biosynthesis, a role of Spx that is currently under investigation (M. Derr, Y. Zhang, and P. Zuber, unpublished data). Our findings uncover an expanded role in global transcriptional control mediated by the Spx-RNAP interaction.

FIG. 8.

Diagram summarizing the proposed role of Spx in the control of organosulfur metabolism. Spx is required for the optimal production of the CymR repressor, which exerts negative control of the organosulfur utilization operons ytmI, yxeI, and ssu, and the methionine utilization pathway that generates cysteine via reactions catalyzed by the products of the yrrT operon. Spx exerts direct positive control of yrrT transcription under conditions that result in elevated Spx protein concentration (oxidative stress) and when cells must produce cysteine from alternative forms of sulfur, such as methionine.