Mycothiol regulates and is regulated by a thiol-specific antisigma factor RsrA and σR in Streptomyces coelicolor


*E-mail; Tel. (+82) 02 880 6706; Fax (+82) 02 888 4911.


Mycothiol (MSH) is a small thiol molecule with a cysteine-ligated disaccharide structure found in actinomycetes that include streptomycetes and mycobacteria. In Streptomyces coelicolor, a model organism for antibiotic production and differentiation, the amount of MSH is under the control of a sigma factor σR, which is regulated by an antisigma factor RsrA with a thiol-disulphide redox switch. We found that the first gene (mshA) in the biosynthetic pathway for MSH and the gene for amidase (mca) that participates in detoxifying mycothiol-reactive drugs are under direct control of σR. The σR target genes are induced not only by a thiol oxidant diamide, but also by alkylating agents that cause a rapid decrease in MSH. Expression of the σR regulon was also elevated in MSH-deficient mutants, suggesting that a decrease in the level of MSH is a natural intracellular trigger for σR activation. We found that MSH was capable of reducing RsrA to bind σR, whereas glutathione was not. These results support a proposal that the RsrA-σR system senses the intracellular level of reduced MSH, and that MSH serves as a natural modulator of the transcription system for its own replenishment in addition to being a redox buffer and drug detoxifier.


Mycothiol (N-acetylcysteinylglucosaminylinositol) is a major low-molecular-weight thiol found in most actinomycetes (Newton et al., 1996) and has been considered as a functional equivalent of glutathione found in many bacteria and almost all eukaryotes. In mycobacteria, where it has been most intensively studied, it serves to protect cells not only by coping with oxidative stress that originates from host immune cells but also by detoxifying incoming drugs (Newton and Fahey, 2002; Rawat and Av-Gay, 2007). The mycothiol biosynthetic and detoxification pathways have been elucidated in detail as described in Fig. 1. The biosynthetic pathway consists of five enzymatic steps involving a glycosyltransferase (MshA), a phosphatase (MshA2), a deacetylase (MshB), a cysteine ligase (MshC) and finally mycothiol synthase (MshD) (Newton et al., 2000a; 2003; 2006; Koledin et al., 2002; Sareen et al., 2002; Rawat et al., 2003).

Figure 1.

The pathways for biosynthesis of mycothiol and its recycling in detoxification cycle. Mycothiol (MSH) is synthesized by successive action of five enzymes (MshA, A2, B, C and D; Newton and Fahey, 2002; Newton et al., 2006). MshA produces a disaccharide moiety N-acetylglucosaminylinositol phosphate (GlcNAc-Ins-P), by combining UDP-N-acetylglucosamine (UDP-GlcNAc) with myo-inositol monophosphate (Ins-P). MshA2 dephosphorylates it to produce N-acetylglucosaminylinositol (GlcNAc-Ins), from which the acetyl group is removed by MshB, forming glucosaminylinositol (GlcN-Ins). MshC links cysteine to GlcN-Ins to form cysteinylglucosaminylinositol (Cys-GlcN-Ins), which is acetylated by MshD to form the final product MSH [1d-myo-inosityl 2-(N-acetyl-l-cysteinyl)amido-2-deoxy-α-d-glucopyranoside]. An electrophilic compound (RX) is conjugated with MSH, and the MSH S-conjugate (MSR) is cleaved by an amidase (Mca) to form a mercapturic acid (AcCysR) and GlcN-Ins (Newton and Fahey, 2002).

In the MSH-dependent detoxification pathway, mycothiol is conjugated to electrophilic compounds (RX) through its thiol moiety, forming the MSH S-conjugate (MSR). MSR is then cleaved by an amidase (Mca) to produce glucosaminyl inositol (GlcN-Ins) and the N-acetylcysteine S-conjugate (AcCysR, a mercapturic acid). AcCysR is excreted to the cytoplasm whereas GlcN-Ins is recycled to the biosynthetic pathway for MSH (Newton et al., 2000b; Steffek et al., 2003; Rawat et al., 2004). Mca is known to have a broad substrate specificity including S-conjugates of monobromobimane (mBBr), N-ethylmaleimide (NEM) and rifamycin SV.

Not only electrophilic but also non-polar compounds that contain an electrophilic carbon, nitrogen or sulphur atom, including aromatic rings, α, β-unsaturated ketones, epoxide rings, arene oxides and quinones, are thought to conjugate with MSH as they do with GSH (Hayes et al., 2005; Rawat and Av-Gay, 2007). Many antibiotics have these chemical structures and diverse N-acetylcysteine S-conjugates of antibiotics were found in the fermentation broth of many actinomycetes, suggesting that MSH can be conjugated to these compounds and cleaved by Mca (Newton and Fahey, 2002; Rawat and Av-Gay, 2007). Accordingly, MSH-deficient mycobacterial mutants were shown to be sensitive to alkylating agents and a wide range of antibiotics such as erythromycin, azithromycin, vancomycin, penicillin G, rifamycin and rifampin (Rawat et al., 2002; Buchmeier et al., 2003). This suggests that the MSH-dependent detoxification system might be important in other actinomycetes such as streptomycetes, rhodococci and corynebacteria.

Streptomyces coelicolor is a model organism for studying antibiotic production and morphological differentiation. It is a soil-dwelling aerobic bacterium that produces a variety of secondary metabolites when it undergoes fungus-like morphological differentiation. For an effective survival in complex soil environment against numerous physical and chemical stresses, it is equipped with a large variety of gene products and regulatory circuits as reflected by the presence of more than 7800 protein coding genes and 66 predicted sigma factors for transcription (Bentley et al., 2002; Hahn et al., 2003). Among many challenges, oxidative and osmotic stress responses have been studied in some detail (Kim et al., 1998; Cho et al., 2000; Hahn et al., 2002; Viollier et al., 2003; Lee et al., 2005). σR, an ECF (extracytoplasmic function or group 4) family sigma factor, is responsible for transcribing genes that cope with thiol-oxidative stress (Paget et al., 1998). Its activity is modulated by an antisigma factor RsrA which belongs to a widespread group of zinc-containing antisigma factors (Campbell et al., 2007), and is sensitive to thiol redox change. RsrA binds and sequesters σR (SigR) in its reduced state, but dissociates from σR upon oxidation of its two cysteine thiols (C11 and C44) to form a disulphide bond (Kang et al., 1999; Li et al., 2003; Bae et al., 2004). The released σR induces expression of over 30 gene products including thioredoxin (Paget et al., 2001) which reduces protein disulphide bonds. The oxidized RsrA is reduced by thioredoxin to re-bind σR, returning cell physiology to its prestimulus state (Kang et al., 1999). The sigR mutant grows and differentiates normally but shows sensitivity toward the thiol oxidant diamide (Paget et al., 1998) and has a reduced level of MSH (Paget et al., 2001).

The genes in MSH biosynthetic pathway (mshA, B, C, and D) have been identified in S. coelicolor (Park et al., 2006). Disruption of mshA, C and D resulted in no synthesis of MSH whereas the mshB mutant synthesized MSH at about 10% of the wild-type level. The mshB and a putative amidase gene (SCO4967) show extensive sequence similarity. The mca gene which functions primarily for MSH recycling in the detoxification pathway has not been characterized. The role of MSH in protecting cells against various antibiotic compounds has not been investigated in streptomycetes. In this paper, we identify the gene for MSH S-conjugate amidase (mca), characterize the role of MSH in protecting cells against drugs and disclose the interplay between the σR-RsrA regulatory system and MSH metabolism.


SCO4967 encodes an amidase (Mca) that recycles mycothiol in detoxification pathway

In S. coelicolor, the mycothiol biosynthetic genes mshA, B, C and D have been identified, and a putative gene (SCO4967) for MSH S-conjugate amidase (Mca) has been predicted (Park et al., 2006). In order to test the function of SCO4967, we created a deletion mutant and compared its amidase activity with the wild type. To measure amidase activity, MSH was labelled with fluorescence by treating cells with mBBr that conjugates with MSH. The amount of mycothiol bimane (MSmB), the fluorescent bimane (mB) derivative, was determined by HPLC. When MSmB is cleaved by Mca, it yields N-acetylcysteinyl bimane (AcCysmB) that is secreted out of the cell into the medium. Figure 2 shows the amount of MSmB and AcCysmB in the cytoplasm of wild type, the SCO4967 mutant, and the mutant complemented with mca. In the wild type, the amount of MSmB is low relative to the cleaved product AcCysmB (0.13 versus 2.17 μmol g−1), whereas in the mutant the relative amount is reversed. This demonstrates that the product of SCO4967 is indeed an MSH S-conjugate amidase (Mca). Introduction of the wild-type gene to the mutant through the att site in the chromosome reduced the amount of MSmB, and increased AcCysmB, confirming the activity of amidase. The amount of cleaved product in the medium was estimated to be about 1.24 μmol g−1 in the wild type, whereas in the mutant it was only 0.25 μmol g−1, consistent with the observation made for the cytoplasm. The low residual amidase activity in the mca mutant suggests the presence of an alternative enzyme with lower efficiency. MshB, a sequence homologue of Mca reported to contain a low amidase activity in vitro (Newton et al., 2000a), is a likely candidate.

Figure 2.

The amidase activity of Mca. The level of reactant (MSmB) and product (AcCysmB) of the amidase following conjugation of mBBr with MSH. Cells grown in YEME were treated with monobromobimane (mBBr), to form a fluorescent bimane derivative of mycothiol (MSmB). Amidase cleaves MSmB to AcCysmB (N-acetylcystein S-conjugate of bimane) that is excreted from the cells. The amounts of MSmB (black bar) and AcCysmB (white bar) in the cytoplasm and the medium were measured by HPLC and are presented as μmol g−1 dried cell weight (WT, the wild type; Δmca, mca mutant; +mca, the mutant complemented with the wild-type mca gene).

Sensitivity of the mca mutant to various alkylating agents and antibiotics

If Mca cleaves MSH S-conjugates of a variety of alkylating agents and antibiotics (Newton et al., 2000b; Steffek et al., 2003; Rawat et al., 2004), it can serve to protect S. coelicolor cells from the toxicity of these compounds as observed in some mycobacteria (Rawat et al., 2004). To test this possibility, we examined the sensitivity of the mca mutant grown on NA plates containing various alkylating agents and antibiotics. 1-chloro-2,4-dinitrobenzene and NEM were used as alkylating agents. The antibiotics used in the assays represent different chemical categories and targets; tetracycline (tetracycline, 30S ribosome), rifamycin SV (ansamycin, RNA polymerase), lincomycin (sugar amide, 50S ribosome), novobiocin (coumerine glycoside, DNA gyrase) and erythromycin (macrolide, 50S ribosome). The wild-type and the complemented strains were grown in parallel. As shown in Fig. 3, the mca mutant demonstrated increased sensitivity to all the agents we tested except erythromycin (data not shown). Complementation of the mutant with mca cloned in pSET152 mitigated sensitivity in all cases, except for novobiocin. As the complementing gene is integrated to the chromosome through the att site that is not the original location of the gene, it might not create the wild-type phenotype in its entirety. For novobiocin, which affects DNA topology, it can be postulated that pSET152 integration into the chromosome could have caused some topological problem that is aggravated in the presence of novobiocin. Overall, the results of sensitivity assays supported the role of Mca in detoxifying a number of drugs in streptomycetes.

Figure 3.

Sensitivity of the mca mutant to various thiol-reactive drugs and antibiotics. Serially diluted spore solutions from M145 (wild type), mca mutant (Δmca) and the complemented strain (+mca) were spotted on NA plates containing CDNB (1-chloro-2,4-dinitrobenzene, 50 μM), NEM (N-ethylmaleimide, 200 μM), tetracycline (40 μg ml−1), rifamycin SV (50 μg ml−1), lincomycin (90 μg ml−1) and novobiocin (7.5 μg ml−1). In all experiments, the number of spores were 3 × 105 (3E5), 3 × 104 (3E4) and 3 × 103 (3E3), except for rifamycin SV where 10-fold more spores were spotted.

σR-dependent induction of msh and mca genes

As σR responds to thiol-oxidative stress and a sigR mutant exhibited lowered production of MSH (Paget et al., 2001), we examined whether the genes in MSH metabolism are regulated by σR. Transcripts from mshA, B, C and D as well as mca were analysed by S1 mapping in cells either untreated or treated with a thiol oxidant diamide (0.5 mM) for different lengths of time. A transcript from SCO1545, which encodes an mshD homologue, was assessed in parallel as a control. As demonstrated in Fig. 4A, the msh and mca genes were either induced transiently (mshA and mca) or gradually (mshB, C and D) by diamide. None of these genes was induced in the sigR mutant. The rapid and transient induction of mshA and mca coincides with the direct induction pattern of the σR regulon. Examination of the promoter regions of mshA and mca revealed that they match highly with the σR consensus sequence as demonstrated in Fig. 4B. The sequence was different by one nucleotide in the spacer length (mshAp1) or in the −10 box (mca) from the previously determined consensus sequence (GGAAT-N18-GTT; Paget et al. 2001). Therefore, it is very likely that mshA and mca are direct targets of σR, whereas the other genes are indirect targets.

Figure 4.

Regulation of msh and mca genes by σR.
A. σR-dependent induction of mycothiol metabolic genes by the thiol oxidant diamide (DA). The mshA, B, C, D and mca transcripts were analysed by S1 mapping at various time points (up to 60 min) after treating cells (wild-type and sigR mutant) with 0.5 mM diamide. Transcripts from SCO1545 encoding an mshD homologue, were analysed in parallel as a control. For mshA, transcripts from the proximal p1 promoter were induced in response to diamide.
B. The promoter sequences of mshA and mca recognized by σR.
C. Redox-dependent transcription of mshAp1 and mca promoters in vitro by RNA polymerase containing σR in the presence of antisigma factor RsrA. The core RNA polymerase from S. coelicolor, σR and RsrA proteins were incubated with varying concentrations of DTT (0, 0.5, 1, 2.5 and 5 mM) before mixing with transcription buffer containing mshA or mca promoter DNAs. The transcription activity of either core polymerase alone or core plus σR holoenzyme was analysed in parallel.

To determine whether mshA and mca are directly regulated by σR, in vitro transcription assay was performed using purified components (Fig. 4C). Transcripts of expected sizes were synthesized from the mshA and mca promoter templates by RNA polymerase containing σR, whereas none was made by core RNA polymerase alone. Inhibition of σR-specific transcription by RsrA was observed in the presence of DTT (at 0.5 mM or higher), confirming that these promoters are indeed controlled by σR and its redox-sensitive antisigma factor RsrA. Taking the sequences of mshAp1 and mca promoters into consideration, the σR consensus sequence needs be revised to GGAAT-N18-19-GTY (Fig. 4B).

Previous study on the σB-regulon that responds to osmotic stress demonstrated that mshA has a σB-specific promoter (mshAp2; Lee et al., 2005). The σR-specific promoter (mshAp1) lies 160 nt downstream from the σB-specific promoter. The mshAp1 promoter produces transcripts whose 5′ end lies about 60 nt downstream from the annotated start codon in the database ( An in-frame downstream start codon with a good ribosome binding sequence (AGGAGG) predicts MshA 39 amino acids shorter than the annotated one. The extended N-terminal 39 amino acids is not conserved among MshA homologues in actinomycetes, and is not preceded with any predictable ribosome binding site. Therefore, it is most likely that MshA starts with a sequence MSQYVSRLGR at nucleotide position 4613253.

The change in the level of MSH by various drugs

The thiol oxidant diamide we routinely use for oxidative treatment forms disulphide bonds between thiols in both proteins and small molecules such as MSH. Therefore, we determined the intracellular level of MSH after diamide treatment by HPLC. Within 10 min after 0.5 mM diamide treatment, the MSH level decreased by about half, and it began to be restored after 30 min (data not shown). Diamide could have caused the decrease in reduced MSH by (i) direct oxidation of MSH to MSSM disulphides, (ii) formation of mixed disulphides between proteins and MSH, as well as subsequent formation of MSSM through the action of thiol-disulphide redox enzymes, and/or (iii) conjugation of MSH with an electrophilic by-product from oxidatively damaged DNA and membrane lipids that arise indirectly from depletion of reducing equivalents.

As the induction of the σR regulon by diamide occurs concomitantly with a decrease in MSH, we needed to delineate whether MSH depletion that arises independently of oxidation can still induce σR regulon. In order to test this, we examined the effect of NEM, which alkylates MSH, on σR induction. The results in Fig. 5A demonstrate that MSH was depleted rapidly by NEM treatment within 20 min, but restored to near normal level after 80 min. When σR-specific transcripts in NEM-treated cells were examined by S1 mapping, we found that all the σR-target genes examined (sigR, mshA, mca) were induced, concomitantly with MSH depletion (Fig. 5B). To rule out the possibility that NEM affects σR–RsrA interaction through direct alkylation of exposed thiols in RsrA, we examined the effect of NEM on σR–RsrA complex formation on native PAGE, and found that NEM did not disrupt the σR–RsrA complex (data not shown). Even when reduced RsrA was alkylated with NEM before mixing with σR in the presence of DTT, it retained σR binding activity, consistent with a previous observation made with another alkylating agent iodoacetamide (IAA; Li et al., 2003). Therefore, it is apparent that depletion of reduced MSH by alkylating agent can also induce the σR regulon.

Figure 5.

The change in the amount of MSH and the induction of the σR regulon by conjugating agents.
A. Changes in MSH level by NEM. The wild-type M145 cells grown in YEME were treated with 40 μM NEM and harvested at intervals of 20 min. The amount of MSH in untreated (▵) and NEM-treated samples (●) was analysed by HPLC as described in the text. The unit represents micromole of MSH per gram of dried cell weight.
B. S1 mapping analysis of σR-specific transcripts upon NEM treatment. RNA samples were prepared from cells treated with NEM as in (A) and analysed by gene specific probes for sigRp2, mca and mshAp1 transcripts.
C. Changes in MSH level by mBBr. Cells were treated with 20 μM of mBBr for up to 140 min and analysed as in (A).
D. S1 mapping analysis of σR-specific transcripts upon mBBr treatment.

To extend this observation further, we examined the effect of mBBr, another MSH conjugating agent. The results in Fig. 5C show that 20 μM mBBr decreased the amount of reduced MSH to less than 30% of the untreated level within 20 min. The MSH level was restored rather sluggishly compared with NEM, but was completed by 140 min. The σR-target genes were induced accordingly as monitored by S1 mapping (Fig. 5D). The quick induction coincided well with the rapid depletion of MSH within 20 min. The shut-off of the induction pulse was almost complete by 80 min as observed for other stimuli (diamide, NEM), even though the level of MSH was still low. Taken all together, it seems very likely that the depletion of MSH, either by oxidation or by conjugation to various electrophilic compounds, caused induction of σR-regulon genes.

Expression of σR-regulon genes is elevated in msh mutants

As the depletion of MSH by chemical conjugation can trigger σR induction, we examined whether MSH-biosynthetic mutations have similar effect. The sigR, trxC and trxB transcripts were analysed by S1 mapping in mshA, B, C and D mutants. Results in Fig. 6 demonstrate that the amounts of σR-specific transcripts were all elevated relative to the wild-type, except in mshD mutant. The effect was most pronounced in mshA and C mutants where no MSH is synthesized (Park et al., 2006). In mshB mutant where MSH is produced at about 10% of the wild-type level, the σR-target genes were slightly induced. In mshD mutant where no detectable MSH is synthesized, it is likely that Cys-GlcN-Ins and its derivatives could serve as alternative thiols to maintain near-wild-type thiol environment (Newton et al., 2005). When mutant strains were complemented with wild-type genes, the amounts of transcripts decreased to the wild-type level (data not shown), confirming the effect of mutations.

Figure 6.

Expression of σR-target genes in msh mutants. Transcripts from sigR, trxC and trxBA promoters in the wild-type and msh mutants were analysed by S1 mapping. Transcripts from σR-specific promoters were presented.

MSH is a natural reductant for RsrA

It has been known that only the reduced form of RsrA binds σR, and the reduction of RsrA was achieved by DTT and thioredoxins in vitro (Kang et al., 1999). As MSH is the major thiol buffer in actinomycetes including Streptomyces, and as the level of MSH is tightly interconnected with σR-regulated genes, we examined whether purified MSH reduces RsrA in vitro. The functional reduced status of RsrA was monitored through its ability to form a complex with σR. When the binding reaction was resolved on native polyacrylamide gel by electrophoresis, we found that MSH allowed σR–RsrA complex formation as DTT did, whereas GSH did not (Fig. 7). In contrast to DTT which reduced all disulphide bonds (up to three in oxidized RsrA), treatment with MSH generated a functional reduced form that binds σR in addition to some non-functional form with slower mobility than the oxidized form. The nature of this slow-moving form requires further investigation.

Figure 7.

Formation of the σR–RsrA complex in the presence of MSH. Purified σR and RsrA proteins were incubated in the binding buffer containing no reducing agent (lane 1), 3 mM DTT (lane 2), 2 mM MSH (lane 3), and 10 mM GSH (lane 4) at 30°C for 3 h. Samples were electrophoresed on 12% native PAGE. Band positions for σR, reduced and oxidized forms of RsrA, and σR–RsrA complexes are indicated.


Mycothiol (MSH) found in actinomycetes is regarded as a functional equivalent of glutathione that is found in some bacteria and in nearly all eukaryotes. We have shown in our work that the formation of MSH S-conjugates and their subsequent cleavage by an amidase (Mca) plays a critical part in protecting S. coelicolor cells against MSH-reactive drugs. As mca homologues are often found in antibiotic synthetic gene clusters of streptomycetes, it has been proposed that MSH-dependent detoxification would provide protection against the antibiotics, their electrophilic precursors or oxidized derivatives in antibiotic-producing organisms (Newton and Fahey, 2002). In Mycobacterium smegmatis, the mca mutant showed increased sensitivity to alkylating agents and antibiotics (Rawat et al., 2004). Although mca in S. coelicolor is not linked to antibiotic gene clusters, it has been expected to be involved in detoxification of antibiotics, and our work demonstrated that it is indeed true. The spectrum of protection against antibiotics is not the same between S. coelicolor and M. smegmatis. For example, the mca mutant of S. coelicolor was not sensitive to erythromycin and streptomycin (data not shown), whereas the mca mutant of M. smegmatis was sensitive to these antibiotics. On the contrary, M. smegmatis mutant was not sensitive to lincomycin and could survive even in the presence of over 90 μg ml−1 of lincomycin (Rawat et al., 2004). This different sensitivity could result from various physiological factors including cell wall structure and MSH S-transferases that are yet to be discovered (Rawat and Av-Gay, 2007). Apart from mca, genes known to contribute to drug resistance in actinomyctes include those of wbl (whiB-like) family. The wblC in S. lividans and its orthologue whiB7 in M. tuberculosis are known to confer multiple drug resistance (Morris et al., 2005). Recently, a novel function was suggested for some Wbl proteins as disulphide reductases (Alam et al., 2007). As the presence of Wbl proteins and MSH is confined to actinomycetes, a possible relationship between MSH and Wbl proteins would be an interesting question to be investigated.

The observation that the synthesis and detoxification pathways of MSH are regulated by σR and its antisigma factor RsrA is intriguing. Previous observation suggested that MSH production must be under the control of σR, as in the sigR mutant the level of MSH is reduced (Paget et al., 2001). Our work demonstrated that the first step in MSH-biosynthetic pathway is under direct control of σR, whereas the rest of the steps are also under its control in indirect ways. The detoxification pathway is also under direct control by σR, with mca being the direct target. The environmental signal for σR induction in S. coelicolor has been a subject of discussion. The discovery that non-oxidative alkylating agents, in addition to the thiol-oxidizing diamide, induce the σR regulon led us to hypothesize that a decrease in the amount of MSH could be a natural physiological signal to turn on the σR system. Bacteria in soil environment are exposed to various pollutants and chemical products from animals, plants and insects. Therefore, xenobiotics in soil and antibiotics produced by actinomycetes and other microorganisms could be additional triggers to induce the σR system. Some antibiotics produced by S. coelicolor (actinorhodin, methylenomycin A and calcium-dependent lipopeptide antibiotic) have electrophilic carbon atoms in quinones and epoxide rings. When antibiotics are produced in S. coelicolor, they would be conjugated to MSH, resulting in a perturbation in MSH thiol buffer and a subsequent induction of MSH/Mca detoxification system. More detailed investigations are needed to test this possibility.

The proposal that a decrease in MSH could trigger σR induction is consistent with the observation that the σR-regulon genes are upregulated in MSH-deficient mutants in the absence of other external treatments. According to the MSH-dependent σR induction model we propose in this study, the σR system is induced as soon as MSH is depleted either through oxidation or chemical conjugation, resulting in an increase in the thiol-reducing systems (such as thioredoxins) as well as the MSH biosynthetic and recycling system. The increased thioredoxin and replenished MSH will take part in re-reducing the oxidized RsrA, leading to shut-off of the σR induction response. Even when the MSH level is not restored in due course, the induced thioredoxins could still shut off the σR response by reactivating RsrA as implied from the results in Fig. 5C and D. Therefore, we propose that it is not the absolute MSH level itself, but the initial decrease in MSH that triggers induction of the σR response.

Then, how could the initial depletion of MSH induce σR activation? Under ordinary reducing environment with sufficient MSH, σR is bound by reduced RsrA (Kang et al., 1999). Depletion of MSH inside the cell would result in accumulation of oxidized thiol forms and depletion of reducing equivalents such as NADPH and other reducing thiols such as thioredoxins. This will increase the fraction of oxidized RsrA, and release σR to transcribe its target genes. In other words, the thiol-disulphide redox switch of RsrA can be turned on (oxidized) not only through direct oxidation but also through an indirect path of MSH depletion by conditions that perturb redox homeostasis. The σR regulon contains a glutaredoxin-like protein (mycoredoxin), which is likely to use MSH for its electron donor (Paget et al., 2001). Whether this protein functions in modulating RsrA activity remains to be determined.

The finding that MSH can reduce RsrA in vitro, whereas GSH cannot, adds further support that MSH is a natural modulator of the σR system in vivo. Whether oxidized RsrA is directly reduced by MSH in vivo remains to be resolved, along with the contribution of thioredoxin in reducing RsrA in vivo. Further genetic and biochemical experiments on the role of thioredoxins and mycoredoxin are needed. Apart from these future assignments, the intricate modulation of σR by MSH, and vice versa, certainly reveals an elegant system of genetic regulation which has evolved in this family of bacteria known to be exposed to myriad secondary metabolites from both inside and outside. Whether redox-responsive regulators with sensitive thiols as found in the σR-RsrA system regulate and are regulated by MSH metabolism in a similar way in other actinomycetes is an interesting question for future investigation.

Experimental procedures

Strains and growth media

Streptomyces coelicolor A3(2) M145 was used as the wild-type strain. The disruption mutants of sigR, mshA, mshB, mshC and mshD were used as described previously (Park et al., 2006). Disruption of mca (SCO4967) was carried out by PCR targeting mutagenesis by replacing the wild-type gene in cosmid 2K31 (from John Innes Centre) with apramycin-resistance cassette (Gust et al., 2003). For gene replacement and conjugal gene transfer, Escherichia coli strains BW25113/pIJ790 (Gust et al., 2003) and ET12567/pUZ8002 (Flett et al., 1997) were used respectively. Complementation of mutants with wild-type genes was carried out by introducing cloned genes in the pSET152 plasmid through chromosome integration as described (Park et al., 2006). E. coli DH5α strain was used for routine gene cloning. For growth of S. coelicolor, YEME (yeast extract-malt extract medium) was routinely used for liquid culture, NA (nutrient agar) medium for spotting assay and preparation of genomic DNA, and MS (mannitol soya flour) agar medium for spore preparations (Kieser et al., 2000).

Quantification and preparation of mycothiol

The mycothiol levels were determined by fluorescent labeling as described by Fahey and colleagues (Fahey and Newton, 1987; Newton et al., 1996; Koledin et al., 2002) with minor modifications. Cells grown in YEME liquid medium to mid-exponential phase (OD 0.7–0.8) were sonicated in 40 mM HEPES (pH 7.6) and mixed with equal volume of acetonitrile containing 4 mM mBBr (Fluka). The mixture was incubated at 60°C for 20 min in the dark, and then acidified with 5 μl of 5 N methanesulphonic acid. Following centrifugation, the supernatants were injected into an HPLC column (4.6 × 25 cm Beckman Ultrasphere ODS IP column No. 235335). MSH was eluted with gradients of buffer A (0.25% acetic acid in distilled water titrated to pH 3.6 with NaOH) and buffer B (HPLC grade methanol, Merk). The proportion of buffer B in continuous gradients was as follows; 20% at 0–5 min, 27% at 20 min, 100% at 22 min, 20% at 24 min, 20% at 24–30 min and reinjection. Control samples were treated with final 5 mM NEM at 60°C for 5 min to allow NEM conjugation to the thiol group of MSH, blocking mBBr conjugation. The mBBr-MSH standard was kindly provided by Dr Yossef Av-Gay at the University of British Columbia, Canada. In order to prepare MSH, S. colicolor cells were extracted with acetonitrile (50% final), which was subsequently removed by rotary evaporation. After adjusting pH with Tris base (pH 8.0), the remaining suspension was loaded on thiopropyl Sepharose 6B column (Sigma) and eluted with 20 mM DTT in ammonium bicarbonate buffer as described previously (Unson et al., 1998). Ethyl acetate extraction was repeated more than five times to remove DTT. The purity and concentration of MSH in the preparation was assessed by HPLC, which indicated no detectable traces DTT or other thiol forms.

S1 mapping analysis

Streptomyces coelicolor cells grown in YEME liquid culture to mid-exponential phase (OD 0.7–0.8) were treated with diamide, NEM and mBBr for varying lengths of time. Cells were sonicated in Kirby mix and RNA was isolated through extraction with phenol/chloroform according to a standard procedure (Kieser et al., 2000). For each sample, 100 μg of RNA was analysed with gene-specific DNA probes labelled with [γ-32P]-ATP. Hybridization was carried out at 50°C overnight, followed by S1 nuclease treatment for 1 h. The protected DNA probes were loaded on 5% polyacrylamide gel containing 7 M urea.

In vitro transcription assay

Transcription assays were performed as described previously (Bae et al., 2004). Pre-incubation of RsrA with σR and then with core RNA polymerase was carried out in storage buffer (10 mM HEPES, pH 7.6, 50 mM NaCl, 50% glycerol) with various concentrations of DTT at 30°C for 30 min. The protein solution was then mixed with transcription buffer containing mshAp1 and mca promoter DNAs. For each reaction, 15 pmol of RsrA, 1.5 pmol of σR, 0.75 pmol of core RNA polymerase and 0.15 pmol of DNA template were mixed in 25 μl of reaction. RNA synthesis was initiated by adding 3 μl of substrate mixture containing 2 μCi [α-32P]-CTP and 0.4 mM each of UTP, ATP and GTP. Heparin (3 μl, 0.1 mg ml−1 final concentration) was added after 2 min to prevent reinitiation and the incubation was continued for 5 min before adding 2 μl of cold CTP (1 mM final concentration). After 10 min of incubation, transcripts were precipitated following termination of the reaction, and analysed on 5% polyacrylamide gel containing 7 M urea.

Amidase activity assay

Amidase activity was measured as described previously (Rawat et al., 2004). Triplicate samples (10 ml) of exponentially grown cells were chilled on ice for 20 min, followed by addition of mBBr (0.5 mM final) in acetonitrile and further incubation on ice for 30 min. Excess amount of 2-mercaptoethanol (final 1 mM) was added to scavenge unreacted mBBr. Cells were harvested and extracted with 50% acetonitrile at 60°C for 10 min. After acidification with methanesulphonic acid, cell debris were removed by centrifugation, and the supernatant was analysed by HPLC. To determine the amount of acetylcysteinyl bimane (AcCysmB) in the medium, the culture supernatant was injected to HPLC. The concentration of bimane derivatives was presented as μmol g−1 dried cell weight.

Gel binding assay

Purified σR (4 μM) and RsrA (8 μM) were incubated in 25 μl binding buffer (20 mM HEPES, 50 mM NaCl, 20% glycerol) with or without DTT, NEM, MSH or GSH at 30°C for 30 min (Kang et al., 1999). Samples were loaded on 12% native PAGE which was pre-run for 30 min before loading. The binding and electrophoresis buffers were purged with nitrogen gas before use.


We are grateful to Dr Robert C. Fahey, Dr Yossef Av-Gay and Dr Chang-Jun Cha for their advice on preparing MSH, providing MSH standard sample, and advice on MSH measurement by HPLC analysis respectively. This work was supported by a grant from the Ministry of Science and Technology to J.H. Roe, for the National Research Laboratory of Molecular Microbiology at Seoul National University. J.H. Park was a recipient of the second stage of BK21 fellowship for graduate students to Life Sciences at SNU.