SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Alternate sigma factors provide an effective way of diversifying bacterial gene expression in response to environmental changes. In Streptomyces coelicolor where more than 65 sigma factors are predicted, σR is the major regulator for response to thiol-oxidative stresses. σR becomes available when its bound anti-sigma factor RsrA is oxidized at sensitive cysteine thiols to form disulphide bonds. σR regulon includes genes for itself and multiple thiol-reducing systems, which constitute positive and negative feedback loops respectively. We found that the positive amplification loop involves an isoform of σRR′) with an N-terminal extension of 55 amino acids, produced from an upstream start codon. A major difference between constitutive σR and inducible σR′ is that the latter is markedly unstable (t1/2 ∼ 10 min) compared with the former (> 70 min). The rapid turnover of σR′ is partly due to induced ClpP1/P2 proteases from the σR regulon. This represents a novel way of elaborating positive and negative feedback loops in a control circuit. Similar phenomenon may occur in other actinomycetes that harbour multiple start codons in the sigR homologous gene. We observed that sigH gene, the sigR orthologue in Mycobacterium smegmatis, produces an unstable larger isoform of σH upon induction by thiol-oxidative stress.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

In order to regulate gene expression at the level of transcriptional initiation, bacteria utilize transcription factors, multiple sigma factors and a number of small regulatory metabolites. Sigma factors of the σ70 family provide specificity to the associated RNA polymerase (RNAP), through DNA-binding regions of the conserved domains 2 and 4, to recognize −10 and −35 elements of promoters respectively (Gruber and Gross, 2003; Paget and Helmann, 2003). Most bacteria utilize a primary sigma factor (group 1) that recognizes housekeeping promoters, and additional sigma factors (groups 2, 3 and 4) whose promoter-binding activity is regulated in response to physiological and environmental cues (Helmann, 2002). Modulation of sigma factor activity is achieved in most cases by their cognate anti-sigma factor, and by post-translational modifications (Campbell et al., 2008).

Group 4 sigma factors, also called ECF (extracytoplasmic function) sigma factors (Lonetto et al., 1994; Helmann, 2002), constitute about 60% of all predictable σ70 family members (Campbell et al., 2007). It is predicted that about one-third of all ECF sigma factor genes are located next to genes encoding anti-sigmas which share a structurally conserved sigma-binding domain (Campbell et al., 2007). The best-studied example of ECF family sigma and anti-sigma pair is σE-RseA in Escherichia coli that responds to envelope stress (De Las Penas et al., 1997; Missiakas et al., 1997; Hayden and Ades, 2008). In response to oxidative stress, σE-ChrR in Rhodobacter sphaeroides (Anthony et al., 2004) and σR-RsrA in Streptomyces coelicolor (Kang et al., 1999; Park and Roe, 2008) have been revealed in some detail to respond to singlet oxygen and thiol-oxidative stress respectively.

In addition to anti-sigma factors, different types of modulation can be applied directly on sigma factors. For example, their stability can be modulated as observed in the degradation of starvation-specific σS, where phosphorylatable adaptor protein RssB binds directly to σS and targets it to ClpXP, an ATP-dependent protease (Zhou et al., 2001). Phosphorylation can occur as observed in attenuating σH in E. coli (Klein et al., 2003), AlgU in Pseudomonas aeruginosa (Schurr et al., 1995) and SigH in Mycobacterium tuberculosis (Park et al., 2008). Processing of an inactive precursor has been observed in σK of Bacillus subtilis, where the N-terminal 20 amino acids prevent σK from binding to RNAP and sequester the pro-σK to the membrane (Kroos et al., 1999). In other cases, multiple initiation codons have been observed to generate multiple isoforms of σH in S. coelicolor, in response to osmotic and heat stresses (Viollier et al., 2003). SigH of 51 and 52 kDa sizes are synthesized from upstream start codons in early phase of growth. Upon stress and in the stationary phase, SigH of 37 kDa is translated from a downstream start codon. In the stationary phase, SigH-σ51/52 is cleaved by a stationary phase-specific protease, yielding 34 and 38 KDa isoforms. This proteolytic control of σH is not related to the stress response but closely co-ordinated to the developmental process, even though its role is not known.

Streptomyces coelicolor A3(2) serves as a model organism to study morphological and physiological differentiation which promotes the production of a wide variety of secondary metabolites that include antibiotics (Kieser et al., 2000; Hopwood, 2007). It contains a large chromosome of about 8.67 Mbp (Bentley et al., 2002), encoding more than 700 transcriptional regulators (Rodionov, 2007), and 64 sigma factors (Bentley et al., 2002; Hahn et al., 2003). Among 64 chromosomally encoded sigma factors, 50 are predicted to belong to group 4 (ECF). σR was initially identified as an abundant sigma factor associated with RNAP (Kang et al., 1997), and later found to control response against disulphide stress by inducing genes for itself, thioredoxins, mycothiol metabolism, proteases, and others that constitute a disulphide stress response in S. coelicolor (Paget et al., 1998, 2001a; Park and Roe, 2008). The activity of σR is regulated by its anti-sigma factor RsrA, which contains a conserved HX3CX2C (HCC) motif that contributes in binding zinc and forming disulphide bonds (Kang et al., 1999; Paget et al., 2001b). In its reduced state, RsrA binds σR and thus limits its availability. Upon oxidative stress, RsrA is inactivated through forming intramolecular disulphide bond(s), releasing σR and zinc (Li et al., 2003; Bae et al., 2004). σR then binds to core RNAP and directs transcription of its target genes. Induction of its own gene constitutes a positive amplification loop to ensure maximal response. The oxidized RsrA returns to its reduced state by the action of thioredoxins and mycothiol systems whose synthesis is induced by σR, constituting a negative feedback loop to turn off the response (Kang et al., 1999; Park and Roe, 2008). This system gives a fine example of sensing redox changes through thiol-disulphide switches that involve multiple cysteines, often co-ordinated with zinc. Analogous examples are found in oxidation-sensitive heat shock protein Hsp33 in E. coli (Ilbert et al., 2006), peroxide-sensing Orp1–Yap1 in Saccharomyces cerevisiae (D'Autreaux and Toledano, 2007), and Keap1–Nrf2 system that responds to electrophiles and ROS in higher eukayotes (Kensler et al., 2007).

In this study we demonstrate that the σR-RsrA regulatory circuit contains additional regulatory modules that involve an unstable isoform of σR. We discovered that there exist two forms of σR, a stable constitutively synthesized one and an unstable induced one with an N-terminal extension of 55 amino acids. We present an additional negative feedback loop of proteases that turns off the σR-mediated response, and propose that this type of regulation is likely to occur in other actinomycetes that include mycobacteria.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

A larger variant of σRR′) appears upon oxidative stress

Even though transcription from the sigR promoter (p2) dramatically increases upon thiol-oxidative stress, the amount of 31 kDa σR protein in cell extracts has not been observed to change, contrary to expectations. When the immunoblot experiments were performed with cell extracts, the presence of non-specific cross-reactive bands impaired accurate interpretation of the data. We therefore partially purified RNAP holoenzyme from cells treated with thiol oxidant diamide and examined the amount of σR associated with RNAP. RNAP holoenzymes were purified from S. coelicolor J1980 strain that contains a histidine-tagged rpoC gene (Babcock et al., 1997). Instead of loading directly on Ni-NTA column, cell extracts were precipitated with polymin P and ammonium sulphate for effective concentration of RNAP (Burgess and Jendrisak, 1975). This procedure allowed a dramatic increase in the recovery and activity of RNAP.

Western blot analysis of RNAP preparations from cells untreated or treated with 0.5 mM diamide for 20, 40 and 80 min produced an unexpected result (Fig. 1). A specific σR protein band of expected 31 kDa size was observed unchanged, whereas a larger band appeared transiently upon induction. This inducible protein band was also visible in cell extract, migrating closely with the non-specific cross-reacting band, which previously masked the detection of this RNAP-associated inducible protein.

image

Figure 1. Expression of σR and σR-related inducible band after diamide treatment. A. The level of σR was analysed by Western blotting. Exponentially growing cells were treated with 0.5 mM diamide (DA) and harvested at indicated time points after treatment. Crude cell extracts (20 μg) or purified RNAP (4 μg) were analysed by using polyclonal antibody against σR. B. The same samples were resolved on 9% SDS-PAGE, and the proteins were stained with Coomassie Brilliant R-250.

Download figure to PowerPoint

The identity of this inducible protein was explored by two-dimensional (2-D) gel electrophoresis of RNAP holoenzyme preparations, followed by MALDI-TOF mass spectrometry (Fig. 2). Silver staining of the 2-D gel demonstrated the presence of inducible protein spots with an expected size, but with multiple isoelectric point (pI) values (spots 1–3, Fig. 2A). When the 2D gel was further analysed by Western blot, the similar-sized multiple spots all turned out to react with anti-σR antibody (Fig. 2B). We found that even σR itself is present as multiple forms with different pI values, implying some post-translational modifications such as phosphorylation as observed for SigH in M. tuberculosis (Park et al., 2008). The induced spots (spots 1–3) and the biggest unchanged spot (spot 4) were cut out of the gel and subjected to tryptic peptide finger printing by MALDI-TOF mass analysis. Spots 1, 2 and 4 gave meaningful results by Mascot analysis (http://www.matrixscience.com), and all of them turned out to be coded by the sigR gene. Therefore, we named the larger induced gene product as σR′ in contrast to the smaller σR, which is the major form under non-stressed condition and exists constitutively (lanes 1 and 5, Fig. 1).

image

Figure 2. Identification of σR and σR′ by 2-D gel electrophoresis and MALDI-TOF peptide mapping. Purified RNAP samples were separated by isoelectric focusing and then by SDS-PAGE. Gels were stained with silver nitrate (A) or transferred to nitrocellulose membrane for Western blot analysis with anti-σR antibody (B). Gel spots of σR and the diamide-induced protein (σR′) indicated by arrows were cut out and subjected to peptide mapping by MALDI-TOF mass spectrometry. Spots 1, 2 and 4 matched significantly with σR.

Download figure to PowerPoint

The primary structure of σR′

Inspection of nucleotide sequences near the constitutive (p1) and inducible (p2) promoters of the sigR gene revealed that at least three in-frame start codons (S1, S2, S3) are present between p1 and p2 promoters, which can extend the N-terminus of σR by 37, 55 and 70 amino acids respectively (Fig. 3). Therefore, induction of transcription from p2 promoter is able to produce larger σR proteins with 37–70 additional amino acids at the N-terminus. There is a good ribosome binding site in front of the second start codon (S2).

image

Figure 3. Distribution of multiple in-frame start codons between inducible (p2) and constitutive (p1) promoters of the sigR gene. Between sigRp1 and p2 promoters, three start codons (S1, S2, S3; boxed) are found in-frame with the previously annotated one (S0; boxed). Transcription start sites from the sigRp1 and p2 promoters were indicated with bent arrows. Upstream start codons were changed to stop codons (TAG or TGA) in mutant strains MK11, 12 and 13. The−35 and −10 regions of the sigRp1 and p2 promoters were marked in bold letters. Putative ribosomal binding sequences (RBS) preceding S0 and S2 start codons are underlined.

Download figure to PowerPoint

In order to find out which upstream start codon could have produced σR′, we changed each start codon to a stop by site-directed mutagenesis. The mutated sigR genes with flanking promoter regions and the downstream rsrA gene were introduced into the chromosome of the ΔsigRrsrA mutant (MK1) through the att site, to generate strains MK11, 12 and 13 that have mutations in S1, S2 and S3 codons respectively. Western blot analysis of cell extracts prepared at 0, 40, 80 and 120 min after diamide treatment revealed that MK11 (S1 to stop) and MK12 (S2 to stop) did not produce any inducible σR′, whereas MK13 (S3 to stop) produced it as observed in the wild type (Fig. 4). This indicates that neither S3 nor S1 was used as a start codon. If S1 was used as a start codon, MK12 should also behave like a wild type. Therefore it is most likely that S2 with a good Shine–Dalgarno sequence is used as a start codon for σR′, adding 55 amino acids to the N-terminus of σR. We confirmed the production of the right-sized σR′ in E. coli by expressing recombinant pET-sigR plasmid whose sigR coding region starts from S2 codon (data not shown).

image

Figure 4. Identification of S2 as the start codon for σR′. Either wild type or mutant sigR gene with substitution of upstream start codons (S1, S2, S3) with stop codons was introduced into ΔsigRrsrA (MK1) strain through the att site to create MK10 (no mutation, wild type), MK11 (S1 to stop), MK12 (S2 to stop) and MK13 (S3 to stop) strains. Cells grown exponentially in YEME were treated with 0.5 mM diamide for different lengths of time (min), and cells extracts were subjected to Western blot analysis with anti-σR antibody.

Download figure to PowerPoint

Contribution of σR′ on the induction of σR regulon by oxidative stress

N-terminal addition of 55 amino acids is not expected to change the specificity for promoter recognition. We performed in vitro transcription with σR′ and verified that it produces σR-specific transcripts from the sigRp2 promoter as σR does (Fig. S1). In order to find out how much σR′ contributes to the induction of SigR regulon, we examined the induction profile of three representative σR-target genes, sigR, trxB and trxC, in the mutant strain (MK11; S1 to stop) that does not produce σR′. S1 mapping analysis indicated that the extent of induction was reduced to about half to one-third of the level observed in the wild type (MK10) (Fig. 5). In MK13 mutant strain where normal σR′ is synthesized, the induction profile of σR-target genes were indistinguishable from the wild type (data not shown). Therefore, production of σR′ from sigRp2-driven transcripts contributes significantly to the full induction of its target genes.

image

Figure 5. Contribution of σR′ to the diamide induction of σR-target genes. Induction of transcripts from σR-target promoters (sigRp2, trxBp1 and trxC) was examined in MK10 (WT; white bar) and MK11 (S1 to stop mutation; black bar) strains following diamide treatment. MK10 produces both σR and σR′ whereas MK11 produces only σR. Results of S1 nuclease mapping for each transcript from three independent analyses were quantified and presented with an average induction fold and a standard deviation. Each transcript signal measured by Fuji BAS2500 Image Plate Reader and quantified with Multi-Gauge program was normalized to 23S rRNA band intensity in the same sample (data not shown).

Download figure to PowerPoint

σR′ is rapidly degraded by a component(s) of its regulon

Whereas the level of σR stays constant during oxidative stress response, that for σR′ changes dramatically by a sharp increase and then a decrease (Fig. 1A). We therefore examined the stability of both σR and σR′, following inhibition of translation by an antibiotic hygromycin. Western blot analysis of cell extracts prepared after treating hygromycin for up to 2 h indicated that σR′ is more rapidly degraded than σR with a half-life of less than 25 min whereas σR stays nearly constant with a half-life longer than 2 h (Fig. 6A). Transcripts from p1 and p2 promoters following hygromycin treatment were examined by S1 mapping (Fig. 6B). Intriguingly, the level of both sigRp1 and p2 transcripts increased in the presence of hygromycin. Therefore, in the presence of hygromycin, which stimulates transcription from sigR promoters but inhibits protein synthesis, the instability of σR′ is clearly demonstrated.

image

Figure 6. Stability of σR and σR′ in the cell. A. Western blotting with σR antibody. Wild type (M145) cells were grown to exponential phase in YEME. Hygromycin, an antibiotic that inhibits translation, was added at 0.05 mg ml−1 to the medium, and cells were sampled at 0, 40, 80 and 120 min after hygromycin treatment to prepare cell extracts. Cell extracts of ΔsigRrsrA (MK1) were loaded in parallel to indicate non-specific bands. B. The level of transcripts from sigRp1 and p2 promoters after hygromycin treatment. S1 nuclease protection analysis was performed for the same cell samples used in (A).

Download figure to PowerPoint

We then examined whether the rapid degradation of σR′ is correlated with the induction of its target genes. For this purpose, the blocking of translation by hygromycin was initiated either right after diamide treatment (at 10 min) or at 40 min, allowing partial and full induction of the response respectively. We found that the half-life of σR′ differs extensively depending on the timing of hygromycin chase (Fig. 7). When translation was blocked right after the induction, σR′ became more stable (t1/2 > 50 min) in contrast to the short half-life (t1/2 ∼ 10 min) observed when hygromycin chase started following the full induction of the response. Therefore, it is most likely that some component(s) of the induced response contributed to the degradation of σR′.

image

Figure 7. Dependence of the stability of σR′ on the products of diamide induction. A. Changes in the level of σR and σR′ in extracts of cells treated with diamide (DA) up to 80 min without (left panel) or with hygromycin (0.1 mg ml−1) added at 10 min after diamide treatment (right panel). Western blot analysis was performed with anti-σR antibody. In the right panel, numbers above bar indicate the duration time of hygromycin chase. B. The same procedures were followed as in (A) except that hygromycin was added at 40 min after diamide treatment (right panel) to ensure full induction of the diamide response.

Download figure to PowerPoint

ClpP1P2 proteases regulate the stability of σR′

From genome sequence information, six protease candidates could be retrieved, whose coding genes are preceded by σR-specific promoter consensus sequence predicted by RSAT (regulatory sequence analysis tools, http://rsat.ulb.ac.be/rsat; Table S1). These include Lon, ClpP proteases in addition to PepN that has been proposed previously (Paget et al., 2001b). We examined mutants lacking lon, clpX or clpP1P2 genes for the rate of σR′ degradation, and found that the lon and clpX mutations did not alter the rate (data not shown). On the other hand, the ΔclpP1P2 mutation significantly increased the stability of σR′ (from t1/2 of 11 min to 23 min) (Fig. 8). We confirmed that the p2 promoter of the clpP1P2 genes that contain σR consensus sequence is indeed induced by diamide in σR-dependent manner (Fig. 8B). Therefore, the induced σR′ is degraded at least partly by ClpP1/P2 proteases that are part of the induced response. This adds another negative feedback mechanism to an already reported one that involves sequestration of the sigma factor by an anti-sigma factor re-activated (reduced) by the components of the induced response.

image

Figure 8. Modulation of σR′ stability by ClpP1/P2 protease. A. Organization of the clpP1P2 and clpX genes. The p2 promoter of clpP1 contains σR-specific promoter sequence. B. Induction of clpP1 transcripts by diamide. RNAs were prepared from the wild-type (M145) and ΔsigRrsrA (MK1) strains and S1 nuclease protection analysis was performed for clpP1 transcripts. C. Western blot analysis of σR and σR′ in cell extracts. The wild-type (M145) or ΔclpP1P2 mutant cells were treated with 0.5 mM diamide for up to 85 min. Hygromycin (0.1 mg ml−1) was added at 40 min after diamide treatment to ensure full induction of the diamide stress response. The duration of hygromycin chasing up to 45 min was indicated above the duration time of diamide. The position of a non-specific (NS) band just above σR′ was indicated. D. Quantification of the stability of σR′. The level of σR′ at the start of hygromycin chasing (40 min after diamide treatment) was set as 100%. All the data were normalized to the intensity of the non-specific band (NS).

Download figure to PowerPoint

Inducible unstable isoform of SigR orthologues in other actinomycetes

We examined whether there are any multiple start codons that can produce isoforms of SigR homologues in other actinomycetes. For this purpose, we searched for σR-specific promoter sequence that is followed by a start codon in the upstream of coding region for SigR homologues. We found that in nearly all cases (with two exceptions out of 37; DIP0709, cu0493), the coding region of SigR homologues have multiple in-frame start codons, which can potentially produce multiple isoforms with different N-terminal extension (Fig. S2). Out of 37 genes, we were able to add longer N-terminal extension (longer than 15 amino acids) to the annotated start site of 21 genes. Some conserved residues were found within the N-terminal extension, especially among those from the same genus. Figure 9A demonstrates the presence of such conserved sequences among streptomycetes and mycobacteria. Predicted extensions of N-terminal residues from the annotated start in the public database were marked with underlines. The sequences from Mycobacterium flavescence (Mfav_4698), M. avium (Map3324c) and M. tuberculosis CDC1551 (MT3320) have already been annotated with the extended N-terminus in the genome database.

image

Figure 9. Possible N-terminally extended isoforms of σR orthologues in Mycobacterium and Streptomyces spp. A. Predicted N-terminal amino acid sequences of σR orthologues in some Mycobacterium and Streptomyces species. The sigR orthologue genes that contain σR-specific promoter sequence were collected from fully sequenced genome information of Mycobacterium avium paratuberculosis K.10 (Mavi; MAP3324c), M. bovis AF2122 97 (Mbov; Mb3250c), M. flavescens PYR.GCK (Mfla; Mflv_4689), M. marinum M (Mmar; MMAR_1334), M. smegmatis MC2-155 (Msme; MSMEG_ 1914), M. tuberculosis CDC1551 (Mtb-CDC; MT3320), M. ulcerans Agy99 (Mulc; MUL_2545), M. vanbaalenii PYR.1 (Mvan; Mvan_ 1778), M. tuberculosis H37Rv (Mtb-Rv; Rv3223c), Streptomyces avermitilis MA.4680 (Save; SAV3038), S. coelicolor A3(2) (Scoe; SCO5216) and S. griseus NBRC 13350 (Sgri; SGR_2306). Possible N-terminal extension starting from upstream in-frame start codons was underlined, with the annotated start codons in the genome database in boxes. When multiple start codons exist, the one closest to the σR-specific promoter sequence was chosen. Sequences were aligned by Align X program provided in Vector NTI package (Advance 10.3.0). B. Existence of diamide-inducible unstable isoform of σH in M. smegmatis. The wild type (mc2-155) or ΔsigHrshA mutant M. smegmatis (a kind gift from T. Song) cells were treated with diamide (DA; 0.5 mM) for 0, 30 and 60 min. Cell extracts were prepared and analysed by Western blotting with antibody against σH from M. tuberculosis. σH and possible σH′ were indicated as well as a non-specific (NS) band.

Download figure to PowerPoint

We then examined whether mycobacterium employ similar mechanism of regulation. For this purpose, we chose SigH (an orthologue of SigR) from Mycobacterium smegmatis, and monitored its change by Western blot. Figure 9B demonstrates that SigH from M. smegmatis indeed is produced in two forms: a constitutively produced smaller form and an inducible larger form with shorter half-life. This closely resembles what is observed in S. coelicolor, suggesting that similar mechanism of regulation may exist in other bacteria, especially in actinomycetes.

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Rapid partial purification of RNAP by combining traditional methods (polymin P and ammonium sulphate precipitations; Burgess and Jendrisak, 1975; Hahn et al., 2003) with nickel affinity chromatography through histidine-tagged β′ (RpoC) subunit (Babcock et al., 1997; Kieser et al., 2000) enabled us to compare RNAP preparations from different cell populations. Following thiol-oxidative stress, we found that a larger isoform of RNAP-associated σR appeared shortly after the stress, and that this species is produced from the upstream in-frame start codon that becomes available by the production of sigRp2 transcript (Fig. 3).

Using multiple start codons is not a rare strategy in nature. Among sigma factors, σH of S. coelicolor has been reported to use two independent start codons depending on growth phases (Viollier et al., 2003). It has also been reported that M. tuberculosisσE uses three alternative start codons, whose availability depends on the two promoters that respond to either surface or oxidative stresses (Donàet al., 2008). However, any functional difference among the isoforms of these sigma factors has not been reported. In this study, we demonstrated that σR exists in two forms: a constitutive stable one and an induced unstable one, the latter containing longer N-terminal extension by 55 amino acids.

Group 1 sigma factors that play a primary housekeeping function (Helmann, 2002) contain unstructured N-terminal extension called domain 1 (region 1.1) preceding the structured and conserved domains 2, 3 and 4 (Lonetto et al., 1992; Gruber and Gross, 2003). For E. coli RpoD (σ70), region 1.1 hinders binding of sigma factor to the promoter sequence by itself in the absence of core RNAP. Group 4 sigma factors, of which σR of S. coelicolor is a member, contain only domains 2 and 4 that recognize −10 and −35 promoter elements respectively (Lonetto et al., 1994; De Las Penas et al., 1997; Helmann, 2002; Campbell et al., 2007). Inspection of N-terminal extension in front of the conserved domain 2 of σR′ did not reveal any significant sequence similarity with region 1.1 of group 1 sigma factors. Addition of 55 amino acids in σR′ rendered this isoform more susceptible to protease attack, without compromising interaction with core RNAP significantly.

The stability of σR′ is regulated partly by ClpP1/P2 proteases. Production of these proteases is induced from a single operon by σR, and contributes to turn off the stress response, constituting a negative feedback regulatory loop. Considering that σR′ is still degraded to some extent in the absence of ClpP1/P2, contribution from other proteases, whether dependent on σR or not, still remains. It is possible that some signal in the extended N-terminal region of σR′ may serve to target σR′ to the ATPase partner (AAA+ proteins) of ClpP1/P2. Or it is also possible that the N-terminal extension somehow alters the structure to expose the targeting sequence. In S. coelicolor, four Clp ATPases (one ClpX and three ClpCs) are predicted from the genome (Bellier and Mazodier, 2004). Disruption of ClpX did not change the stability of σR′ significantly, suggesting involvement of ClpC proteins in degradation pathway. The N-terminal sequence of σR′ does not fit to the typical destabilizing residues of the N-end rule well known in E. coli (Mogk et al., 2007). How the extended N-terminal region of σR′ contributes to rapid degradation awaits further studies.

Production of an induced isoform of σR provides a modified positive amplification loop in the regulatory circuit of σR-mediated thiol-oxidative stress response. Its degradation by the components of its own regulon provides a second negative feedback loop in the control circuit (Fig. 10). In the absence of exogenous disulphide stress, the sigR coding region produces two forms of SigR product: σR from p1 transcript and σR′ from p2 transcript. In cell extract, as well as in RNAP-bound state, the level σR′ under unstressed condition is usually lower than that of σR. RsrA can bind both σR and σR′. The sensing of thiol oxidative stress occurs through disulphide bond formation in the anti-sigma factor RsrA and a subsequent dissociation of σR and σR′. The released σR and σR′ then induce expression of SigR-target genes that include sigR gene to produce more SigR in the form of σR′ (positive loop; P). Among the induced gene products, thiol reducers such as thioredoxin and mycothiol systems reduce RsrA to sequester σR again to shut off the response (negative loop N1), and proteases degrade highly amplified σR′ (negative loop N2), whereas various other members constitute cellular response reaction towards thiol oxidation (Kang et al., 1999; Paget et al., 2001b; Park and Roe, 2008).

image

Figure 10. The regulatory loop involving σR′ in the induction of SigR regulon. In the absence of exogenous thiol-oxidative stress, the sigR coding region produces two forms of SigR product: σR from p1 transcript and σR′ from p2 transcript. The amount of σR′ under unstressed condition is usually lower than that of σR. Under reducing environment as occurs normally in the cell, the reduced RsrA binds σR and σR′ and inhibits SigR-directed transcription. Upon oxidative stress, disulphide bonds are formed in RsrA and SigR is released from the complex. The released σR and σR′ direct transcription of target genes (SigR regulon) that includes those for itself (from sigRp2 promoter), several thiol reducing systems (trxBA, trxC, mshA, mrx etc.), proteases (clpP1/P2, lon, pepN etc.). Induction of sigR gene by released σR and σR′ constitutes the initial positive amplification loop (P). The induced σR′ keeps inducing itself, constituting the major part of the positive amplification loop (P). The induced thiol reducers contribute to reducing RsrA which then binds both σR and σR′ and turns off the response. This constitutes a negative feedback loop (N1). An additional feedback regulatory loop ensures rapid turn-off of the response, by degrading over-produced σR′ by induced proteases (N2).

Download figure to PowerPoint

Overall, this system provides a novel example of utilizing an isoform of a single alternative sigma factor, which enables an effective turn-off of the highly amplified response, by providing an additional negative regulatory loop. If the production of stable σR was amplified, more reducing power would be required to restore sufficient amount of active anti-sigma factors to turn off the response, and cells are also under heavier burden in order to avoid adventitious outcome of prolonged stress response. Similar regulatory mechanism appears to exist in other bacteria, especially among actinomycetes, which contain closely related sigR-like genes with σR-specific promoter consensus sequence and multiple start condons upstream of the conserved sigma factor domains. We were able to observe a similar phenomenon occurring in M. smegmatis for σH that is the closest homologue of σR in mycobacteria.

Additional mechanism of regulation that employs modification of σR and its anti-sigma factor by phosphorylation or other reactions was hinted by the presence of multiple σR spots with different pI values (Fig. 2). The nature and role of these modifications await further investigation. Whether these multiple σR and σR′ proteins arose by translational stuttering frequently observed under extreme starvation condition is not certain (Parker et al., 1978). However, the observation that even unstressed cells produced multiple spots of σR and σR′ (Fig. 2B) suggests that thiol-oxidative stress condition we applied in our study is not likely to be the main cause of such modifications.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Bacterial strains and culture conditions

Streptomyces coelicolor A3(2) strain M145 was grown in YEME liquid medium containing 5 mM MgCl2·6H2O and 10% sucrose at 30°C by inoculating spore suspension (Kieser et al., 2000). E. coli DH5α was used for all initial transformation and propagation of plasmids. E. coli ET12567, a non-methylating strain harbouring pUZ8002 to supply the donor function, was used for conjugation. E. coli cells were grown in LB (1% tryptone, 0.5% yeast extract and 1% NaCl). E. coli BW25113 with pIJ790 was used as recommended (Gust et al., 2002). M. smegmatis wild type-strain mc2-155 was grown in Middlebrook 7H9 broth (Sareen et al., 2002) supplemented with 0.05% Tween 80 and 0.4% glucose at 37°C.

Construction of mutant strains and complementation

The ΔsigRrsrA disruptant (MK1) in M145 background was obtained through PCR-targeting mutagenesis system, by replacing the sigRrsrA coding region with an apramycin resistance cassette (Gust et al., 2002). The disruptant was confirmed by PCR and Southern hybridization. Either the wild type or mutant sigR gene with nonsense substitutions in the upstream start codons (S1, S2, S3) was introduced into the chromosome of the ΔsigRrsrA strain (MK11) via the att site, using pSET152H-based recombinant plasmids carrying wild-type or mutated sigRrsrA sequence. The resulting strains were named MK10 for wild type, and MK11, 12, 13 for mutations in S1, S2 and S3 start codons respectively. The clpP1P2, clpX and lon genes in M145 strain were disrupted by PCR targeting. The cosmids harbouring ΔclpP1P2 and ΔclpX disruption were provided by Mark J. Buttner at John Innes Centre, UK.

Plasmids

The 1.7 kb fragment containing hygromycin phosphotransferase (hyg) gene (Kieser et al., 2000) was cloned into the sphI site of pSET152 to yield pSET152H. DNA fragment from −264 to +1190 nt relative to the previously annotated sigR start codon, containing sigRp2, p1 promoters and the entire sigRrsrA operon, was amplified by PCR with primers sigRrsrA5 and sigRrsrA3, using SC7E4 cosmid (from John Innes Centre, UK) as a template. The sigRrsrA5 and sigRrsrA3 primers are mutagenic primers containing a HindIII and a BamHI sites respectively. The amplified DNA cut with HindIII and BamHI was cloned into pUC18, yielding pMK10. Derivatives of pMK10 (pMK11, pMK12, pMK13) were constructed by site-directed mutagenesis. Site-directed mutagenesis was performed using the QuickChange site-directed mutagenesis kit from Stratagene. The sticky ends of the 1.45 kb HindIII/BamHI fragment of pMK10 were made blunt ends by Klenow. The 1.45 kb fragment was inserted into the EcoRV site of pSET152H yielding pMKH10. In similar way, the 1.45 kb HindIII/BamHI fragments of pMK11, 12, 13 were cloned into pSET152H, yielding pMKH 11, 12, 13 respectively.

S1 nuclease mapping

Exponentially growing cells were treated with 0.5 mM diamide or 0.05–0.1 mg ml−1 hygromycin for designated length of duration. Harvested cells were re-suspended in Kirby mix and disrupted by sonication. RNA isolation and S1 nuclease mapping was performed according to a standard procedure (Kieser et al., 2000). PCR products containing sigR (−265 to +80 nt relative to the sigR start condon), trxB (−318 to +25 nt relative to the trxB start condon), trxC (−432 to +118 nt relative to the trxC start condon), clpP1 (−520 to +60 nt relative to the clpP1 start condon) promoter regions were used for probes. RNA (25–100 μg) was hybridized at 50°C with gene-specific probes labelled with [γ-32P]-ATP. S1 nuclease was treated for 1 h and the protected DNA probes were loaded on 5% polyacrylamide gel containing 7 M urea. The signal was detected and quantified by BAS-2500 (Fuji film).

Purification of RNAP from J1980

RNA polymerase was purified from cell pellets according to the procedures developed for the purification of J1980 RNAP with some modifications (Burgess and Jendrisak, 1975; Babcock et al., 1997; Hahn et al., 2003). All purification steps were carried out at 4°C. J1980 cells were disrupted with ultrasonicator (Sonics and Materials) and re-suspended in lysis buffer [20 mM Tris-HCl (pH 7.9), 10% (v/v) glycerol, 5 mM EDTA, 0.1 mM DTT, 10 mM MgCl2, 1 mM PMSF, 0.15 M NaCl]. The cell suspension was treated with 1 mg ml−1 lysozyme for 30 min on ice and disrupted by sonication at 25% amplitude until the viscosity of the lysate was greatly reduced. The homogenates were clarified by centrifugation at 12 000 r.p.m. for 30 min. A 5% (v/v) solution of PEI (pH 7.9, polymin P, Sigma) was added slowly with thorough stirring to a final concentration of approximately 0.3%. After stirring for 5 min, the mixture was centrifuged at 12 000 r.p.m. for 5 min and the supernatant discarded. The drained pellet was crushed with plastic homogenizer and re-suspended in TGED buffer [10 mM Tris-HCl (pH 7.9 at 4°C), 0.1 mM EDTA, 0.1 mM DTT, 10% glycerol] containing 0.5 M NaCl. The suspension was centrifuged at 12 000 r.p.m. for 5 min in a microcentrifuge and the supernatant was discarded. The washed pellet was again crushed and re-suspended in TGED buffer containing 1 M NaCl. The suspension was centrifuged at 12 000 r.p.m. for 10 min and the supernatant collected (1 M NaCl eluate). Sixty-six per cent ammonium sulphate solution (pH 7.9) was added to 1 M NaCl eluate with stirring. After 30 min, it was centrifuged for 30 min at 12 000 r.p.m. and the drained pellet was dissolved in buffer I [10 mM Tris-HCl pH 7.9, 0.1 M NaCl, 1 mM β-mercaptoethanol, 5% (v/v) glycerol] containing 2.5 mM imidazole (Kieser et al., 2000) and applied to Ni-NTA column (Novagen) which had been previously equilibrated with buffer I containing 2.5 mM imidazole. After washing with buffer I containing 20 mM imidazole, proteins were eluted with buffer I containing 300 mM imidazole.

Western blot analysis

Harvested cells were re-suspended in lysis buffer and disrupted by sonication with ultrasonicator (Sonics and Materials). The suspension was centrifuged to obtain total cell extract from the supernatant. The protein concentration was determined using Bradford reagent solution (Bio-Rad) using BSA as a standard. Twenty-five micrograms of crude cell extract or 5 μg of purified RNAP was resolved on 9% SDS-PAGE. Following electrophoresis, the gel was soaked in transfer buffer [25 mM Tris, 192 mM glycine, 20% (v/v) methanol] for 10 min, and then electrotransferred to nitrocellulose membrane Protran (Schleicher & Schuell) at 170 mA for 60 min in Trans-Blot Cell (Bio-Rad). Membrane was blocked in Tris-buffered saline buffer containing 0.1% Triton X-100 (TBST) supplemented with 0.5% BSA, for more than 1 h. The blocked membrane was incubated with antibody diluted in the same buffer for 1 h, and then washed with TBST for 10 min twice. Washed membrane was incubated with anti-mouse IgG secondary antibody diluted to 1:5000 in TBST, and washed with TBST twice for 10 min each. Signal was detected using Western ECL detection system (Amersham Life Science) and LAS-3000 (Fuji film).

Two-dimensional PAGE

Exponentially growing J1980 cells were treated with 0.5 mM diamide (DA). RNAP purified from diamide-treated or control cells were further clarified by 2-D Clean-Up Kit (Amersham Biosciences). The clarified RNAP (100 μg) were resolved through ImmobilineTM DryStrip gels (pH 4–7, 13 cm, Amersham Biosciences) for the first dimensional isoelectric focusing as recommended by the manufacturer. Gel strips were then equilibrated in SDS equilibration buffer (50 mM Tris-HCl, pH 8.8, 6 M urea, 30% glycerol, 2% SDS) and subjected to second dimensional electrophoresis on a 12.5% SDS-acrylamide gel. Following electrophoresis, gels were stained with silver nitrate as recommended by manufacturer, or subjected for Western analysis.

MALDI-TOF mass spectroscopy

The gel spots visualized by silver staining were excised, followed by destaining and reduction/alkyation with 10 mM DTT and 50 mM iodoacetamide (Shevchenko et al., 1996). Following dehydration of gel slices with 100% acetonitrile, two volumes of freshly prepared trypsin (20 ng ml−1 in 25 mM NH4HCO3) were added and incubated at 37°C overnight. Peptides were extracted with 5% trifluoroacetic acid/50% acetonitrile solution and went through ZipTips containing C18 resin (Millipore). The washed peptides were eluted with saturated matrix solution (α-hydroxy-4-cinnamic acid in 60% acetonitrile, 0.1% trifluoroacetic acid). Mass analysis was carried out on a Voyager-DETM STR spectrometer (Applied Biosystems). External and internal calibrations were performed with autolytic trypsin peptides. For data analysis, mascot peptide mass fingerprint program (http://www.matrixscience.com) was used.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

We are grateful to Yann S. Dufour (UW-Madison, WI, USA) for help in finding sigR homologous genes in actynomycetes, Dr. Mark J. Buttner and Ngat Tran for providing ΔclpP1/P2,ΔclpX mutated cosmid, Dr Taeksun Song for M. smegmatis strains. This work was funded by a KOSEF grant for NRL (0427-20080009) to J.-H. Roe.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information
  • Anthony, J.R., Newman, J.D., and Donohue, T.J. (2004) Interactions between the Rhodobacter sphaeroides ECF sigma factor, sigma(E), and its anti-sigma factor, ChrR. J Mol Biol 341: 345360.
  • Babcock, M.J., Buttner, M.J., Keler, C.H., Clarke, B.R., Morris, R.A., Lewis, C.G., and Brawner, M.E. (1997) Characterization of the rpoC gene of Streptomyces coelicolor A3(2) and its use to develop a simple and rapid method for the purification of RNA polymerase. Gene 196: 3142.
  • Bae, J.B., Park, J.H., Hahn, M.Y., Kim, M.S., and Roe, J.H. (2004) Redox-dependent changes in RsrA, an anti-sigma factor in Streptomyces coelicolor: zinc release and disulfide bond formation. J Mol Biol 335: 425435.
  • Bellier, A., and Mazodier, P. (2004) ClgR, a novel regulator of clp and lon expression in Streptomyces. J Bacteriol 186: 32383248.
  • Bentley, S.D., Chater, K.F., Cerdeno-Tarraga, A.M., Challis, G.L., Thomson, N.R., James, K.D., et al. (2002) Complete genome sequence of the model actinomycete Streptomyces coelicolor A3(2). Nature 417: 141147.
  • Burgess, R.R., and Jendrisak, J.J. (1975) A procedure for the rapid, large-scall purification of Escherichia coli DNA-dependent RNA polymerase involving Polymin P precipitation and DNA-cellulose chromatography. Biochemistry 14: 46344638.
  • Campbell, E.A., Greenwell, R., Anthony, J.R., Wang, S., Lim, L., Das, K., et al. (2007) A conserved structural module regulates transcriptional responses to diverse stress signals in bacteria. Mol Cell 27: 793805.
  • Campbell, E.A., Westblade, L.F., and Darst, S.A. (2008) Regulation of bacterial RNA polymerase sigma factor activity: a structural perspective. Curr Opin Microbiol 11: 121127.
  • D'Autreaux, B., and Toledano, M.B. (2007) ROS as signalling molecules: mechanisms that generate specificity in ROS homeostasis. Nat Rev Mol Cell Biol 8: 813824.
  • De Las Penas, A., Connolly, L., and Gross, C.A. (1997) SigmaE is an essential sigma factor in Escherichia coli. J Bacteriol 179: 68626864.
  • Donà, V., Rodrigue, S., Dainese, E., Palu, G., Gaudreau, L., Manganelli, R., and Provvedi, R. (2008) Evidence of complex transcriptional, translational, and posttranslational regulation of the extracytoplasmic function sigma factor sigmaE in Mycobacterium tuberculosis. J Bacteriol 190: 59635971.
  • Gruber, T.M., and Gross, C.A. (2003) Multiple sigma subunits and the partitioning of bacterial transcription space. Annu Rev Microbiol 57: 441466.
  • Gust, B., Kieser, T., and Chater, K.F. (2002) Redirect Technology: PCR-targeting System in Streptomyces coelicolor. Norwich: John Innes Centre.
  • Hahn, M.Y., Bae, J.B., Park, J.H., and Roe, J.H. (2003) Isolation and characterization of Streptomyces coelicolor RNA polymerase, its sigma, and antisigma factors. Methods Enzymol 370: 7382.
  • Hayden, J.D., and Ades, S.E. (2008) The extracytoplasmic stress factor, sigmaE, is required to maintain cell envelope integrity in Escherichia coli. PLoS ONE 3: e1573.
  • Helmann, J.D. (2002) The extracytoplasmic function (ECF) sigma factors. Adv Microb Physiol 46: 47110.
  • Hopwood, D.A. (2007) Streptomyces in Nature and Medicine. Oxford: Oxford University Press.
  • Ilbert, M., Graf, P.C., and Jakob, U. (2006) Zinc center as redox switch – new function for an old motif. Antioxid Redox Signal 8: 835846.
  • Kang, J.G., Hahn, M.Y., Ishihama, A., and Roe, J.H. (1997) Identification of sigma factors for growth phase-related promoter selectivity of RNA polymerases from Streptomyces coelicolor A3(2). Nucleic Acids Res 25: 25662573.
  • Kang, J.G., Paget, M.S., Seok, Y.J., Hahn, M.Y., Bae, J.B., Hahn, J.S., et al. (1999) RsrA, an anti-sigma factor regulated by redox change. EMBO J 18: 42924298.
  • Kensler, T.W., Wakabayashi, N., and Biswal, S. (2007) Cell survival responses to environmental stresses via the Keap1–Nrf2–ARE pathway. Annu Rev Pharmacol Toxicol 47: 89116.
  • Kieser, T., Bibb, M.J., Buttner, M.J., Chater, K.F., and Hopwoodk, D.A. (2000) Practical Streptomyces Genetics. Norwich: The John Innes Foundation.
  • Klein, G., Dartigalongue, C., and Raina, S. (2003) Phosphorylation-mediated regulation of heat shock response in Escherichia coli. Mol Microbiol 48: 269285.
  • Kroos, L., Zhang, B., Ichikawa, H., and Yu, Y.T. (1999) Control of sigma factor activity during Bacillus subtilis sporulation. Mol Microbiol 31: 12851294.
  • Li, W., Bottrill, A.R., Bibb, M.J., Buttner, M.J., Paget, M.S., and Kleanthous, C. (2003) The role of zinc in the disulphide stress-regulated anti-sigma factor RsrA from Streptomyces coelicolor. J Mol Biol 333: 461472.
  • Lonetto, M., Gribskov, M., and Gross, C.A. (1992) The sigma 70 family: sequence conservation and evolutionary relationships. J Bacteriol 174: 38433849.
  • Lonetto, M.A., Brown, K.L., Rudd, K.E., and Buttner, M.J. (1994) Analysis of the Streptomyces coelicolor sigE gene reveals the existence of a subfamily of eubacterial RNA polymerase sigma factors involved in the regulation of extracytoplasmic functions. Proc Natl Acad Sci USA 91: 75737577.
  • Missiakas, D., Mayer, M.P., Lemaire, M., Georgopoulos, C., and Raina, S. (1997) Modulation of the Escherichia coli sigmaE (RpoE) heat-shock transcription-factor activity by the RseA, RseB and RseC proteins. Mol Microbiol 24: 355371.
  • Mogk, A., Schmidt, R., and Bukau, B. (2007) The N-end rule pathway for regulated proteolysis: prokaryotic and eukaryotic strategies. Trends Cell Biol 17: 165172.
  • Paget, M.S., and Helmann, J.D. (2003) The sigma70 family of sigma factors. Genome Biol 4: 203.
  • Paget, M.S., Kang, J.G., Roe, J.H., and Buttner, M.J. (1998) sigmaR, an RNA polymerase sigma factor that modulates expression of the thioredoxin system in response to oxidative stress in Streptomyces coelicolor A3(2). EMBO J 17: 57765782.
  • Paget, M.S., Bae, J.B., Hahn, M.Y., Li, W., Kleanthous, C., Roe, J.H., and Buttner, M.J. (2001a) Mutational analysis of RsrA, a zinc-binding anti-sigma factor with a thiol-disulphide redox switch. Mol Microbiol 39: 10361047.
  • Paget, M.S., Molle, V., Cohen, G., Aharonowitz, Y., and Buttner, M.J. (2001b) Defining the disulphide stress response in Streptomyces coelicolor A3(2): identification of the sigmaR regulon. Mol Microbiol 42: 10071020.
  • Park, S.T., Kang, C.M., and Husson, R.N. (2008) Regulation of the SigH stress response regulon by an essential protein kinase in Mycobacterium tuberculosis. Proc Natl Acad Sci USA 105: 1310513110.
  • Park, J.H., and Roe, J.H. (2008) Mycothiol regulates and is regulated by a thiol-specific antisigma factor RsrA and sigma(R) in Streptomyces coelicolor. Mol Microbiol 68: 861870.
  • Parker, J., Pollard, J.W., Friesen, J.D., and Stanners, C.P. (1978) Stuttering: high-level mistranslation in animal and bacterial cells. Proc Natl Acad Sci USA 75: 10911095.
  • Rodionov, D.A. (2007) Comparative genomic reconstruction of transcriptional regulatory networks in bacteria. Chem Rev 107: 34673497.
  • Sareen, D., Steffek, M., Newton, G.L., and Fahey, R.C. (2002) ATP-dependent l-cysteine: 1d-myo-inosityl 2-amino-2-deoxy-alpha-d-glucopyranoside ligase, mycothiol biosynthesis enzyme MshC, is related to class I cysteinyl-tRNA synthetases. Biochemistry 41: 68856890.
  • Schurr, M.J., Yu, H., Martinez-Salazar, J.M., Hibler, N.S., and Deretic, V. (1995) Biochemical characterization and posttranslational modification of AlgU, a regulator of stress response in Pseudomonas aeruginosa. Biochem Biophys Res Commun 216: 874880.
  • Shevchenko, A., Wilm, M., Vorm, O., and Mann, M. (1996) Mass spectrometric sequencing of proteins silver-stained polyacrylamide gels. Anal Chem 68: 850858.
  • Viollier, P.H., Weihofen, A., Folcher, M., and Thompson, C.J. (2003) Post-transcriptional regulation of the Streptomyces coelicolor stress responsive sigma factor, SigH, involves translational control, proteolytic processing, and an anti-sigma factor homolog. J Mol Biol 325: 637649.
  • Zhou, Y., Gottesman, S., Hoskins, J.R., Maurizi, M.R., and Wickner, S. (2001) The RssB response regulator directly targets sigma(S) for degradation by ClpXP. Genes Dev 15: 627637.

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information
FilenameFormatSizeDescription
MMI_6824_sm_Table_S1_and_Fig_S1-S2.pdf302KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.