SEARCH

SEARCH BY CITATION

Summary

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

In Escherichia coli, Fe-S clusters are assembled by gene products encoded from the isc and suf operons. Both the iscRSUA and sufABCDSE operons are induced highly by oxidants, reflecting an increased need for providing and maintaining Fe-S clusters under oxidative stress conditions. Three cis-acting oxidant-responsive elements (ORE-I, II, III) in the upstream of the sufA promoter serve as the binding sites for OxyR, IHF and an uncharacterized factor respectively. Using DNA affinity fractionation, we isolated an ORE-III-binding factor that positively regulates the suf operon in response to various oxidants. MALDI-TOF mass analysis identified it with IscR, known to serve as a repressor of the iscRSUA gene expression under anaerobic condition as a [2Fe-2S]-bound form. The iscR null mutation abolished ORE-III-binding activity in cell extracts, and caused a significant decrease in the oxidant induction of sufA in vivo. OxyR and IscR contributed almost equally to activate the sufA operon in response to oxidants. Purified IscR that lacked Fe-S cluster bound to the ORE-III site and activated transcription from the sufA promoter in vitro. Mutations in Fe-S-binding sites of IscR enabled sufA activation in vivo and in vitro. These results support a model that IscR in its demetallated form directly activates sufA transcription, while it de-represses isc operon, under oxidative stress condition.


Introduction

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

Proteins that contain Fe-S as a cofactor are widely distributed and play critical roles in metabolic reactions as well as in gene regulation (Beinert et al., 1997; Kiley and Beinert, 2003). The biogenesis of Fe-S clusters have been studied extensively in bacteria, and the mitochondria and plastids of eukaryotes (Frazzon and Dean, 2003; Johnson et al., 2005; Lill and Mühlenhoff, 2005). Bacteria possess three biosynthetic systems to generate Fe-S clusters. The first discovered system consists of NifS, a cysteine desulphurase, and NifU, an Fe-S scaffold, providing the Fe-S cluster to nitrogenase in nitrogen-fixing bacteria Azotobacter vinelandii (Jacobson et al., 1989; Dean et al., 1993). The second one called Isc system for their role in iron sulphur cluster assembly are found in many bacteria with or without nitrogen-fixing ability (Zheng et al., 1998; Schwartz et al., 2000). In many bacteria, the isc operon is found in a highly conserved gene cluster, iscRSUA-hscBA-fdx, where IscR is the negative regulator of gene expression (Schwartz et al., 2001). The iscS gene, homologous to nifS, encodes a pyridoxal phosphate-containing cysteine desulphurase that provides atomic sulphur from cysteine during Fe-S cluster formation (Zheng et al., 1993; 1998; Zheng and Dean, 1994; Schwartz et al., 2000).

The third Fe-S assembly system found in bacteria, called the Suf system encoded by sufABCDSE genes, is known as an alternative system for iron sulphur assembly, which rescues the growth defect of isc mutants when overexpressed in Escherichia coli (Takahashi and Tokumoto, 2002). The Suf system is present in a variety of bacteria, Archaea and plastids of plants. SufA and SufS are homologous to IscA and IscS serving as an iron-delivery as well as scaffold protein, and a desulphurase respectively. (Patzer and Hantke, 1999; Ollagnier-de-Choudens et al., 2001, 2003; 2004; Loiseau et al., 2003; Ding and Clark, 2004; Ding et al., 2004; 2005). SufBCD are homologous to components of ABC transporter complex. SufC in a plant pathogen Erwinia chrysanthemi and a plant plastid (Arabidopsis thaliana) have been demonstrated to be ATPases that function in the assembly and maintenance of Fe-S cluster (Nachin et al., 2003; Xu and Moller, 2004). A recent report demonstrated that in E. coli SufBCD complex together with SufE enhances cystein desulphurase activity of SufS (Outten et al., 2003). In bacteria such as E. coli where both Isc and Suf systems are present, it is thought that Isc system plays a primary roleto assemble Fe-S cluster whereas the Suf system serves as a back-up and/or a specified system adapted to pro-oxidative and iron-depleted conditions (Takahashi and Tokumoto, 2002; Outten et al., 2004). A recent study proposes a primary role of IscS even in repairing Fe-S clusters damaged under oxidative stress conditions (Djaman et al., 2004).

Both the sufABCDSE and iscRSUA operons are induced under oxidative stress conditions (Zheng et al., 2001; Lee et al., 2004; Outten et al., 2004). One direct regulator of the iscRSUA operon is IscR that contains a [2Fe-2S] cluster and represses the operon under anaerobic condition (Schwartz et al., 2001). Even though its underlying mechanism is not clearly understood, oxidative stress could cause inactivation of IscR as a repressor, leading to derepression of the operon. The sufABCDSE operon is directly regulated by OxyR, IHF, Fur and an unidentified oxidant-responsive (ORE) activator (Lee and Roe, 1997; Lee et al., 2003; 2004; Outten et al., 2004). Three ORE cis-acting DNA sites have been mapped and found to bind their cognate regulators; ORE-I (from −236 to −197 nt from the transcriptional start site) for OxyR, ORE-II (from −156 to −127) for IHFs, and ORE-III (from −56 to −35) for an unidentified activator responding to both superoxide generators and hydrogen peroxide (Lee et al., 2004). Fur binds between −32 and −3 and thus could inhibit binding of RNA polymerase to the sufA promoter. Removal of OxyR, IHF, ORE-I, or ORE-II, did not abolish the ORE-III mediated oxidative induction of the sufA gene (Lee et al., 2004).

In this study, we present our work to isolate the ORE-III-binding ORE factor, and its identification with IscR. IscR binds to the promoter region of sufA and activates its transcription as a demetallated apo-form.

Results

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

Isolation of an ORE-III-binding factor using its specific DNA affinity

In order to isolate the factor that binds to ORE-III and positively regulates sufA induction in response to oxidants, DNA affinity fractionation using the ORE-III DNA fragment was employed. The DNA fragments containing ORE-III were generated by polymerase chain reaction (PCR) (Fig. 1A), and were cloned into pUC18. A clone with five tandem insertion of the 70 bp ORE-III fragment was selected and the 350 bp insert DNA was re-amplified by PCR using biotinylated primers, and attached to magnetic beads coated with streptavidins. Cell extracts from exponentially grown ΔoxyR cells treated with 30 μM phenazine methosulphate (PMS) for 1 h were partially purified through heparin and Q-sepharose column chromatographies as described in Experimental procedures. Pooled eluants from Q-sepharose column that contain active ORE-III-binding factor were applied to the ORE-III-DNA affinity beads. Beads were washed with TGED buffer with different NaCl concentrations, and the remaining ORE-III-binding proteins were eluted with SDS from 0.01 to 0.5% as described by Jeong et al. (2004). The ORE-III-binding activity in each fraction was monitored by gel mobility shift assay as shown in Fig. 1B. The majority of ORE-III-binding activity was present in the fraction E1 eluted with 0.01% SDS when the eluant fractions were directly tested for DNA-binding activity. When SDS was removed by dialysis to avoid its inhibitory effect on DNA binding, similar amount of ORE-III-binding activity was recovered in the fraction E2 eluted with 0.1% SDS (see E2′ lane in Fig. 1B). Pooled eluant fractions were concentrated by freeze-drying and proteins were resolved on SDS-PAGE stained with silver nitrate (Fig. 1C). The characteristic bands (1 and 2) in E1 and E2 fractions that migrated at about 17 kDa position were excised, digested with trypsin in gel, and subjected to MALDI-TOF mass spectrometry. The analysis identified both bands from E1 and E2 with IscR, known as a transcriptional repressor of its own gene, iscR, and the Fe-S cluster assembly genes of the iscRSUA oepron (Schwartz et al., 2001). The iscR gene encodes a protein of 162 amino acids, a member of the MarA/Rob/SoxS family proteins. The molecular mass of 17 326 Da matches with the electrophoretic mobility of the purified ORE-III-binding factor.

image

Figure 1. Identification of an ORE-III-binding factor through DNA affinity chromatography. A. The regulatory elements of the sufA operon. Three oxidant-responsive elements (OREs), and promoter sites (−35, −10 and +1) are indicated. Sites of action of activators (OxyR, IHF and an unidentified factor) and a repressor (Fur) were marked. The boundary positions of the ORE-III DNA probe used for making affinity beads were marked by arrows at the BamHI cut site. B. The DNA-binding activity of fractions from affinity beads. Cell-free extracts were prepared from exponentially grown ΔoxyR cells treated with 30 μM PMS for 1 h. Cell extracts partially purified through heparin and Q-sepharose chromatographies, were subsequently applied to DNA affinity beads coated with pentameric repeat of 70-mer ORE-III DNAs. Washed and eluted fractions from DNA affinity beads were examined for ORE-III-binding activity by GMSA. The samples tested were: lane 1, 10 μg of crude extracts; lane 2, Q-sepharose pool (load on); lane 3, flow through; lanes 4–5, the first and second wash with TGED buffer; lane 6, wash with TGED plus 0.2 M NaCl; lanes E1–E4, eluants with 0.01% (E1), 0.1% (E2 and E3)and 0.5% SDS (E4); lanes E1′–E4′, E1–E4 dialysed against TGED to remove SDS before GMSA. C. Silver-stained protein profile of eluants from DNA affinity chromatography. Proteins in wash (lanes 2, 3, 4, 6) and SDS-eluant (E1–E4) fractions were resolved on SDS-PAGE, and stained with silver nitrate as described in Experimental procedures. The arrows point the band of IscR, which was identified by MALDI-TOF mass spectrometry, as summarized in the right panel.

Download figure to PowerPoint

Effect of the iscR gene deletion on sufA induction in vivo

We then created ΔiscR null and ΔiscRoxyR double mutants in GC4468 strain background to observe the effect of iscR mutation on sufA induction by oxidants. The suf transcripts (sufAp) were analysed by S1 mapping in strains of various genetic backgrounds before (N) and after treatment with PMS (P) or hydrogen peroxide (H) (Fig. 2). The transcripts from the iscR promoter (iscRp) were also analysed in parallel (Schwartz et al., 2001). The oxyR and fur mutations affected sufA transcription as observed previously, exhibiting partial positive action of OxyR and partial negative action of Fur (Lee et al., 2004). The transcripts from the iscR promoter was not affected by the oxyR and fur mutations as observed previously (Zheng et al., 2001). Deletion of the iscR gene reduced sufA induction to the level comparable to that of the oxyR mutant. This demonstrates that IscR positively contributes to induce the sufA gene as much as OxyR does in response to oxidative stress. Consistent with this, ΔoxyRiscR double mutation caused nearly no induction of the sufA gene. In contrast, the iscR mutation derepressed the level of iscR gene expression in the absence of any oxidant treatment, consistent with the proposal that IscR acts as a repressor of its own operon (Schwartz et al., 2001). Additional oxyR mutation did not affect the iscR gene expression as expected.

image

Figure 2. Effect of ΔiscR mutation on sufA and iscR expression – S1 mapping analysis. RNA samples were prepared from the wild type (GC4468), ΔoxyR (JL102), Δfur (JL101), ΔiscR (WS101) and ΔoxyRiscR (WS102) cells either untreated (N) or treated with 0.1 mM PMS (P) or 1 mM H2O2 (H) for 10 min. The protected S1 probes were visualized and quantified by a Phosphor Image analyzer (FLA2000, Fuji). Below each lane, average values for the relative fold induction from five independent experiments were presented. For comparison, the iscR transcripts in the same sample were analysed in parallel.

Download figure to PowerPoint

We then examined the effect of iscR null mutation on sufA expression using the chromosomal sufA–lacZ fusion strains containing different lengths of the sufA upstream sequences (Fig. 3A). As demonstrated in Fig. 3B, deletion of iscR reduced the sufA–lacZ induction by PMS to about a third of the wild-type level in full-length construct (SPD256). It also abolished the remaining induction in SPD184 and SPD61 constructs that are mediated via ORE-III. The uninduced basal levels were all decreased in ΔiscR mutant. In comparison, ΔoxyR mutation reduced the sufA–lacZ induction to about 40% of the wild-type level in SPD256, and retained induction in SPD184 and SPD61 constructs, representing the contribution from IscR. These results support the proposal that IscR positively acts to induce sufA expression in response to PMS via ORE sites, most likely ORE-III. Exogenously provided IscR on a multicopy plasmid in ΔiscR mutant restored the induction, consistent with the proposal.

image

Figure 3. Effect of ΔiscR mutation and IscR overproduction on the sufAp–lacZ expression. A. The structure of the sufAp–lacZ fusion constructs containing various lengths of sufA upstream region. B. Expression of sufAp–lacZ in wild type, ΔoxyR, ΔiscR background and in ΔiscR provided with pTac-IscR. Exponentially grown cells were either untreated or treated with 0.1 mM PMS for 1 h. The β-galactosidase activity was presented in Miller units with average values from three independent experiments marked on top of each bar.

Download figure to PowerPoint

Effect of iscR mutation on ORE-III DNA binding in vitro

To confirm direct involvement of IscR in ORE-III-mediated regulation, we examined the effect of iscR deletion and complementation by IscR overproduction in vivo and in vitro. As demonstrated in Fig. 4A, S1 mapping analysis of sufA transcripts revealed that heterologous overproduction of IscR in iscR null mutant restored PMS-dependent induction of sufA in the mutant. We also noted that IscR overproduction caused a slight increase in the uninduced level of sufA transcripts, consistent with observations with the lacZ fusion constructs. Using extracts from the same cell samples, gel mobility shift assay was performed with the ORE-III DNA probe. The result presented in Fig. 4B demonstrated that the sufA ORE-III-specific bound complexes that become abundant upon oxidant (PMS) treatment (lanes 2–4) in wild-type cells are absent in ΔiscR mutant cell extract (Lanes 8–10). When the iscR gene was provided on a multicopy plasmid (pTac-IscR), bound complexes reappeared in ΔiscR mutant even in the absence of PMS treatment (Fig. 2, lanes 11–13). Slower-migrating, putative multimeric form of complexes were detected when higher amount of IscR was provided through IPTG-induction of the multicopy cloned gene (data not shown). These results support the proposal that IscR binds directly to the ORE-III region. When the iscR promoter region from −96 to −27 was examined for IscR binding in parallel, some weak binding complexes were also formed with cell extracts from oxidant-treated cells, suggesting that IscR can also bind to its own promoter region under aerobic condition, consistent with previous observation by Schwartz et al. (2001) (data not shown).

image

Figure 4. IscR-dependent formation of ORE-III-bound complexes. RNA samples and cell extracts from the wild type, ΔiscR, and ΔiscR with pTac-IscR were prepared for S1 mapping analysis (A) and gel mobility shift assay (B). Cells grown to OD600 = 0.2–0.5 were either untreated or treated with 20 or 50 μM PMS for 30 min before harvest. For each assay, 50 μg of RNA or 10 μg of crude extracts were used. In B, F denotes free probes and C indicates DNA-protein complex.

Download figure to PowerPoint

Binding of purified IscR to the sufA promoter upstream fragments – footprint analysis

We overproduced and isolated IscR protein using pET21b with a histidine tag at C-terminus through a standard air-exposed purification procedure. Assessment of metals in the purified IscR sample by inductively coupled plasma-atomic emission spectroscopy (ICP-AES) revealed that it does not contain any detectable level of iron. We then examined the direct binding of purified IscR to the sufA ORE-III fragment by GMSA and found bound complexes with similar electrophoretic mobility as observed with cell extract (data not shown). To determine the binding site of purified apo-IscR on the sufA promoter region, DNase I footprinting analysis was performed. The results presented in Fig. 5 demonstrates that IscR binds to at least three sites in the upstream of the sufA promoter (Fig. 5A, top and bottom strands). It binds to a region between −26 and −60 (site 1) most efficiently. This region overlaps closely with ORE-III site (−35 to −56) determined previously. IscR also binds efficiently to a region between −132 and −164 (site 2) which overlaps with ORE-II site. A weak binding was observed at high concentrations of IscR between −88 and −118 (site 3) which lies between ORE-II and ORE-III. The protected sites 1 and 2 contain several DNase I-hypersensitive points on both strands, suggesting that IscR binding to the sufA promoter upstream could cause some distortion in DNA structure. The footprinting pattern of IscR on the sufA gene is summarized in Fig. 5B. No obvious conserved sequence pattern was detected among the binding sites.

image

Figure 5. Binding of IscR binding to the sufA promoter region – DNase I footprinting analysis. A. The sufA DNA probes, labelled at the 5′ end of either the top or bottom strand, were incubated with increasing concentrations of IscR, followed by DNase I treatment. The samples were run on 8% polyacrylamide sequencing gel with Maxam-Gilbert G + A sequencing ladders. The regions protected by IscR were indicated with brackets. Arrows indicate a position of enhanced DNase I cleavage site. IscR was added to each binding reaction at 5 pmol (125 nM, lanes 1), 10 pmol (250 nM, lanes 2), 20 pmol (500 nM, lanes 3), 50 pmol (1.25 μM, lanes 4), 100 pmol (2.5 μM, lane 5) and 200 pmol (5 μM, lane 6). B. Summary of footprinting pattern. Square boxes indicate previously determined ORE sites (Lee et al., 2004). Solid lines indicate strongly protected region whereas dotted lines represent weakly protected one. Arrows represent DNase I-hypersensitive sites.

Download figure to PowerPoint

IscR directly activates sufA transcription in vitro

Effect of purified IscR on sufA transcription was examined by in vitro transcription assay. As demonstrated in Fig. 6, addition of IscR dramatically increased the amount of sufA transcripts, confirming its role as a direct activator of sufA transcription (Fig. 6A). Half-maximal level of induction was obtained at about 5 pmol (200 nM) IscR.

image

Figure 6. IscR-directed activation of the sufA transcription in vitro. A. Single-round transcription assays in vitro were carried out with linear DNA templates encompassing the sufA (from −256 to +90, A) promoter region. The binding reaction contained increasing amount of purified IscR (0–10 pmol) in 25 μl reaction as indicated. The first lane (-R) indicates a control reaction without RNA polymerase. RNA1 transcripts were monitored as an internal control. Transcripts separated on a 8% polyacrylamide gel were visualized by Phosphor Imager analyzer. B. Quantification of in vitro sufA transcription from two independent experiments. The transcription reaction contained IscR from 0 to 20 pmol (0–800 nM).

Download figure to PowerPoint

Dispensability of Fe-S-binding sites in IscR for activating sufA transcription

IscR homologues in bacteria contain three highly conserved cysteine residues (C92, C98 and C104 in E. coli IscR) which is thought to coordinate Fe-S cluster. We created the mutant iscR genes by replacing each or all of these cysteines into alanines and tested whether these mutations affect sufA induction. The mutant iscR genes were cloned into pTac1 vector and introduced to ΔoxyRiscR mutant where nearly no sufA induction occurs. The sufA transcripts were analysed by S1 mapping along with the transcripts from the iscR promoter. Results in Fig. 7 demonstrate that the exogenously provided wild-type IscR restored sufA induction in ΔoxyRiscR mutant to the level comparable to that in ΔoxyR mutant (see Fig. 2). The C to A substitution mutants, whether in single or triple sites, all restored sufA induction as the wild-type IscR. Compared with the wild-type IscR, the basal level of the sufA transcripts slightly increased when mutant IscR proteins were provided. Effect of IscR on transcripts from the iscR promoter was as expected. Provision of IscR from the multicopy plasmid lowered the basal level of the iscR transcripts most likely due to increased amount of IscR. The mutant forms of IscR derepressed the iscR transcripts to the full level in the absence of oxidative treatments in accordance with the model proposed by Schwartz et al. (2001).

image

Figure 7. Effect of cysteine to alanine substitutions in IscR on sufA induction in vivo. The transcripts from the sufA and iscR promoters were analysed by S1 mapping in ΔoxyRiscR mutant provided with pTac1 plasmid (–) or the iscR genes cloned in pTac1; wild-type iscR (IscR), single substitution mutants (C92A, C98A, C104A) or a triple mutant (C92/98/104 A). RNA samples were prepared from cells either untreated (N) or treated with 0.1 mM PMS (P) or 1 mM H2O2 (H) for 10 min as described in Fig. 2. Transcripts from the wild-type E. coli (GC4468) were analysed in parallel.

Download figure to PowerPoint

The observation that C to A substitutions in IscR still allowed sufA induction in ΔoxyR background confirms that IscR in its apo-form is able to activate sufA transcription. The low level of the sufA transcripts in the absence of oxidants is attributable partly to repression by Fur, suggesting that sufA induction occurs only in the absence of Fur repression and in the presence of apo-IscR as demonstrated in Fig. 8. When the C92/98/104 A triple mutant form of IscR was purified and assessed for its activity in transcriptional activation in vitro, similar amount of activation was observed as described in Fig. 6 (data not shown).

image

Figure 8. A model for the opposing role of IscR on the sufA and iscR promoters. Action of IscR as an activator in a demetallated apo-form on the sufA promoter is presented along with the actions of other regulators. Repression of the iscR promoter by the Fe-S form of IscR is presented as proposed by Schwartz et al. (2001).

Download figure to PowerPoint

Discussion

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

The sufABCDSE operon is induced by oxidants such as superoxide generators and H2O2 and also by iron starvation (Lee et al., 2003; 2004; Outten et al., 2004). Part of the oxidant induction is mediated by OxyR that binds to ORE-I site (−236 to −197 nt upstream from the transcription start site) and probably contacts RNA polymerase through a loop formation facilitated by IHF bound at ORE-II (−156 to −127). The other part of oxidant induction is mediated by IscR, as demonstrated in this study, which binds to ORE-III site (−56 to −35) and directly activates sufA transcription when cells are under oxidative stress. The binding of Fur at the promoter site (−32 to −3) represses sufA transcription, which are de-repressed under iron-starvation (Lee et al., 2003; Outten et al., 2004). The fur mutant, however, does not derepress sufA to the maximal level achieved by oxidant induction (Lee et al., 2004), suggesting that the combined action of both the activation and de-repression is necessary to fully induce sufA transcription.

Inspection of sequences in ORE-III and the sites 1 and 2 of sufA upstream region does not pull out any putative consensus sequence. The sequence also does not match significantly with a Marbox (AYNGCACNNWNNRYYAAAYN) suggested to be recognized by [2Fe-2S]-IscR based on its close similarity with MarA (Schwartz et al., 2001). Out of 22 nt of ORE-III sequence, only six matched the consensus Marbox. Comparison of ORE-III sequence with the IscR-binding sites within the iscR promoter region from −80 to +49 did not reveal any significant matches. Further systematic studies with extensive mutations are necessary to reveal the recognition sequences of apo-IscR.

IscR has been reported to act as a repressor for its own operon iscRSUA (Schwartz et al., 2001). The isc operon is also induced by oxidants and iron starvation (Zheng et al., 2001; Lee et al., 2003; 2004; Outten et al., 2004). The regulatory mechanism for iscR transcription is not well characterized, except that the repressor activity of IscR depends on the Fe-S cluster formation, most likely a [2Fe-2S], as observed in anaerobically purified IscR. The observation that the apo-form of IscR also retains yet another regulator activity, in this case an activator for the sufABCDSE operon, another iron-sulphur assembly system in E. coli, adds a new dimension to the complexity of action modes of the regulator.

It is fascinating to note that the two functionally overlapping Fe-S cluster assembly systems in E. coli are regulated by a common regulator, acting in opposite ways, to ensure full induction under emergency conditions of oxidative stress as well as iron deprivation. The status of Fe-S cluster including its redox state could affect the activity of IscR as a repressor and an activator, thereby sensing the intracellular state of Fe-S assembly and maintenance, and thus coordinating gene expression as cell needs. In the absence of any additional oxidative stress, except that endogenously generated from aerobic respiration, only the basal level of isc expression will suffice for E. coli cells to survive. When cells are under attack by exogenous reactive oxygen species as encountered in phagocytic environment, oxidation of Fe-S cluster and its subsequent disassembly in IscR will actively induce the sufABCDSE operon, and will derepress the iscRSUA operon at the same time, maximizing the cellular capacity for Fe-S assembly and maintenance. When sufficient amount of Fe-S cluster is provided, as sensed by forming [Fe-S]-IscR, iscR expression gets repressed to a certain basal level, and sufA expression is no longer activated and returns to near ground level. This mode of regulation appears to ensure maximal induction in response to different environmental ques and the optimally economic maintenance as needed. The contribution from the Suf system to the biogenesis and maintenance of Fe-S cluster pool is indispensable under oxidative stress condition, as reflected in the high sensitivity of suf mutants to superoxide-generating oxidants (Lee et al., 2004). It is not clear why E. coli requires this additional Suf system under oxidative stress condition. There is not much biochemical evidence to support the advantageous function of Suf compared with Isc system under oxidative condition. The fact that cyanobacteria contain only the Suf system for Fe-S assembly could be an evolutionary evidence to support the hypothesis that the Suf system may be advantageous under pro-oxidative condition, as might occur under photosynthetic electron transfer processes.

Experimental procedures

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

Bacterial strains and culture conditions

The E. coli strains (GC4468 for wild type, ΔoxyR, Δfur, ΔiscR and ΔoxyRiscR mutants in GC4468 background) were grown as described in Lee et al. (2004). Antibiotics were used at the following concentrations: ampicillin, 50 μg ml−1; kanamycin, 25 μg ml−1; apramycin, 50 μg ml−1. For oxidant treatment, cells were grown in Luria–Bertani (LB) broth up to OD600 = 0.2–0.5 and treated with PMS (20, 50 or 100 μM) or H2O2 (1 mM). To create ΔiscR mutant, the iscR coding region was replaced with an apramycin cassette in E. coli DY330 (Yu et al., 2000). The apramycin cassette (aminoglycoside acetyltransferase gene; aac(3)IV) was amplified from pIJ773 (Gust et al., 2003) by PCR using a pair of primers that contain 50 nt iscR flanking sequences at 5′ sides; IscRAprF (5′-TTTTACAATAAAAAACCCCGGGCAGGGGCGAGTTTGAGGTGAAGTAAGACGTGCAATACGAATGGCGAAA-3′) and IscRAprR (5′-CCGTGTTTACGGAGTATTTAGCACTCCGGCCTGATTCTGAATTCTTT TTATCATGAGCTCAGCCAATCGA-3′). The PCR product was introduced by electroporation to DY330 where the linear DNA recombination system was induced by shifting temperature from 32°C to 42°C for 15 min. The recombinants were selected by apramycin resistance. The iscR::aac(3)IV gene in DY330 was transduced to GC4468 background by P1 phage to create WS101 (GC4448 iscR::aac(3)IV), which was confirmed by PCR and Southern hybridization. The ΔoxyR::kan gene was transduced from JL102 (Lee et al., 2003) by P1 to WS101 to create ΔoxyRiscR double mutant WS102 (GC4468 ΔoxyR::kan ΔiscR::aac(3)IV). For complementation, the full-length iscR gene was cloned in a multicopy pTac1-based plasmid (Koo et al., 2003). The sufA–lacZ fusion strains were used to monitor sufA expression as described by Lee et al. (2004).

Isolation of ORE-III-binding factor

The ΔoxyR mutant (JL102, Lee et al., 2004) cells were grown to OD600 of 0.2–0.5 in LB medium and treated with 30 μM PMS for 1 h. Cells were harvested, resuspended in the binding buffer and disrupted with a sonicator (MISONIX XL-2020) at 30% of its full power. The cell-free extracts clarified at 12 000 r.p.m. for 30 min were further centrifuged at 27 000 r.p.m. (100 000 g) for 1 h. The soluble fractions were applied to heparin column, eluted with NaCl (0–1 M) and monitored for ORE-III-binding activity by gel mobility shift assay. The active fractions were pooled, concentrated by amicon and dialysed against TGED buffer [50 mM Tris-HCl, pH 8.0, 5% glycerol (v/v), 1 mM EDTA and 1 mM DTT]. They were subsequently applied to Q-sepharose column and eluted with NaCl (0–1 M). The active fractions were finally subjected to DNA affinity fractionation.

The 83 bp DNA fragment (−98∼−16) that contain ORE-III site (ORE-III-70) was amplified by PCR using B93-UP (5′-GGGTGGATCCGATTAAAAGCAGGCAA-3′; BamHI site underlined) and B24-DOWN (5′-ATAATGAGGATCCGTTCAAC-3′; BamHI site underlined) primers, digested with BamHI and cloned into pUC18. A plasmid clone that contains five tandem copies of ORE-III-70 was selected, and further used to generate 350 bp-long biotinylated DNA affinity probe (ORE-III-70 × 5) by PCR using a primer pair; pUC-UP (5′-CAGGAAACAGTATGAC-3′) and pUC-DOWN (biotin-5′-GTTTTCCCAGTCACGA-3′). The PCR mixture contained 250 units (1 unit μl−1) of Taq polymerase, 0.2 mM of each NTP, and 100 pmol of each primer in a total volume of 10 ml distributed into 100 tubes. The amplified DNA (∼100 μg ml−1) was then purified by PCR purification kit (Qiagen) and eluted with TEN100 (Tris-HCl, pH 7.5, 1 mM EDTA, 0.1 M NaCl) from the column. The biotinylated ORE-III-70 × 5 DNA (∼2 μmol) was incubated with streptavidin-coated magnetic beads (Roche, 1 ml) at room temperature for 30 min with gentle agitation on a rotator, and applied to a magnetic particle separator to remove unbound DNA. The DNA affinity beads were washed twice with TEN1000 (10 mM Tris-HCl, pH 7.5, 1 mM EDTA, 1 M NaCl) and equilibrated with the binding buffer (50 mM Tris-HCl, pH 8.0, 1 mM MgCl2, 50 mM KCl, 1 mM EDTA, 5% glycerol and 1 mM DTT).

Active eluant fractions from the Q-sepharose column were incubated with the DNA affinity beads coated with biotinylated ORE-III probe. The beads were separated from unbound proteins by magnetic particle separator, washed with 1 ml each of binding buffer containing no added salt, 0.1 M NaCl, and finally 0.2 M NaCl respectively. The bound proteins were eluted with 1 ml of 0.01% SDS (E1), 1 ml each of 0.1% SDS for two consecutive times (E2 and E3), and 1 ml of 0.5% SDS (E4) once. The eluted fractions, with or without further dialysis to remove SDS, were monitored for ORE-III-binding activity by gel mobility shift assay. Aliquots of 150 μl from each of washed and eluted fractions were freeze-dried and resuspended in 15 μl of deionized water to analyse proteins by SDS-PAGE and silver staining.

MALDI-TOF mass analysis of in-gel digested peptides

The protein bands from polyacrylamide gel stained with silver nitrate were excised, and followed by destaining, reduction and alkylation with 10 mM DTT and 50 mM iodoacetamide (Shevchenko et al., 1996). Following dehydration with 100% acetonitrile, the gel pieces were completely dried in a vacuum centrifuge, and the polypeptides in the gel fragments were digested overnight with 100 ng of trypsin (Promega) in 20 μl of 50 mM NH4HCO3. The peptides digested by trypsin were eluted with 10 μl of 5% trifluoroacetic acid and 95% acetonitrile for two consecutive times and concentrated to 4–5 μl in vacuum. MALDI-TOF mass spectra of the tryptic peptides were acquired using a Voyager-DETM STR Biospectrometry (Applied Biosystems) equipped with a nitrogen laser (337 nm) at the National Center for Inter-University Research Facilities (NCIRF) at Seoul National University. The peptide fingerprints were analysed through Mascot program (http://www.matrixscience.com) linked to the primary sequence database of E. coli.

Purification of recombinant IscR

The entire coding region of the wild type or the triple C to A substitution mutant form of the iscR gene was amplified by PCR, cloned into hexa-histidine-tagging plasmid, pET21b, and introduced into E. coli strain BL21-Gold (DE3). Purification was performed according to the manufacturer's recommendations. His-tagged IscR proteins bound to Ni-NTA were eluted by step gradient using 10 volumes of elution buffer containing 100, 200 and 300 mM imidazole. The eluted fractions were pooled and dialysed against TGEN (50 mM Tris-HCl, pH 8.0, 10% glycerol, 1 mM EDTA, 200 mM NaCl) to remove imidazole and nickel. Finally, purified IscR was dialysed against renaturation buffer (50 mM Tris-HCl, pH 8.0, 20% glycerol, 1 mM EDTA, 1 mM DTT, 200 mM NaCl) and subsequently against a storage buffer (50 mM Tris-HCl, pH 8.0, 50% glycerol, 1 mM EDTA, 200 mM NaCl). Concentrations of purified IscR proteins were determined by the Bradford method.

Site-directed mutagenesis of cysteine-to-alanine substitutions in iscR

Site-directed mutagenesis of the cysteine residues in IscR was carried out with QuikChangeTM Site-directed Mutagenesis Kit (Stratagene) according to the manufacturer's instruction. Each substitution of C92, C98 and C104 to alanine was generated by PCR using primers that contain an alanine docon (GCC) in place of cyteine (TGT or TGC) as well as a suitable restriction enzyme sites (C92A, ApaI; C98A, NarI; C104A, HaeIII). The recombinant plasmid (pWS10) that contains the iscR gene cloned in pUC18 background was used as a template. The primers used are; C92A-UP (5′-ATGCCACCCGGGCCCAGGGTAAAG-3′), C92A-DOWN (5′-CTTTACCCTGGGCCCGGGTGGCAT-3′), C98A-UP (5′-GTAAAGGC GGCGCCCAGGGCGGCG-3′), C98A-DOWN (5′-CGCCGCCCTGGGCGCCGCCTTTAC-3′), C104A-UP (5′-GCGGCGATAAGGCCCTGACCCACG-3′), C104A-DOWN (5′-CGT GGGTCAGGGCCTTATCGCCGC-3′), with the converted codons underlined. The mutated clones were selected and confirmed for mutations by nucleotide sequencing. The single mutant iscR gene cloned in pUC18 (pWS11, pWS12 and pWS13 for C92A, C98A and C104A, respectively) were used as a template to create double and triple mutants. The C92/104 A double mutant was constructed with pWS11 as a template for PCR using a primer set, C104A-UP and C104A-DOWN. The triple mutant was generated by using C92/104 A construct as a template and a primer set of C98A-UP and C98A-DOWN. All the mutations were confirmed by nucleotide sequencing. The mutated iscR genes were cut out with BamHI and recloned into pTac1.

Inductively coupled plasma-atomic emission spectrometry

The metal content of purified IscR protein was analysed by inductive coupled plasma-atomic emission spectrometer with ICPS-1000IV (SHIMADZU) in the NCIRF at Seoul National University.

Gel mobility shift assay

Cell-free extracts were obtained from cells grown in LB broth to an OD600 of 0.2–0.5 and either untreated or treated with various concentrations of PMS for 30 min. Cells from 20 ml of culture were collected, washed and resuspended in 0.3 ml of TGED buffer [50 mM Tris-HCl, pH 8.0, 5% glycerol (v/v), 1 mM EDTA and 1 mM DTT]. ORE-III DNA probes (−98∼−16, 83 bp) were amplified by PCR using B93-UP (5′-GGGTGGATCCGATTAAAAGCAGGCAA-3′) and B24-DOWN (5′-ATAATGAGGATCCGTTCAAC-3′) primers. To prepare the iscR promoter DNA probe, a pair of PCR primers were used to amplify 70 bp region from −96 to −27 nt from the transcription start site of the iscR gene (Schwartz et al., 2001); UP primer (5′-GCCACGATAAAAAAATGG-3′) and DOWN primer (5′-AATTGGTCAACTATTTAC-3′). The PCR products were labelled with [γ-32P] ATP using T4 polynucleotide kinase and gel mobility shift assays were performed as described previously (Lee et al., 2004). For purified protein, 0.1–50 pmol of His-tagged IscR protein was used in each binding reaction.

DNase I footprinting analysis

End-labelled DNA probes were generated by PCR with a 5′-radiolabelled primer with [γ-32P]-dATP and a second non-radiolabelled primer as described by Lin and Shiuan (1995). For the sufA probe, the forward (Suf-UP2, 5′-CTTATTTCTGAACGCTGC-3′) and the reverse primer (Suf-DF DOWN, 5′-AAATCTTGTGGGTTAAA-3′) generated 326 bp probe encompassing nts from −256 to +70. To produce DNA probes labelled at either top or bottom strand, only one of the two PCR primers was labelled at the 5′ end with 32P using [γ-32P] ATP. DNA probes were purified by non-denaturing PAGE and eluted using the standard crush and soak method. The binding reactions were carried out in 40 μl containing 20 000–30 000 cpm of labelled DNA fragment (about 5–10 nM) and various concentrations of purified IscR, as described above for gel mobility shift assays. After preincubation for 10 min at room temperature, DNase I (0.2 unit in 4 μl, Promega) was added to the binding mixture, and incubated at room temperature for 90 s. The reaction was terminated by adding 90 μl of stop solution (0.2 M sodium chloride, 30 mM EDTA, 1% SDS and 100 μg ml−1 yeast tRNA). Following phenol extraction and ethanol precipitation, the pellet was dissolved in loading buffer and heated at 90°C for 2 min prior to loading onto 8% polyacrylamide gel containing 7 M urea. To identify the binding sites, the Maxam and Gilbert G + A sequencing ladder was generated.

S1 nuclease mapping

RNA samples were prepared from the wild type (GC4468) and various mutant cells freshly grown to OD600 = 0.2–0.5 and treated with 0.1 mM PMS or 1 mM H2O2 for 10 min. For each S1 analysis, 50 μg of RNA was incubated with the sufA or iscR probes prepared by PCR. The sufA probe DNA was prepared by PCR from the sufA gene cloned in pRS415 using a forward (pRS415-UP, 5′-TTGGGGATCGAATTCC-3′) and a reverse primer (Suf-DOWN2, 5′-GCGTTAAGCCTTGCCAGGCG-3′), generating 364 bp fragment that contain 18 bp of pRS415 and the sufA sequence (346 bp, from −256∼+90). For the iscR probe DNA (−96∼+66), the forward (IscR-UP, 5′-GCCACGATAAAAAAATGG-3′) and the reverse primer (IscR-DOWN, 5′-CTTACTTCACCTCAAACT-3′) generated 162 bp probe. The 5′ ends were labelled with 32P, and the standard S1 mapping procedure was employed.

In vitro transcription assay

Linear templates for in vitro transcription were generated by PCR. For suf templates, the forward (Suf-UP2, 5′-CTTATTTCTGAACGCTGC-3′) and the reverse primer (Suf-DOWN2, 5′-GCGTTAAGCCTTGCCAGGCG-3′) generated 346 bp sufA promoter fragment encompassing nts from −256 to +90 relative to the transcription start site. To provide a control promoter template, RNA1 promoter was generated by PCR using the RNA1-UP (5′-TAAGACACGACTTATC-3′) and the RNA1-DOWN (5′-CTGCGCGTAATCTGCT-3′) primers that generated 108 bp RNA1 promoter fragment encompassing nts from −119 to +131. Amplified DNA fragments were recovered from polyacrylamide gel through electroelution using a standard protocol. DNA template (0.1 pmol) was incubated with purified IscR (0, 0.75, 1.25, 2.5, 5, 10, 20 pmol) for 5 min at 37°C and then with RNA polymerase holoenzyme (0.6 pmol; Amersham) for another 5 min in 21.5 μl of transcription buffer (4 mM Tris-HCl, pH 8.0, 1 mM MgCl2, 0.06 mM EDTA, 0.04 mM KH2PO4-K2HPO4, 0.15 mM DTT, 0.25 mg ml−1 BSA, 20% glycerol). Single-round transcription reactions were allowed for 10 min at 37°C by adding 3.5 μl of unlabelled nucleotide triphosphates (ATP, UTP and GTP to 400 μM; CTP to 40 μM), heparin (100 μg ml−1) and 5 μCi of [α-32P] CTP (400 Ci mmole−1, Amersham). The reaction was terminated by adding 50 μl precooled stop solution (375 mM sodium acetate, pH 5.2, 15 mM EDTA, 0.1 mg ml−1 calf thymus DNA). The transcripts were precipitated with ethanol, washed with 70% ethanol, dissolved in the sample buffer [80% (v/v) formamide, 8% glycerol, 8 mM EDTA, 0.01% bromophenol blue, 0.01% xylene cyanol) and electrophoresed on 8% polyacrylamide gel containing 7 M urea. Gels were dried and exposed to X-ray film overnight at −80°C or quantified with Phosphor Image analyser (FLA2000, Fuji).

Acknowledgements

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

We thank Dr Y.J. Seok for giving helpful advice for DNA affinity chromatography and Dr Tricia Kiley for helpful discussions. This work was supported by a KOSEF Grant (M10400000017-04J0000-01710) to J.H. Roe for the National Research Laboratory for Molecular Microbiology at the Institute of Microbiology, Seoul National University. W.S. Yeo and K.C. Lee were recipients of BK21 fellowships for graduate students from the Ministry of Education and Human Resources.

References

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