Regulation of iron-sulfur cluster homeostasis through transcriptional control of the Isc pathway by [2Fe-2S]-IscR in Escherichia coli

SUMMARY

Fe-S clusters are essential across the biological world, yet how cells regulate expression of Fe-S cluster biogenesis pathways to cope with changes in Fe-S cluster demand is not well understood. Here, we describe the mechanism by which IscR, a [2Fe-2S] cluster-containing regulator of Escherichia coli, adjusts the synthesis of the Isc Fe-S biogenesis pathway to maintain Fe-S homeostasis. Our data indicate that a negative feedback loop operates to repress transcription of the iscRSUA-hscBA-fdx operon, encoding IscR and the Isc machinery, through binding of [2Fe-2S]-IscR to two upstream binding sites. IscR was shown to require primarily the Isc pathway for synthesis of its Fe-S cluster, providing a link between IscR activity and demands for Fe-S clusters through the levels of the Isc system. Surprisingly, the isc operon was more repressed under anaerobic conditions, indicating increased Fe-S cluster occupancy of IscR and decreased Fe-S cluster biogenesis demand relative to aerobic conditions. Consistent with this notion, overexpression of a Fe-S protein under aerobic conditions, but not under anaerobic conditions, led to derepression of P_iscR. Together, these data show how transcriptional control of iscRSUA-hscBA-fdx by [2Fe-2S]-IscR allows E. coli to respond efficiently to varying Fe-S demands.

INTRODUCTION

Iron-sulfur (Fe-S) proteins are nearly ubiquitous across all domains of life and carry out many important cellular functions such as respiration, photosynthesis, nitrogen fixation, and gene regulation (Beinert, 2000, Fontecave, 2006, Kiley and Beinert, 2003). In Escherichia coli, the majority of Fe-S cluster biogenesis under non-stress conditions is proposed to be catalyzed by the highly conserved, housekeeping Isc (iron-sulfur cluster) pathway, encoded by the iscRSUA-hscBA-fdx (isc) operon (Johnson et al., 2005, Py and Barras, 2010, Py et al., 2011). The current model for Isc-mediated Fe-S biogenesis in E. coli is that IscS, a cysteine desulfurase, provides the sulfur to build a transient Fe-S cluster on the scaffold protein IscU. HscB and HscA form a complex with IscU to apparently facilitate transfer of a Fe-S cluster from IscU to a subset of apo-proteins. Alternatively, IscA, an A type carrier, promotes cluster transfer by delivering the Fe-S cluster from the scaffold to target apo-proteins. Other genes in the operon encode a [2Fe-2S] ferredoxin (Fdx) and the transcription factor IscR. Although transcription of the isc operon is repressed by IscR (Schwartz et al., 2001), a [2Fe-2S] cluster-containing protein encoded by the first gene of this operon, a detailed understanding of how this autoregulatory mechanism occurs and ultimately how Fe-S biogenesis is regulated in vivo remains to be elucidated.

While IscR was first discovered for its role in regulating transcription of the Isc system (Schwartz et al., 2001), subsequent genome-wide transcription profiling data revealed that IscR controls the expression of more than 40 genes in E. coli (Giel et al., 2006).Among these genes are those encoding for anaerobic respiratory Fe-S enzymes (periplasmic nitrate reductase, hydrogenases-1 and −2) and additional proteins with roles in Fe-S formation, including ErpA, NfuA, and the Suf system. IscR also directly activates the expression of the alternate Fe-S biogenesis pathway encoded by the sufABCDSE (suf) operon (Nesbit et al., 2009, Yeo et al., 2006), which includes proteins that exhibit functional similarity to those of the Isc pathway (Py and Barras, 2010, Py et al., 2011, Johnson et al., 2005). However, Suf is proposed to function primarily under oxidative and nitrosative stress and iron-limiting conditions (Justino et al., 2005, Mukhopadhyay et al., 2004, Outten et al., 2004, Zheng et al., 2001). Nevertheless, the transcription profiling results implicated a broader role for IscR in sensing the cellular Fe-S status and thus controlling the expression of the Isc and Suf systems accordingly.

Further analysis of promoters directly regulated by IscR revealed two classes of DNA target sites for IscR site-specific binding, denoted as Type 1 and Type 2 sites (Giel et al., 2006). The nucleotide sequences vary between Type 1 (Giel et al., 2006) and Type 2 (Nesbit et al., 2009) sites, and it is not yet clear how a dimer of IscR distinguishes between them (Nesbit et al., 2009).Of the IscR-dependent promoters identified thus far, most contain a single Type 1 or Type 2 site (Giel et al., 2006). In contrast, DNase I footprinting has revealed that IscR binds to three individual sites within the iscR promoter region (P_iscR) (Giel et al., 2006). One region of protection, which encompasses nucleotides −67 to −14 relative to the +1 transcription start site, includes two Type 1 IscR binding sites: site A (−65 to −41) and site B (−40 to −16) (Giel et al., 2006). The remaining IscR site, located from +9 to +26, does not share sequence similarity with other known IscR binding sites and is referred to as P_iscR site C (Giel et al., 2006). The contribution of each site to repression of the isc operon in vivo is not known.

A previous study provided important insight that repression of the isc operon by IscR may allow cells to homeostatically regulate the synthesis of the Isc Fe-S biogenesis pathway (Schwartz et al., 2001). In this model, IscR [2Fe-2S] cluster occupancy would be a reporter of the cellular Fe-S status, tuning synthesis of the Isc pathway in response to changing demands for Fe-S biogenesis. In support of the model, the presence of [2Fe-2S]-IscR was demonstrated in vivo by Mössbauer spectroscopy and repression of P_iscR was shown to require [2Fe-2S]-IscR, since clusterless IscR mutants failed to repress P_iscR (Fleischhacker et al., 2012, Yeo et al., 2006). In addition, strains lacking IscS and HscA showed decreased IscR repression, indicating a role of the Isc pathway in maintaining the [2Fe-2S] form of IscR (Schwartz et al., 2001). Nevertheless, the effect of other Isc proteins or the Suf pathway on [2Fe-2S]-IscR activity has not been examined. In addition, whether the decrease in in vivo repression observed with the clusterless mutants can be explained by a lack of DNA binding is of particular interest because IscR target promoters containing Type 2 sites (such as those in the sufABCDSE, napFDAGHBC, hyaABCDEF, and hybOABCDEFG operons) do not require the [2Fe-2S] cluster form of IscR either for binding DNA in vitro (Nesbit et al., 2009) or for transcriptional regulation in vivo (Nesbit et al., 2009, Yeo et al., 2006). Together, the unique features of the [2Fe-2S] cluster-dependent and -independent modes of IscR regulation have led to the proposal that conditions that increase the Fe-S demand will lead to decreased [2Fe-2S]-IscR levels, de-repression of the isc operon, elevated apo-IscR levels, and accordingly, increased regulation of promoters with Type 2 sites by apo-IscR. Such differences in [2Fe-2S] occupancy of IscR may explain why the isc operon is expressed at higher levels under aerobic conditions (Giel et al., 2006) since the demand for Fe-S biogenesis is expected to be greater because of the O₂ sensitivity of some Fe-S clusters (Imlay, 2006, Imlay, 2008). Thus, determining whether the [2Fe-2S] cluster occupancy of IscR differs between aerobic and anaerobic conditions is an important question to address.

In this study, we dissected the mechanism by which IscR negatively regulates P_iscR to better understand how cells maintain Fe-S homeostasis. We first established roles for the three predicted IscR binding sites in repressing P_iscR transcription in vivo. The requirement of the [2Fe-2S]-containing form of IscR for site-specific binding in vitro and regulation of P_iscR under aerobic and anaerobic conditions in vivo was analyzed using IscR mutants that cannot ligate an Fe-S cluster but are otherwise functional. We report the role of other Isc pathway components and the Suf system in providing Fe-S clusters to IscR under aerobic and anaerobic growth conditions. To address whether differences in [2Fe-2S] cluster occupancy for IscR exist between aerobic and anaerobic growth, an in vivo titration experiment was performed to determine the level of IscR protein required for P_iscR repression for both conditions. Finally, to examine whether IscR responds to cellular demands for Fe-S clusters, we tested whether overexpression of an Fe-S protein affected transcription from P_iscR. Together, our data indicate that [2Fe-2S]-IscR regulates expression of the Isc pathway via a negative feedback mechanism based on the cellular Fe-S demand.

RESULTS

The two Type 1 sites play a role in negative autoregulation of the iscR promoter

To examine the contribution of each of the three IscR binding sites (A, B, and C; Fig. 1A) within the iscR promoter region to negative autoregulation, mutations in each individual site were constructed and their effect on IscR binding was determined. Since sites A and B are representative of the Type 1 IscR binding site, the first AT of each site, which are among the most highly conserved bases within the Type I site (Giel et al., 2006), were substituted with GC (Fig. 1A). Using a wild-type DNA fragment for comparison (Fig. 1B), DNase I footprinting showed that when site A was mutated, anaerobically isolated [2Fe-2S]-IscR no longer protected the region (−67 to −41 bp relative to the +1 iscR transcription start site) encompassing site A, but still protected the regions (−40 to −14 bp and +9 to +26 bp) encompassing sites B and C in a protein concentration-dependent manner (Fig. 1C). In an analogous fashion, the DNA fragment containing the mutated site B showed IscR-mediated protection in the regions encompassing sites A and C, but not site B (Fig. 1D). Site C is not as well-conserved among enterobacteria and does not resemble other known IscR binding sites (Giel et al., 2006). Nevertheless, upon mutating the base pairs located at positions +9, +10, +12, and +14 within this site (Fig. 1A), protection of site C from DNase I cleavage was eliminated, while there was no effect on IscR binding to sites A and B (Fig. 1E). Thus, our data demonstrate that mutation of sites A, B, and C within P_iscR decreases in vitro binding of IscR to each individual site but does not affect IscR binding to the remaining intact sites.

Fig. 1.

In vitro binding of anaerobically purified [2Fe-2S]-IscR to wild-type and mutated IscR binding sites within P_iscR. A) The IscR binding sites A (underlined), B (double underlined), C (underlined), the transcriptional start site (bent arrow), the −35 and −10 promoter elements (shaded),and the bases substituted in this study (bold italics) are indicated. DNase I footprints of IscR bound to DNA fragments containing the wild-type P_iscR region (panel B) or promoter regions with mutations in sites A, B, or C (panels C, D, and E, respectively). The amount of IscR protein (nM) present in each reaction and the extent of the footprint relative to the +1 transcription start site are denoted. Samples were electrophoresed with Maxam-Gilbert (A + G) ladders. The lane order in panel B was digitally rearranged for presentation purposes.

We determined the effects of the mutated IscR sites on in vivo P_iscR repression via β-galactosidase assays using aerobically or anaerobically grown strains with chromosomal P_iscR -lacZ fusions containing the mutations in sites A, B, or C (Fig. 2). As expected from previous results (Giel et al., 2006), IscR repressed P_iscR expression more under anaerobic (17-fold) than aerobic (6-fold) growth conditions. Nevertheless, under both growth conditions, mutation of either site A or B caused a 3- to 4-fold loss in repression compared to the intact promoter, indicating a defect in P_iscR repression. When sites A and B were both mutated, repression was eliminated, similar to that observed in a strain lacking IscR. Surprisingly, mutation of site C had no effect on P_iscR expression. Furthermore, mutation of site C did not result in any further significant defect in P_iscR repression when sites A or B were also mutated. Thus, our data suggest that both Type 1 sites A and B are necessary for full IscR-mediated repression of P_iscR in vivo. Since sites A and B overlap sequences (−65 to −16 bp) of the −35 and extended −10 promoter elements required for σ⁷⁰ binding (Record et al., 1996) and a region potentially encompassing an UP element important for recognition by αCTD (Gourse et al., 2000), it appears that IscR binding of both Type 1 sites is required to fully prevent RNA polymerase from binding P_iscR.

Fig. 2.

In vivo effects of mutations in IscR binding sites A, B, and C within P_iscR. β-galactosidase activity from wild-type (PK7571) or mutated (PK8521, PK8528, PK8582, PK8551, PK8801, PK8808, PK8820) P_iscR-lacZ fusions (recombined at the chromosomal lac region) was measured in cells grown under A) aerobic or B) anaerobic conditions in MOPS minimal media with glucose (0.2%). The amount of IscR-dependent repression (Fold repression) was determined by dividing the β-galactosidase activity present in the strain lacking IscR (PK7572) by the β-galactosidase activity measured for each strain. The presence (+) or absence (−) of wild-type IscR binding sites A, B, and C in the P_iscR-lacZ fusion is indicated below the figure. Error bars represent the propagation of standard errors for three biological replicates.

IscR requires its Fe-S cluster to repress expression from the iscR promoter in vivo under aerobic and anaerobic conditions

To establish the extent of the IscR [2Fe-2S] cluster requirement for negative autoregulation under both aerobic and anaerobic growth conditions, we measured P_iscR repression in strains expressing IscR mutants that are defective in cluster binding. These mutants contain one or more alanine substitutions of three of the four IscR [2Fe-2S] cluster ligands (C92, C98, C104) (Fleischhacker et al., 2012) and were previously shown to contain no detectable Fe-S cluster upon isolation (Nesbit et al., 2009). Under either aerobic or anaerobic conditions, strains containing chromosomally encoded IscR-C92A, IscR-C98A, IscR-C104A, or IscR-C92A/C98A/C104A exhibited a nearly complete defect in P_iscR-lacZ repression (Fig. 3), indicating that under both growth conditions, IscR requires its [2Fe-2S] cluster to repress P_iscR. Measurements of IscR protein in these strains showed that each of the IscR variants was present at levels ∼2-fold and ∼8-fold greater than wild-type IscR under aerobic and anaerobic conditions, respectively, (Table 2) (Nesbit et al., 2009), indicating that the loss in P_iscR repression did not result from a defect in protein accumulation. Furthermore, the same mutants were previously shown to be otherwise functional at a subset of IscR-regulated promoters that contain the Type 2 IscR binding motif and do not require the Fe-S form of IscR for binding (Nesbit et al., 2009). Taken together, our data support the notion that IscR requires its [2Fe-2S] cluster to negatively regulate P_iscR under both aerobic and anaerobic conditions and this occurs by [2Fe-S]-IscR binding to the two Type 1 sites within the iscR promoter region.

Fig. 3.

β-galactosidase activity from P_iscR fused to lacZ integrated at the λatt site was determined in strains containing wild-type IscR (PK6364) or chromosomal mutants defective in [2Fe-2S] cluster binding [IscR-C92A (PK7854), IscR-C98A (PK7855), IscR-C104A (PK7856), and IscR-C92A/C98A/C104A (PK7898)]. Strains were grown under A) aerobic or B) anaerobic conditions in MOPS minimal media containing glucose (0.2%). Fold repression was determined by dividing the β-galactosidase activity present in the strain lacking IscR (PK6512) by the β-galactosidase activity measured for each strain and error bars represent the propagation of standard errors for three biological replicates.

Table 2.

In vivo levels of IscR protein measured via quantitative Western blotting.

	IscR monomers, µM		IscR monomers/cell
Chromosome	Aerobic	Anaerobic	Aerobic	Anaerobic
wild-type IscR	17.1 ± 1.2	1.8 ± 0.1	10,300 ± 750	1060 ± 80
IscR- C92A/C98A/C104A	29.9 ± 2.0	13.1 ± 1.2	18,000 ± 1200	7900 ± 750

[2Fe-2S]-IscR binds with greater affinity to the P_iscR sites A and B in vitro than does apo-IscR

Previous in vitro studies demonstrated that [2Fe-2S]-IscR and IscR-C92A/C98A/C104A bind with equal affinity to P_hyaA (Nesbit et al., 2009), which contains a Type 2 IscR binding site. However, the above in vivo data raise the question as to whether [2Fe-2S]-IscR binds with higher affinity than apo-IscR to the Type 1 sites within P_iscR. To address this question, fluorescence anisotropy assays were performed under anaerobic conditions with DNA fragments containing either site A or B from P_iscR (Fig. 4). Curve fitting the fraction bound with the Hill equation indicates that the apparent K_d of monomeric [2Fe-2S]-IscR is 93 ± 5 nM for P_iscR site A and 112 ± 2 nM for P_iscR site B (Fig. 4C). Since the purified wild-type IscR used in this experiment was ≥50% occupied with [2Fe-2S] clusters, it is likely that the actual K_d values for P_iscR sites A and B may be lower than the apparent K_d values determined. In contrast, K_d values could not be determined for IscR-C92A/C98A/C104A because saturation was not obtained even at the highest protein concentration tested (1.0 µM)(Fig. 4B). Thus, our data suggest that [2Fe-2S]-IscR has a much higher binding affinity for the P_iscR A and B sites than does clusterless IscR. Furthermore, the observed difference in [2Fe-2S]-IscR and clusterless IscR binding affinity for the two P_iscR Type 1 sites provides an explanation as to why the [2Fe-2S] cluster is required for IscR to repress P_iscR in vivo.

Fig. 4.

Binding isotherms of [2Fe-2S]-IscR and apo-IscR for the two Type 1 sites within the P_iscR region. A) Sequences of IscR binding sites A (underlined) and B (double underlined) within P_iscR. Numbers indicate the distance relative to the +1 transcription start site. B) DNA binding isotherms of wild-type [2Fe-2S]-IscR (open symbols) and the clusterless mutant protein IscR-C92A/C98A/C104A (closed symbols) measured as a change in anisotropy under anaerobic conditions. Both forms of IscR protein were incubated with 5 nM fluorescently labeled DNA containing either site A (circles) or site B (triangles) in 40 mM Tris-Cl (pH 7.9) and 150 mM KCl. C) Fraction bound corrected for fluorescence quenching of labeled P_iscR sites A or B bound by [2Fe-2S]-IscR as a function of protein concentration. Error bars represent the standard errors of triplicate experiments.

IscR requires a functional Isc system to fully repress P_iscR in vivo

P_iscR repression requires [2Fe-2S]-IscR and is decreased in strains lacking iscS or hscA (Schwartz et al., 2001), supporting the model that the Isc machinery synthesizes the [2Fe-2S] cluster for IscR. However, since iscS and hscA are reported to have roles beyond Fe-S cluster assembly (Py and Barras, 2010, Johnson et al., 2005), we characterized the role of other members of the Isc pathway in supplying IscR its [2Fe-2S] cluster. Strains lacking individual components such as IscS, IscU, or Fdx yielded a similar defect in repression as the strain lacking the entire pathway (ΔiscSUA-hscBA-fdx), suggesting that these proteins are necessary for the biogenesis of [2Fe-2S]-IscR (Fig. 5). Strains lacking IscA had a smaller effect (∼2-fold) on IscR activity (Fig. 5).Quantitative Western blots revealed that under aerobic growth conditions, levels of IscR are increased about 4- and 2-fold in ΔiscU and ΔiscA strains, respectively (data not shown), indicating that IscR protein was present but unable to fully repress P_iscR. Finally, we found that elimination of the Isc pathway or individual components showed a similar defect in P_iscR repression under both aerobic and anaerobic growth conditions (Fig. 5). Interestingly, these results contrast those of a previous study that revealed that the [4Fe-4S] cluster-containing transcription factor FNR had a more stringent requirement for the Isc pathway under aerobic conditions than anaerobic conditions (Mettert et al., 2008).

Fig. 5.

β-galactosidase activity was measured in wild-type (PK6364) or mutant (PK8120, PK8122, PK7783, PK6828, PK7759, PK6564, PK6826, PK6848, PK7540, PK7751) strains containing P_iscR fused to lacZ integrated at the λatt site. Cultures were grown under A) aerobic or B) anaerobic conditions in MOPS minimal media supplemented with glucose (0.2%); for the wild-type, ΔiscSUA, ΔiscSUA-hscBA-fdx, ΔiscS, ΔiscU, ΔiscA strains, media were also supplemented with nicotinic acid (12.5 µg ml⁻¹), and thiamine (2 µg ml⁻¹). Fold repression was determined by dividing the β-galactosidase activity present in the strain lacking IscR (PK6512), grown in the presence or absence of nicotinic acid and thiamine, by the β-galactosidase activity measured for each strain and error bars represent the propagation of standard errors for three biological replicates. ND, not determined.

Other Fe-S biogenesis proteins are not necessary for [2Fe-2S]-IscR activity when the Isc pathway is functional

We also examined whether other Fe-S biogenesis proteins contribute to IscR cluster assembly since a small amount of [2Fe-2S]-IscR activity remained in the strain lacking the Isc pathway (Fig. 5). Under either aerobic or anaerobic conditions, deletion of the suf operon had no effect on P_iscR repression (Fig. 5), indicating that the Suf pathway plays no apparent role in IscR function under nonstress conditions when the Isc system is present. In a strain lacking NfuA, an IscA/NifU-like protein that facilitates Fe-S cluster transfer (Bandyopadhyay et al., 2008, Angelini et al., 2008)and whose expression is repressed by IscR (Giel et al., 2006), there was a small defect (∼2-fold) in IscR repression, but only under anaerobic conditions (data not shown). In addition, P_iscR repression was not affected in strains lacking CsdA, a cysteine desulfurase (Trotter et al., 2009), or YtfE, a protein implicated in Fe-S cluster assembly or repair (Justino et al., 2009)(data not shown). Finally, deletion of the genes encoding IscX, PepB, and SseB, which are located downstream of iscRSUA-hscBA-fdx and are repressed by IscR (Giel et al., 2006), were also tested for their effects on IscR function, although only IscX has been suggested to have a possible function in Fe-S biogenesis (Pastore et al., 2006, Shimomura et al., 2005). Only the strain lacking IscX showed any defect, slightly less than a 2-fold effect under either aerobic or anaerobic conditions (Fig. 5). Thus, in the presence of the Isc pathway, other Fe-S assembly proteins do not have a major effect on IscR-dependent repression of P_iscR, thus supporting our model that the Isc pathway provides the [2Fe-2S] cluster required for IscR to fully repress P_iscR expression under aerobic and anaerobic conditions.

In the absence of the Isc machinery, elevated expression of the Suf pathway can fully restore [2Fe-2S]-IscR activity

Unlike regulation of P_iscR, regulation of the suf operon by IscR does not require IscR to ligate a [2Fe-2S] cluster (Nesbit et al., 2009, Yeo et al., 2006). Thus, we considered the possibility that upregulation of the Suf machinery by apo-IscR and subsequent transfer of a Suf-generated [2Fe-2S] cluster to IscR may explain the residual [2Fe-2S]-IscR-dependent activity measured in the ΔiscSUA-hscBA-fdx strain. To address whether IscR could serve as a substrate for the Suf machinery, P_iscR expression was measured in strains that contained the ΔiscSUA-hscBA-fdx or ΔiscSUA-hscBA-fdx ΔsufABCDSE alleles and a plasmid with the suf operon under control of P_BAD. As expected, the strain lacking both the Isc and the Suf pathways was viable only when the suf plasmid, not the vector, was present since deletion of both pathways is lethal (Py et al., 2011). Furthermore, when 10 mM arabinose was present in the growth medium to induce suf operon expression, strains containing the suf plasmid had P_iscR activity similar to that of the wild-type strain under aerobic and anaerobic growth conditions (Fig 6). This activity correlated with at least 30-fold higher levels of SufD protein compared to that of the wild-type strain as determined by Western blot analyses (data not shown). Thus, the Suf pathway, when expressed at elevated levels, appears to be able to substitute for the Isc pathway by inserting a [2Fe-2S] cluster into IscR, and this may account for the residual [2Fe-2S]-IscR activity measured in the strain lacking the Isc pathway.

Fig. 6.

β-galactosidase activity from λ-P_iscR-lacZ was determined in aerobically (black) or anaerobically (white) grown wild-type (PK6364), ΔiscSUA-hscBA-fdx [Δisc (PK8122, PK8614, PK8615)], and ΔiscSUA-hscBA-fdxΔsufABCDSE [ΔiscΔsuf (PK8618)] strains. The ΔiscSUA-hscBA-fdx strain derivatives contained either the plasmid expressing the sufABCDSE operon from P_BAD (pGS0164), the vector alone (pBAD/Myc-hisC), or no vector. All cultures were grown in M9 minimal media containing glucose (0.2%), nicotinic acid (12.5 µg ml⁻¹), and thiamine (2 µg ml⁻¹). The plasmid-containing strains were also grown in the presence of Ap (50 µg ml⁻¹) and where indicated, arabinose (0.2%). Error bars represent the standard errors of three biological replicates.

IscR [2Fe-2S] cluster occupancy is greater under anaerobic growth conditions

In the wild-type strain, repression of P_iscR was ∼3–5-fold greater under anaerobic conditions than aerobic conditions despite the fact that there is ∼10-fold less IscR protein under anaerobic conditions (Figs. 2, 3, and 5, Table 2). Since in vitro DNA binding by [2Fe-2S]-IscR is not sensitive to the O₂-dependent change in the [2Fe-2S] cluster oxidation state (Fleischhacker et al., 2012), a simple interpretation of our in vivo results is that the fraction of IscR containing a [2Fe-2S] cluster may be higher under anaerobic conditions. This, in turn, would lead to more P_iscR repression and less IscR protein. This notion was tested by measuring the levels of IscR protein required for P_iscR repression under both growth conditions. A strain containing chromosomally-encoded iscR and iscSUA-hscBA-fdx under control of P_tac and P_BAD, respectively, was grown in medium containing 20 mM arabinose and various concentrations of IPTG (0–640 µM). We subsequently measured P_iscR-lacZ repression which, according to our in vivo data, requires the [2Fe-2S] form of IscR, and the amount of IscR present at each IPTG concentration was quantified by Western blots. Under both growth conditions, the amount of P_iscR repression increased with increasing levels of IscR; however, more IscR protein was needed to fully repress P_iscR under aerobic conditions compared to anaerobic conditions (Fig S1). In fact, half-maximal P_iscR repression required ∼9-fold more IscR protein aerobically than anaerobically (955 ± 93 nM vs. 107 ± 14 nM monomeric IscR). Assuming that the concentration of [2Fe-2S]-IscR is equivalent to the IscR protein concentration required for 50% repression, our data indicate that under aerobic conditions IscR is at least 9-fold less occupied with [2Fe-2S] clusters than under anaerobic conditions. Thus, we conclude that the difference in the amount of P_iscR repression under aerobic and anaerobic conditions can be explained by O₂-dependent changes in the cluster occupancy of IscR.

The iscR promoter is derepressed in response to increased Fe-S demand

Because the Isc requirement for biogenesis of the [4Fe-4S] cluster of FNR is greater under aerobic than anaerobic conditions (Mettert et al., 2008), we considered the possibility that acquisition of the [2Fe-2S] cluster by IscR is reduced under aerobic conditions because of the competition with other substrates for the Isc machinery. Such a competition mechanism could explain how IscR senses the cellular Fe-S demands and correspondingly adjusts expression of the Isc pathway. If IscR senses the cellular Fe-S status through its [2Fe-2S] cluster, then conditions that increase the demand for Fe-S clusters are predicted to affect [2Fe-2S]-IscR activity. This notion was tested by measuring P_iscR-lacZ activity upon overexpression of the [4Fe-4S] protein FNR from P_tac. Since FNR does not regulate P_iscR (Kang et al., 2005, Giel, 2007), we expected that changes in iscR transcription may result from the competition between FNR and IscR for the Isc machinery. Furthermore, because the [4Fe-4S] cluster of FNR is not stable under aerobic conditions (Khoroshilova et al., 1997), we also overexpressed the mutant protein FNR-L28H, whose [4Fe-4S] cluster is more resistant to destruction by oxygen (Bates et al., 2000). Under aerobic conditions, P_iscR was derepressed within 20 minutes of addition of IPTG to a final concentration of 100 µM to the strain containing either the P_tac-fnr or the P_tac-fnr-L28H expression plasmid but not the vector control (Fig. 7A). Furthermore, no derepression was observed for the P_tac-fnr strain when no IPTG was added (data not shown). Thus, under aerobic conditions, an increase in the Fe-S cluster demand via overexpression of FNR relieves repression of the isc operon, presumably due to the competition between FNR and IscR for acquiring Fe-S clusters from the Isc machinery.

Fig. 7.

Activity from P_iscR-lacZ integrated at the λatt site was measured in strains (PK7329, PK7333, PK7328) containing plasmids with IPTG-inducible wild-type fnr (open circles), IPTG-inducible fnr-L28H (triangles) or the vector control (closed circles). Strains were grown under A) aerobic or B) anaerobic conditions in M9 minimal media containing glucose (0.2%), Ap (50 µg ml⁻¹), and Cm (20 µg ml⁻¹). When cells reached an OD₆₀₀ of 0.2 (0 minutes), IPTG was added to a final concentration of 100 µM, samples were taken at various time points, and β-galactosidase activity (Miller units) was assayed and normalized for cell number as explained in Experimental Procedures. Error bars represent the standard errors of three biological replicates.

In contrast, derepression of P_iscR was not observed when FNR was overexpressed under anaerobic conditions (Fig. 7B), despite the accumulation of FNR protein to similar levels as under aerobic conditions (data not shown). This suggests that the Isc machinery can meet an increase in Fe-S demand under anaerobic conditions such that no competition between FNR and IscR was detected. Thus, the Isc pathway seems to be more buffered against changes in Fe-S demand under anaerobic conditions, likely due to less general cluster turnover under these conditions. In summary, this apparent difference in the cellular Fe-S demand between aerobic and anaerobic conditions highlights the significance of negative feedback regulation of the isc operon.

DISCUSSION

The findings of this study have revealed new mechanistic insight as to how IscR regulates transcription of P_iscR to maintain Fe-S cluster homeostasis in E. coli. We have defined the requirement for the two Type 1 IscR binding sites within P_iscR for full repression of the isc operon and have shown that these sites are preferentially bound by the [2Fe-2S] form of IscR. Examination of P_iscR repression under aerobic and anaerobic growth has revealed that the increased repression of isc under anaerobic conditions is due to increased IscR [2Fe-2S] cluster occupancy under these conditions. Furthermore, decreased isc repression under aerobic conditions suggests that there may be more competition between IscR and substrate proteins for the Isc machinery when O₂ is present. Consistent with this notion, overexpression of an Fe-S protein under aerobic conditions, but not under anaerobic conditions, led to derepression of P_iscR. In particular, our results shed new insight on how IscR responds to the presence of O₂ to satisfy the cellular need for Fe-S cluster biogenesis.

IscR binding sites A and B are critical for IscR-mediated repression of P_iscR

Of the three IscR binding sites present within P_iscR (Giel et al., 2006), we found that only the two adjacent Type 1 IscR binding sites, A and B, are responsible for P_iscR repression in vivo under standard growth conditions. The requirement for the [2Fe-2S] form of IscR for repression also correlated well with the increased affinity of [2Fe-2S]-IscR for either site A or B. Nevertheless, the in vivo concentration of the clusterless mutant IscR-C92A/C98A/C104A is sufficiently high (≥17 µM; Table 2) such that some binding of the mutant might be expected even though its binding characteristics in vitro were poor. Perhaps an additional mechanism exists in vivo to prevent repression of P_iscR by IscR apo-protein. Because apo-IscR binds to many locations on the E. coli chromosome (Myers and Kiley, unpublished), one possibility is that apo-IscR is bound at other higher affinity sites across the genome. Thus, like many transcription factors, regulation of promoters by IscR is likely dependent on a hierarchy of binding affinities.

Since both IscR sites A and B span the region that RNA polymerase binds, the mechanism of IscR repression can be simply explained by promoter occlusion. Indeed, DNase I footprinting experiments at wild-type P_iscR showed that addition of IscR at the same time as RNAP resulted in a protection pattern virtually identical to the one for IscR alone (Giel, 2007), indicating that binding of IscR to P_iscR occludes RNAP. Since most IscR-regulated promoters contain just a single IscR binding site (Giel et al., 2006), perhaps the presence of multiple IscR binding sites within the P_iscR region may be important to extend the dynamic range to regulate Fe-S biogenesis in vivo. While our footprinting experiments indicate that [2Fe-2S]-IscR could bind independently to these two sites, the use of linear DNA in these experiments does not rule out the possibility that some protein-protein interactions occur when present on supercoiled chromosomal DNA.

The Isc pathway is the major system for IscR [2Fe-2S] cluster biogenesis

Deletion of the iscSUA-hscBA-fdx operon resulted in decreased P_iscR repression, suggesting that IscR primarily receives its [2Fe-2S] cluster from the Isc pathway. When the Isc pathway was absent, however, a small amount of [2Fe-2S]-IscR activity remained, suggesting that IscR could acquire [2Fe-2S] clusters from another source. Similar results were also observed for the transcription factor FNR and the enzymes isopropylmalate isomerase, NADH dehydrogenase I, and IspG/H, which all require Fe-S clusters for function (Mettert et al., 2008, Jang and Imlay, 2010, Vinella et al., 2009). Since apo-IscR is sufficient to activate P_sufA transcription (Nesbit et al., 2009, Yeo et al., 2006), we propose that in the ΔiscSUA-hscBA-fdx strain, upregulation of the Suf pathway by apo-IscR may explain the residual [2Fe-2S]-IscR activity. Consistent with this notion, sufA promoter expression and SufD protein levels were elevated 2- to 3-fold in a mutant strain lacking the Isc pathway compared to the wild-type strain (Mettert et al., 2008). Furthermore, our studies indicate that IscR can in fact serve as a substrate for the Suf machinery in the ΔiscSUA-hscBA-fdx mutant when the suf operon is overexpressed. Since the Isc system has been shown to be present but non-functional during H₂O₂ stress (Jang and Imlay, 2010), the ability of an Isc substrate protein to also receive Fe-S clusters from the Suf pathway would be particularly advantageous for E. coli under conditions of oxidative stress.

IscR senses and responds to the different Fe-S demands that exist between aerobic and anaerobic growth conditions

Our finding that IscR is more sensitive to changes in Fe-S demand under aerobic conditions makes physiological sense. Considering that greater than 150 Fe-S proteins exist in E. coli (Py et al., 2011) and that some Fe-S clusters are sensitive to O₂ and/or reactive oxygen species (Imlay, 2006, Imlay, 2008), it is likely that Fe-S clusters are continually being damaged or destroyed during aerobic growth, thereby increasing the levels of substrate proteins that need Fe-S biogenesis or repair. As a result, there would be more competition between IscR and substrate proteins for the Isc machinery under aerobic conditions than anaerobic conditions. In support of this notion, we showed that aerobic but not anaerobic overexpression of either wild-type [4Fe-4S]-FNR or the [4Fe-4S]-FNR-L28H mutant derivative that contains an O₂-resistant cluster (Bates et al., 2000) relieved P_iscR repression. Why IscR may not compete as well with other substrate proteins is not known but may be related to the fact that IscR binds its [2Fe-2S] cluster with an atypical ligation scheme of three cysteines and one histidine (Fleischhacker et al., 2012) perhaps making it a poor substrate for the Isc machinery.

In sum, we propose that under aerobic conditions, the sensitivity of some Fe-S clusters to O₂ and/or reactive oxygen species results in a higher rate of cluster turnover, which in turn leads to increased synthesis of the isc operon to meet the demands for Fe-S biogenesis. In contrast, under anaerobic conditions, the Isc machinery appears to satisfy the Fe-S demand more efficiently due to decreased general cluster turnover, and thus less competition among substrate proteins, under these conditions (Fig. 8). However, we cannot rule out the possibility that additional mechanisms may also contribute to regulating Isc Fe-S cluster biogenesis under anaerobic growth. Taken together, the different needs for Isc Fe-S biogenesis under aerobic and anaerobic conditions emphasize the important role of IscR as an Fe-S sensor, enabling E. coli to respond appropriately to environments of varying O₂ tension.

Fig. 8.

A model for the differential demand in Isc-mediated Fe-S cluster biogenesis between aerobic and anaerobic growth conditions is shown. In the presence of O₂, the need for Fe-S clusters is predicted to be high due to increased rates of general Fe-S cluster turnover under aerobic conditions, and thus elevated levels of apoprotein substrates (gray squares). As a result, there is competition between these substrates and IscR in acquiring Fe-S clusters from the Isc pathway, resulting in low IscR [2Fe-2S] cluster occupancy and thus less repression of iscR-SUA-hscBA-fdx. Under anaerobic conditions, the general rate of cluster turnover is decreased compared to aerobic conditions. The low demand for Fe-S biogenesis in the absence of O₂ leads to increased IscR [2Fe-2S] cluster occupancy, and in turn, more repression of the Isc pathway.

EXPERIMENTAL PROCEDURES

Bacterial strain and plasmid construction

Strains and plasmids described in this work are listed in Table 1 and sequences of primers used are available upon request. The λ-P_iscR-lacZ reporter fusion (−476 to +74 bp relative to the IscR start codon) (Schwartz et al., 2001) is located within the λatt site, whereas the P_iscR-lacZ reporter fusion (−428 to +144 bp relative to the IscR start codon) replaces the wildtype lac promoter and was constructed by a method previously described (Giel et al., 2006, Kang et al., 2005). In the latter case, a lacI-kan-P_iscR-lacZ fragment from pPK8508 was PCR-amplified and recombined onto the chromosome in the lac promoter region of BW25993/pKD46. P_iscR-lacZ fusions containing mutations within P_iscR were recombined onto the chromosome in the same manner after introducing mutations into pPK8508 using QuikChange (Stratagene). The resulting kan-promoter-lacZ constructs were transduced into the appropriate strain backgrounds using P1vir and confirmed by colony PCR and DNA sequencing.

Table 1.

Strains and plasmids used in this work.

Strain	Relevant Genotype or Phenotype	Source
MG1655	λ− F− rph-1	Laboratory stock
PK8039	MG1655 yhgI-lacZ	(Giel et al., 2006)
PK9110	PK8039 nadA∷Tn10λcI857 Δ(cro-bioA)	This study
PK9116	PK9110 but nadA+, TetS	This study
PK9120	PK9116 mutS104∷mini-Tn10	This study
PK9133	PK9120 cat-FRT-araC-PBAD	This study
PK9520	PK9133 bla-Ptac replacing −100 to +28 bp of PiscR relative to the +1 transcription start site	This study
PK6364	MG1655 λ-iscR’-lacZ	(Schwartz et al., 2001)
PK6512	PK6364 ΔiscR	This study
PK8120	PK6364 ΔiscSUA∷cat	This study
PK8122	PK6364 ΔiscSUA-hscBA-fdx∷cat	This study
PK8614	PK8122 + pGS0164	This study
PK8618	PK8614 ΔsufABCDSE∷kan	This study
PK8615	PK8122 + pBAD/Myc-hisC	This study
PK7783	PK6364 ΔiscS	This study
PK6828	PK6364 ΔiscU	This study
PK7759	PK6364 ΔiscA	This study
PK6564	PK6364 ΔsufABCDSE∷kan	This study
PK6826	PK6364 fdx∷kan	This study
PK6848	PK6364 ΔiscX	This study
PK7540	PK6364 ΔpepB	This study
PK7751	PK6364 sseB∷Tn5KAN2	This study
PK7854	PK6364 iscR-C92A zfh-3601∷FRT	This study
PK7855	PK6364 iscR-C98A zfh-3601∷FRT	This study
PK7856	PK6364 iscR-C104 zfh-3601∷FRT	This study
PK7898	PK6364 iscR-C92A/C98A/C104A zfh-3601∷FRT	This study
PK8616	PK6364 ΔcsdA∷cat	This study
PK8617	PK6364 ΔytfE∷kan	This study
PK7325	PK6364 + pACYClacIQ-CAM	This study
PK7329	PK6364 + pACYClacIQ-CAM + pDHB60	This study
PK7333	PK6364 + pACYClacIQ-CAM + pPK7332	This study
PK7328	PK6364 + pACYClacIQ-CAM + pPK7307	This study
PK7571	MG1655 PiscR-lacZ (in lac region)	This study
PK8151	PK7571 ΔlacY	This study
PK9523	PK8151 bla-Ptac-iscR-FRT-cat-FRT-araC-PBAD-iscS + pACYClacIQ	This study
PK8677	PK8151 ΔiscR + pACYClacIQ	This study
PK8521	MG1655 PiscR (site A mutated)-lacZ	This study
PK8528	MG1655 PiscR (site B mutated)-lacZ	This study
PK8582	MG1655 PiscR (sites A & B mutated)-lacZ	This study
PK8551	MG1655 PiscR (site C mutated)-lacZ	This study
PK8801	MG1655 PiscR (sites A & C mutated)-lacZ	This study
PK8808	MG1655 PiscR (sites B & C mutated)-lacZ	This study
PK8820	MG1655 PiscR (sites A, B, & C mutated)-lacZ	This study
PK4854	MG1655 ΔiscR	(Schwartz et al., 2001)
PK7572	PK4854 PiscR-lacZ	This study
PK8523	PK4854 PiscR (site A mutated)-lacZ	This study
PK8530	PK4854 PiscR (site B mutated)-lacZ	This study
PK8584	PK4854 PiscR (sites A & B mutated)-lacZ	This study
PK8553	PK4854 PiscR (site C mutated)-lacZ	This study
PK8803	PK4854 PiscR (sites A & C mutated)-lacZ	This study
PK8810	PK4854 PiscR (sites B & C mutated)-lacZ	This study
PK8822	PK4854 PiscR (sites A, B, & C mutated)-lacZ	This study
PK7599	BL21 himA∷tet + pPK6161	(Nesbit et al., 2009)
PK7881	BL21 himA∷tetΔiscR∷kan + pPK7862	(Nesbit et al., 2009)
DY329	W3110 ΔlacU169nadA∷Tn10gal490 λcI857 Δ(cro- bioA)	(Yu et al., 2000)
CSH115	F-, ara-600, Δ(gpt-lac)5 λ−, mutS104∷mini-Tn10 relA1?, spoT1?, thi-1	The Coli Genetic Stock Center (Yale)
BW25993	lacIqhsdR514ΔaraBADAH33ΔrhaBADLD78	(Datsenko and Wanner, 2000)
Plasmid
pPK6161	iscR cloned into pET-11a	(Schwartz et al., 2001)
pPK7862	iscR-C92A/C98A/C104A cloned into pET-11a	(Nesbit et al., 2009)
pPK823	fnr cloned into pET-11a	(Lazazzera et al., 1993)
pPK1868	fnr-L28H cloned into pET-11a	(Bates et al., 2000)
pDHB60	ApR, Ptac (pBR322-derived)	(Boyd et al., 2000)
pPK7332	XbaI-HindIII fnr fragment from pPK823 cloned into XbaI and HindIII sites of pDHB60	This study
pPK7307	XbaI-HindIII fnr-L28H fragment from pPK1868 cloned into XbaI and HindIII sites of pDHB60	This study
pKD13	FRT-kan-FRT	(Datsenko and Wanner, 2000)
pKD32	Same as pKD13 but with FRT-cat-FRT	B. L. Wanner
pCP20	ApR	(Datsenko and Wanner, 2000)
pKD46	Phage λgam-bet-exo genes under ParaB control	B. L. Wanner
pBR322	ApR	(Bolivar et al., 1977)
pPK7035	pBR322 with kan from pHP45Ω and BamHI-NdeI fragment from pRS1553	(Kang et al., 2005)
pPK8508	pPK7035 with PiscR (−428 to +144 bp relative to the IscR start codon) cloned into BamHI and XhoI sites	This study
pBAD/Myc- hisC	ApR	Invitrogen
pGS0164	sufABCDSE cloned into pBAD/Myc-hisC	(Outten et al., 2003)
pPK9125	FRT-cat-FRT from pKD32 cloned into SphI site of pBAD/Myc-hisC	This study
pACYC184	TetR, CmR cloning vector	(Chang and Cohen, 1978)
pPK5960	pACYC184 with StuI-EcoRV 985 bp fragment containing iscR and part of iscS	(Schwartz et al., 2001)
pPK9112	iscRS fragment from pPK5960 cloned into EcoRV site of pBR322	This study
pPK9127	FRT-cat-FRT-araC-PBAD fragment from pPK9125 cloned into EcoRI site of pPK9112	This study
pRZ7411	HindIII-BamHI fragment of fnr in pACYC184	(Lazazzera et al., 1993)
pPK9003	pRZ7411 based vector with bla from pBR322 at AflII site and Ptac from pPK7332 at AflII and NdeI sites	This study
pACYClacIQ	lacIQ in pACYC184, TetR	(Boyd et al., 1987)
pACYClacIQ CAM	lacIQ in pACYC184, CmR	(Derman et al., 1993)
pPK6511	PiscR cloned into pUC19-spf’	(Schwartz et al., 2001)
pPK8515	pPK6511 site A mutated	This study
pPK6806	pPK6511 site B mutated	This study
pPK8547	pPK6511 site C mutated	This study

In frame deletions of iscR, iscSUA, iscSUA-hscBA-fdx, iscS, iscU, iscA, iscX, pepB, csdA, and lacY were constructed by replacing the coding region(s) with a Cm^R or Kan^R cassette flanked by FLP recognition target (FRT) sites from plasmid pKD32 or pKD13, respectively, as described previously (Datsenko and Wanner, 2000). Transduction with P1 vir was used to move the cat or kan allele to the appropriate strain backgrounds. In some cases, the Cm^R or Kan^R cassette was removed by transforming strains with pCP20, encoding FLP recombinase (Datsenko and Wanner, 2000) and by screening for Cm or Kan sensitivity. All gene deletions were confirmed by colony PCR. Previously constructed alleles encoding IscR mutants in which cysteine residues were substituted with alanine (Nesbit et al., 2009)were transduced into PK6364 using P1 vir and selecting for Kan^R. The Kan^R cassette was subsequently removed by transforming strains with pCP20. The following existing alleles were also transduced into appropriate strain backgrounds using P1vir∷sseB with a Tn5KAN2 insertion at position 150, fdx∷kan, ΔsufABCDSE∷kan, and ΔytfE∷kan.

Construction of bla-P_tac-iscR-cat-araC-P_BAD-iscSUA on the chromosome of strain PK9520 was performed in several steps. First, a MG1655 recombineering strain (PK9120) was constructed by transducing nadA∷Tn10 [λcI857Δ(cro-bioA)] from DY329 into PK8039 using P1 vir and selecting for Tet^R at 30°C to form PK9110. The nadA∷Tn10 was removed by transducing to nicotinic acid prototrophy using P1 vir MG1655 and screening for Tet^S at 30°C to make PK9116. Colonies retaining the biotin requirement were then transduced with P1 vir mutS104∷mini-Tn10 from CSH115 to form PK9120. Next, FRT-cat-FRT was PCR-amplified from pKD32 and cloned into the SphI site of pBAD/Myc-hisC to form pPK9125. FRT-cat-FRT-araC-P_BAD was amplified and inserted between iscR and iscS in pPK9112, using the EcoRI site 6 nt after the iscR stop codon to form pPK9127. To obtain FRT-cat-FRT-araC-P_BAD upstream of iscS on the chromosome, this construct was amplified with primers containing ends with homology to this region of the chromosome, electroporated into PK9120, and selected for Cm^R. The bla-P_tac region from pPK9003 was amplified and electroporated into the above PK9120 derivative (PK9133) to replace −100 to +28 bp of P_iscR (relative to the +1 transcription start site) with bla-P_tac, forming PK9520. After verification by DNA sequencing, bla-P_tac-iscR-FRT-cat-FRT-araC-P_BAD-iscS was transduced into PK8151 using P1 vir and selected for Cm^R and Ap^R with 10 mM arabinose to derepress P_BAD-iscSUA. Finally, the strain was transformed with pACYClacI^Q to form PK9523.

Wild-type fnr and fnr-L28H were cloned into pDHB60 by digesting pPK823 or pPK1868, respectively, with HindIII and XbaI, followed by ligation of these fragments into the HindIII and XbaI sites of pDHB60. PK6364 was transformed with pACYClacI^Q-CAM, and the resulting strain was transformed with pDHB60, pPK7332 (fnr), or pPK7307 (fnr-L28H).

β-galactosidase assays

Strains were grown aerobically by shaking at 250 rpm to an OD₆₀₀ ∼0.2 or anaerobically in screw-capped tubes to an OD₆₀₀ ∼0.1 at 37° C in MOPS minimal medium (Neidhardt et al., 1974) supplemented with 0.2% glucose (w/v) or M9 minimal medium (Miller, 1972) containing 0.2% glucose (w/v), 1mM MgSO₄, 2.5 µg ml⁻¹ ferric ammonium citrate, 2 µg ml⁻¹ thiamine, and 0.02% ammonium molybdate (w/v). As indicated in the figure legends, media were also supplemented with 12.5 µg ml⁻¹ nicotinic acid and 2 µg ml⁻¹ thiamine for some experiments. For the FNR overexpression experiment, strains were grown under aerobic or anaerobic conditions by sparging (Sutton and Kiley, 2003). For all β-galactosidase assays, Cm or Tet were added to culture samples at final concentrations of 20 or 10 µg ml⁻¹, respectively, to terminate cell growth and any further protein synthesis; cells were placed on ice until assayed for β-galactosidase activity as previously described (Miller, 1972). Because of aerobic and anaerobic cell count differences, the Miller units from aerobically grown strains were multiplied by 1.55, similar to previous studies (Mettert and Kiley, 2007). Assays were repeated at least three independent times and the standard errors for data plotted as “Fold Repression” were calculated using a propagation of standard error formula (Ku, 1966).

For the FNR overexpression experiment, strains were grown in M9 minimal medium (described in the above paragragh) containing 50 µg ml⁻¹ Ap, and 20 µg ml⁻¹ Cm under aerobic or anaerobic conditions by sparging (Sutton and Kiley, 2003). FNR expression was induced by adding a final concentration of 100 µM IPTG at OD₆₀₀ of 0.2. Samples were removed at various time points and processed as above, and β-galactosidase assays were performed in triplicate for each time point. The experiment was repeated on at least two separate occasions.

Western blot analysis

Strains were grown as for β-galactosidase assays, and IscR, SufD, or FNR levels were measured by Western blots as described previously (Nesbit et al., 2009, Mettert et al., 2008, Sutton et al., 2004). IscR was quantified using isolated proteins as standards, followed by imaging as described (Nesbit et al., 2009, Mettert et al., 2008). The cytoplasmic concentration of IscR (µM) was calculated using the molecular weight determined for monomeric IscR (17,336 Da) (Schwartz et al., 2001), the number of cells per ml of culture sampled as previously determined via viable cell counts (Mettert et al., 2008), and the estimated cell volume of ∼1 × 10⁻¹⁵ L (Kubitschek, 1990).

Protein purification

Purifications of [2Fe-2S] cluster-containing wild-type IscR (Giel et al., 2006) and IscR-C92A/C98A/C104A (Nesbit et al., 2009) were performed and the protein concentration and iron and sulfide content were determined as previously described (Beinert, 1983, Kennedy et al., 1984, Khoroshilova et al., 1995). For all in vitro experiments, wild-type IscR was ≥50% occupied with [2Fe-2S] clusters. For simplicity, all protein concentrations herein are reported as monomers, although IscR is mainly a dimer in solution (Nesbit et al., 2009).

DNase I footprinting assays

Mutations within P_iscR were introduced into pPK6511 using QuikChange (Stratagene), generating pPK8515, pPK6806, and pPK8547, and DNA fragments containing the wild-type or mutated iscR promoter region were isolated from their respective plasmids after digestion with XbaI and HindIII. Klenow fragment (NEB) and [α-³²P]dCTP were used to label the XbaI end of the fragment (comprised of bases −161 to +38 relative to the P_iscR transcription start site). Assays were performed as described (Giel et al., 2006) except the incubation time of protein with DNA prior to the addition of DNase I was 10 minutes.

Fluorescence anisotropy

DNA binding isotherms were generated under anaerobic conditions by measuring changes in fluorescence polarization when IscR bound dsDNA as described (Nesbit et al., 2009), and assays were repeated on three independent occasions. Briefly, 30-mer dsDNA contained the following Type 1 sites (underlined): P_iscR site A (5'-AAATACCCGACTAAATCAGTCAAGTAAATA-3') or P_iscR site B (5'-AAATAGTTGACCAATTTACTCGGGAATGTC-3'). [2Fe-2S]-IscR (20 to 718 nM) or IscR-C92A/C98A/C104A (35 to 1000 nM) was incubated with 5 nM Texas Red labeled-DNA, 40 mM Tris-Cl (pH 7.9), and 150 mM KCl for 10 min at room temperature under anaerobic conditions. For wild-type IscR, the fraction bound was determined as previously described (Nesbit et al., 2009).