RsrR : a novel redox sensitive Rrf 2 family transcription factor in Streptomyces 1 venezuelae 2 3

RsrR: a novel redox sensitive Rrf2 family transcription factor in Streptomyces 1 venezuelae 2 3 John T. Munnoch, M Teresa Pellicer Martinez, Dimitri A. Svistunenko, Jason C. Crack, 4 Nick E. Le Brun and Matthew I. Hutchings 5 6 School of Biological Sciences, University of East Anglia, Norwich, Norwich Research Park, 7 NR4 7TJ 8 Centre for Molecular and Structural Biochemistry, School of Chemistry, University of East 9 Anglia, Norwich, Norwich Research Park, NR4 7TJ 10 School of Biological Sciences, University of Essex, Wivenhoe Park, Colchester CO4 3SQ 11 12 Running Head: RsrR is a novel redox sensor in S. venezualae. 13 14 Address correspondence to: n.le-brun@uea.ac.uk; m.hutchings@uea.ac.uk 15 16 Abstract length: 225 words. 17

Introduction.Filamentous Streptomyces bacteria produce bioactive secondary metabolites that account for more than half of all known antibiotics as well as anticancer, anti-helminthic and immunosuppressant drugs (1,2).More than 600 Streptomyces species are known and each encodes between 10 and 50 secondary metabolites but only 25% of these compounds are produced in vitro so there is huge potential for the discovery of new natural products from Streptomyces and their close relatives.This is revitalizing research into these bacteria and Streptomyces venezuelae has recently emerged as a new model for studying their complex life cycle, in part because of its unusual ability to sporulate to near completion when grown in submerged liquid culture.This means the different tissue types involved in the progression to sporulation can be easily separated and used for tissue specific analyses such as RNA and ChIP-seq (3,4).Streptomyces species are complex bacteria that grow like fungi, forming a branching, feeding substrate mycelium in the soil that differentiates upon nutrient stress into reproductive aerial hyphae that undergo cell division to form spores (5).Differentiation is closely linked to the production of antibiotics which are presumed to offer a competitive advantage when nutrients become scarce in the soil.
Streptomyces bacteria are well adapted for life in the complex soil environment with more than a quarter of their ~9 Mbp genomes encoding one and two-component signaling pathways that allow them to rapidly sense and respond to changes in their environment (6).
They are facultative aerobes and have multiple systems for dealing with redox, oxidative and nitrosative stress.Most species can survive for long periods in the absence of O 2 , most likely by respiring nitrate, but the molecular details are not known (7).They deal effectively with nitric oxide (NO) generated either endogenously through nitrate respiration (7) or in some cases from dedicated bacterial NO synthase (bNOS) enzymes (8) or by other NO generating organisms in the soil (9).We recently characterised NsrR, which is the major bacterial NO stress sensor in Streptomyces coelicolor (ScNsrR).NsrR is a dimeric Rrf2 family protein with peer-reviewed) is the author/funder.All rights reserved.No reuse allowed without permission.
The copyright holder for this preprint (which was not .http://dx.doi.org/10.1101/050989doi: bioRxiv preprint first posted online Apr. 29, 2016; to class 2 sites that are not identified using MEME.This is supported by the fact that RsrR binds weakly to artificial half sites in vitro.Class 1 target genes include NAD(P) dependent oxidoreductases and the NADH and NADPH dehydrogenase operons consistent with a primary role for RsrR in regulating NAD(P)H metabolism in response to redox changes in the cell.

RsrR regulates genes involved in NAD(P)H metabolism. To investigate RsrR function in S.
venezuelae we decided to first identify target genes for RsrR in S. venezuelae.We constructed an S. venezuelae ∆rsrR mutant that expresses an N-terminally 3xFlag-tagged RsrR protein and performed ChIP-seq against this strain and wild-type S. venezuelae (ChIP-Seq accession number -TBC).The sequencing reads from the wild-type (control) sample were subtracted from the experimental sample before ChIP peaks were called (Fig. 1A).With no defined cut-off or arbitrary cut-offs of ≥200 reads or ≥500 sequencing reads per peak we identified >2700, >600 and 119 enriched target sequences, respectively (Supplementary File S1).A subset of the 119 targets can be found in Table 1.Working with the shortlist of 119 targets we confirmed the peaks by visual inspection of the data using Integrated Genome Browser (17).Fourteen of the ChIP peaks are in intergenic regions between divergent genes giving a total of 133 possible targets (Table S3).The core MEME suite tool, MEME (18) was used to search for a consensus RsrR binding site in all 119 sequences and identified a single conserved motif present in all 126 sequences (Fig. 1B and Supplementary File S4).In 14 of these 119 peaks a motif is present corresponding to an inverted 11-3-11bp repeat, which is characteristic of full-length Rrf2 binding sites and we called these class 1 targets (Fig. 1C, Table 1 and Supplementary File S5).Previous studies of E. coli NsrR have suggested that target genes with full 11bp inverted repeat binding sites are most strongly repressed and peer-reviewed) is the author/funder.All rights reserved.No reuse allowed without permission.
The copyright holder for this preprint (which was not .http://dx.doi.org/10.1101/050989doi: bioRxiv preprint first posted online Apr. 29, 2016; therefore most physiologically relevant, thus perhaps giving clues about the primary function of RsrR.Two of the class 1 sites are between divergent genes (sven3827/8 and sven6562/3the rsrR peak).The 107 bp intergenic region between sven6562 and rsrR, contains two putative class 1 RsrR binding sites separated by a single base pair.sven6562 encodes a LysR family regulator with an NmrA-type ligand binding domain predicted to sense redox poise by binding NAD(P)+ but not NAD(P)H (19).From hereon we refer to sven6562 as nmrA.The positions of the two RsrR binding sites relative to the transcript start sites (TSS) of sven6562 and rsrR suggests that RsrR represses transcription of both genes by blocking the RNA polymerase binding site.Other class 1 targets include the nuo (NADH dehydrogenase) operon sven4265-78 (nuoA-N) which contains an internal class 1 RsrR site (upstream of sven4272, nuoH), the putative NADP+ dependent dehydrogenase Sven1847 and the quinone oxidoreductase Sven5174 which converts quinone and NAD(P)H to hydroquinone and NAD(P)+ (Table 1).Based on this data we suggest RsrR plays a primary role in regulating NAD(P)H metabolism and possibly senses redox poise in the cell.Intriguingly, however, dRNA-seq (expression data for the regulon is available in File S1.5 and TSS data in S3. [6][7][8][9] (dRNA-seq accession number -TBC) suggests only a single class 1 targets is induced in an rsrR mutant and that is the divergent gene nmrA.This is probably because the other genes are subject to more complex regulation from multiple transcription factors.The remaining 105 genes on the RsrR shortlist were classified as class 2 targets because they have single copies of the class 2 motif, which we call half sites (Fig. 1B).Half sites have been observed for other Rrf2 proteins including E. coli NsrR and these half-site promoters are subject to much weaker repression and their physiological relevance is unclear (20)(21)(22).
Purified RsrR contains a redox active [2Fe-2S] cluster.RsrR contains three C-terminal cysteine residues which is characteristic of Rrf2 proteins that ligate Fe-S clusters.To peer-reviewed) is the author/funder.All rights reserved.No reuse allowed without permission.
The copyright holder for this preprint (which was not .http://dx.doi.org/10.1101/050989doi: bioRxiv preprint first posted online Apr. 29, 2016; investigate the cofactor and DNA binding activity of RsrR we over-expressed the rsrR gene in E. coli and purified the protein under strictly anaerobic conditions.Upon purification, the fractions containing RsrR were pink in colour but rapidly turned brown when exposed to O 2 , suggesting the presence of a redox-active cofactor.The UV-visible absorbance spectrum of the as isolated protein, Fig. 2A, revealed broad weak bands in the 300-640 nm region. Following exposure to O 2 , the spectrum changed significantly, with a more intense absorbance band at 460 nm and a pronounced shoulder feature at 330 nm (Fig. 2A).The form of the reduced and oxidized spectra are similar to those previously reported for [2Fe-2S] clusters that are coordinated by three Cys residues and one His (23,24).The anaerobic addition of dithionite to the previously air-exposed sample (at a 1:1 ratio with [2Fe-2S] cluster as determined by iron content) resulted in a spectrum very similar to that of the as isolated protein (Fig. 2A), demonstrating that the cofactor undergoes redox cycling.
Because the electronic transitions of iron-sulfur clusters become optically active as a result of the fold of the protein in which they are bound, CD spectra reflect the cluster environment (25).The near UV-visible CD spectrum of RsrR (Fig. 2B) for the as isolated protein contained three positive (+) features at 303, 385 and 473 nm and negative features at (-) 343 and 559 nm.When the protein was exposed to ambient O 2 for 30 min, significant changes in the CD spectrum were observed, with features at (+) 290, 365, 500, 600 nm and (-) 320, 450 and 534 nm (Fig. 2B).The CD spectra are similar to those reported for Riesketype [2Fe-2S] clusters (23,26,27), which are coordinated by two Cys and two His residues.

Anaerobic addition of dithionite (1 equivalent of [2Fe-2S] cluster) resulted in reduction back
to the original form (Fig. 2B) consistent with the stability of the cofactor to redox cycling.The absorbance data above indicates that the cofactor is in the reduced state as isolated.[2Fe-2S] clusters in their reduced state are paramagnetic (S = ½) and therefore should give rise to an EPR signal.The EPR spectrum for the as isolated protein contained peer-reviewed) is the author/funder.All rights reserved.No reuse allowed without permission.
The copyright holder for this preprint (which was not .http://dx.doi.org/10.1101/050989doi: bioRxiv preprint first posted online Apr. 29, 2016; signals at g = 1.997, 1.919 and 1.867 (Fig. 2C).These g-values and the shape of the spectrum are characteristic of a [2Fe-2S] 1+ cluster.The addition of excess sodium dithionite to the as isolated protein did not cause any changes in the EPR spectrum (Fig. 2C) indicating that the cluster was fully reduced as isolated.Exposure of the as isolated protein to ambient O 2 resulted in an EPR-silent form, with only a small free radical signal typical for background spectra, consistent with the oxidation of the cluster to the [2Fe-2S] 2+ form (Fig. 2C), and the same result was obtained upon addition of the oxidant potassium ferricyanide (data not shown).
To further establish the cofactor that RsrR binds, native ESI-MS was employed.Here, a C-terminal His-tagged form of the protein was ionized in a volatile aqueous buffered solution that enabled it to remain folded with its cofactor bound.The deconvoluted mass spectrum contained several peaks in regions that corresponded to monomer and dimeric forms of the protein, (Fig. S6).In the monomer region (Fig. 3A), a peak was observed at 17,363 Da, which corresponds to the apo-protein (predicted mass 17363.99Da), along with adduct peaks at +23 and +64 Da due to Na + (commonly observed in native mass spectra) and most likely two additional sulfurs (Cys residues readily pick up additional sulfurs as persulfides (28), respectively.A peak was also observed at +176 Da, corresponding to the protein containing a [2Fe-2S] cluster.As for the apo-protein, peaks corresponding to Na + and sulfur adducts of the cluster species were also observed (Fig. 3A).A significant peak was also detected at +120 Da which corresponds to a break down product of the [2Fe-2S] cluster (from which one iron is missing, FeS 2 ).In the dimer region, the signal to noise is significantly reduced but peaks are still clearly present (Fig. 3B).The peak at 34,726 Da corresponds to the RsrR homodimer (predicted mass 34727.98Da), and the peak at +352 Da corresponds to the dimer with two [2Fe-2S] clusters.A peak at +176 Da is due to the dimer containing one [2Fe-2S] cluster.A range of cluster breakdown products similar to those peer-reviewed) is the author/funder.All rights reserved.No reuse allowed without permission.
The copyright holder for this preprint (which was not .http://dx.doi.org/10.1101/050989doi: bioRxiv preprint first posted online Apr. 29, 2016; detected in the monomer region were also observed (Fig. 3B).Taken together, the data reported here demonstrate that RsrR contains a [2Fe-2S] cluster that can be reversibly cycled between oxidised (+2) and reduced (+1) states.The reduced cluster was obtained by reduction with sodium dithionite (confirmed by UVvisible absorbance) and the reduced state was maintained using EMSA running buffer containing an excess of dithionite.The resulting EMSAs, Fig. 4B and C, show that, in both cases, DNA-binding occurred but the oxidised form bound significantly more tightly.Tight binding could be restored to the reduced RsrR samples by allowing it to re-oxidise in air (data not shown).We cannot rule out that the apparent low affinity DNA binding observed for the reduced sample results from partial re-oxidation of the cluster during the electrophoretic experiment.Nevertheless, the conclusion is unaffected: oxidised, [2Fe-2S] 2+ RsrR is the high affinity DNA-binding form.

Cluster-and oxidation state dependent binding of
peer-reviewed) is the author/funder.All rights reserved.No reuse allowed without permission.

Oxidised [2Fe-2S] RsrR binds strongly to class 1 and 2 promoters in vitro.
To further investigate the DNA binding activities of [2Fe-2S] 2+ RsrR EMSAs were performed on two class 2 promoters sven0247 and sven519 (Fig. 5A).Both class 2 promoters we tested were shifted by oxidized [2Fe-2S] RsrR thus showing that RsrR binds strongly to both full and half site (class 1 and 2) promoters.To further test the idea of full and half site binding, we constructed a series of mutated nmrA-rsrR promoter fragments carrying both natural class 1 sites (Fig. 5B), or artificial half sites (Fig. 5C).The results show that RsrR binds strongly to both full class 1 binding sites at the nmrA-rsrR promoters (Fig. 5B) but RsrR binds only weakly to artificial half sites (Fig. 5C).This suggests that although MEME only calls half sites in most of the RsrR target genes identified by ChIP-seq they must contain sufficient sequence information in the other half to enable strong binding.
Mapping RsrR binding sites using ChIP-exo and dRNA-seq.MEME analysis of the ChIP-seq data detected only 14 full sites out of the >600 target sites bound by RsrR in S. venezuelae.However, ChIP-Seq and EMSAs show that RsrR binds tightly to target promoters whether they contain predicted class 1 or class 2 sites.To gain more information about RsrR recognition sequences and the positions of these binding sites at target promoters we combined dRNA-seq, which maps the start sites of all expressed transcripts, with ChIPexo, which uses Lambda exonuclease to trim excess DNA away from ChIP complexes leaving only the DNA which is actually bound and protected by RsrR (ChIP-exo accession number -TBC).For dRNA-seq, total RNA was prepared from cultures of wild type S. venezuelae grown for 16 hours and for the ∆rsrR mutant.ChIP-exo was performed on the ∆rsrR strain producing Flag-tagged RsrR at a single 16 hour time point.The targets identified using ChIP-exo matched the previously identified ChIP-seq targets, with 630 target genes.
Howeevr, the ChIP-exo peaks are on average only ~50bp wide.MEME analysis using all 630 peer-reviewed) is the author/funder.All rights reserved.No reuse allowed without permission.
The copyright holder for this preprint (which was not .http://dx.doi.org/10.1101/050989doi: bioRxiv preprint first posted online Apr. 29, 2016; ChIP-exo sequences identified the same class 2 binding motif in every sequence.We identified transcript start sites (TSS) for 261 of the 630 RsrR target genes using dRNA-seq data from the 16h time point (File S3.10).Fig. 6 shows a graphical representation of the class 1 targets that have clearly defined TSS, indicating the centre of the ChIP peak, the associated TSS and any genes within the ~200 bp frame.Based on the RsrR binding site position, transcription repression is most likely either by obstruction of RNA polymerase binding or blockage of transcription elongation where they are inside the coding sequence.This is consistent with a primary role for RsrR as a transcriptional repressor.
Discussion.In this work we have characterised a new member of the Rrf2 protein family, which is mis-annotated as an NsrR homologue in the S. venezuelae genome.The purified protein contains a [2Fe-2S] cluster, which is stable in the presence of O 2 and can be reversibly cycled between reduced (+1) and oxidized (+2) states.The [2Fe-2S] 2+ form binds strongly to both class 1 and class 2 bonding sequences in vitro, whereas the [2Fe-2S] 1+ form exhibited, at best, significantly weaker binding and the apo form does not bind to DNA at all.Given these observations and the stability of the Fe-S cluster to aerobic conditions, we propose that the activity of RsrR is modulated by the oxidation state of its cluster, becoming activated for DNA binding through oxidation and inactivated through reduction.Exposure to O 2 is sufficient to cause oxidation, but other oxidants may also be important in vivo.The properties of RsrR described here are reminiscent of an E. coli [2Fe-2S] cluster containing transcription factor called SoxR, which controls the regulation of another regulator, SoxS, through the oxidation state of its cluster (29).However, SoxR is a transcriptional activator that switches on soxS transcription upon oxidation of the cluster to its [2Fe-2S] 2+ state (29).
ChIP-seq and ChIP-exo analysis show that RsrR binds to a large regulon of ~630 genes in S. venezuelae and approximately 2% of these contain obvious class 1 binding sites, peer-reviewed) is the author/funder.All rights reserved.No reuse allowed without permission.
The copyright holder for this preprint (which was not .http://dx.doi.org/10.1101/050989doi: bioRxiv preprint first posted online Apr. 29, 2016; with an 11-3-11 bp inverted repeat.The fact that class 1 target genes are involved in either signal transduction and / or NAD(P)H metabolism also points to a link with redox poise and recycling of NAD(P)H to NAD(P).The >600 class 2 target genes likely bind to a full site sequence based on our EMSA results however 1 half of the site is less conserved resulting in MEME artificially reporting half site sequences.In addition to the genes involved directly in NADH/NAD(P)H metabolism, class 2 targets include 22 transcriptional regulators, genes involved in both primary and secondary metabolism, RNA/DNA replication and modification genes, transporters (mostly small molecule), proteases and a large number of genes with no known function.One of the most strongly induced target promoters in the ∆rsrR mutant is the divergent nmrA gene that encodes a LysR family regulator with an N terminal NAD(P)+ binding domain.NmrA proteins are thought to control redox poise in fungi by sensing the levels of NAD(P), which they can bind, and NAD(P)H, which they cannot (30).This is intriguing since RsrR presumably senses redox stress through reduction of its [2Fe-2S] cluster and this induces expression of NmrA which could sense redox poise via the ratio of NAD(P)/NAD(P)H and modulate expression of its own (unknown) target genes.It will be interesting to identify the overlap in target genes between the RsrR and NmrA gene regulons.
The ∆rsrR mutant has no obvious phenotype and is no more sensitive to redox active compounds or oxidative stress that the wild-type (not shown).This is not surprising given the number of systems in Streptomyces bacteria that can deal with reactive oxygen species and redox stress including detoxifying enzymes: Catalases, peroxidases (31) and superoxide dismutases (32) and associated regulators such as OxyR (33), SigR (34), OhrR (35), Rex (19) and SoxR (36).
NmrA proteins are similar in function to the Rex protein in Gram-positive bacteria for which NAD+ and NADH compete for binding.NAD+ enhances the DNA binding activity of Rex and NADH switches it off (37).Intriguingly, Rex is well conserved in Gram-positive peer-reviewed) is the author/funder.All rights reserved.No reuse allowed without permission.
The copyright holder for this preprint (which was not .http://dx.doi.org/10.1101/050989doi: bioRxiv preprint first posted online Apr. 29, 2016; bacteria but is missing from most actinomycetes, with the exception of Streptomyces species where it was first characterised (19).The reverse is true of NmrA and RsrR, which are both conserved (as back to back genes) in most filamentous actinomycetes but are missing from other Gram-positive bacteria, including most Streptomyces species.The majority of the RsrR regulon genes identified here must be subject to more complex regulation because they are not induced in the ∆rsrR background.For example, the nuo (NADH dehydrogenase) operon sven_4265-78 (nuoA-N) contains an internal class 1 RsrR site (upstream of sven_4272, nuoH) but is not expressed in the ∆rsrR strain.Nuo is alsdo known to be repressed by Rex in S. coelicolor (19) and probably other streptomycetes.It will be interesting to further investigate the potential co-regulation of RsrR and NmrA target genes and to further elucidate the global network controlled by RsrR.

Materials and Methods
Bacterial strains, plasmids, oligonucleotides and growth conditions.Bacterial strains and plasmids are listed in Table S6 and oligonucleotides are listed in Table S7.For ChIP-seq experiments, S. venezuelae strains were grown at 30 o C in MYM liquid sporulation medium (38) made with 50% tap water and supplemented with 200μl trace element solution (39) per 100ml and adjusted to a final pH.of 7.3.Disruption of rsrR was carried out following the PCR-targeting method (40) as described previously described (41,42).Primers JM0109 and JM0110 were used to PCR amplify the apramycin disruption cassette from pIJ773.Cosmid SV-5-F05 was used as the template cosmid.The disruption cosmid (pJM026) was checked by PCR using primers JM0111 and JM0112.Antibiotic marked, double crossover exconjugants, were identified as previously described and confirmed once more with JM0111 and JM0112.
The 3x Flag tag copy of rsrR was synthesized by genescript (sequence is available in peer-reviewed) is the author/funder.All rights reserved.No reuse allowed without permission.
The copyright holder for this preprint (which was not Supplementary File S6) and subcloned into pMS82 using HindIII/KpnI and confirmed by PCR using primers JM0113 and JM0114.

ChIP (chromatin immunoprecipitation) -seq and exo.
ChIP-Seq was carried out as previously described (43).A spore inoculum (~5-10 ul of 1x10 8 spores) sufficient to reach an OD600 of 0.35 after 8 hours of growth was added to 35ml of MYM tap media in 250 ml flasks containing springs.Following growth to the chosen time point, the entire content of the flask was transferred to a 50 ml falcon tube for crosslinking, which was carried out by incubation at 30 o C for 30 mins with 1% final concentration of formaldehyde (v/v).Crosslinking was quenched by incubation at room temperature with glycine (final concentration of 125 mM).Mycelium was harvested by centrifugation 4000 rpm at 4 o C for 10 minutes and washed twice with ice cold PBS before transfer to a 2 ml centrifuge tube.Pellets were resuspended in 0.75 ml lysis buffer (10 mM Tris-HCl pH 8.0, 50 mM NaCl, 10 mg/ml lysozyme, 1x protease inhibitor-Roche1186170001) and incubate at 37C for 10-25 mins.Then 0.75 ml 1x IP buffer (100mM tris-HCl pH 8.0, 250 mM NaCl, 0.5% Triton x-100, 0.1% SDS, 1x protease inhibitor (Roche)) was added and samples mixed by pipetting up and down.Samples were sonicated 7x at 50Hz, 10 sec/cycle with a 1 min incubation on ice after each cycle.DNA fragmentation was checked by agarose gel electrophoresis following phenol extraction of 25 µl of the crude lysate mixed with 75 µl of TE buffer with 100-200 µl of phenol/chloroform.Contaminating RNA was removed with 2 µl RNase (1mg/ml) added to extracted DNA followed by an incubation for 30 min at 37°C.A smear of DNA from 200 to 1000 bp with the majority of DNA 200-400 bp should be visible.
Crude lysate was centrifuged at 13,000 rpm for 15 minutes at 4°C to clear the sample of cell debris.M2 affinity beads (Sigma-Aldrich #A2220) were prepared by washing in ½IP buffer peer-reviewed) is the author/funder.All rights reserved.No reuse allowed without permission.
The copyright holder for this preprint (which was not .http://dx.doi.org/10.1101/050989doi: bioRxiv preprint first posted online Apr. 29, 2016; following manufacturers instructions.The cleared lysate was incubated with 40 µl of washed M2 beads and incubated for 4 h at 4C in a vertical rotor.The lysate was removed and the beads pooled into one 1.5 microfuge tube and washed in 0.5 IP buffer.The beads were transferred to a fresh microfuge tube and washed a further 3 times removing as much buffer as possible without disturbing the beads.The DNA-protein complex was eluted from the beads with 100 µl elution buffer (50 mM Tris-HCl pH7.6, 10mM EDTA, 1% SDS) by incubating at 65°C overnight.Removing the ~100µl elution buffer, an extra 50 µl of elution buffer was added and further incubated at 65 o C for 5 min.To extract the DNA 150 µl eluate, 2 µl proteinase K (10 mg/ml) was added and incubated 1.5 hrs at 55 o C. To the reaction 150 µl phenol-chloroform was added.Samples were vortexed and centrifuged at full speed for 10 min.The aqueous layer was extracted and purified using the Qiaquick column from Qiagen with a final elution using 50 µl EB buffer (Qiagen).The concentration of samples were determined using Quant-iT™ PicoGreen ® dsDNA Reagent (Invitrogen) or equivalent kit or by nanodrop measurement.
DNA sequencing of ChIP-Seq samples was carried out by GATC.ChIP-exo following sonication of lysates was carried out by Peconic LLC (State College, PA) adding an additional exonuclease treatment to the process as previously described (44).

RNA-seq
Mycelium was harvested at experimentally appropriate time points and immediately transferred to 2 ml round bottom tubes, flash frozen in liquid N 2 , stored at -80 o C or used immediately.All apparatus used was treated with RNaseZAP (Sigma) to remove RNases for a minimum of 1 hour before use.RNaseZAP treated mortar and pestles were used, the pestle being placed and cooled on a mixture of dry ice and liquid N 2 with liquid N 2 being poured into the bowl and over the mortar.Once the bowl had cooled the mycelium samples were peer-reviewed) is the author/funder.All rights reserved.No reuse allowed without permission.
The copyright holder for this preprint (which was not added directly to the liquid N 2 and thoroughly crushed using the mortar leaving a fine powder of mycelium.Grindings were transferred to a pre-cooled 50 ml Falcon tube and stored on dry ice.Directly to the tube, 2 ml of TRI reagent (Sigma) was added to the grindings and mixed.
Samples are then thawed while vortexing intermittently at room temperature for 5-10 minutes until the solution cleared.To 1 ml of TRI reagent resuspension, 200 µl of chloroform was added and vortexed for 15 seconds at room temperature then centrifuged for 10 minutes at 13,000 rpm.The upper, aqueous phase (clear colourless layer) was removed into a new 2 ml tube.The remainder of the isolation protocol follows the RNeazy Mini Kit (Qiagen) instructions carrying out both on and off column DNase treatments.On column treatments were carried out following the first RW1 column wash.DNaseI (Qiagen) was added (10 µl enzyme, 70 µl RDD buffer) to the column and stored at RT for 1 hour.The column was washed again with RW1 then treated as described in the manufacturer's instructions.Once eluted from the column, samples were treated using TURBO DNA-free Kit (Ambion) following manufacturer's instructions to remove residual DNA contamination.
Data analysis was carried out as described in the ChIP-Seq/exo section for visualisation, as well as expression profiling using CLC genomics workbench 8 and the TSSAR webservice for dRNA transcription start site analysis (45).In addition a manual visual processing approach was carried out for each.

Purification of RsrR.
5 L Luria-Bertani medium (10 × 500 mL) was inoculated with freshly transformed BL21 (DE3) E. coli containing a pGS-21a vector with the prsrR-His insert.100 µg/mL ampicillin and 20 µM ammonium ferric citrate were added and the cultures were grown at 37 °C, 200 rpm until OD 600 nm was 0.6-0.9.To facilitate in vivo iron-sulfur cluster formation, the flasks were placed on ice for 18 min, then induced with 100 µM IPTG and incubated at 30 ºC and peer-reviewed) is the author/funder.All rights reserved.No reuse allowed without permission.
105 rpm.After 50 min, the cultures were supplemented with 200 µM ammonium ferric citrate and 25 µM L-Methionine and incubated for a further 3.5 h at 30 ºC.The cells were harvested by centrifugation at 10000 × g for 15 min at 4 ºC.Unless otherwise stated, all subsequent purification steps were performed under anaerobic conditions inside an anaerobic cabinet (O 2 < 2 ppm).Cells pellets were resuspended in 70 mL of buffer A (50 mM TRIS, 50 mM CaCl 2 , 5% (v/v) glycerol, pH 8) and placed in a 100 mL beaker.30 mg/mL of lysozyme and 30 mg/mL of PMSF were added and the cell suspension thoroughly homogenized by syringe, removed from the anaerobic cabinet, sonicated twice while on ice, and returned to the anaerobic cabinet.The cell suspension was transferred to O-ring sealed centrifuge tubes (Nalgene) and centrifuged outside of the cabinet at 40,000 × g for 45 min at 1 °C.
The supernatant was passed through a HiTrap IMAC HP (1 x 5mL; GE Healthcare) column using an ÄKTA Prime system at 1 mL/min.The column was washed with Buffer A until A 280 nm <0.1.Bound proteins were eluted using a 100 mL linear gradient from 0 to 100% Buffer B (50 mM TRIS, 100 mM CaCl 2 , 200mM L-Cysteine, 5% glycerol, pH 8).A HiTrap Heparin (1 x 1mL; GE Healthcare) column was used to remove the L-Cysteine, using buffer C (50 mM TRIS, 2 M NaCl, 5% glycerol, pH 8) to elute the protein.Fractions containing RsrR-His were pooled and stored in an anaerobic freezer until needed.RsrR-His protein concentrations were determined using the method of Bradford (Bio-Rad Laboratories) (46), with BSA as the standard.Cluster concentrations were determined by iron assay (47), from which an extinction coefficient, ε, at 455 nm was determined as 3450 ±25 M-1 cm-1, consistent with values reported for [2Fe-2S] clusters with His coordination (23).

Preparation of Apo-RsrR
peer-reviewed) is the author/funder.All rights reserved.No reuse allowed without permission.
The copyright holder for this preprint (which was not Apo-RsrR -His was prepared from as isolated holoprotein by aerobic incubation with 1 mM EDTA overnight.

Spectroscopy and mass spectrometry
UV-visible absorbance measurements were performed using a Jasco V500 spectrometer, and CD spectra were measured with a Jasco J810 spectropolarimeter.EPR measurements were performed at 10 K using a Bruker EMX EPR spectrometer (X-band) equipped with a liquid helium system (Oxford Instruments).Spin concentrations in the protein samples were estimated by double integration of EPR spectra with reference to a 1 mM Cu(II) in 10 mM EDTA standard.For native MS analysis, His-tagged RsrR was exchanged into 250 mM ammonium acetate, pH 8, using PD10 desalting columns (GE Life Sciences), diluted to ~21 µM cluster and infused directly (0.

Electrophoretic Mobility Shift Assays (EMSAs)
DNA fragments carrying the the intergenic region between sven1847 and sven1848 of the S. venezualae chromosome were PCR amplified using S. venezualae genomic DNA with 5' 6-FAM modified primers (Table S4).The PCR products were extracted and purified using a QIAquick gel extraction kit (Qiagen) according to the manufacturer's instructions.Probes peer-reviewed) is the author/funder.All rights reserved.No reuse allowed without permission.
The copyright holder for this preprint (which was not .http://dx.doi.org/10.1101/050989doi: bioRxiv preprint first posted online Apr. 29, 2016; were quantitated using a nanodrop ND2000c.The molecular weights of the double stranded FAM labelled probes were calculated using OligoCalc (48).
Polyacrylamide gels were pre-run at 30 mA for X min prior to use.For investigations of [2Fe-2S] 1+ RsrR DNA binding, in order to maintain the cluster in the reduced state, 5 mM of sodium dithionite was added to the isolated protein and the running buffer (de-gassed for 50 min prior to running the gel).Analysis by UV-visible spectroscopy confirmed that the cluster remained reduced under these conditions.The copyright holder for this preprint (which was not  The copyright holder for this preprint (which was not RsrR to RsrR-regulated promoter DNA.To determine which form of RsrR is able to specifically bind DNA, EMSA experiments using a highly enriched ChIP target sven1847/8.Increasing ratios of [2Fe-2S] RsrR to DNA resulted in a clear shift in the mobility of the promoter DNA from unbound to bound, see Fig. 4A.Equivalent experiments with cluster-free (apo) RsrR did not result in a mobility shift, demonstrating that the cluster is required for the observed DNA-binding activity.These experiments were performed aerobically and so the [2Fe-2S] cofactor would have been in its oxidised state.To determine if oxidation state affects DNA binding activity, EMSA experiments were performed with [2Fe-2S] 2+ and [2Fe-2S] 1+ forms of RsrR.The oxidised cluster was generated by exposure to air and confirmed by UV-visible aborbance.
3 mL/h) into the ESI source of a Bruker micrOTOF-QIII mass spectrometer (Bruker Daltonics, Coventry, UK) operating in the positive ion mode.Full mass spectra (m/z 700-3500) were recorded for 5 min.Spectra were combined, processed using the ESI Compass version 1.3 Maximum Entropy deconvolution routine in Bruker Compass Data analysis version 4.1 (Bruker Daltonik, Bremen, Germany).The mass spectrometer was calibrated with ESI-L low concentration tuning mix in the positive ion mode (Agilent Technologies, San Diego, CA).

Funding information .
We are grateful to the Natural Environment Research Council for a PhD studentship to John Munnoch, to the Biotechnology and Biological Sciences Research Council for the award of grant BB/J003247/1 (to NLB and MIH), to the UEA Science Faculty for a PhD studentship to Maria Teresa Pellicer Martinez.The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Table 1 .
Combined ChIP-Seq and dRNA-Seq data for selected RsrR targets.a -Genes flanking the ChIP peak.b -Distance to the translational start codon (bp).c -Distance to the transcriptional start site (bp).

FIG. 1 .
FIG. 1. Defining the regulon and binding site for RsrR.Top panel (A) shows the whole

a-
Genes flanking the ChIP peak.b -Distance to the translational start codon (bp).c -Distance to the transcriptional start site (bp).d -Fold change WT vs.Mutant, values normalised per million reads per sample e -EMSA shift reactions have been carried out successfully and specifically f -A fold change <-2 or >2 _ -Class 1 targets.peer-reviewed) is the author/funder.All rights reserved.No reuse allowed without permission.

Table 1 . Combined ChIP-Seq and RNA-Seq data for selected RsrR targets.
SpoU rRNA Methylase family, RNA 2'-O ribose methyltransferase substrate bindingDoxX peer-reviewed) is the author/funder.All rights reserved.No reuse allowed without permission.