IscR Is Essential for Yersinia pseudotuberculosis Type III Secretion and Virulence

Type III secretion systems (T3SS) are essential for virulence in dozens of pathogens, but are not required for growth outside the host. Therefore, the T3SS of many bacterial species are under tight regulatory control. To increase our understanding of the molecular mechanisms behind T3SS regulation, we performed a transposon screen to identify genes important for T3SS function in the food-borne pathogen Yersinia pseudotuberculosis. We identified two unique transposon insertions in YPTB2860, a gene that displays 79% identity with the E. coli iron-sulfur cluster regulator, IscR. A Y. pseudotuberculosis iscR in-frame deletion mutant (ΔiscR) was deficient in secretion of Ysc T3SS effector proteins and in targeting macrophages through the T3SS. To determine the mechanism behind IscR control of the Ysc T3SS, we carried out transcriptome and bioinformatic analysis to identify Y. pseudotuberculosis genes regulated by IscR. We discovered a putative IscR binding motif upstream of the Y. pseudotuberculosis yscW-lcrF operon. As LcrF controls transcription of a number of critical T3SS genes in Yersinia, we hypothesized that Yersinia IscR may control the Ysc T3SS through LcrF. Indeed, purified IscR bound to the identified yscW-lcrF promoter motif and mRNA levels of lcrF and 24 other T3SS genes were reduced in Y. pseudotuberculosis in the absence of IscR. Importantly, mice orally infected with the Y. pseudotuberculosis ΔiscR mutant displayed decreased bacterial burden in Peyer's patches, mesenteric lymph nodes, spleens, and livers, indicating an essential role for IscR in Y. pseudotuberculosis virulence. This study presents the first characterization of Yersinia IscR and provides evidence that IscR is critical for virulence and type III secretion through direct regulation of the T3SS master regulator, LcrF.


Introduction
Type III secretion systems (T3SS) are important components in the progression of disease for a number of clinically relevant human pathogens, including those in the genera Shigella, Salmonella, Escherichia, Chlamydia, Vibrio, Pseudomonas, and Yersinia [1,2]. The T3SS functions as an injectisome that delivers bacterial effector proteins directly into the host cell cytoplasm [2]. While the T3SS apparatus itself is structurally conserved, the repertoire of T3SS effector proteins used by each group of pathogens is distinct [2]. Thus, the effect of the T3SS on the host is unique to the needs of the pathogen [2]. While the T3SS is generally essential for a T3SS-expressing pathogen to cause disease, several aspects of the T3SS may be detrimental to bacterial growth [2]. For example, T3SS components are recognized by the host immune system [3,4]. In addition, expression of the T3SS is energetically costly and, in some organisms, T3SS induction correlates with growth arrest [5]. Therefore, regulation is essential for proper T3SS function in order to ensure that it occurs only during host cell contact in the appropriate host tissue [2,6].
Members of the genus Yersinia that utilize a T3SS are important human pathogens: Y. pestis, the causative agent of plague, and the enteropathogens Y. enterocolitica and Y. pseudotuberculosis. The Y. pseudotuberculosis Ysc T3SS is encoded on a 70-kb plasmid termed pYV [7][8][9] and is made up of approximately 25 known proteins comprising three main structures: the basal body, the needle apparatus, and the translocon [10,11]. The basal body, which displays a high degree of similarity to the flagellar basal body, is made up of rings that span the inner and outer membranes and a rod that traverses the periplasmic space [12]. Basal body associated proteins include YscN, an ATPase that aids in the secretion and translocation of effector proteins [13]. The needle complex, which is thought to act as a molecular channel for effector protein translocation, is a straight hollow appendage approximately 60 nm in length and is made up of helical polymerized subunits of YscF [12]. The translocon is comprised of three proteins: YopD, YopB and LcrV, which are essential for pore formation in the target host membrane and proper translocation of effector proteins YopHEMOJTK to the host cytoplasm [12,14]. Also encoded on pYV are chaperones important for efficient translocation of a subset of effector proteins [15]. Lastly, several transcriptional and post-transcriptional regulators of the T3SS are found on pYV, including the AraClike transcriptional regulator LcrF. LcrF is responsible for expression of a number of T3SS structural genes and Yop effectors, specifically the virC and lcrGVH-yopBD operons as well as genes encoding effector Yops, the adhesin YadA, and the lipoprotein YlpA [16][17][18][19][20][21][22]. LcrF itself is thermoregulated at both the transcriptional and translational levels through the action of the histone-like protein YmoA and a cis-acting RNA thermosensor located on the lcrF transcript, respectively [23,24]. This enables Yersinia to express T3SS genes at 37uC within the mammalian host, but not at lower temperatures [23,24]. Importantly, proper LcrF-mediated control of T3SS expression is important for Y. pseudotuberculosis virulence [24].
IscR belongs to the Rrf2 family of winged helix-turn-helix transcription factors [25,26] and has been studied extensively in E. coli, where its DNA-binding activity is dependent on coordination of an iron-sulfur [2Fe-2S] cluster through three conserved cysteines and a histidine [27][28][29][30]. E. coli IscR recognizes two distinct DNA motifs, type 1 and type 2, depending on the Fe-S status of the protein [31]. Holo-IscR coordinating an Fe-S cluster binds both type 1 and type 2 motifs, while clusterless apo-IscR recognizes only the type 2 DNA-binding motif [27,32,33]. As iron starvation, oxidative stress, and oxygen limitation affect the holo-IscR/apo-IscR ratio, these environmental cues are thought to have a direct effect on gene expression through IscR in E. coli [28][29][30]. For example, holo-IscR represses transcription of the housekeeping iscRSUA-hscBA-fdx Fe-S cluster biogenesis operon [32,34], while either holo-or apo-IscR promotes transcription of the inducible sufABCDSE Fe-S cluster biogenesis operon [33,35]. Both pathways function to insert Fe-S clusters onto proteins involved in a range of metabolic processes including electron transfer, substrate binding/activation, iron/sulfur storage, regulation, and enzyme activity [36]. In addition, E. coli IscR is also known to regulate transcription of other Fe-S cluster assembly genes such as erpA (yadR) as well as genes integral to oxidative stress resistance, biofilm formation, and anaerobic respiration [28][29][30]34]. IscR is widely conserved among bacteria [25] and its regulatory activity is integral to the infectious process of the plant pathogen Erwinia chrysanthemi [37]. Furthermore, IscR plays an important role in the virulence of the human pathogens Pseudomonas aeruginosa through modulation of the catalase katA [38], Burkholderia mallei through resistance to reactive nitrogen species [39], and Vibrio vulnificus through induction of several virulence-associated pathways [39,40]. While the iron-dependent transcriptional repressor Fur has been shown to control T3SS expression in Salmonella and Shigella [41,42], IscR has never been linked to regulation of the T3SS in any organism and has not been studied in Yersinia.
In this study, we isolated two independent IscR transposon insertion mutants in a novel screen for Y. pseudotuberculosis genes important for T3SS function. We assessed the impact of iscR deletion on Y. pseudotuberculosis in vitro and in vivo growth, type III secretion, and global gene expression. We found IscR to be essential for full T3SS function and virulence in a mouse model of infection. In addition, we provide evidence that IscR control of the T3SS stems from direct transcriptional regulation of the T3SS master regulator LcrF.

IscR is required for Y. pseudotuberculosis Ysc T3SS function
To identify regulators of the Y. pseudotuberculosis T3SS, we utilized a novel screen to isolate transposon mutants with defects in T3SS function. We previously showed that Y. pseudotuberculosis expressing a functional T3SS induces NFkB activation in HEK293T cells [43], enabling us to use host cell NFkB activation as a readout for T3SS function in Y. pseudotuberculosis transposon mutants. As some T3SS effector proteins inhibit NFkB signaling [44], we performed the screen using a Y. pseudotuberculosis transposon mutant library in a genetic background that lacked the known T3SS effector proteins YopHEMOJT (Dyop6; [43]). We identified several transposon mutants with defects in triggering activation of NFkB in HEK293T cells (L. Kwuan, N. Herrera, H. Ramirez, V. Auerbuch, data not shown), suggesting defective T3SS function. Among these were two strains with unique transposon insertions in YPTB2860 ( Figure 1A), encoding a protein with 79% identity to the E. coli iron-sulfur cluster regulator IscR, part of the iscRSUA-hscBA-fdx operon involved in Fe-S cluster biogenesis ( Figure 1B). Importantly, the helix-turn-helix DNA binding domain as well as the three cysteines and histidine known to coordinate an iron-sulfur (Fe-S) cluster in E. coli IscR are conserved in all three Yersinia species ( Figure 1B). These data indicate that Yersinia IscR may coordinate an Fe-S cluster and, as in E. coli, may regulate gene transcription.
To validate that loss of IscR in Y. pseudotuberculosis leads to T3SS defects, we isolated the two iscR transposon mutants (iscR::Tn1 and iscR::Tn2) from our library and again measured their ability to trigger NFkB activation in HEK293T cells compared to the Dyop6 parental strain and a DyscNU T3SS-null mutant [43]. In addition, we constructed an in-frame iscR deletion mutant in the Dyop6 genetic background (Dyop6/DiscR) and tested it in this assay. We found that disruption of iscR led to ,2-fold less NFkB activation relative to the Dyop6 T3SS + parental strain, although NFkB activation levels were still ,5-fold higher than a strain with complete lack of T3SS function (DyscNU; Figure 2A), suggesting that loss of iscR leads to partial T3SS loss.
To verify further that deletion of iscR leads to alterations in T3SS function, we assessed the ability of the Dyop6/DiscR mutant to insert YopBD pores in target host cell membranes by measuring

Author Summary
Bacterial pathogens use regulators that sense environmental cues to enhance their fitness. Here, we identify a transcriptional regulator in the human gut pathogen, Yersinia pseudotuberculosis, which controls a specialized secretion system essential for bacterial growth in mammalian tissues. This regulator was shown in other bacterial species to alter its activity in response to changes in iron concentration and oxidative stress, but has never been studied in Yersinia. Importantly, Y. pseudotuberculosis experiences large changes in iron bioavailability upon transit from the gut to deeper tissues and iron is a critical component in Yersinia virulence, as individuals with iron overload disorders have enhanced susceptibility to systemic Yersinia infections. Our work places this ironmodulated transcriptional regulator within the regulatory network that controls virulence gene expression in Y. pseudotuberculosis, identifying it as a potential new target for antimicrobial agents. entry of ethidium bromide (EtBr) inside Y. pseudotuberculosis-infected bone marrow derived macrophages [45,46]. Pore formation by the Dyop6/DiscR mutant was decreased by 7-fold (p,0.05) relative to the Dyop6 parental strain, which could be restored upon complementation with plasmid-encoded iscR ( Figure 2B). To determine whether loss of iscR affects T3SS function in a wild type genetic background, we constructed an in-frame iscR deletion (DiscR) in the wild type Y. pseudotuberculosis IP2666 strain expressing six of the seven known T3SS effector proteins YopHEMOJK [47]. We then visualized the secretome of the DiscR mutant relative to wild type. Deletion of iscR led to a dramatic decrease in secretion of T3SS cargo relative to the wild type background, which can be restored upon complementation with plasmid-encoded iscR ( Figure 2C). Importantly, this lack of type III secretion did not result from a defect in growth of the mutant, as the DiscR mutant actually grew better than wild type bacteria under T3SS-inducing conditions ( Figure S1A). This is consistent with a T3SS defect in this strain, as wild type Yersinia display a characteristic growth arrest upon T3SS expression [5,48,49]. Collectively, these data demonstrate that Y. pseudotuberculosis IscR is required for proper T3SS function.

IscR is required for full virulence of Y. pseudotuberculosis
Based on the knowledge that the T3SS plays an important role in the virulence of human pathogenic Yersinia, we sought to investigate whether the diminished type III secretion observed in the Y. pseudotuberculosis DiscR strain would lead to a reduction in the infectious capacity of this mutant. Mice were orogastrically infected with 2610 8 CFU of either the Y. pseudotuberculosis wild type or isogenic DiscR mutant strains. At Figure 1. Alignment of protein sequences shows a high level of conservation between E. coli and Yersinia IscR. (A) The Y. pseudotuberculosis DNA sequence, which displays the unique insertions sites for the two transposon mutants generated from our genetic screen. A space in the DNA sequence and a solid black line indicate the site of insertion for either iscR::Tn1 or iscR::Tn2. (B) Multiple sequence alignment was performed on the IscR protein sequence from E. coli K12-MG1655 and each of the three human pathogenic Yersinia spp., Y. pseudotuberculosis IP 32953 (Y. pstb), Y. enterocolitica 8081 (Y. ent) and Y. pestis CO92 (Y. pestis) using ClustalW [86]. The N-terminal helix-turn-helix DNA-binding motif is indicated by a black box. The three conserved cysteine residues (C92, C98 and C104) responsible for coordinating an Fe-S cluster are in bold and identified by black arrows [33]. Asterisks indicate nucleotides that are conserved across all four species. doi:10.1371/journal.ppat.1004194.g001 5 days post-inoculation, mice infected with the DiscR mutant displayed significantly decreased colonization of Peyer's patches and mesenteric lymph nodes (MLN) as well as diminished systemic colonization ( Figure 3). Specifically, we noted 10-and 130-fold reductions in CFU recovered from the Peyer's patches and MLNs, respectively, in mice infected with the DiscR mutant strain relative to wild type. Notably, we observed a 1000-to 10,000-fold decrease in bacterial burden in the spleen and liver respectively. The diminished ability of the DiscR mutant strain to colonize deep tissue sites is underscored by the fact that bacteria were not detected in seven of the nine livers analyzed. These findings suggest that IscR is essential for Y. pseudotuberculosis virulence in an oral infection model.

IscR deletion leads to global misregulation of gene expression in Y. pseudotuberculosis
To begin to understand the mechanistic contribution of IscR to Y. pseudotuberculosis pathogenesis, we performed high throughput transcriptome sequencing (RNAseq) analysis to determine the Y. pseudotuberculosis genes directly and indirectly controlled by IscR under iron replete, T3SS-inducing conditions. Total RNA was collected from wild type Y. pseudotuberculosis as well as the DiscR mutant strain grown in M9 at 37uC for 3 h, a point at which the DiscR and wild type strains display comparable growth rates ( Figure S1A).
For the DiscR mutant relative to the wild type, a total of 226 genes demonstrated a statistically significant fold change of $2 (Table S1). Of these, 134 genes were up-regulated in the DiscR mutant relative to the wild type (Table 1 & Figure 4A), while 92 were down-regulated (Table 2 & Figure 4B). Genes found to be up-regulated in the DiscR mutant include key elements of Fe-S cluster biosynthesis, cellular detoxification, metabolism, and protein fate ( Figure 4A). The most notable increases in transcription were observed for genes encoding Fe-S cluster biosynthesis proteins including those encoded in the isc operon, iscS (18.7-fold), iscU (21.7-fold) and iscA (13-fold) (Table 1 & Figure S2A). Additional genes encoding proteins involved in Fe-S cluster assembly and their respective fold increases include iscX/yfhJ (10.8), fdx (10.9), hscB (10), hscA (9.3), yadR/erpA (6.8), pepB (10.1) and nfuA (7.0). To validate these findings, we performed qRT-PCR analysis on the second gene encoded in the iscRSUA operon, iscS, as well as on the gene encoding the Fe-S biosynthesis protein ErpA. Transcription of iscS was increased by 30-fold, while erpA expression was increased 5-fold (p,0.05; Figure 5A). Bioinformatic analysis identified two IscR type 1 motifs upstream of the iscRSUA-hscBA-fdx operon ( Figure S2B) as well as one site each located upstream of both erpA and nfuA (data not shown). Based on this data, we propose that Y. pseudotuberculosis IscR modulates Fe-S cluster biosynthesis expression in a manner akin to that of E. coli IscR. (A) HEK293T cells expressing an NFkB luciferase reporter gene were infected with either T3SS-positive Y. pseudotuberculosis lacking the six known effector proteins YopHEMOJT (Dyop6), two isogenic iscR transposon mutants (Dyop6/iscR::Tn1 and Dyop6/iscR::Tn2), the iscR deletion mutant (Dyop6/DiscR), or a T3SS null mutant (DyscNU). At 4 h post-inoculation, T3SS functionality was determined by assessing the ability of the mutants to trigger NFkB activation in host cells by measuring bioluminescence. Shown are the average raw luminescence values of the sample compared to uninfected 6 standard error of the mean (SEM) from five independent experiments. *p#0.05, as determined by one-way ANOVA followed by Bonferroni post hoc test where each indicated group was compared to the appropriate negative (DyscNU) and positive (Dyop6) controls. (B) To analyze T3SS-dependent pore formation in macrophages, C57Bl/6 immortalized BMDMs were infected with Dyop6, a T3SS-defective mutant lacking the translocator protein YopB (Dyop6/DyopB), the iscR deletion mutant (Dyop6/iscR), or the iscR complemented strain (Dyop6/iscR pIscR), or were left uninfected. At 2 h post-inoculation, pore formation was determined by assessing the number of cells that took up ethidium bromide (EtBr) compared to the total number of Hoechst-stained cells. Shown are the averages 6 SEM from three independent experiments. *p#0.05 relative to both Dyop6 and Dyop6/iscR pIscR, as determined by one-way ANOVA followed by Bonferroni post hoc test where each indicated group was compared to the appropriate negative (Dyop6/DyopB) and positive (Dyop6) controls. (C) Y. pseudotuberculosis IP2666 wild type (WT), iscR deletion (DiscR), iscR complemented (DiscR pIscR), apo-locked IscR (apo-IscR), and apo-IscR complemented (apo-IscR pIscR) strains were grown in 2xYT low calcium media at 37uC to induce type III secretion in the absence of host cells. Proteins in the bacterial culture supernatant were precipitated and visualized alongside a protein molecular weight marker (L) on a polyacrylamide gel using commassie blue. Sample loading was normalized for OD 600 of each culture. Results are representative of three independent experiments. doi:10.1371/journal.ppat.1004194.g002 IscR is required for transcription of T3SS genes In total, 92 genes were significantly down-regulated in the DiscR mutant relative to wild type Y. pseudotuberculosis (Table 2). These data demonstrate that the majority of pYV-encoded genes are decreased in the DiscR mutant relative to the wild type strain, including genes essential for proper T3SS expression and function. The virC and lcrGVH-yopBD operons as well as genes encoding the T3SS cargo YopHEMOJTK were the most affected upon deletion of iscR: the effector proteins YopJ (23.4-fold), YopM (25.3-fold) and YopT (25.5-fold), the effector protein and translocation regulator YopK (29.3-fold), as well as a number of genes encoding T3SS structural proteins. Genes encoding regulators that control T3SS expression and function were decreased in the mutant including lcrQ (22.1-fold), lcrF (23.3-fold), lcrG (22.8-fold) and lcrH (23.9-fold). To verify that T3SS gene expression was indeed decreased in the DiscR mutant, we measured the transcript levels of the genes encoding T3SS structural proteins YscN, YscF, and the T3SS transcriptional regulator LcrF via qRT-PCR. As detailed in Figure 5B, we observed fold decreases of 2.8-fold (p,0.05), 6.9fold (p,0.001), and 5.4-fold (p,0.0001) for yscN, yscF, and lcrF, respectively. These data support our RNAseq analysis and confirm that IscR is required for robust transcription of Y. pseudotuberculosis T3SS genes.
In addition to T3SS genes, 25 other pYV-encoded genes were decreased in the mutant, but these are annotated as hypothetical proteins, transposases, and pseudogenes. Analysis of the relative abundance of pYV in the Y. pseudotuberculosis wild type and DiscR strains was performed in order to verify that the decreases in pYV-encoded genes were not a result of plasmid loss ( Figure S3). The concentration of plasmid isolated from the wild type and DiscR mutant was comparable, suggesting that the decreased transcription of pYV-encoded genes, including those encoding the T3SS, are not a result of decreased stability of the pYV plasmid.   Y. pseudotuberculosis expressing only apo-locked IscR has a proton motive force defect and cannot secrete Yops To assess the contribution of Fe-S cluster ligation to IscR control of the T3SS, we constructed an IscR mutant strain in which the three conserved cysteines were substituted with alanines (C92A, C98A, C104A; apo-locked IscR). Identical mutations in E. coli IscR render the protein unable to coordinate an iron-sulfur cluster, yet able to bind type 2 DNA binding motifs and to regulate target gene transcription [28][29][30]. We analyzed the secretome of the Y. pseudotuberculosis apo-locked IscR strain under T3SSinducing conditions and found that the mutant was just as defective as the DiscR strain in Yop secretion ( Figure 2C). This defect could be complemented with plasmid-encoded wild type IscR. As apo-locked IscR is insufficient to promote type III secretion, holo-IscR-mediated regulation of gene expression through one or more type 1 motifs may be specifically involved in regulating T3SS gene expression. Alternatively, forcing all IscR expression within the cell to the clusterless form, which leads to IscR overexpression, may lead to alterations of bacterial pathways that indirectly affect type III secretion.
Consistent with this latter explanation, the apo-locked IscR mutant exhibited decreased colony size on LB agar, slower growth in rich media ( Figure S1), and decreased motility ( Figure 6A). The flagellar basal body is a T3SS itself, indicating that the defect in the Ysc T3SS for this strain may be a result of gross abnormalities in secretion systems. Based on these findings, we set out to examine whether the apo-locked IscR mutant demonstrated alterations in membrane potential, as this has been shown to be important for both motility and Ysc T3S in Y. enterocolitica [50]. To this end, we examined bacterial membrane potential under T3SS inducing conditions. As demonstrated in Figure 6B, there is a notable decrease in membrane potential in the apo-locked IscR mutant relative to the wild type strain, which can be complemented upon addition of wild type iscR on a plasmid. Furthermore, the membrane potential of the DiscR mutant strain is comparable to that of the wild type. Collectively, these data suggest that the apo-locked IscR mutant has a proton motive force defect, leading to decreased type III secretion and motility. These findings highlight the importance for Yersinia to maintain appropriate levels of holo-IscR relative to apo-IscR in order maintain normal membrane potential.
IscR recognizes a type 2 motif upstream of the yscW-lcrF operon To begin to understand the nature of the T3SS defect in the presence of only apo-IscR, we carried out RNAseq analysis on the Y. pseudotuberculosis apo-locked IscR mutant grown under T3SSinducing conditions and compared the results with data from the wild type and DiscR strains. Curiously, the apo-locked IscR mutant displayed aberrant expression of genes involved in stress response, transport, cell envelope, as well as electron transport (data not shown). Of note, the Fe-S cluster biosynthesis proteins encoded in the iscRSUA-hscBA-fdx-iscX-pepB-sseB, yadR/erpA and nfuA operons are significantly increased in this background, similar to that of the DiscR mutant strain ( Figure S2A), indicating that holo-IscR represses expression of these genes under the aerobic, iron-replete conditions used. In contrast, increases in the sufABCDS Fe-S cluster biogenesis operon were observed for the apo-locked IscR strain when compared to both the wild type and DiscR strains ( Figure  S4). As IscR is overexpressed by 30-fold (p,0.05) in the apolocked iscR mutant compared to wild type ( Figure S2A), we speculate that the suf operon is positively regulated by IscR in Yersinia as in E. coli. In contrast, the extensively studied E. coli IscR target, hyaABCDEF, is not encoded in the Y. pseudotuberculosis genome.
Importantly, our RNAseq analysis demonstrated that transcription of genes within the virA, virB, virC, yscW-lcrF, and lcrGVH-yopBD operons was restored in the apo-locked IscR mutant compared to the DiscR mutant ( Figure 7 and Table S2). However, we observed a decrease in transcription of genes encoding the T3SS effector proteins YopH (24.4-fold), YopM (23.0-fold), YopK (27.1-fold), and YopE (22.1-fold) in the apo-locked IscR mutant compared to wild type. Transcription of yopE has been shown to be regulated by Yop secretion through a positive feedback loop [51,52], suggesting that the defect in YopHEMK transcription observed in the apo-locked IscR mutant may be caused by the lack of Yop secretion we observed in this strain. Together, these data suggest that both holo-and apo-IscR can promote T3SS gene transcription, possibly through binding to one or more type 2 DNA motifs.
To determine whether IscR might directly regulate T3SS gene expression, we carried out bioinformatic analysis to search pYV for sequences resembling the E. coli IscR type 2 motif (xxWWWWCCxYAxxxxxxxTRxGGWWWWxx) [30,31,33], as the DNA-binding domain of Yersinia IscR is 100% identical to that of E. coli IscR ( Figure 1A). We searched within the 150 nucleotides upstream of the 99 genes encoded on the pYV plasmid and obtained a ranked list of putative type 2 motifs (data not shown). Among these was a site located within the yscW-lcrF promoter region ( Figure 8A) [24]. To test whether IscR bound specifically to this site, we performed equilibrium DNA competition assays utilizing purified E. coli IscR-C92A (apo-locked IscR) [33], with a fluorescently-labeled E. coli hya type 2 site previously identified by Nesbit et al. [33]. Purified E. coli IscR was utilized in this assay, as complementation of the Y. pseudotuberculosis DiscR mutant strain with IscR of E. coli encoded on a plasmid fully restored secretion of T3SS cargo ( Figure 8B). Competitor DNA included unlabeled E. coli hya as a positive control, the identified site within the Yersinia yscW-lcrF promoter region, a mutated version of this sequence (mlcrF), where nucleotides previously demonstrated in E. coli to be important for type 2 motif binding were altered [33], as well as one of the Y. pseudotuberculosis isc type 1 motif sites we identified as a negative control ( Figure S2B & Figure 8C). We found that unlabeled lcrF DNA competed as well as unlabeled hya DNA (IC 50 27 nm and 61 nm, respectively), suggesting that IscR can indeed bind to the identified type 2 motif upstream of lcrF ( Figure 8D). Furthermore, mutation of key nucleotides in the lcrF promoter sequence led to alleviation of competition and increased the IC 50 to greater than 1000 nM, a level comparable to that of the isc negative control type 1 motif site  ( Figure 8D). These findings suggest that IscR may regulate transcription of the T3SS through a type 2 motif within the yscW-lcrF promoter region.

Discussion
In this study, we present the first characterization of the ironsulfur cluster regulator, IscR, of Yersinia. Initially identified through a genetic screen for modulators of Ysc T3SS function, iscRdeficient Y. pseudotuberculosis had a dramatic defect in secretion of T3SS effector proteins and in targeting macrophages through their T3SS, yet displayed normal growth in broth culture and wild type flagellar motility. Bioinformatic and DNA binding analysis revealed an IscR binding site upstream of the operon encoding the T3SS master regulator LcrF, indicating that IscR controls expression of the Ysc T3SS. Collectively, these findings indicated that IscR is a central component of the Y. pseudotuberculosis T3SS regulatory cascade.
Both E. coli holo-and apo-IscR are active transcription factors with distinct DNA binding targets. Holo-IscR can bind both type 1 and 2 motifs whereas apo-IscR can only bind type 2 motifs. IscR of E. coli autoregulates the isc operon, iscRSUA-hscBA-fdx, through binding to type 1 motifs within the isc promoter region [34]. In addition, Giel et al. described increased transcription of the genes located immediately downstream of the isc operon, yfhJ-pepB-sseB, in an iscR mutant, suggesting a negative regulatory effect on these genes as well [30]. We observed derepression of the iscRSUA-hscBA-fdx operon and the yfhJ-pepB-sseB locus in the Y. pseudotuberculosis DiscR mutant as well as the mutant expressing apo-locked IscR. Furthermore, we identified two sites within the Y. pseudotuberculosis isc promoter that closely match the E. coli IscR motif I consensus sequence. These data indicate that the iscRSUA-hscBA-fdx operon, and possibly the yfhJ-pepB-sseB locus, are negatively regulated by holo-IscR in Yersinia as they are in E. coli ( Figure 9A). IscR in E. coli is known to activate transcription of the sufABCDSE operon through binding to a type 2 motif [29]. Our analysis revealed that the Y. pseudotuberculosis apo-locked IscR mutant overexpresses the sufABCDS operon compared to the wild type and DiscR strains, which we predict results from the overexpression of IscR observed in the apo-locked mutant as found in E. coli [32,33]. We identified a site within the Y. pseudotuberculosis suf promoter region that closely resembles an E. coli IscR type 2 motif (data not shown). Together, these data indicate that the suf operon is positively regulated by IscR in Yersinia as in E. coli. Thus, we propose that IscR of Y. pseudotuberculosis modulates transcription of both the isc and suf Fe-S cluster biosynthesis pathways via mechanisms established for its E. coli ortholog.
In addition to control of Fe-S cluster biogenesis pathway expression, we present evidence that IscR controls expression and function of the Y. pseudotuberculosis T3SS. Bioinformatic analysis revealed a type 2 motif within the promoter of the T3SS master regulator LcrF that contained all nine bases previously found to be important for IscR binding ( Figure 8A) [33]. Indeed, DNA binding assays demonstrated that IscR is able to specifically recognize this type 2 motif, suggesting that IscR may be acting directly to promote transcription of lcrF ( Figure 9B). In support of this, we observed a marked decrease in transcription of numerous T3SS genes in the DiscR mutant strain. These include the gene that encodes LcrF, as well as a number of LcrF-regulated genes including the virC operon, yopK, yopT, yopM, yopH, yopJ, and lcrGVH-yopBD [17,20,22,53,54]. The lcrF type 2 motif is further upstream of the -10/-35 region previously identified by Böhme et al. [24] than other IscR binding sites that promote transcription [33], as we propose this site does. However, there may be an alternative 210/235 region closer to the identified motif 2 site that might be used under specific growth conditions. Together, these data suggest that IscR is required for full expression of lcrF and LcrF-regulated genes through binding to a type 2 motif in the yscW-lcrF promoter ( Figure 9B). Based on these findings, an IscR mutant unable to coordinate an Fe-S cluster (apo-locked IscR) should lead to restoration of T3SS expression. Indeed, transcription of the yscW-lcrF and virC operons, as well as the majority of genes in the lcrGVH-yopBD operon, were no longer significantly decreased in the apo-locked IscR mutant compared to the DiscR strain. However, decreased transcription of yopE, yopK, yopM, and yopH as well as a severe defect in secretion of Yops was still observed. This could be explained by a deficiency in the apo-locked mutant's membrane potential, but not in the DiscR strain ( Figure 9B). Wilharm et al., demonstrated that Y. enterocolitica motility and type III secretion requires the proton motive force [50]. Indeed, the apo-locked Y. pseudotuberculosis strain displayed a significant motility defect while the DiscR mutant was fully motile. Therefore, the type III secretion defect of the Y. pseudotuberculosis apo-locked IscR mutant can be explained by a deficiency in the proton motive force. Furthermore, the defect in YopHEMK transcription in the apo-locked IscR mutant may be explained by the fact that Yop secretion has a positive regulatory effect on Yop transcription [51,52]. Together, these data suggest that apo-IscR can promote LcrF transcription, but that locking iscR is the apo form causes a proton motive force defect that prevents effector Yop transcription and secretion ( Figure 9B).
It is unclear why locking IscR in the apo-locked form leads to a proton motive force defect. We observed ,9-fold more suf transcript in the apo-locked IscR mutant compared to the DiscR strain that does not have a proton motive force defect, whereas the isc operon was expressed to the same degree in both mutants. Ezraty et al. recently showed that expression of the suf, but not the isc, operon in E. coli leads to a proton motive force defect, possibly as a result of impaired loading of Fe-S clusters into aerobic respiratory complexes [55]. Although the isc operon is expressed in the apo-locked Y. pseudotuberculosis mutant, perhaps overexpression of the suf pathway leads to misassembly of the Fe-S complexes of the electron transport chain that drive the proton motive force.
Both holo-and apo-IscR are predicted to bind to the type 2 motif within the yscW-lcrF promoter [33]. Based on previous data on E. coli IscR [28][29][30]34,56], low iron, aerobic growth, or high oxidative stress conditions are predicted to result in high expression of IscR through derepression of the isc operon, which in turn should increase T3SS gene expression. Likewise, high iron, anaerobic, or low oxidative stress conditions should lead to decreased IscR levels and therefore lower T3SS expression. Under normal aerobic culture conditions, we do not observe a change in wild type Y. pseudotuberculosis type III secretion when iron levels are altered (data not shown). However, in vivo, bacteria may be present in microaerophilic or anaerobic niches, where changes in iron bioavailability and reactive oxygen species production may impact iscR and T3SS gene expression. Upon ingestion by a host animal, Y. pseudotuberculosis enters the lumen of the intestine, which receives approximately 15 mg of iron per day [57,58]. In the small intestine, Y. pseudotuberculosis can cross the gut barrier and enter the bloodstream and deeper tissues, which have very low iron bioavailability (,10 224 M free serum iron) [59][60][61]. Sequestration of iron by iron carriers in mammalian tissues is an important host defense mechanism to prevent growth of bacterial pathogens, the majority of which require iron for growth [62]. The Ysc T3SS has been shown to be required for Y. pseudotuberculosis pathogenesis in these deep tissue sites that are low in iron bioavailability [44]. Perhaps Y. pseudotuberculosis uses IscR to sense iron, O 2 , and/or ROS concentration in order to optimally control T3SS expression in vivo.
Consistent with the severe T3SS expression defect displayed by the Y. pseudotuberculosis DiscR strain, this mutant was deficient in colonization of the Peyer's patches, spleen, and liver. Interestingly, the DiscR mutant was also defective in colonization of the mesenteric lymph nodes (MLN), yet T3SS mutants were previously shown to persist in the MLN and chromosomally-encoded factors were found to be important for Y. pseudotuberculosis survival in this tissue [24,63,64]. These results indicate that the virulence defect of the Y. pseudotuberculosis DiscR strain may not be due solely to misregulation of the T3SS, suggesting the existence of other IscR gene targets important for virulence. IscR of Pseudomonas aeruginosa has been shown to be important for full virulence through its ability to upregulate KatA, encoding a catalase that protects against oxidative stress [38,[65][66][67]. In Vibrio vulnificus, IscR upregulates two genes encoding the antioxidants peroxiredoxin and glutaredoxin 2, and is Figure 6. The apo-IscR mutant strain displays decreased motility and disruption of electrical potential. (A) Motility was analyzed by spotting 1 ml aliquots of either a nonmotile strain (Dyop6/flhD Y.pestis ), WT, DiscR, or apo-locked IscR Y. pseudotuberculosis onto motility agar plates. The diameters of the colonies were determined one day later and used to calculate percent motility relative to WT, which was set at 100%. Shown is the average percent motility 6 SEM and is representative of three independent experiments. ***p#0.0001 as determined by one-way ANOVA followed by Bonferroni post hoc test where each indicated group was compared to the appropriate negative (Dyop6/flhD Y.pestis ) and positive (WT) controls. (B) Proton motive force (PMF) was measured using JC-1 dye for Y. pseudotuberculosis IP2666 wild type (WT), iscR deletion mutant (DiscR), iscR complemented (DiscR pIscR), apo-IscR, and apo-IscR complemented (apo-IscR pIscR) strains grown in M9 at 37uC for 3 hours. The protonophore CCCP was added to a WT sample as a negative control (CCCP). Decreases in PMF were measured as a decrease in red (590 nm) fluorescent cells relative to green (530 nm). The data is presented as total fluorescence intensities at 590 (red) relative to 530 (green) 6 SEM and is representative of three independent experiments. *p#0.05, as determined by one-way ANOVA followed by Bonferroni post hoc test where each indicated group was compared to the appropriate negative (CCCP) and positive (WT) controls. doi:10.1371/journal.ppat.1004194.g006 essential for survival during exposure to reactive oxygen species [40]. Interestingly, our analysis suggests that Y. pseudotuberculosis IscR plays an opposite regulatory role, as IscR negatively affects expression of the genes encoding cellular detoxification proteins KatY, Tpx, SodC and SodB. Furthermore, hydrogen peroxide sensitivity assays showed comparable levels of survival between the Y. pseudotuberculosis wild type and DiscR strains ( Figure S5). This suggests that the virulence defect observed for the DiscR Y. pseudotuberculosis mutant is not due to increased susceptibility to oxidative stresses encountered during infection. Pathways other than the T3SS, such as the hmu hemin uptake system, were found to be misregulated in the Y. pseudotuberculosis DiscR strain (Table 2 & Figure 4B). While the hmu operon was shown to not affect Y. pestis virulence, it is possible that IscR control of the Y. pseudotuberculosis hmu pathway is important for virulence.
In summary, we present the first characterization for the ironsulfur cluster regulator, IscR, of Yersinia. We reveal that IscR regulates genes involved in Fe-S cluster assembly in a manner akin to that of E. coli. Most notably, we demonstrate that mutation of IscR leads to decreased function of the Y. pseudotuberculosis T3SS and that this is due to a decrease in transcription of genes encoding structural, regulatory, and effector proteins. Furthermore, we present evidence showing that IscR is essential for the virulence of Y. pseudotuberculosis and that this attenuation is likely due, in part, to direct regulation of the T3SS by IscR. Collectively, this study argues for the important and novel role of IscR in the virulence of  [83] representations of log 2 -ratios (log 2 (RPKM mutant /RPKM wt ) for each gene on the pYV plasmid for both the DiscR (middle ring) and apo-IscR (inner ring) mutants relative to wild type. Outer ring: pYV base coordinate position from Y. pseudotuberculosis IP32953. Known genes are identified and the virA, virB and virC operons highlighted by black arcs. On the interior right side is the color bar legend displaying log 2ratios from 23.5 to 2. Using this scale, orange/red colorations represent genes with decreased transcription in the mutant relative to the wild type strain and blue/green coloring represents increases in gene transcription for the mutant relative to the wild type. Tan/cream denotes no change. doi:10.1371/journal.ppat.1004194.g007 Y. pseudotuberculosis as well as regulation of the Ysc T3SS, and identifies IscR as a potential target for novel antimicrobial agents.

Materials and Methods
All animal use procedures were in strict accordance with the NIH Guide for the Care and Use of Laboratory Animals and were approved by the UCSC Institutional Animal Care and Use Committee.

Bacterial strains, plasmids and growth conditions
All strains used in this study are listed in Table 3. Y. pseudotuberculosis strains were grown in either 2xYT or M9 minimal media supplemented with casamino acids [68], referred to here as M9, at 26uC with shaking at 250 rpm, unless otherwise indicated. Where stated, Yop synthesis was induced via back-dilution of cultures into either M9 or low calcium media (2xYT plus 20 mM sodium oxalate and 20 mM MgCl 2 ) to an OD 600 of 0.2 and grown for 1.5 h at 26uC/ shaking followed by 2 h at 37uC/shaking as previously described [69].  [24]. The IscR type 2 DNA-binding site is indicated by a black box. The nine bases previously found to be important for IscR binding are indicated by asterisks [35]. (B) Y. pseudotuberculosis IP2666 wild type (WT), iscR deletion (DiscR), DiscR complemented with Y. pseudotuberculosis iscR (DiscR pIscR Y.pstb ), and DiscR complemented with E. coli iscR (DiscR pIscR E.coli ) strains were grown in 2xYT low calcium media at 37uC to induce type III secretion in the absence of host cells. Proteins in the bacterial culture supernatant were precipitated and visualized alongside a protein molecular weight marker (Ladder) on a polyacrylamide gel using commassie blue. Sample loading was normalized for OD 600 of each culture. These results are representative of three independent experiments. (C) The competitor DNA sequences used for the competition assay and the resulting IC 50 concentrations are displayed. Nucleotides in bold and underlined correspond to those that were changed in the mlcrF sequence and have been found to be important for IscR binding in E. coli [33]. (D) Competition assay utilizing 59 nM E. coli apo-locked IscR (IscR-C92A) and 5 nM TAMRA labeled hya DNA [33]. Assay were performed using a range of 8 to 1000 nM unlabeled competitor DNA, including the known E. coli hya site competitor (closed triangles), the in silico identified Y.
The resulting ,1 kb fragments were cloned into the TOPO TA cloning vector (Invitrogen) and further subcloned into a BamHIand NotI-digested pSR47s suicide plasmid (lpir-dependent replicon, kanamycin R (Kan R ), sacB gene conferring sucrose sensitivity) [71,72]. Recombinant plasmids were transformed into E. coli S17-1 lpir competent cells and later introduced into Y.  [32,34] and on data shown here. In wild type bacteria, the Isc Fe-S cluster biogenesis pathway loads a [2Fe-2S] cluster onto IscR (holo-IscR) [32], which recognizes a type 1 DNA-binding motif in the isc promoter to repress transcription in a negative feedback loop. Expression of the apo-locked IscR allele (***, IscR-C92A/C98A/C104A) results in loss of holo-IscR-mediated repression, thereby increasing transcription of the isc operon relative to wild type, resulting in a 30-fold increase in iscR. (B) Model depicting the mechanism by which IscR controls the Y. pseudotuberculosis Ysc T3SS. Holo-and apo-IscR are predicted to bind a newly identified type 2 DNA-binding site within the yscW-lcrF operon encoding the LcrF T3SS master regulator. Subsequently, LcrF expression leads to transcription of the LcrF regulon, which includes the lcrGVH-yopBD and virC operons and yop genes [17,20,22,53,54]. These genes encode the majority of T3SS structural, regulatory, and effector proteins. However, through an as yet undefined mechanism, overexpression of apo-locked IscR leads to a decrease in the proton motive force, which is required for type III secretion [50]. As Yop secretion positively regulates yop gene transcription [51,52], the secretion defect of the apolocked IscR mutant is predicted to lead to a decrease in effector yop transcription. doi:10.1371/journal.ppat.1004194.g009 pseudotuberculosis IP2666 via conjugation. The resulting Kan R , irgansan R (Yersinia selective antibiotic) integrants were grown in the absence of antibiotics and plated on sucrose-containing media to select for clones that had lost sacB (and by inference, the linked plasmid DNA). Kan S , sucrose R , congo red-positive colonies were screened by PCR and subsequently sequenced to verify loss of the intended iscR coding region. The iscR complement construct was generated by insertion of a fragment containing the iscR coding region as well as 530 bp of 59 upstream sequence. This was PCR amplified using primer pair FiscRC and RiscRC, and cloned into the vector pACYC184 via BamHI/SalI restriction sites [73,74]. Recombinant plasmids were transformed into E. coli S17-1 lpir competent cells and later introduced into Y. pseudotuberculosis IP2666 DiscR via a modified transformation method [75]. Briefly, recipient Yersinia strains were grown overnight in LB containing 2% glucose at 26uC. Cultures were centrifuged at 3,500 rpm for 3 min then washed with 750 ml of ice-cold sterile diH 2 O and repeated for a total of three washes. Washed pellets were resuspended in 100 ml of sterile diH 2 O, combined with 3 ml of plasmid and electroporated at EC2. Cells were allowed to recover in 1 mL SOC media for 1 h at 26uC followed by plating on LB containing carbenicillin to select for Yersinia bearing the plasmid of interest. Clones were confirmed by PCR analysis, using a combination of gene-and vector-specific primers, to construct both the DiscR complemented strain (DiscR pIscR) and the apo-IscR complemented strain (apo-IscR pIscR).
The nonmotile Dyop6/flhDC Y.pestis mutant was generated by crossing in the Y. pestis flhDC gene into Y. pseudotuberculosis. Y. pestis flhD has a frameshift mutation, resulting in suppression of flagellin production [76]. The suicide plasmid pSB890 encoding a partial flhC gene and the full flhD gene from the Y. pestis KIM strain, generously provided by Dr. Brad Cookson, was conjugated into Y. pseudotuberculosis Dyop6 and nonmotile, recombinant mutants isolated as previously described [46].

Transposon screen generation and insertion site identification
Transposon mutagenesis was preformed similarly to Crimmins et al. [64]. Briefly, E. coli SM10lpir harboring pSC189, which encodes Himar1 [77], was mated with Y. pseudotuberculosis Dyop6. Mating culture was then pelleted, resuspended, spread out evenly among six 150 mm615 mm petri plates containing LB supplemented with 2 mg mL 21 irgasan and 30 mg mL 21 Kan, and incubated for 3 days at room temperature. Colonies were patched onto LB supplemented with 100 mg mL 21 carbenicillin to ensure insertion of the transposon. Colony patches were used to grow 2xYT overnight cultures in 96-well plates, which were then frozen down to preserve the library. HEK293T cells were plated in 96well white clear bottom plates (Corning) and transfected with a plasmid encoding a luciferase reporter gene fused to an NFkBdependent promoter (Stratagene). Mutants from the transposon library were grown overnight in 96 well plates in M9 at 26uC and used to infect the transfected HEK293T cell monolayers. After 4 h incubation at 37uC, 100 ml of 1:1 NeoLite:PBS solution was added to each well of the 96-well clear-bottom white plate (Corning), and luminescence was measured using a Victor 3 plate reader (Perki-nElmer). Each transposon mutant was assayed in duplicate. The positions of the transposons in the iscR::Tn1 and iscR::Tn2 mutants were determined by plasmid rescue, as previously described [78], except BamHI was used for digestion of genomic DNA.

NFkB activity assay
Validation of the transposon screen was performed through the use of an NF K B activity assay, which is based on our previous work showing that Y. pseudotuberculosis induces NFkB activation in HEK293T cells dependent on expression of a functional T3SS [43]. Briefly, HEK293T cells were transfected with a plasmid encoding a luciferase reporter gene fused to an NFkB-dependent promoter (Stratagene). Bacterial strains were grown overnight in 2xYT and subcultured to an OD 600 of 0.2 into low calcium media and grown at 26uC for 1.5 h followed by a shift to 37uC for an additional 1.5 h to induce the T3SS. Bacterial cultures were resuspended in prewarmed (37uC) DMEM and 200 ml aliquots were then used to infect the HEK293T cells containing the luciferase reporter plasmid at an MOI of 10. After 4 h incubation at 37uC, 100 ml of 1:1 NeoLite:PBS solution was added to each well of the 96-well clear-bottom white plate (Corning), and luminescence was measured using a Victor 3 plate reader (PerkinElmer). Data from three separate wells were averaged for each independent experiment.

Type III secretion assay
Visualization of T3SS cargo secreted in broth culture was performed as previously described [46]. Briefly, Y. pseudotuberculosis in M9 low calcium media (M9 plus 20 mM sodium oxalate and 20 mM MgCl 2 ) was grown for 1.5 h at 26uC followed by growth at 37uC for 2 h. Cultures were normalized by OD 600 and pelleted at 13,200 rpm for 10 min at room temperature. Supernatants were removed and proteins precipitated by addition of trichloroacetic acid (TCA) at a final concentration of 10%. Samples were incubated on ice for 20 min and pelleted at 13,200 rpm for 15 min at 4uC. Resulting pellets were washed twice with ice-cold 100% EtOH and subsequently resuspended in final sample buffer (FSB) containing 20% dithiothreitol (DTT). Samples were boiled for 5 min prior to running on a 12.5% SDS PAGE gel.

Ethidium bromide entry assay
Evaluation of pore formation was performed via the ethidium bromide (EtBr) entry assay as previously described [46]. Briefly, 2610 4 immortalized C57Bl/6 BMDMs were plated in a 96 well clear bottom black plate (Corning) in 100 uL DMEM +10% FBS. Infection was performed in triplicate at an MOI of 25. Plates were centrifuged at 7506g at 4uC for 5 min to facilitate contact. Infections were carried out at 37uC with 5% CO 2 for 2 h, at which point media was aspirated and replaced with 30 mL of PBS containing 25 mg mL 21 ethidium bromide (EtBr) and 12.3 mg mL 21 Hoechst dye. The cell monolayer was visualized using an ImageXpressMICRO automated microscope and MetaXpress analysis software (Molecular Devices). The percent of EtBrpositive cells was calculated by dividing the number of EtBrstained cells by the number of Hoechst-stained cells. Data from three separate wells was averaged for each independent experiment.

Growth curves
Y. pseudotuberculosis strains were cultured overnight in 2xYT or M9 at 26uC and sub-cultured to an OD 600 of 0.2 in 25 mL of either 2xYT or M9. Cultures were incubated at either 26uC or 37uC with shaking at 250 rpm and optical density measured at 600 nm every hour for 9 h.

Mouse infections
All animal use procedures were in strict accordance with the NIH Guide for the Care and Use of Laboratory Animals and were approved by the UC Santa Cruz Institutional Animal Care and Use Committee. Eleven to twelve-week-old 129S6/SvEvTac mice from our breeding facilities were used for oral infections as previously described [79]. Briefly, mice were orogastrically inoculated with 2610 8 CFU in a 200 ml volume using a feeding needle. Mice were given food and water ad libitum and were euthanized at 5 days post-inoculation. Peyer's patches, mesenteric lymph nodes, spleens, and livers were isolated and homogenized for 30 s in PBS followed by serial dilution and plating on LB supplemented with 1 mg mL 21 irgasan for CFU determination.

RNAseq analysis
RNA was isolated from the IP2666 wild type and isogenic DiscR and apo-IscR strains grown for 3 h at 37uC in M9, using the RNeasy Mini Kit (Qiagen) as per the manufacturer's protocol. We chose M9 media for our RNASeq analysis because this condition enables expression of T3SS genes and secretion of T3SS cargo at 37uC [68]. Contaminating DNA was removed from the RNA samples using a DNA-free kit (Life Sciences). Samples were subjected to removal of contaminating rRNA via the Ribo-Zero Magnetic Kit for Gram-negative bacteria (Epicentre). The cDNA library was created using the NEBNext Ultra Directional RNA Library Prep Kit for Illumina (NEB). These studies were performed with three biological replicates per condition. Six indexed samples were sequenced per single lane using the HiSeq2500 Illumina sequencing platform for 50 bp single reads (UC Davis Genome Center) and subsequently analyzed and visualized via the CLC Genomics Workbench version 5.5.1 (CLC bio). Samples were normalized for both sequence depth and gene size by determining RPKM (Reads Per Kilo base per Million reads) and mapped to the Y. pseudotuberculosis genome (IP32953). Differentially regulated genes were identified as those displaying a fold change with an absolute value of 2 or greater. Statistical significance was determined by baySeq test with a corrected FDR post hoc test where p,0.05 was deemed significant [80].

Real-time PCR
Total RNA generated from our RNAseq analysis at a concentration of 2 mg was used to make cDNA as previously described [81]. SYBR Green PCR master mix (Applied Biosystems) was used for qPCR reactions according to the manufacturer's instructions and a 60uC annealing temperature. Primers used are listed in Table S4. Control primers were for the 16S rRNA as described previously [82]. Results were analyzed using the Bio-Rad CFX software.

Virulence plasmid map generation
Average RPKM values generated from RNAseq analysis for the wild type, DiscR and apo-IscR mutants were converted to log 2ratios (log 2 (RPKM mutant /RPKM wt ) for each gene encoded on the virulence plasmid, pYV. These values were converted to a Circos heatmap [83] and plotted against the respective pYV base coordinate positions from Y. pseudotuberculosis IP32953.

Motif identification and search
Position specific scoring matrix (PSSM) was generated by the alignment of the known E. coli IscR type 2 motifs (Table S4) (Maverix Biomics, Inc) [31]. PSSM of type 2 was used to scan against the 150-nt upstream of 99 genes encoded on the Y. pseudotuberculosis pYV plasmid and obtained a ranked list of putative type 2 motifs.

DNA binding fluorescence anisotropy assays
Fluorescence anisotropy was measured similar to Nesbit et al., [33]. E. coli apo-IscR lacking the [2Fe-2S] cluster (IscR-C92A) was isolated anaerobically following the protocol described previously for wild type IscR [30]. Competition assays were performed using 5nM of 30-mer dsDNA of the known E. coli hyaA type 2 motif containing a 59 TAMRA fluorophore (IDT) on the top strand and unlabeled competitor dsDNA concentrations ranged from 8 to 1000 nM (IDT , Table  S4). DNA was annealed by heating equimolar concentrations of complementary DNA strands in annealing buffer (40 mM Tris (pH 7.9), 30 mM KCl) to 95uC for 5 min followed by slow cooling to room temperature over 2 hours. Annealed DNA was incubated with 90 nM apo-IscR in anisotropy buffer (40 mM Tris pH 7.9, 150 mM KCl, 100 ng ul 21 Salmon Sperm DNA) for 12 min at room temperature and anisotropy was measured using an EnVision 2103 Multilabel Reader (Perkin Elmer) with Wallac EnVision Manager software. Data is representative of experiments performed on three separate days.

Motility assay
Motility was analyzed by spotting 1 ml aliquots of either a nonmotile strain bearing an inactive, Y. pestis allele of flhD (Dyop6/ flhD Y.pestis ), WT, DiscR, or apo-IscR strains onto motility agar plates (1% tryptone, 0.25% agar) from overnight cultures standardized to an OD 600 of 2.5. Plates were incubated at room temperature for 1 day, at which point the diameters of the colonies were determined and used to calculate percent motility relative to WT, which was set at 100%.

Measurement of the membrane potential
The electrical potential was measured similar to the JC-1 red/ green dye assay previously described for E. coli [84]. JC-1 is a membrane-permeable dye that emits green fluorescence (,530 nm) upon excitation when the dye is in the monomeric form. Due to the membrane potential of the bacterial cell, JC-1 dye will form J aggregates which emit red fluorescence (,590 nm). If the membrane potential decreases, there will be a decrease in J aggregate formation and subsequently a decrease in red fluorescence. As such, membrane potential can be displayed as a ratio of red/green fluorescence. Briefly, Y. pseudotuberculosis wild type and isogenic DiscR, DiscR complemented (DiscR pIscR), apo-IscR and apo-IscR complemented (apo-IscR pIscR) strains were grown overnight in M9 at 26uC. Strains were subcultured to an OD 600 of 0.2 in M9 and grown at 37uC for 3 hours. A negative control containing a sample of wild type Y. pseudotuberculosis treated with 40 mM of the protonophore, CCCP (carbonyl cyanide mchlorophenylhydrazone), during the last 30 min of growth was included in each experiment. After incubation at 37uC, 1 mL aliquots were harvested for each strain and pelleted at 4,5006g for 3 min followed by resuspension in 1 mL of permeabilisation buffer (10 mM Tris-HCl, pH 7.6, 1 mM EDTA and 10 mM glucose). Post resuspension, 2 ml of the membrane-permeable JC-1 dye (5 mg/mL) was added and the samples were incubated at room temperature for 30 min. Samples were then pelleted at 4,5006g for 3 min and resuspended in 500 ml of permeabilisation buffer. Slides were prepared by first coating with Poly-L-Lysine solution through addition of 100 ml aliquots of a 0.01% solution followed by a 5 min incubation at room temperature. Slides were washed a total of 3 times with sterile diH 2 O. Once dry, 10 ml of prepared sample was added to the slide and allowed to adhere for 5 min. Unattached bacteria were removed by washing with PBS and excess liquid removed via aspiration. A coverslip was applied and the cells were imaged using a LSM 5 PASCAL laser scanning microscope (Zeiss) fitted with a Plan-Apochromat 63x/1.4 Oil DIC objective and analyzed using the LSM 510 software (Zeiss). Quantification of image intensities was performed using ImageJ [85]. Figure S1 IscR does not affect Y. pseudotuberculosis growth under non-T3SS-inducing conditions, but partially alleviates T3SS-associated growth restriction. The Y. pseudotuberculosis WT, DiscR, apo-IscR and, where applicable, DiscR and apo-IscR complemented strains (DiscR pIscR and apo-IscR pIscR, respectively) and Y. pseudotuberculosis lacking the virulence plasmid pYV (pYV 2 ), were grown (A) in M9 at 37uC, (B) in 2xYT at 37uC, (C) in M9 at 26uC or (D) in 2xYT at 37uC. Optical density of the cultures were monitored at 600 nm every hour for 9 h. The averages 6 SEM from three independent experiments are shown. * p,0.05, **p,0.01, ***p,0.001 as determined by a Student t test relative to the wild type. (TIF) Figure S2 Deletion of IscR leads to increased transcription of Fe-S cluster biogenesis genes. (A) RPKM expression levels generated from RNAseq analysis of Y. pseudotuberculosis DiscR and apo-IscR mutants relative to WT for 12 genes involved in Fe-S cluster biogenesis are displayed. *p,0.001 as determined by Bayseq test with a corrected FDR post hoc test from three independent experiments. (B) Displayed is the nucleotide sequence of a region 130 bp upstream of the putative IscR start codon in Y. pseudotuberculosis IP 32953 including the putative transcriptional start site (arrow; UCSC Microbial Genome Browser) and putative sigma70 promoter elements (210) and (235), as well as the two putative IscR type I binding sites (brackets). (TIF) Figure S3 Mutation of iscR does not affect pYV virulence plasmid yield. Relative amounts of the virulence plasmid, pYV, were analyzed from standardized cultures of the wild type (WT), iscR mutant (DiscR) and pYV 2 strains grown in M9 at 37uC for 3 hours through midiprep analysis (Promega) according to the manufacturer's protocol. Plasmid yield was quantified via spectrophotometric analysis (Nanodrop). The data is displayed as mg of plasmid isolated per mL of culture 6 SEM and is an average of 3 independent experiments. *p#0.05, as determined by Student t test. (TIF) Figure S4 Expression of the suf operon is increased in the apo-locked IscR mutant strain. RNAseq analysis was performed on WT, DiscR and apo-IscR Y. pseudotuberculosis strains after growth in M9 at 37uC for 3 h (T3SS-inducing conditions). The data is presented as mean RPKM 6 SEM and is an average of 3 independent experiments. ***p#0.0001, as determined by Bayseq followed by FDR post hoc test. (TIF) Figure S5 IscR is not required for survival postexposure to hydrogen peroxide stress. Hydrogen peroxide assays were performed similar to Schiano et al. [87]. Y. pseudotuberculosis wild type (WT), DiscR, and iscR complemented (DiscR pIscR) strains were grown overnight in 2xYT at 26uC. Cultures were standardized to an OD 600 of 0.1 and grown at 26uC with shaking to mid-log phase, at which point they were diluted 1:10 into fresh 2xYT. Samples were supplemented with 50 ml of either sterile water (negative control) or hydrogen peroxide to a final concentration of 50 mM. Samples were incubated with shaking at 26uC and CFU determined via serial dilution and plating 10 min after the start of treatment. The data is displayed as percent survival (CFU H 2 O 2 /CFU H 2 O)*100) 6 SEM and is an average of 3 independent experiments. (TIF)