All strains used in this study are listed in Table 1. Y. pseudotuberculosis strains were grown in either 2xYT or M9 minimal media supplemented with casamino acids referred to here as M9, at 26°C with shaking at 250 rpm, unless otherwise indicated (Cheng et al., 1997).

The ΔhmuRSTUV and ΔhmuSTUV mutants were generated via splicing by overlap extension (Warrens et al., 1997). Primer pairs F5/R5ΔhmuRSTUV and F5/R5ΔhmuSTUV (Table 2) were used to amplify ~1,000 bp 5′ of hmuRSTUV and hmuSTUV, respectively. Primer pair F3/R3ΔhmuRSTUV were used to amplify ~1,000 bp 3′ of ΔhmuRSTUV (Table 2). Amplified PCR fragments served as templates in an overlap extension PCR using the outside primers F5/R3ΔhmuRSTUV and F5ΔhmuSTUV/R3ΔhmuRSTUV for ΔhmuRSTUV and ΔhmuSTUV, respectively. The resulting ~2 kb fragments were cloned into the TOPO TA cloning vector (Invitrogen) and further subcloned into a BamHI- and NotI-digested pSR47s suicide plasmid (λpir-dependent replicon, kanamycin^R (Kan^R), sacB gene conferring sucrose sensitivity) (Merriam et al., 1997; Andrews et al., 1998). Recombinant plasmids were transformed into E. coli S17-1 λpir competent cells and later introduced into Y. 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 sequencing.

E. coli IscR-C92A was anaerobically purified as previously described (Giel et al., 2006; Nesbit et al., 2009). E. coli Fur was purified as described in Lee et al. (2003) except that a pET-11a overexpression vector containing fur was used (pPK11241). Cultures were grown in LB containing ampicillin (50 μg/mL) at 37°C to an OD₆₀₀ of 0.6 and Fur synthesis was subsequently induced by addition of 400 μM IPTG for 2.5 h at 37°C.

DNA fragments containing the intergenic promoter region between hmuR and hmuSTUV (−143 to +72 bp relative to the hmuS start codon), and a control DNA fragment (−280 to −123 bp relative to the hmuS start codon), were respectively isolated from pPK12468 and pPK12469 after digestion with HindIII and BamHI. These fragments and linearized plasmid (which served as competitor DNA in the EMSAs) were purified with Elutip-d columns (Schleicher and Schuell). IscR-C92A was incubated with DNA fragments (~5–10 nM) for 30 min at 37°C in 40 mM Tris (pH 7.9), 30 mM KCl, 100 μg/mL bovine serum albumin (BSA), and 1 mM DTT. Samples were loaded onto a non-denaturing 6% polyacrylamide gel in 0.5× Tris-borate-EDTA (TBE) buffer and run at 100 V for 90 min. The gel was stained with SYBR Green EMSA nucleic acid gel stain (Molecular Probes) and visualized using a Typhoon FLA 900 imager (GE). These assays (along with other in vitro assays performed with IscR-C92A described below) were carried out under aerobic conditions with IscR-C92A that was diluted immediately before use.

Assays performed with E. coli Fur were carried out in a similar manner except that Fur (which was pre-equilibrated with 100 μM MnCl₂ at room temperature for 20 min to form active Mn²⁺-Fur) was incubated with DNA fragments in 20 mM BisTris (pH 7.5), 1 mM MgCl₂, 40 mM KCl, 100 μM MnCl₂, 100 mM DTT, 5% glycerol, 100 μg/mL BSA. DNA fragments encompassing the hmuR promoter (−100 to −1 bp relative to the hmuR transcriptional start site) or the intergenic promoter region between hmuR and hmuSTUV (−123 to −1 bp relative to the hmuS start codon) were PCR amplified from pFU99::p1 and pFU99::p2, respectively, and purified with Elutip-d columns (Schleicher and Schuell).

The intergenic promoter region between hmuR and hmuSTUV (−123 to −1 bp relative to the hmuS start codon) was isolated from pPK12440 after digestion with HindIII and BamHI. Sequenase version 2.0 (USB Scientific) was used to 3′-end radiolabel the BamHI end of the fragment (top strand) with [α-³²P]dGTP (Perkin Elmer). The labeled fragment was isolated from a non-denaturing 5% acrylamide gel and purified with Elutip-d columns (Schleicher and Schuell). Footprinting was performed by incubating IscR-C92A with labeled DNA (~5 nM) for 30 min at 37°C in 40 mM Tris (pH 7.9), 30 mM KCl, 100 μg/mL BSA, and 1 mM DTT followed by the addition of 2 μg/mL DNase I (Worthington) for 30 s. The reaction was terminated by the addition of sodium acetate and EDTA to final concentrations of 300 and 20 mM, respectively, and the reaction mix was ethanol precipitated, resuspended in urea loading dye, heated for 60 s at 90°C, and loaded onto a 7 M urea-8% polyacrylamide gel in 0.5× Tris-borate-EDTA (TBE) buffer. An A+G ladder was made by formic acid modification of the radiolabeled DNA, followed by piperidine cleavage (Maxam and Gilbert, 1980). The reaction products were visualized by phosphorimaging.

The effect of IscR-C92A on σ⁷⁰-dependent promoter activity from candidate control regions was determined by incubating IscR-C92A with 2 nM supercoiled pPK12440 [purified with the QIAfilter Maxi kit (Qiagen)], 0.25 μCi of [α-³²P]UTP (3,000 μCi/mmol; Perkin Elmer), 20 μM UTP, and 500 μM each of ATP, GTP, and CTP for 30 min at 37°C in 40 mM Tris (pH 7.9), 30 mM KCl, 10 mM MgCl₂, 100 μg/mL bovine serum albumin (BSA), and 1 mM DTT. Eσ⁷⁰ RNA polymerase (NEB) was added to a final concentration of 50 nM and the reaction was terminated after 5 min by addition of Stop Solution (USB Scientific). Samples were heated for 60 s at 90°C, and loaded onto a 7 M urea-8% polyacrylamide gel in 0.5× Tris-borate-EDTA (TBE) buffer. The reaction products were visualized by phosphorimaging.

For RNA isolated from samples subjected to iron starvation, Y. pseudotuberculosis WT and ΔiscR strains were grown in 5 mL of M9 minimal media overnight. Cultures were then diluted to OD₆₀₀ of 0.1 in 20 mL of M9 media treated with Chelex (Bio-Rad) to remove iron and allowed to grow for 8 hrs. The cultures were then diluted again to OD₆₀₀ 0.1 into 20 mL Chelex treated M9 media and allowed to grow for 12 h. Cultures were then diluted to OD₆₀₀ of 0.1 into 20 mL Chelex treated M9 media with either no iron, 5 μM hemin, or 1 mg/L FeSO_4. Before addition, stock hemin solution was Chelex-treated overnight. After 3 h of growth at 37°C, 5 mL of culture from each condition was pelleted by centrifugation for 5 min at 4,000 rpm.

For all samples, the supernatant was removed and pellets were resuspended in 500 μL of media and treated with 1 mL Bacterial RNA Protect Reagent (Qiagen) according to the manufacturer's protocol. Total RNA was isolated using the RNeasy Mini Kit (Qiagen) per the manufacturer's protocol. Contaminating DNA was removed using the TURBO DNA-free Kit (Life Technologies/Thermo Fischer). rRNA was removed using the RiboZero Magnetic Kit for Gram Negative Bacteria (Illumina). The cDNA library was prepared using the NEB Ultra Directional RNA Library Prep Kit for Illumina. Quality of total RNA, mRNA after rRNA depletion and cDNA libraries were assessed using an Agilent 2000 Bioanalyzer.

These studies were performed with three biological replicates per condition. RNA-Seq analysis on bacteria that were not starved for iron (Figure 2A) was previously published (Miller et al., 2014). For cultures depleted of iron and either left iron depleted or given FeSO₄ or hemin (Figures 2B–D), indexed samples were sequenced using a MiSeq Illumina sequencing platform for 150 bp end reads (UC Davis Genome Center). The full RNA-Seq data set from these iron depleted cultures will be published separately (Wei, Schwiesow, Balderas, and Auerbuch, data not shown). All sequencing data was analyzed and visualized via the CLC Genomics Workbench version 9.5.3 (CLC bio). Reads were mapped to the Yersinia 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 using an EdgeR-like analysis within the RNAseq plugin in CLC workbench with a corrected FDR post-hoc test where p ≤ 0.01 was deemed significant.

After harvesting total RNA from wild-type or ΔiscR strains cultured under the indicated conditions via the procedure described above, genomic DNA was removed via the TURBO-DNA-free kit (Life Technologies/Thermo Fisher). cDNA was generated for each sample by using the M-MLV Reverse Transcriptase (Invitrogen) according to the manufacturer's instructions, as we previously described (Miller et al., 2014). Each 20 μl qRT-PCR assay contained 5 μl of 1:10 diluted cDNA sample, 10 μl of Power CYBR Green PCR master mix (Thermo Fisher Scientific), and primers (Table 2) with optimized concentrations. The expression levels of each target gene were normalized to that of 16S rRNA present in each sample, and calculated by the ΔΔCt method. Three independent biological replicates were harvested for each tested condition. For each target transcript, significant differential expression between different bacterial strains were defined by p-value < 0.05 of one-way analysis of variance (one-way ANOVA).

The effect of addition of hemin on bacterial growth was assayed in liquid cultures similar to Thompson et al. (1999). Y. pseudotuberculosis was grown as described in the RNA-Seq section above. OD₆₀₀ was measured every hour for 9 h. Data is representative of three independent experiments.

Strains were grown overnight in 2xYT at 26°C and subsequently standardized to an OD₆₀₀ of 0.2. Standardized cultures were centrifuged at 3,500 rpm for 3 min and the supernatants discarded. Bacterial pellets were resuspended in 2 mL of whole sheep's blood in sodium heparin (Hemostat, Inc.) and CFU determined via serial dilution and plating at hour 0. Samples were then placed in a rollerdrum at 37°C and CFU determined at hours 2, 4, 6, and 8. Data is representative of three independent experiments. We observed blood batch to batch variability in terms of the dynamics of bacterial survival; however, the difference between WT and ΔiscR survival were observed for all batches of blood.

Non-heat inactivated sheep serum did not exhibit killing of E. coli K12 or Yersinia (data not shown), so bovine serum was used. Y. pseudotuberculosis or E. coli K12 was grown in 5 mL of M9 minimal media overnight. In the morning, cultures were diluted to OD₆₀₀ 0.05 in non-heat treated bovine serum (innovative technologies), or bovine serum that had been heat-killed for 30 min at 60°C. OD₆₀₀ was measured every hour for 5 h.

The genetic structure of the hmuRSTUV hemin uptake locus has been characterized in Yersinia pestis (Thompson et al., 1999). This locus contains two promoter regions: one upstream of hmuR (p1), and one in the intergenic region between hmuR and hmuS (p2; Figure 1) (Thompson et al., 1999). Previous work suggested that p1, but not p2, promoter activity was increased by iron limitation through a Fur-dependent mechanism (Thompson et al., 1999). Here we extend this study by using RNA-seq analysis in the related pathogen Y. pseudotuberculosis to measure expression of hmuRSTUV under conditions when cells had been depleted of iron or when inorganic iron or hemin were added back as iron sources after a period of iron starvation (Figures 2B–D, Supplementary Dataset). In the presence of added inorganic iron (1.0 mg/L FeSO₄; iron replete conditions), levels of hmuR mRNA (RPKM values) were significantly less than those of hmuSTUV (Figure 2B). This pattern of RNA expression was similar to what we previously observed (Miller et al., 2014) when Yersinia was grown in standard M9 minimal media with no period of iron starvation (Figure 2A). Surprisingly, addition of hemin to iron deprived cells actually decreased expression of hmuSTUV compared to cells with added inorganic iron (Figure 2D). In contrast, in the absence of added iron following a period of iron starvation (iron depletion), hmuR mRNA levels were upregulated (Figure 2C), consistent with previous reports in Y. pestis (Thompson et al., 1999). Furthermore, the RPKM values now showed that hmuR expression was nearly equivalent to hmuSTUV (Figure 2C). This finding was also validated by qPCR (Figures 3A,B). In addition, reads spanning the intergenic region between hmuR and hmuS were detected by RNA-Seq, and a transcript containing hmuR and hmuS was detected by RT-PCR under iron depleted conditions, indicating some contiguous transcription under these conditions (data not shown). This may be a result of read-through transcription from the hmuR promoter in the absence of iron stemming from Fur de-repression (see below). In summary these data suggest complex transcriptional regulation of the hmuRSTUV locus in Y. pseudotuberculosis.

Previous work in Y. pestis suggested that Fur regulates the promoter upstream of hmuR and that hmuSTUV genes were expressed from a weak, constitutive promoter in the intergenic region between hmuR and hmuS (Thompson et al., 1999). However, Fur binding to these regions was not tested biochemically. To determine if Fur directly binds to the hmuR promoter and/or the intergenic promoter between hmuR and hmuS in Y. pseudotuberculosis, we performed gel shift assays using purified E. coli Fur protein. When incubated with increasing amounts of Fur protein, a shift was observed for the promoter region upstream of hmuR at 31.25 nM protein, while no binding was detectable for the intergenic region between hmuR and hmuS until the protein concentration was increased 8-fold (Figure 4A). These data suggest that Fur directly regulates the hmuRSTUV heme uptake locus through direct binding to the promoter upstream of hmuR, and not through binding the intergenic region between hmuR and hmuS. Thus, the increased hmuRSTUV RNA levels observed under iron depletion conditions (Figure 2) are likely explained by the loss of Fur binding to the hmuR promoter.

Previous analysis of RNA-Seq data from our group suggested that IscR also plays a role in the regulation of the hmuRSTUV hemin uptake system (Miller et al., 2014). While hmuR levels were not altered by deletion of iscR as measured by RNA-Seq and qPCR, hmuSTUV levels were significantly decreased in the ΔiscR mutant compared to WT and an apo-locked mutant variant strain under iron replete conditions (Figure 2A). This pattern of expression is consistent with IscR activation of hmuSTUV by binding to an IscR type II motif since both WT and apo-locked mutant were able to increase expression. As expected for a type II motif that binds either apo-IscR or holo-IscR, the requirement for IscR was also observed when cultures were starved of iron for several days and then given an inorganic iron source (Figures 2B, 3B). However, the presence of hemin appeared to disable the increased expression by IscR since RNA levels were not further enhanced when IscR was present (Figures 2D, 3C). Moreover, the difference in hmuSTUV RNA levels between the WT and ΔiscR strains was eliminated when the bacteria remained starved for iron (Figures 2C, 3A). This is in part likely a result of increased transcription from the hmuR promoter stemming from Fur de-repression, which apparently transcribes through hmuSTUV under these conditions. Collectively, these data suggest that the source of iron plays a critical role in determining the relative amount and ratios of hmuR to hmuSTUV mRNAs and that IscR and Fur play distinct roles in hmu regulation.

We asked if IscR could directly bind to the intergenic region between hmuR and hmuS (Figure 4B). E. coli IscR has been shown to complement a Y. pseudotuberculosis ΔiscR mutant and is 100% identical in its DNA binding domain to Yersinia IscR (Miller et al., 2014). We used E. coli IscR-C92A protein that has one Fe-S cluster coordinating cysteine mutated to assay DNA binding because it is in the apo-locked conformation, and binds type II sites in the absence of the [2Fe-2S] cluster allowing DNA binding to be assayed under normal laboratory conditions. We found that purified IscR-C92A was bound toward the 3′ end of the p2 fragment used in our Fur binding studies (Figure 4A; data not shown). To characterize the IscR binding site further, we used a larger fragment containing the intergenic region between hmuR and hmuS and a control fragment containing a region within the 3′ end of the hmuR coding region. When incubated with increasing concentrations of IscR-C92A protein, a shift was observed for the intergenic region DNA at 62.5 nM protein, while a shift for the control DNA fragment was not observed until 250 nM protein was used (Figure 4B). To further define the IscR binding site, we performed DNase footprinting assays with E. coli IscR-C92A. A region of protection was observed from −40 to −1 nucleotides upstream of the translational start site of hmuS (Figure 5A). From this area of protection, we identified two potential IscR type II motif sites, termed site A and site B (Figure 5B). Comparison of these binding sites with the consensus type II motif shows that they contain six and five of the nine bases in the consensus binding motif, respectively (Figure 5B) (Giel et al., 2006). These data suggest that IscR directly binds to the intergenic region between hmuR and hmuS through one or two type II motifs.

We asked whether apo-IscR binding to the intergenic region between hmuR and hmuS directly increased hmuSTUV transcription as predicted from the in vivo RNA analysis. We performed in vitro transcription assays with E-σ70 RNA Polymerase to determine if IscR could drive transcription from the intergenic promoter. We observed no increase in transcript levels over RNA polymerase alone when IscR-C92A was added to the assay (Figure 6). Given these data and the fact that IscR directly binds to this region at one or two putative type II motifs, we hypothesize that IscR functions with another transcriptional regulator to affect transcription of hmuSTUV.

Our previous work showed that Y. pseudotuberculosis lacking iscR is severely attenuated in disseminated infection following oral inoculation of mice (Miller et al., 2014). As IscR regulates expression of the Hmu heme uptake locus, we assessed whether IscR was important for Yersinia survival in the heme rich environment of blood by measuring the ability of Y. pseudotuberculosis to survive in whole sheep's blood. Even in the absence of starving the bacteria of iron, colony forming units (CFU) of the ΔiscR mutant were decreased by 1-2-logs compared to the WT strain at 4 and 8 h (Figure 7A). This suggests that IscR plays a role in the survival of Y. pseudotuberculosis in blood.

We previously characterized IscR as important for expression of the T3SS through the master regulator LcrF in Y. pseudotuberculosis (Miller et al., 2014). The Yersinia T3SS, encoded on the pYV virulence plasmid, is important for the evasion of phagocytic cells, which may be present in our ex vivo blood model (Raymond et al., 2013). Additionally, the pYV-encoded protein YadA, which is controlled by LcrF, has been shown to be important in the evasion of complement in Yersinia enterocolitica through exploitation of C3 and iC3b to sequester large amounts of Factor H (Skurnik and Toivanen, 1992; Schindler et al., 2012). Therefore, we assessed the requirement for IscR-mediated T3SS and YadA control in the survival of Y. pseudotuberculosis in blood. We observed that neither a pYV⁻ strain of Y. pseudotuberculosis that lacks the T3SS and YadA, nor a ΔyscNU mutant that lacks critical T3SS basal body components, display survival defects in whole blood compared to WT (Figure 7B). These data suggest that IscR-mediated T3SS and YadA regulation is not important for survival of Y. pseudotuberculosis in our blood model. Furthermore, the ΔiscR mutant and WT strains survived equally well in heat inactivated and non-heat inactivated bovine serum (Figures 7C–E), ruling out the involvement of complement in killing of the ΔiscR mutant. These data indicate that either the cells or clotting factors present in whole blood but not in serum are responsible for killing the ΔiscR mutant. The response regulator PhoP is required for Y. pseudotuberculosis to replicate in macrophages (Grabenstein et al., 2004). We found that a ΔphoP mutant does not display a survival defect in blood (Figure 7B), suggesting that the inability of the ΔiscR mutant to survive in blood is not due to an inability to replicate in macrophages.

HmuRSTUV have been shown to be necessary for the utilization of heme under iron depleted conditions in Y. pestis (Hornung et al., 1996; Thompson et al., 1999). Therefore, we examined whether IscR and the Hmu locus played a role in heme utilization in Y. pseudotuberculosis. We assessed the ability of pYV⁻, ΔiscR/pYV⁻, ΔhmuRSTUV/pYV⁻, and ΔhmuSTUV/pYV⁻ mutant strains to grow in the presence of hemin or inorganic iron as the sole iron sources, following several days of iron starvation to deplete iron stores. We chose to use the pYV⁻ background because WT Y. pseudotuberculosis undergoes growth restriction in M9 minimal media at 37°C due to T3SS activity, while a strain lacking pYV does not. While all strains could utilize FeSO₄ as an iron source, ΔhmuRSTUV/pYV⁻ and ΔhmuSTUV/pYV⁻ were defective in growth on hemin compared to the pYV⁻ parental strain (Figure 8A), suggesting that the Hmu system promotes Y. pseudotuberculosis growth in heme as the sole iron source. Surprisingly, growth of the ΔiscR/pYV⁻ strain was indistinguishable from the pYV⁻ parental strain, indicating that hmu expression levels in the ΔiscR/pYV⁻ mutant were sufficient to sustain Yersinia growth on the amount of hemin provided. Indeed, there was no significant difference in hmuSTUV RPKM levels between the WT and ΔiscR strains 3 h after hemin addition (Figures 2D, 3C). Interestingly, Yersinia could still utilize the iron in the hemin provided in our experiment in the absence of the Hmu pathway, as the ΔhmuRSTUV/pYV⁻ and ΔhmuSTUV/pYV⁻ strains still had residual growth in heme compared to no iron source added (Figure 8B). Our hemin stocks had to be Chelex-treated to remove free iron prior to start of the growth curve in order to observe a significant difference in growth between the parental strain and the hmu mutants (Figure 8A, unpublished observations). This suggests that residual iron-stimulated growth may occur from breakdown of heme followed by iron uptake by non-heme uptake systems under our growth assay conditions.

While the Hmu locus contributes to Yersinia growth on M9 plus hemin following iron starvation, the ΔhmuRSTUV and ΔhmuSTUV mutants did not display a defect in our ex vivo blood model, rich in heme (Figure 7F). This finding may stem from the lack of iron starvation of the mutants before the survival assays in whole blood were performed. Indeed, to observe the ΔhmuRSTUV and ΔhmuSTUV growth defect in M9 media plus hemin, the bacteria had to be iron-starved to deplete iron stores prior to hemin exposure (data not shown). These data also indicate that the survival defect of the ΔiscR mutant in blood does not stem from diminished heme utilization under our experimental conditions.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.