Transcriptional and posttranscriptional regulation of Shigella shuT in response to host‐associated iron availability and temperature

Abstract Like most bacteria, Shigella must maintain a precise balance between the necessity and toxicity of iron; a balance that is achieved, at least in part, by regulating the production of bacterial iron acquisition systems in response to specific environmental signals. Using the Shigella heme utilization (Shu) system, S. dysenteriae is able to acquire iron from heme, a potentially rich source of nutritional iron within the otherwise iron‐limited environment of the human host. Investigations presented within reveal two distinct molecular mechanisms underlying previously uncharacterized transcriptional and translational regulation of shuT, a gene encoding the periplasmic‐binding component of the Shu system. While shuT transcription is regulated in response to iron availability via a process dependent upon the global regulator Fur and a Fur‐binding site located immediately downstream of the promoter, shuT translation is regulated in response to environmental temperature via the activity of an RNA thermometer located within the 5′ untranslated region of the gene. Such complex regulation likely increases the fitness of S. dysenteriae by ensuring maximal ShuT production when the pathogen is within the iron‐limited and relatively warm environment of the infected host, the only environment in which heme will be encountered as a potential source of essential iron.

safe antibiotic regimen to treat the infection makes understanding the molecular mechanisms underlying the infection with S. dysenteriae of utmost importance. It is only after such an understanding is achieved that this information can be used to guide the development of novel therapeutics designed to disrupt specific virulence-associated processes, and by doing so, lessen or eliminate the ability of Shigella to cause human disease.
Following entry into the host via the fecal-oral route, a successful Shigella infection requires sequential invasion into colonic epithelial cells, intracellular replication, and cell-to-cell spread by the pathogen.
Collectively, these processes disrupt water balance within the colon and promote the formation of lesions within the colonic epithelium; processes that result directly in the bloody diarrhea associated with Shigella infections (Schroeder & Hilbi, 2008). During each stage of the infection process, both within the host and during transmission, Shigella must acquire iron. Like that of most bacterial species, iron is essential to survival of Shigella due to its critical role as a cofactor to enzymes involved in such processes as respiration and DNA replication. In the nonhost environment, iron is relatively abundant and bioavailable to Shigella and other bacteria. In the human body, however, as a result of host iron utilization and an innate host immune defense against infection, the level of bio-available iron is maintained at exceedingly low concentrations. One mechanism by which low free iron levels are maintained within the human host is be sequestration of the element within host compounds, such as heme. Binding approximately 95% of all iron within the human body, heme represents the most abundant potential source of nutritional iron for an invading pathogen if the pathogen is able to utilize the iron bound within (Otto, Verweij-van Vught, & MacLaren, 1992). The utilization of heme-bound iron is one mechanism by which some pathogenic bacteria overcome the extreme iron limitation encountered within the host. Examples of heme utilization systems in pathogenic bacteria include, but are not limited to, the Chu system present in some strains of pathogenic Escherichia coli, the Has system in Serratia marcescens, the Bhu system in Bordetella species, the Hut system in Bartonella quintana, and multiple heme acquisition systems in Haemophilus influenzae (Ghigo, Letoffe, & Wandersman, 1997;Murphy et al., 2002;Parrow, Abbott, Lockwood, Battisti, & Minnick, 2009;Vanderpool & Armstrong, 2001;Whitby, Seale, VanWagoner, Morton, & Stull, 2009;Wyckoff et al., 1998). Importantly, the ability of a bacterial pathogen to utilize heme as a source of nutritional iron is often positively associated with virulence (Otto et al., 1992).
The ability of S. dysenteriae to utilize heme as an iron source is strictly dependent upon the activity of the Shigella heme uptake (Shu) system (Mills & Payne, 1995;Wyckoff et al., 1998). Components of the Shu system are encoded within the shu locus, a 9.1 kb region predicted to contain eight genes under the control of four predicted promoters.
The predicted arrangement of genes and promoters within the shu locus would result in the generation of two monocistronic transcripts (shuA and shuS), and two polycistronic transcripts (shuTWXY and shuUV) (Wyckoff et al., 1998). The essential roles that Shu proteins play in the ability of S. dysenteriae to utilize heme as a sole source of iron has been demonstrated in multiple studies (Burkhard & Wilks, 2007; Eakanunkul et al., 2005;Wyckoff, Lopreato, Tipton, & Payne, 2005;Wyckoff et al., 1998). Despite a well-established role in the essential process of iron acquisition, direct measurement of the contribution of the Shu system to S. dysenteriae pathogenesis is limited given the lack of availability to relevant in vivo infection models and the limitations of in vitro tissue culture-based virulence assays. However, studies have indicated a positive association between the Shu or orthologous systems and virulence (Mills & Payne, 1995;Torres, Redford, Welch, & Payne, 2001;Wyckoff et al., 1998). Specifically, inactivation of chuA, the shuA ortholog in uropathogenic E. coli, diminishes the ability of the strain to colonize mice as compared to that of the wild-type strain (Torres et al., 2001). Additionally, the fact that only pathogenic bacterial strains encode the shu or orthologous genes, such as chu genes in pathogenic Escherichia coli and hmu genes in Yersinia species, support the conclusion that the ability to utilize heme as a source of nutritional iron facilitates bacterial virulence including, most likely, that of Shigella species (Mills & Payne, 1995;Wyckoff et al., 1998).
While iron is essential, the element is toxic when in excess (Everse & Hsia, 1997). As a means to maintain the critical balance between the necessity and toxicity of iron, the production of bacterial iron-uptake systems is tightly regulated. One such regulatory mechanism utilized by a wide variety of bacteria is that of transcriptional modulation by Fur, an iron-responsive transcriptional regulator (Bagg & Neilands, 1987). In these cases, Fur functions to repress transcription by binding in an irondependent manner to specific sequences within the promoter region of each target gene (Troxell & Hassan, 2013). Moreover, as potential sources of nutritional iron vary depending upon the environment, the production of different iron acquisition systems by pathogenic bacteria is often regulated in response to multiple environment-specific signals.
In addition to iron availability, the production of specific bacterial iron acquisition systems has been shown to be influenced by such environmental signals as oxygen levels and temperature (Battisti, Sappington, Smitherman, Parrow, & Minnick, 2006;Carpenter & Payne, 2014;Kouse, Righetti, Kortmann, Narberhaus, & Murphy, 2013;Vanderpool & Armstrong, 2003;Wei & Murphy, 2016a). Such multi-factored regulation likely ensures the production of a given iron acquisition system when, and only when, a bacterium is within an environment likely to contain the corresponding iron source.
Production of S. dysenteriae ShuA, the outer-membrane heme receptor of the Shu system, is known to be regulated in response to multiple environmental-specific signals. Specifically, shuA has been demonstrated to be subject not only to iron-dependent transcriptional regulation mediated by Fur, but also to temperature-dependent posttranscriptional regulation mediated by an RNA thermometer located within the 5′ untranslated region (5′UTR) of the gene (Kouse et al., 2013;Mills & Payne, 1997). RNA thermometers are cis-encoded ribo-regulators that inhibit translation of the target gene in which they are housed. At nonpermissive temperatures, an inhibitory hairpin is formed within the regulated transcript that physically occludes the Shine-Dalgarno (SD) sequence, thus preventing ribosomal binding and translation initiation (Kortmann & Narberhaus, 2012). RNA thermometers are most often implicated in regulating the production of heat shock proteins and specific bacterial virulence determinants (Kortmann & Narberhaus, 2012;Wei & Murphy, 2016b). The characterization of RNA thermometer-mediated regulation of S. dysenteriae shuA was the first demonstration of such regulation affecting the expression of a gene required for nutrient acquisition (Kouse et al., 2013).
Aside from shuA, little is known about the expression of any other gene within the S. dysenteriae shu locus. This study aims to identify and characterize the regulatory mechanism(s) controlling the expression of shuT, a gene encoding the periplasmic heme-binding component of the Shu system. The findings of this study provide the first direct evidence that shuT expression is regulated in response to two different host-associated environmental cues, and elucidates the molecular mechanism governing each. Specifically, it is demonstrated that shuT is subject to: (1) iron-dependent transcriptional regulation that is mediated by Fur and dependent upon a Fur-binding site located immediately downstream of the transcriptional start site of the gene, and (2) temperature-dependent posttranscriptional regulation mediated by an RNA thermometer located within the shuT 5′ UTR. Such multilevel regulation likely increases the fitness of S. dysenteriae by ensuring that ShuT production is limited to times within the infection cycle when the pathogen is within the human host, the only environment where it will encounter heme as a potential source of essential iron.

| Strains and culture conditions
All bacterial strains and plasmids used in this research are shown in Table S1. Escherichia coli was cultured in Luria-Bertani (LB) broth (1% tryptone, 0.5% yeast extract, and 1% NaCl) or on LB agar plates (LB with 1.6% [w/v]) at 37°C. S. dysenteriae was cultured in LB broth or on Tryptic soy broth agar plates (Becton, Dickenson and Company, Sparks, MD) containing 0.01% (w/v) Congo red dye (ISC BioExpress, Kaysville, UT) at the indicated temperatures. "Iron-rich" media (+Fe) refers to the LB broth, whereas "iron-poor" media (−Fe) refers to LB broth containing 150 μmol L −1 2,2′-bipyridine (Alfa Aesar, A Johnson Matthey Company) as an iron chelating reagent. Chloramphenicol and Ampicillin were used at a final concentration of 30 (μg ml −1 ) and 150 (μg ml −1 ), respectively, for the growth of bacterial strains carrying plasmids. Specifically for the heme utilization assay, 2,2′-bipyridine with a final concentration of 225 μ mol L −1 was added to LB broth to chelate iron from the culturing media, generating "iron-poor" conditions, under which about 90% of wild type S. dysenteriae growth is inhibited.
To generate the growth condition in which heme represents the major source of iron (heme), heme (Sigma) was added to the -Fe media at a final concentration of 100 (μg ml −1 ), a minimal concentration determined to permit maximal recovery of wild-type S. dysenteriae growth.

| Oligonucleotide primers
All oligonucleotide primers were designed based on the chromosomal sequences of S. dysenteriae and synthesized by Integrated DNA Technologies. Primer sequences used in this research are summarized in Table S2.

| Reporter plasmid construction
pT-lacZ: Complementary oligonucleotides containing the nucleic acid sequences of the putative shuT promoter region (Table S2) were hydrated in STE buffer (0.1 mol L −1 NaCl, 10 mmol L −1 Tris-HCl, 1 mmol L −1 EDTA, pH 8.0). The putative shuT promoter region was reconstructed by combining the complementarity oligonucleotides, boiling the molecules in a water-bath for 10 min and allowing them to cool slowly to room temperature. The annealed double-stranded DNA molecule was then ligated into plasmid pHJW20 (Castellanos et al., 2009) digested with restriction enzymes SalI (New England Biolabs inc.) and XbaI (New England Biolabs Inc.). Such cloning placed the DNA fragment containing the putative shuT promoter region immediately upstream of the report lacZ gene of pHJW20.
pT-UTR: Primers were designed to amplify the region containing the promoter and full 5′-UTR of shuT (Table S2) by PCR using genomic DNA of wild-type S. dysenteriae as template. The amplified fragment was purified by gel extraction and digested with restriction enzymes AatII (New England Biolabs Inc.) and NheI (New England Biolabs Inc.).
The digested fragment was then purified and ligated into plasmid pXG10 (Urban & Vogel, 2007) that had been digested with AatII (New England Biolabs Inc.) and NheI (New England Biolabs Inc.) to generate plasmid pT-UTR. Such cloning removed the PLtetO-1 constitutive promoter of pXG-10 and placed the promoter and full 5′ UTR of shuT immediately upstream of the reporter gfp gene.
pF-UTR, pS-UTR, and pD-UTR: Complementary oligonucleotide primers containing the indicated nucleic acid sequences (Table   S2) were combined and annealed in STE buffer (0.1 mol L −1 NaCl, 10 mmol L −1 Tris-HCl, 1 mmol L −1 EDTA, pH 8.0) by boiling in a waterbath for 10 min followed by slow cooling to room temperature. The generated double-stranded DNA products were then ligated into plasmid pXG10 (Urban & Vogel, 2007) that had been digested with restriction enzymes NsiI (New England Biolabs Inc.) and NheI (New England Biolabs Inc.). Such cloning placed the DNA fragment encoding the putative RNA thermometer being investigated between the PLtetO-1 constitutive promoter and the reporter gfp gene of pXG-10.

| DNA sequencing
All DNA sequencing was completed at the Ohio University Genomics Facility using Big Dye chemistry and analyzed on an Applied Biosystems 3130xL Genetic Analyzer.

| RNA extraction and DNA removal
Total RNA was harvested from wild-type S. dysenteriae cultured to the midlogarithmic phase under the indicated growth condition. Following growth to midlogarithmic phase, RNA preserving buffer (95% ethanol and 5% phenol, pH 4.5) was added to each culture at a culture:buffer ratio of 4:1, and the mixture incubated at 4°C overnight. Following overnight incubation, cells present in 3 ml of bacterial culture were pelleted by centrifugation at 17,000 g for 2 min. After discarding the supernatant, bacterial cells were resuspended in 357.3 μl DEPC-treated ddH 2 O, 40 μl 10% (w/v) sodium dodecyl sulfate (SDS), and 2.67 μl 3 mol L −1 sodium acetate (pH 5.2) by vortexing for 15 s.
Lysis of the bacterial cells was achieved by incubating each resuspended sample at 90°C for 7 min. After lysis, 1 ml of Trizol reagent (Ambion) was added to each sample, the sample transferred to a 2 ml phase-lock gel tube (5 PRIME Inc., Gaitherburg, MD) and the reaction incubated at room temperature (~25°C) for 5 min. Trizol reagent was extracted from each sample by the addition of 250 μl chloroform followed by vigorous shaking for 30 s to 1 min and then incubation at room temperature for 2 min. Each sample was then subjected to centrifugation for 2 min at 17,000 g and the nucleic acid containing aqueous phase transferred to a clean 2 ml microfuge tube. One milliliter of 100% ethanol was added to each sample and the samples incubated overnight at −80°C. Following overnight incubation at −80°C, nucleic acid present in each sample was pelleted by centrifugation at 17,000 g for 15 min at 4°C. Each pellet was washed by the addition of 1 ml cold 70% ethanol followed by centrifugation as described above. The supernatant was then removed and each pellet dried by centrifugation in a vacufuge (Eppendorf) for approximately 2 min in Alcohol Mode.
Finally, each RNA pellet was resuspended in 53 μl of nuclease-free water.
Following DNA removal from each RNA sample using TURBO DNA-free kit (Ambion) as directed, the absence of contaminating DNA was confirmed by PCR analysis; using the RNA sample as template, the lack of amplification by a known primer set indicates the absence of DNA in the RNA sample. Finally, the concentration of the total RNA present in each sample was measured using a ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE).

| Reverse transcriptase PCR
After harvesting total RNA from wild-type S. dysenteriae cultured under the indicated conditions via the procedure dictated above, a cDNA library was generated by using the iScript cDNA Synthesis Kit (Bio-Rad) as directed. PCR amplification was conducted to identify the approximate shuT transcriptional start site by using the generated cDNA library as template. Sequences of the four forward primers (shuT-F1, shuT-F2, shuT-F3, and shuT-F4) and one conserved reverse downstream primer (shuT-R) are detailed in Table S2. Validation of each primer set was achieved by using each in a PCR reaction containing genomic DNA of wild-type S. dysenteriae as template. Total RNA was used as PCR template with each primer set to confirm the absence of contaminating DNA; and double-distilled sterilized water was used as PCR template to confirm the absence of DNA contamination in reagents used for the PCR reaction. PCR products were detected by gel electrophoresis and UV light detection.

| 5′-RACE analysis
Following isolation of total RNA from S. dysenteriae cultured to the midlog phase under iron-poor conditions at 37°C and DNA removal (as detailed above), ribosomal RNA was depleted from each sample using the RiboMinus Transcriptome Isolation Kit for bacteria (Ambion) as directed, and the RNA sample concentrated by ethanol precipitation. 5′ RACE analysis was completed using the FirstChoice RLM-RACE Kit (Ambion) with the following modifications. First, approximately 250 ng of mRNAs was treated with the Tobacco Acid Pyrophosphatase provided by the FirstChoice RLM-RACE Kit prior to adaptor ligation. Secondly, a shuT-specific primer and SuperScript III Reverse Transcriptase (Invitrogen) were used to generate shuT-specific cDNA separately from adaptor-ligated RNA (experiment group) and RNA without adaptors (control group), which were then used as the templates for PCR amplification. The first round of PCR amplification was performed using the outer primer pair that binds within shuT and to the 5′ region of the adaptor sequence (Table S2). Amplified products that were present in the experiment group and absent in the control group were gel purified and used as the templates for a second round of PCR. The second PCR round was completed to increase the specific amplification of shuT using the inner primer pair: an upstream primer that binds to shuT at a location upstream to that bound by the primer used in the first amplification and a downstream primer that binds to the 3′ region of the adaptor sequence (Table S2).
Lastly, the amplified products of the second round of PCR were purified by gel extraction and ligated directly into the plasmid pGEM-T Easy (PROMEGA, Madison, MI) for nucleic acid sequencing.

| Beta-galactosidase assay
The Miller method was used to complete all beta-galactosidase analyses (Ausubel et al., 1995;Miller, 1992). Briefly, S. dysenteriae containing a given reporter plasmid (pT-lacZ) or pMic21 control vector was cultured at 37°C to stationary phase in 3 ml LB broth and then subcultured in 3 ml LB broth to the logarithmic phase (OD 600 : 0.28-0.7) at 37°C. 30 (μg ml −1 ) Chloramphenicol was added to each culture to ensure maintenance of the reporter or control plasmid. Following incubation of each culture on ice for 20 min, bacteria present in 1 ml of the culture were pelleted by centrifuging at 5,000g for 5 min. Next, each bacterial cell pellet was resuspended in 1 ml Z buffer (0.06 mol L −1 Na 2 HPO 4 , 0.04 mol L −1 NaH 2 PO 4 , 0.01 mol L −1 KCl, 0.001 mol L −1 MgSO 4 , 0.05 mol L −1 β-mercaptoethanol, pH 7.0) and the OD 600 measured using a ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE). Next, 400 μl of the resuspended cells was diluted 1:1 with Z buffer and the bacterial cells permeabilized by the addition of 50 μl 0.1% SDS and 100 μl chloroform prior to vortexing for 10 s.
Following incubation at 30°C for 15 min, 160 μl of 4 (mg ml −1 ), Ortho-Nitrophenylβ-galactopyranoside (ONPG) was added to each sample and the reaction incubated at 37°C until the sample developed a visible yellow color. Once a given sample developed a detectable yellow color, the time of incubation was noted, 400 μl of 1 mol L −1 Na 2 CO 3 was added to stop the reaction and the sample was subjected to centrifugation at maximum speed for 2 min. The supernatant of each sample was then collected and the OD 420 and OD 550 measured using the ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE). Miller units were calculated using the following equation: t×V×OD 600 (t stands for the reaction time in minutes; V represents the volume of culture used in milliliters, which, in this case, is 0.4 ml). A reaction containing all reaction components except the cell culture was used as a negative control in each beta-galactosidase assay. Additionally, S. dysenteriae carrying plasmid pMic21, a reporter plasmid containing no cloned promoter, was used to measure background beta-galactosidase in the strain.

| Western blot analysis
All western blot analyses were completed using whole cell extracts.
Specifically, S. dysenteriae containing the indicated plasmid was cultured to midlogarithmic phase under the indicated conditions and the OD 600 measured using the ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE). A total of 5 × 10 8 bacterial cells were pelleted by centrifugation at 17,000g for 2 min and suspended in 200 μl Laemmli protein dye (Bio-Rad) containing 5% β-mercaptoethanol. Samples were boiled for 10 min and then stored at −20°C until use.
Proteins present in 15 μl of each sample were separated on a 7.5% gel using sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). PVDF was cut to size, presoaked in methanol for 10 min and rinsed with water prior to transfer of all protein from the acrylamide gel to PVDF membrane at 350 milliamperes for 1 hr. Next, the membrane was blocked in 10% nonfat milk dissolved in PBS with 0.1% (wt ml −1 ) Tween 20 (PBST) overnight at 4°C. After blocking, the membrane was incubated for 1 hr at 4°C in a solution of mouse anti-GFP antibody (Roche) diluted 1:1000 in 5% milk with PBST. Next, the membrane was washed three times for 15 min each in PBST and blocked in 10% nonfat milk dissolved in PBS with 0.1% (wt ml −1 ) Tween 20 (PBST) prior to incubation for 1 hr at 4°C in a solution of goat antimouse HRP conjugated IgG (Bio-Rad) diluted 1:20,000 in 5% (wt ml −1 ) milk with PBST. Following incubation with the secondary antibody the membrane was washed with PBST three times for 15 min as detailed above. Then the Chemiluminescent HRP Substrate (Millipore Corporation, Billerica, MA) was added to the membrane, the reaction incubated at room temperature for approximately 3 min and the membrane imaged using the Molecular Imager ChemiDoc XRS+ imaging system (Bio-Rad). Total protein present on each membrane was visualized by staining with Ponceau S, following completion of the Western blot procedure to ensure even loading of all lanes.

| Quantitative real-time PCR analysis
For samples to be analyzed by both Western blot and quantitative real-time PCR (qRT-PCR), total RNAs was harvested from the same culture used to generate the corresponding whole cell protein preparations using the procedures detailed above. After the isolation of total RNA and subsequent DNA removal as detailed above, the iScript cDNA Synthesis Kit (Bio-Rad) was used to generate a cDNA library as directed. Each cDNA sample was diluted 1:10 in double-distilled water. Five microliter of diluted cDNA sample was mixed with 10 μl of iTaq Universal SYBR Green Supermix (Bio-Rad) and 5 μl of each primer set at an optimum concentration, making an amplification reaction mixture with a total volume of 20 μl. All amplification reactions were performed in a CFX96 Real-Time System (Bio-Rad) under reaction conditions optimized for each primer set. For each target gene, a six-point standard curve was generated to ensure that acceptable amplification efficiency was achieved and that all experimental samples amplify within the linear portion of the standard curve. Using the ΔΔCt calculation method the expression level of each target gene was normalized to that of rrsA present in each sample and expressed relative to that within a selected control sample. All primers used in qRT-PCR were designed using Beacon Designer 7.5 and are detailed in Table S2.

| Generation of labeled DNA species
Primer pXG10-for was labeled with 32 P at the 5′ end by T4 polynucleotide kinase (New England Biolabs Inc.) based on the factory protocol.
The radio-labeled primer was then used to generate 32 P end-labeled DNA fragments for EMSA via PCR amplification ( Chloramphenicol, and its presence verified by PCR-based screening.
Once constructed, pXG-1 was purified and used as template to amplify a DNA fragment identical to that generated above but lacking all shuT-derived sequences.

| Production and purification of S. dysenteriae Fur
PCR primers were designed based on S. dysenteriae fur gene sequence to generate a DNA fragment that contains a BamHI and EcoRI recognition site at the 5′ and 3′ end, respectively (Table S2)
Following this incubation, protein-bound and unbound DNA fragments were separated by electrophoresis on an 8% polyacrylamide gel and radioactive signals detected using a phosphor-imaging system.

| Construction of the run-off plasmid pT7-T
Complementary oligonucleotides containing the nucleic acid sequences of the T7 promoter, the full shuT 5′ UTR, and the first five codons of shuT (Table S2) were combined and annealed in STE buffer (0.1mol L −1 NaCl, 10 mmol L −1 Tris-HCl, 1 mmol L −1 EDTA, pH 8.0) by boiling in a water bath for 10 min followed by slow cooling to room temperature. The generated double-stranded DNA was then ligated into plasmid pXG10 (Urban & Vogel, 2007) that had been digested by restriction enzymes AatII (New England Biolabs inc.) and NheI (New England Biolabs inc.) to generate the run-off plasmid pT7-T, from which the 5′ portion of the shuT transcript would be generated for subsequent enzymatic RNA structure probing analysis.

| In vitro transcription
E. coli strain DH5α carrying the run-off plasmid pT7-T was cultured to the stationary phase of growth in 3 ml LB broth with 30 (μg ml −1 ) chloramphenicol at 37°C. Following extraction, purified pT7-T was linearized by digestion with restriction enzyme NheI-HF (New England Biolabs Inc.) at 37°C for 2 hr. After purification by gel extraction, the linear DNA was used as template in an in vitro transcription reaction carried out according to the protocol provided by AmpliScribe™ T7-Flash™ Transcription kit (epicenter, an Illumina company). Following completion of the in vitro transcription reaction, 1 μl of DNase was added to the reaction mixture, and the mixture was incubated at 37°C for 15 min. Next, RNA products were precipitated at −80°C for more than 15 min following addition of 1 ml 100% ethanol and 40 μl 3 mol L −1 sodium acetate (pH 5.2). Precipitated RNA was pelleted by centrifugation at 17,000 g for 15 min at 4°C and the RNA pellet washed twice by 1 ml ice-chilled 70% ethanol followed by centrifugation as described above. After drying by centrifugation in a vacufuge set on "Alcohol mode" (Eppendorf) for approximately 2 min, the RNA pellet was rehydrated in 15 μl DEPC-treated double-distilled water.

| Radio-labeling and enzymatic probing
Following synthesis and purification of the shuT transcript by in vitro transcription, the sample was treated with Calf Intestinal Alkaline Phosphatase (CIP) (New England Biolabs inc.) for 1 hr at 37°C to remove the 5′ triphosphate from each RNA molecule. Next, the CIP was removed from sample by phenol extraction and the RNA precipitated and dried as described above. The RNA pellet was rehydrated in 7 μl nuclease-free water and 5′ end radio-labeled as follows: T4 polynucleotide kinase (New England Biolabs Inc.) was used to transfer the γ-32 P-phosphate group from γ-32 P-ATP to the RNA sample. Radiolabeled transcripts were purified and extracted from 8% acrylamide gel (UreaGel-8, National Diagnostics) with elution buffer (10 mmol L −1 ethylenediaminetetraacetic acid (EDTA), 0.5% SDS, and 0.1 mol L −1 sodium acetate, pH 5.6), followed by RNA precipitation. Finally, structure probing assay was conducted using RNase T1 (Ambion), according to the procedures described in previous studies (Waldminghaus, Heidrich, Brantl, & Narberhaus, 2007). RNase T1 is used at either a 5-fold or 10-fold dilution as indicated.

| Statistical analysis
All experimental analyses were performed in biological triplicate. An Ftest was conducted to test whether the comparing groups have equal variances or not, followed by the corresponding two-tailed Student's t tests to determine significance (p ≤ .05).

| shuT expression is regulated in response to iron availability by the transcriptional regulator Fur
To experimentally determine if expression of shuT is regulated in response to iron availability, quantitative real-time PCR (qRT-PCR) was performed to measure the relative amount of shuT transcript in wildtype S. dysenteriae cultured to the midlogarithmic phase of growth under iron-rich and iron-poor conditions. Following growth under ironpoor conditions, the relative amount of shuT transcript measured in wild-type S. dysenteriae was significantly higher than that measured following growth of the strain under iron-rich conditions ( Figure 1). These data confirm that shuT expression is influenced by iron availability, and suggest that this regulation is mediated by an iron-responsive altera- Taken together, these data demonstrate that the expression of shuT is regulated in response to iron availability and that the observed irondependent decrease in shuT transcript levels is mediated, directly or indirectly, by the iron-dependent transcriptional regulator Fur.

| Identification of the transcriptional start site and promoter region of shuT
When direct, Fur regulates target gene transcription via irondependent binding of the protein to specific DNA sequences, termed Fur-binding site(s). For genes whose expression is directly repressed by Fur, the Fur-binding site(s) most often overlap the promoter region of a target gene (Bagg & Neilands, 1987). In these cases, binding by the regulator effectively blocks binding by RNA polymerase, and by doing so, prevents transcription initiation.
The first required step in characterizing the molecular mechanism underlying Fur-dependent regulation of shuT expression was to experimentally identify the transcription start site and promoter region of the gene. Based on the predicted transcript structure, four forward primers (F1, F2, F3, and F4) and one reverse primer (R) were generated and utilized in a series of reverse transcriptase PCR analyses designed to localize the shuT transcription start site (Figure 2a). The ability to amplify a specific product using each upstream primer (F1, F2, F3 (Figure 2b).
The isolated RNA was confirmed to be free of DNA contamination by the lack of amplification by any of the utilized primer pairs when the RNA itself was used as template ( Figure 2b). Together these data demonstrate that the shuT transcription start site falls within the 28-nucleotide region that separates primer F2 from primer F3.
To further define the transcription start site of shuT, rapid amplification of cDNA 5′ end (5′-RACE) analysis was performed using total RNA isolated from wild-type S. dysenteriae grown to the midlogarithmic phase under iron-poor conditions. Results obtained following sequencing of the products generated by 5′-RACE are consistent with those obtained by reverse-transcription PCR analysis and indicate that the transcription start site of shuT is located 42 nucleotides upstream of the start codon (Figures 2a,3a).
Guided by the experimentally confirmed location of the transcriptional start site and by in silico analysis (Bprom) a putative shuT promoter was identified (Solovyev & Salamov, 2011). The activity of the predicted shuT promoter was tested experimentally using beta-galactosidase analyses. Specifically, a transcriptional reporter plasmid (pT-lacZ) was constructed by cloning the predicted shuT promoter directly upstream of a promoter-less lacZ gene. The cloned S. dysenteriae sequence contains only one predicted promoter, and ends with the nucleotide immediately 5′ to the identified transcription start site (Figure 2c). Wild-type S. dysenteriae carrying reporter plasmid pT-lacZ or the empty vector control (pMic21) were cultured to the midlogarithmic phase under iron-poor conditions; and a basic beta-galactosidase assay was used to determine if the sequence cloned into pT-lacZ contains an active promoter. Significantly more beta-galactosidase activity was measured from wild-type S. dysenteriae carrying pT-lacZ as compared to that of the strain carrying the promoter-less negative control (pMic21) (Figure 2d). These data confirmed that the cloned sequences contain an active promoter, data consistent with the predicted localization of the shuT promoter ( Figure 2c).
The expression of shuT is regulated in response to iron by the transcriptional regulator Fur. Quantitative real-time PCR (qRT-PCR) targeting shuT was performed following growth of wildtype and Δfur Shigella dysenteriae to mid-logarithmic phase at 37°C under the indicated culture conditions. "+Fe" indicates "iron-rich" conditions achieved by culturing in Luria-Bertani broth, whereas "−Fe" represents "iron-poor" conditions achieved by culturing in Luria-Bertani broth containing 150 μmol L −1 2,2′-bipyridine. Using the ΔΔCt method of calculation, the expression level of shuT is normalized to that of rrsA present in each sample and is expressed relative to the amount of the transcript present in the first wild-type sample cultured under iron-rich conditions. All analyses were carried out in biological triplicate. Error bars indicate the standard deviation and "*" indicates a statistically significance difference (p-value ≤ .05)

| A functional Fur-binding site is located immediately down-stream of the shuT transcription start site
In order to experimentally determine if the observed irondependent regulation of shuT expression is mediated by specific nucleic acid sequences within the 5′ UTR and/or promoter region of the gene, a transcriptional reporter plasmid (pT-UTR) was con- Genomic DNA and water were used as positive and negative amplification controls, respectively. RNA template was used as template to demonstrate a lack of DNA contamination in each RNA sample. (c) A schematic of the plasmid pT-lacZ. "shuT-P" represents a 62-nucleotide region of the genome of S. dysenteriae that starts from 26 nucleotides upstream of the putative promoter of shuT and extends to the nucleotide before the identified transcription start site. The −35 and −10 promoter regions are highlighted by gray boxes and the nucleotide before the predicted transcription start site of shuT is labeled as "−1″. (d) Measurement of beta-galactosidase activity from wild-type S. dysenteriae containing pT-lacZ or the negative control plasmid pMic21. Plasmid pMic21 contains a promoter-less lacZ, and function as a negative control; whereas plasmid pT-lacZ has the putative promoter of shuT cloned immediately before the lacZ reporter gene. S. dysenteriae carrying the indicated plasmid was cultured to the midlogarithmic phase in LB broth with 150 μmol L −1 2,2′-bipyridine at 37°C. All analyses were carried out in biological triplicate. Error bars indicate one standard deviation and "*" indicates a statistically significance difference (p-value ≤ .05) only three nucleotides from the consensus Fur-binding site sequence was identified immediately downstream of the shuT transcription start site, from +1 to +19 of shuT (Figure 3a) (Lavrrar & McIntosh, 2003).
The prediction of a Fur-binding site within DNA encoding a 5′UTR is unusual, as in most cases of Fur-mediated iron-dependent repression, the Fur-binding site overlaps with the −35 region of the regulated promoter.
To confirm that sequences within the shuT promoter region do not mediate iron-dependent transcriptional regulation, a betagalactosidase assay using the reporter plasmid pT-lacZ was conducted.
This previously constructed reporter plasmid, pT-lacZ, contains nucleic acid sequences composing just the shuT promoter region cloned immediately upstream of the promoter-less lacZ reporter gene, sequences that do not include the putative Fur-binding site (Figure 2c).
F I G U R E 3 Iron-dependent regulation requires a putative Fur-binding site in the 5′ untranslated region of shuT. (a) Full sequence of the 5′ untranslated region of shuT as well as the sequence of a consensus Fur-binding site. Nucleotides within the identified putative Fur-binding site that differ from the consensus Fur-binding site sequence are indicated with an arrow. Regions between the dashed-lines indicate the region of shuT that has been inserted into the corresponding reporting plasmid (pT-UTR, pf-UTR, and pS-UTR). (b and c) Quantitative real-time PCR (qRT-PCR) analyses of the relative levels of gfp transcript present in wild-type and ΔfurShigella dysenteriae carrying the indicated reporter plasmid. Host cells were cultured under 37°C to midlogarithmic phase in either LB broth (+Fe for iron-rich conditions) or LB broth containing 150 μmol L −1 2.2′-bipyridine (−Fe for iron-poor conditions). Using the ΔΔCt method, the relative abundance of gfp is normalized to the amount of rrsA present in each sample and expressed relative to the amount of gfp transcript present in one of the wild-type samples grown under iron-rich conditions (WT+Fe). (d) Western blot analysis detecting Gfp levels in wild-type S. dysenteriae carrying the indicated reporter plasmid following growth under the indicated conditions. Histograms indicate the relative intensity of the detected Gfp-specific bands. (e) Measurement of beta-galactosidase activity from wild-type S. dysenteriae containing plasmid pT-lacZ, a reporter plasmid on which the constitutive plasmid promoter immediately upstream of the lacZ reporter gene has been replaced with the shuT promoter. Cultures were grown to midlogarithmic phase in LB broth (+Fe) or LB broth containing 150 μmol L −1 2.2′-bipyridine (−Fe). All analyses shown in this figure were carried out in biological triplicate. Error bars indicate one standard deviation and "*" indicates a statistically significance difference (p-value ≤ .05) As expected, no significant difference was measured in the amount of beta-galactosidase activity detected from wild-type S. dysenteriae carrying pT-lacZ grown under iron-rich or iron-poor conditions ( Figure 3e). These data confirm that the activity of shuT promoter itself is not influenced by iron availability.
To determine if nucleic acid sequences within the shuT 5′ UTR, the region containing the identified putative Fur-binding site, are sufficient to confer iron-dependent regulation, the reporter plasmid pF-UTR was constructed. pF-UTR contains the full 42-nucleotide shuT 5′ UTR cloned between a constitutive plasmid promoter and the reporter gene gfp (Figure 3a). Following up on the finding that sequences encoding the shuT 5′ UTR harbor a putative Fur-binding site and are sufficient to mediate iron-dependent regulation, electrophoretic mobility shift assays (EMSA) were conducted to determine if Fur binds directly to these nucleic acid sequences. 32 P end-labeled primers were used to generate radio-labeled DNA fragments via PCR amplification. Plasmid pF-UTR, which has been confirmed to mediate iron-dependent regulation, was used as PCR template to generate a DNA fragment containing the putative Fur-binding site ( Figure 3a). As a nonspecific control, an identical DNA fragment (negative control) lacking the shuT-specific sequences was amplified from plasmid pXG-1, which contains the constitutive promoter located immediately upstream of the gfp coding region. As shown in Figure  F I G U R E 4 Fur binds specifically to sequences within the shuT 5′ UTR. Electrophoretic mobility shift assay testing direct binding between Fur and nucleic acid sequence within the shuT 5′ UTR. 32 P-labeled DNA fragments were generated by PCR amplification from plasmid pF-UTR (shuT-UTR) or pXG-1 (Negative Control) using radio-labeled primers. Fragment "shuT-UTR", which is amplified from pF-UTR, contains the full 42-nucleotide shuT 5′ UTR; whereas fragment "Plasmid", amplified from pXG-1, has the same sequence as "shuT-UTR" but is lacking the 42-nucleotide shuT 5′ UTR. For the left five reactions indicated by "shuT-UTR", nonlabeled "Negative Control" fragment serves as the competitor DNA; whereas for the right four reactions indicated as "Negative Control", the competitor DNA is nonlabeled "shuT-UTR". The ratio between Fur protein (10 nmol L −1 ) and each radio-labeled DNA species (<1 nmol L −1 ) is about 10:1; concentrations of the competitor DNA are either 10 times, 50 times, or 100 times higher than that of the labeled DNA as indicated. After incubating the binding reactions at 37°C for 15 min, samples were separated on an 8% acrylamide gel. "+" and "−" represents the existence of the indicated component

| Nucleic acid sequences within the shuT promoter and/or 5′ UTR confer temperaturedependent posttranscriptional regulation
In addition to iron-limitation, environmental temperature also serves as a host-associated environmental signal. Specifically, the increase in temperature from a variable room temperature of approximately 25°C to a constant 37°C corresponds to entry of a pathogen into the human body, and can trigger the production of factors that facilitate bacterial survival and/or virulence in this often harsh environment. Multiple virulence-associated genes in Shigella are regulated in response to environmental temperature ( (Schroeder & Hilbi, 2008). Previously, the gene encoding the outer-membrane heme receptor of the Shu system, shuA, has been identified to be subject to temperature-dependent posttranscriptional regulation via the activity of an RNA thermometer located within the 5′ UTR of the gene (Kouse et al., 2013). Whether temperature changes also influence expression of other shu genes, especially those encoded on transcripts separated from that of shuA, remained unknown. shuT is encoded in the opposite direction of shuA, with more than 300-nucleotides separating the transcriptional start sites of these two genes (Wyckoff et al., 1998). While temperature does not influence the relative abundance of shuT transcript measured in wild-type S. dysenteriae (Figure 5), in silico analysis predicts that the gene may be subject to posttranscriptional temperature-dependent regulation. Specifically, in silico analysis using Mfold predicts the presence of a structure formed by the terminal 45 nucleotides (including the ATG) of the shuT 5′ UTR that, when formed, would occlude the SD site within the shuT transcript ( Figure 7a) (Zuker, 2003). Moreover, the predicted stabilizing energy (ΔG) of this putative inhibitory structure becomes less negative as temperature increases, with a decrease in the absolute value of nearly twofold upon a temperature increase from 25 to 37°C (data not shown). The identification of a putative temperature-sensitive structure within the shuT 5′ UTR that occludes the SD site supports the hypothesis that shuT translation is regulated in response to temperature via the activity of an RNA thermometer.
To experimentally determine if nucleic acid sequences contained within the shuT promoter and 5′ UTR are sufficient to confer temperature-dependent regulation, reporter plasmid pT-UTR ( Figure 3a) was used to determine if changes in environmental temperature result in altered expression of the reporter gene gfp. As described above, pT-UTR carries the shuT promoter and full 5′ UTR region cloned immediately upstream of the start codon of a gfp gene.
As a result of this arrangement, expression of gfp from the pT-UTR is driven by the shuT promoter and subject to regulation via any regulatory element located within the shuT promoter and/or 5′ UTR.
Following growth of wild-type S. dysenteriae carrying pT-UTR in iron-poor media at either 25 or 37°C, the relative levels of both gfp transcript and Gfp protein were measured by qRT-PCR and Western Blot analyses, respectively. Given that sequences within the 5′ UTR of shuT confer iron-dependent regulation, iron-poor media was used in these analyses to minimize the iron-dependent repression of shuT transcription. The data obtained demonstrate that while there was no significant difference in the relative amount of gfp transcript measured following growth of the strain at either temperature tested (Figure 6a), Gfp protein levels were significantly higher following growth of the strain at 37°C as compared to those measured following growth of the strain at 25°C (Figure 6b). Taken together, these data demonstrate that the nucleic acid sequences within the shuT promoter and/or 5′ UTR are sufficient to confer temperature-dependent posttranscriptional regulation onto expression of the reporter gfp gene; data that is consistent with regulation by an RNA thermometer.

| Nucleic acid sequences composing the putative shuT RNA thermometer are sufficient to confer temperature-dependent posttranscriptional regulation
According to the in silico prediction, the 5′ UTR of shuT forms a single hairpin that resembles an RNA thermometer, a regulatory element that would be predicted to confer temperature-dependent posttranscriptional regulation. To determine whether the observed temperaturedependent regulation of gfp expression from pT-UTR is mediated by nucleic acid sequences contained solely within the shuT 5′ UTR, wildtype S. dysenteriae carrying the pF-UTR reporter plasmid (Figure 3a) were grown in iron-poor media at either 25°C or 37°C. As detailed above, pF-UTR contains the full 42-nucleotide shuT 5′ UTR cloned between a constitutive plasmid promoter and the reporter gene gfp ( Figure 3a). As cloned, expression of gfp from pF-UTR is driven by the constitutive plasmid promoter and subject to regulation mediated solely via sequences contained within the cloned shuT 5′ UTR.
Following growth to midlogarithmic phase at 25 or 37°C, the relative amounts of gfp transcript and Gfp protein were measured by qRT-PCR and Western blot analyses, respectively. The data demonstrate that F I G U R E 5 Environmental temperature does not influence the relative abundance of shuT transcript. The relative abundance of shuT transcript was measured using quantitative real-time PCR in wild-type Shigella dysenteriae following growth of the strain to the midlogarithmic phase under iron-poor conditions at either 25 or 37°C. Iron-poor conditions were achieved by culturing in Luria-Bertani broth containing 150 μmol L −1 2,2′-bipyridine. Using the ΔΔCt method of calculation, shuT levels were normalized to that of rrsA present in each sample and are expressed relative to the amount of the transcript present in the first cultured at 37°C. All analyses were carried out in biological triplicate. Error bars indicate the standard deviation Gfp protein levels were significantly higher following growth of the strain at 37°C as compared to those measured following growth of the strain at 25°C (Figure 7b), whereas gfp transcript levels shown no significantly difference (Figure 7c). Taken together, these data demonstrate that nucleic acid sequences within the shuT 5′ UTR, sequences that are predicted to form an RNA thermometer, are sufficient to confer temperature-dependent posttranscriptional regulation.

| A functional RNA thermometer is contained within the shuT 5′ UTR
Site-specific mutagenesis was utilized to determine if the putative regulatory element identified within the shuT 5′ UTR is a functional RNA thermometer. Given that the regulatory activity of an RNA thermometer is dependent on a dynamic secondary structure, destabilizing the inhibitory hairpin structure of a functional RNA thermometer would be expected to alter its regulatory activity, specifically resulting in increased expression of the regulated gene at previously inhibitory temperatures. To destabilize the inhibitory hairpin within the shuT 5′ UTR, the uracils at position +15 and +16 (transcription start site is +1) were replaced by an adenine and a cytosine, respectively ( Figure 7a). As was done for the construction of the wild-type reporter (pF-UTR), the mutated 5′ UTR of shuT was cloned between a constitutive promoter and the reporter gene gfp, making plasmid pD-UTR (Figures 3a,7a). Since plasmid pD-UTR and pF-UTR are identical, with the exception of the destabilizing mutations within the cloned shuT 5′ UTR, plasmid pF-UTR was used as the wild-type control in these analyses. According to the results of Western blot and qRT-PCR analyses, S. dysenteriae carrying pD-UTR produced significantly higher amount of Gfp protein as compared to the strain carrying the wild-type reporter pF-UTR ( Figure 7d) when each strain was cultured at the previously inhibitory temperature of 25°C; under these conditions gfp transcripts levels were equivalent (Figure 7e). Thus, as expected, destabilization of the inhibitory structures within the putative regulatory element results in more efficient translation of the target gene at a nonpermissive temperature. These data demonstrate that sequences within the shuT 5′ UTR of shuT contain a functional RNA thermometer.

| Double-stranded structure containing the shuT ribosomal-binding site gradually opens when environmental temperature increases
A feature that defines an RNA thermometer is a temperatureresponsive alteration in structure that results in increased access of the ribosomal-binding site at permissive temperatures. To determine if the secondary structure of the shuT RNA thermometer truly responds to changes in environmental temperature, enzymatic probing of the in vitro transcribed shuT RNA thermometer was performed at experimentally validated permissive (37°C) and nonpermissive (25°C) temperatures using RNase T1, an RNase that specifically digests 3′ of each single-stranded guanine (Figure 8a,b). In order to account for the potential difference of enzyme activity at different temperatures, the intensity of each detected band resulting from digestion at each native guanine is expressed relative to that of the band resulting from digestion at G27, a site that is predicted and confirmed by this experiment to be single-stranded at each temperature tested (Figure 8a,c).
Completion of the structure probing analyses revealed that as the temperature increased from 25 to 37°C, the band intensity associated with G32 and G34 increased significantly, indicating increased sensitive of these residues to digestion by RNase T1 (Figure 8c).
These results indicate that the double-stranded structure containing the ribosomal-binding site of shuT gradually opens when the environmental temperature increases. Moreover, the relative intensity of the band resulting from digestion at G32 has a higher fold change (about 2.1-fold) when the temperature increases than that associated with digestion at G34 (about 1.8-fold), data suggesting that the dissociation of base pairs initiates from the loop region and proceeds down the length of the hairpin stem. Aside from those associated with digestion at position 32 and 34, none of the detected bands resulting from digestion at other native guanine residues demonstrated a significant change in intensity at the temperatures tested, data suggesting that these regions of the structure are less sensitive to alterations in environmental temperature (Figure 8c). Together these data demonstrate that the region of the shuT RNA thermometer structure containing G32 and G34 is most sensitive to alterations in environmental temperature, a sensitivity that would preferentially expose the SD sequences at permissive temperatures.
F I G U R E 6 shuT is subject to temperature-dependent posttranscriptional regulation. Wild-type Shigella dysenteriae carrying pT-UTR (reporter plasmid on which the constitutive plasmid promoter immediately upstream of gfp has been replaced by the shuT promoter and full-length 5′ UTR) were cultured to midlogarithmic phase at 25 or 37°C in LB broth containing 150 μmol L −1 2′-Bipyridine, and the relative amounts of gfp transcript or Gfp protein measured by quantitative real-time PCR (qRT-PCR) (a) or Western blot analysis (b), respectively. Protein and RNA samples were generated from the same set of cultures to ensure relevant comparisons. For the qRT-PCR analyses, the ΔΔCt method was used to normalize the level of gfp to that of rrsA present in each sample and express it relative to that measured in the first 37°C sample. All analyses were carried out in biological triplicate. Error bars indicate one standard deviation and "*" indicates a statistically significance difference (p-value ≤ .05)

| Heme utilization in S. dysenteriae is regulated by environmental temperature
Together with previous published data, two of the essential components of the Shu system have been characterized to having an RNA thermometer within the 5′ UTR of their transcript and subject to temperature-dependent regulation. Specifically, these two components is outer membrane heme-binding protein ShuA (Kouse et al., 2013) and periplasmic heme-binding protein ShuT (this study). As a means to determine if the observed regulation translates into a physiologically relevant phenotype, a series of growth analyses were completed. Specifically, it was investigated whether temperature-dependent regulation of shu genes translates into an impact on the ability of the pathogen to utilize heme as a sole source of iron at various temperatures. To this end, wild-type S. dysenteriae was cultured to stationary phase under conditions with varied iron availability and the OD 600 value measured as an indication of growth. LB broth alone was used as the iron-rich (+Fe) condition, whereas iron-poor (−Fe) condition was generated by adding 2,2′-bipyridine to the LB broth to a concentration that inhibits the growth of wild-type S. dysenteriae by about 90%. Heme was added to the iron-poor (−Fe) media, generating a growth environment (Heme) in which heme represents the major source of nutritional iron. When cultured at 37°C, the addition of heme to the iron-poor media restored growth of wild-type S. dysenteriae to that observe in the iron-rich media. When cultured at 25°C, however, the addition of heme to the iron-poor growth environment did not return growth of wild-type S. dysenteriae to the level observed for this strain when cultured in an iron-rich environment (Figure 9). These data demonstrate that the ability of S. dysenteriae to utilize heme as a sole source of nutritional iron is influenced by temperature, with increased efficiency observed during growth of the pathogen at the temperature within the human body, the only environment where Shigella will encounter heme as a potential source of nutrient iron.

F I G U R E 7
The 5′ untranslated region of shuT contains a functional RNA thermometer that is sufficient to confer temperature-dependent posttranscriptional regulation. (a) A schematic of the secondary structure formed by the nucleic acid sequence of the Shigella dysenteriae shuT 5′ UTR as predicted by Mfold analysis. Sequences composing the ribosomal-binding site are indicated by a bolded line and the translational start codon is in bold text. Arrows indicate the bases that were changed by site-directed mutagenesis to generate the destabilized version of the element (U15 was changed to A and U16 was changed to C). (b and d) Western blot analysis detecting the relative amount of Gfp in wild-type S. dysenteriae carrying the indicated reporter plasmid and grown under the indicated conditions. Histograms indicate the relative intensity of Gfp-specific bands. (c and e) Quantitative Real-time PCR (qRT-PCR) analyzing the relative amounts of gfp transcript present in wild-type S. dysenteriae carrying the indicated reporter plasmid and grown under the indicated conditions. Using the ΔΔCt method, the relative amount of gfp is normalized to the level of rrsA present in each sample and is expressed relative to the level measured in the first 37°C sample (c) or first pD-UTR samples (e). In all analyses presented, the indicated bacterial strain was cultured to mid-log phase in LB broth containing 150 μmol L −1 2.2′-bipyridine. Plasmid pF-UTR contains a constitutive promoter and the wild-type sequence of the full shuT 5′ UTR, followed by the reporter gene gfp; whereas plasmid pD-UTR has the mutated 5′ UTR of shuT inserted between the constitutive promoter and the gfp reporter gene. Protein and RNA samples were generated from the same set of cultures to ensure relevant comparisons. All analyses were carried out in biological triplicate. Error bars indicate one standard deviation and "*" indicates a statistically significance difference (p-value ≤ .05)

| DISCUSSION
The data presented in this study demonstrate that nucleic acid sequences harbored within the 5′ UTR of S. dysenteriae shuT confer both iron-dependent transcriptional regulation and temperaturedependent posttranscriptional regulation onto expression of the gene by distinct molecular mechanisms. Thus, as is the case of ShuA, ShuT production is limited to environments that mimic those encountered within the human host (low iron availability and relatively high temperature as compared to the nonhost environment), the only environment in which heme will be present as a potential source of nutritional iron ( Figure 10) (Kouse et al., 2013;Mills & Payne, 1997). In addition, having multiple host-associated environmental factors involved in regulating the production of multiple components of the heme acquisition system is likely beneficial to the pathogenic bacterium in at least two ways: (1) ensuring efficient response of the bacteria to changes of environmental factors; and (2) saving energy for other cellular processes when bio-available iron sources are relatively abundant and/ or when heme is not likely to be encountered. Furthermore, the coordinated regulation of multiple components of the S. dysenteriae Shu system likely increases fitness of the pathogen as these factors work in concert with one another and thus incomplete synthesis of the systems would provide limited benefit to the bacterium. Given that it is now established that both shuA and shuT are subject to posttranscriptional temperature-dependent regulation, it remains a distinct F I G U R E 9 Heme utilization in Shigella dysenteriae is more efficient at higher environmental temperature. Wild-type S. dysenteriae cells were grown to stationary phase under indicated conditions, and the optical density (OD 600 ) was measured as an indication of growth. Values were normalized to that of cells cultured in LB broth (+Fe) at each temperature (37 or 25°C). Iron-rich (+Fe) indicates growth of the strain in LB broth, whereas iron-poor (−Fe) indicates growth in LB broth supplemented with 225 μmol L −1 2,2′-bipyridine. To measure growth in the presence of heme as the major source of iron (Heme) 100 (μg ml −1 ) heme was added to iron-poor media. All analyses were carried out in biological triplicate. Error bars indicate one standard deviation and "*" indicates a statistically significance difference (p-value ≤ .05) F I G U R E 8 The inhibitory hairpin within the shuT RNA thermometer gradually opens as environmental temperature increases. In vitro transcribed RNA molecules were radio-labeled at the 5′ end, and then partially digested by RNase T1, which specifically cut immediately 3′ to single-stranded guanines. (a) Schematic of the shuT RNA thermometer with the Shine-Dalgarno (SD) sequence highlighted with a line, the start codon in bold text and sequence from the engineered NheI site bracketed. All potential RNase T1 cutting sites are indicated with arrows. (b) Representative gel showing the radio-labeled bands generated by digestion of the in vitro transcribed shuT RNA molecules. Control lanes: lane C contains RNA samples prior to the enzymatic or alkaline digestion, showing the background digestion of the experiment processes, lane L contains the sequencing ladder, and lane T1 contains the bands generated by RNase T1 digestion of the denatured template. Experimental lanes: lanes labeled 25 and 37°C contain the bands resulting from digestion of the shuT RNA thermometer at the indicated temperature with RNase T1 at a 5-fold dilution. (c) As a means to increase the sensitivity of the assay, it was repeated using a lesser concentration of RNase (10fold diluted) and the results quantified. Relative intensities of each band were normalized to that of the G27-associated band in the same lane. G27 is predicted to be single-stranded and thus subject to equal digestion at each temperature tested. All statistics were generated from three independent repeats. Error bars indicate one standard deviation. "*" indicates a statistically significance difference (p-value ≤ .05) possibility that the expression of the other shu genes is influenced in response to temperature, or another host-associated environmental condition, by an as of yet unidentified regulatory mechanism(s) (Kouse et al., 2013).
The presented studies demonstrate that Fur mediates irondependent transcriptional regulation of shuT expression by directly binding to nucleic acid sequences located immediately downstream of the transcription start site of the gene. The current model of the molecular mechanism underlying Fur-mediated repression of bacterial gene expression is that binding of the Fur protein within the promoter region of the regulated gene prevents the initiation of transcription by physically blocking binding of RNA polymerase (Troxell & Hassan, 2013). For genes whose transcription is repressed by Fur, Fur-binding sites are most commonly located adjacent to, or overlapping with, the −35 region of the gene's promoter (Christoffersen, Brickman, Hook-Barnard, & McIntosh, 2001;Escolar, de Lorenzo, & Pérez-Martín, 1997;Escolar, Pérez-Martín, & de Lorenzo, 1998). Located immediately downstream of the shuT promoter, the Fur-binding site mediating Fur-dependent regulation of shuT expression is in a relatively unique location with respect to the shuT promoter. Although the precise molecular mechanism remains to be characterized, it is possible that, similar to the accepted model of Fur-mediated regulation, binding of Fur to nucleic acid sequences immediately downstream of the promoter region acts to prevent transcription initiation by physically blocking binding by RNA polymerase. This regulatory model is supported by previous studies demonstrating that even with a core binding of only 19 nucleotides, in a DNA foot-printing assay Fur binding protects no <31 nucleotides, and potentially up to 100 nucleotides, due to the fact that Fur can form multimeric complexes (Baichoo & Helmann, 2002;Escolar, Pérez-Martín, & Lorenzo, 2000). A second potential mechanism underlying Fur-mediated repression of shuT expression is that binding by Fur prevents transcription elongation by physically blocking the processivity of RNA polymerase. The possibility of such a mechanism is supported by the identification of functional Fur-binding sites within the coding region of four genes in E. coli: gspC, garP, yahA, and fadD (Chen et al., 2007).
The shuT RNA thermometer is the second RNA thermometer located within the S. dysenteriae shu locus to be identified and characterized to date; the first being that housed within the 5′ UTR of shuA (Kouse et al., 2013). Although these two RNA thermometers are located in the same gene locus and regulate components of the same transportation system, they are distinct from each other in at least two important ways; distinctions that result in them being classified within two different families of RNA thermometers. First, the shuT 5′ UTR contains only 42 nucleotides and forms a single hairpin, a hairpin that functions as the shuT RNA thermometer. As a comparison, the shuA 5′ UTR is approximately 300-nucleotide in length and is predicted to contain eight hairpins, the 3′-most of which having been identified as the functional RNA thermometer (Kouse et al., 2013). The role of the remaining nucleic acid sequences and/or structures within the shuA 5′ UTR in controlling expression of the gene in response to temperature, or other signals, remains unknown. Second, a comparison of shuT and shuA reveals differences in the nucleic acid sequences of both the SD and anti-SD regions of each. Significantly, unlike the shuT RNA thermometer, the shuA RNA thermometer contains four consecutive uracil residues that base pair with sequences of the SD region within the transcript to form the inhibitory structure of the element, a feature that results in its classification as a member of the FourU class of RNA thermometers (Kouse et al., 2013).
When the shuT RNA thermometer is distinct from that of shuA, it does share at least four similarities with the RNA thermometer shown to regulate expression of a small heat shock protein, Hsp17, in cyanobacteria Synechocystis (Kortmann, Sczodrok, Rinnenthal, Schwalbe, & Narberhaus, 2011). The features conserved between the shuT and hsp17 RNA thermometers include: (1) a single stem-loop structure formed by sequences constituting the entire 5′ UTR; (2) complete pairing between sequences of the SD site and anti-SD site; (3) the presence of both an asymmetric internal loop and a large terminal loop within the inhibitory structure; and (4) the vast majority of the of the inhibitory structure being composed of relatively weak A-U or G-U pairs, with only two G-C pairs predicted. These features in the hsp17 RNA thermometer, particularly the presence of an asymmetric internal loop and the low number of G-C pairs ensure the thermosensing property of the hairpin by making it increasingly responsive to alterations in environmental temperature, a functional feature likely conserved in the shuT thermometer (Wagner, Rinnenthal, Narberhaus, & Schwalbe, 2015). Given the conservation of key features, shuT represents a second member of the now formed hsp17-like family of RNA thermometers. The identification and characterization of additional members of this newly formed family of RNA thermometers will advance the foundational understanding of not only how these elements function, but also of what cellular processes they control. Such knowledge is essential to improving the ability to first identify RNA thermometers, and second manipulate their activity for research, industrial and/or therapeutic applications. This study provides an example of two regulatory elements existing within the same sequence region, yet functioning to regulate gene expression via distinct molecular mechanisms. Such complex and multileveled regulation highlights the importance of the regulated F I G U R E 1 0 Model of shuT regulation in response to changes of iron availability and environmental temperature processes to the survival and/or virulence of bacterial pathogens.
Continuing studies to further elucidate the molecular events underlying known regulatory mechanisms, as well as those aimed at revealing novel regulatory mechanisms are critical to gaining a comprehensive understanding of bacterial pathogenesis. Such and understanding has the potential to advance the development of therapeutic agents to disrupt these key processes, and by doing so, limit the morbidity and mortality associated with bacterial infection.

ACKNOWLEDGMENTS
The presented studies were supported by a grant from the American Heart Association and by Ohio University and Ohio University Heritage College of Osteopathic Medicine. The authors of this manuscript also thank Dr. Michael Vasil for providing the Fur antibody, Dr.
Tomohiko Sugiyama for his technical support in EMSA, and Michelle Pate for her expert technical support.