RNA-Mediated Thermoregulation of Iron-Acquisition Genes in Shigella dysenteriae and Pathogenic Escherichia coli

The initiation, progression and transmission of most bacterial infections is dependent upon the ability of the invading pathogen to acquire iron from each of the varied environments encountered during the course of a natural infection. In total, 95% of iron within the human body is complexed within heme, making heme a potentially rich source of host-associated nutrient iron for invading bacteria. As heme is encountered only within the host, pathogenic bacteria often regulate synthesis of heme utilization factors such that production is maximal under host-associated environmental conditions. This study examines the regulated production of ShuA, an outer-membrane receptor required for the utilization of heme as a source of nutrient iron by Shigella dysenteriae, a pathogenic bacterium that causes severe diarrheal diseases in humans. Specifically, the impact of the distinct environmental temperatures encountered during infection within a host (37°C) and transmission between hosts (25°C) on shuA expression is investigated. We show that shuA expression is subject to temperature-dependent post-transcriptional regulation resulting in increased ShuA production at 37°C. The observed thermoregulation is mediated by nucleic acid sequences within the 5′ untranslated region. In addition, we have identified similar nucleotide sequences within the 5′ untranslated region of the orthologous chuA transcript of enteropathogenic E. coli and have demonstrated that it also functions to confer temperature-dependent post-transcriptional regulation. In both function and predicted structure, the regulatory element within the shuA and chuA 5′ untranslated regions closely resembles a FourU RNA thermometer, a zipper-like RNA structure that occludes the Shine-Dalgarno sequence at low temperatures. Increased production of ShuA and ChuA in response to the host body temperature allows for maximal production of these heme acquisition factors within the environment where S. dysenteriae and pathogenic E. coli strains would encounter heme, a host-specific iron source.


Introduction
Shigella species are gram-negative, facultative intracellular bacteria and the causative agents of shigellosis, a severe diarrheal disease in humans. The World Health Organization conservatively estimates the global health burden of shigellosis to be 165 million infections and 1.1 million deaths annually [1,2]. With an infectious dose estimated to be as small as ten bacteria, shigellosis is commonly spread directly from person-to-person via the fecaloral route of infection, although intermediates such as contaminated food or water have been shown to transmit the infection [3]. Once ingested, Shigella transit the gastrointestinal system of the human host to the colon, where a shigellosis infection is initiated by invasion of the bacterium into cells within the colonic epithelium [2]. In addition to profuse bloody diarrhea and intense inflammation, the primary symptoms of infection, shigellosis is also associated with the development of hemolytic uraemic syndrome (HUS) and Reiter's syndrome, causing sequelae to the kidneys, cardiovascular system and joints of the infected host [4,5].
During infection within and transit between hosts, Shigella species, like most pathogenic bacteria, must acquire iron to survive. Binding of iron within molecules such as heme, hemoglobin, myoglobin, transferrin and ferritin, effectively reduce the concentration of bioavailable iron within the human host to levels that are too low to support the growth of most bacterial pathogens [6,7]. In response, bacteria have evolved to produce specialized high affinity uptake systems to mediate the acquisition of nutrient iron from heme and other host-associated iron sources [7,8]. Shigella species contain several conserved iron uptake systems [9,10,11,12,13,14,15,16,17,18,19,20]; however, unique to S. dysenteriae is the Shu (Shigella Heme Uptake) system, a system dedicated to the utilization of heme and heme containing proteins as sources of nutrient iron [13,14,21,22].
The S. dysenteriae Shu system is encoded within a single chromosomal locus predicted to contain two monocistronic (shuA and shuS) and two polycistronic (shuTWXY and shuUV) transcripts driven by four iron-regulated promoters [13]. The shuA gene encodes an outer-membrane heme receptor that binds and imports heme into the periplasm in a TonB dependent manner [14,21]. Once inside the periplasm, the heme moiety is bound by ShuT and is transported to the cytoplasmic heme binding protein, ShuS, via the ABC transporter containing ShuU and ShuV in a coupled action [23]. Within the cytoplasm, it is proposed that ShuS either transports heme to a heme-degradation protein to utilize the molecule as a source of nutrient iron or to a hemebinding protein to prevent heme induced toxicity [24]. The functions of the remaining genes within the shu locus, shuWXY has not yet been deduced yet these genes are highly conserved across S. dysenteriae strains. Within pathogenic strains of E. coli the Chu system, an ortholog to the Shu system, has also been shown to mediate the utilization of iron from heme [25].
Given the abundance of heme within the human host (approximately 95% of the total iron [26]), and the importance of iron for survival, the ability of pathogenic bacteria to utilize heme as a source of nutrient iron is often correlated with virulence [27,28,29,30,31,32,33]. While the impact of the Shu system on the ability of S. dysenteriae to cause disease in the human host has not been directly tested, several lines of evidence suggest this system promotes virulence. First, the entire shu locus is conserved between S. dysenteriae and several pathogenic strains of Escherichia coli but is absent in non-pathogenic E. coli K-12 [13,29,31]. Furthermore, the presence of the shuA ortholog chuA in uropathogenic, enterohemorrhagic and enteroaggregative E. coli, has been positively correlated with virulence by both in vitro analysis and by epidemiological studies of naturally occurring human infections [13,29,31,34]. Finally, the pathophysiology of shigellosis, specifically the resulting physical damage and bleeding from the colonic epithelium, creates an environment in which S. dysenteriae will encounter heme. Together, these facts strongly suggest that the ability of S. dysenteriae to utilize heme as a source of nutrient iron provides the pathogen a competitive advantage within the human host.
While iron is essential for the survival of S. dysenteriae, and heme represents a potentially rich source of iron for the bacterium, both iron and heme are toxic at high concentrations [35]. The potential toxicity of iron and heme forces the bacterium to maintain a precise balance between the nutritional requirements for iron and the toxic effects of over-accumulation; an iron disequilibrium can lead to death of the bacterium [35]. Furthermore, it is energetically advantageous for the pathogen to produce a heme acquisition system only when it is within the host, as this is the sole environment in which the organism will encounter heme. For these reasons, the production of bacterial heme acquisition and utilization systems is often regulated in response to multiple hostassociated environmental conditions including iron limitation, the presence of heme and/or host body temperature [25,27,30,36,37,38,39,40,41].
The influence of environmental conditions in controlling the expression of genes within the shu locus has focused primarily on the impact of iron availability on expression of shuA, the gene encoding an outer-membrane heme receptor [13,14,21]. Previous studies have demonstrated that shuA transcription is regulated in response to iron availability by the activity of the iron-dependent transcriptional repressor Fur, such that maximal expression is observed under iron-limited conditions [14,21]. While these previous studies implicated iron-availability in modulating shuA expression, the impact of additional host-associated environmental conditions on shuA expression had not been investigated.
Several lines of evidence suggest that temperatures may influence shuA expression. First, temperature-dependent regulation of iron-acquisition systems has been observed in several other bacterial species [27,40]. Second, temperature varies significantly between the heme-containing environment of the human-host (37uC) and the non-heme containing environment encountered during transmission from one host to the next (25uC). Lastly, the prediction of a promoter located over 300 bases from the translational start site suggests the presence of an extended 59 untranslated region (utr) within the shuA mRNA molecule [13,14]. An extended 59 utr is a feature common among cis-acting RNA thermometers where RNA base pairing is known to regulate the expression of virulence and heat-shock genes in response to environmental temperature [42]. Such zipper-like RNA thermometers form inhibitory structures that function to occlude the ribosome binding site until the bacterium invades into a warmblooded host or sudden heat shock causes melting of the inhibitory structure and permits translation initiation. [43,44,45,46,47,48,49,50,51,52].
The goals of this study were to determine if the expression of S. dysenteriae shuA, and the E. coli ortholog chuA, is regulated in response to environmental temperature and to determine the molecular mechanism underlying any observed thermoregulation.

Bacterial Strains, Plasmids, Primers and Culture Conditions
All bacterial strains and plasmids used in this study are shown in Table 1. E. coli was routinely cultured in Luria-Bertani (LB) broth (1% tryptone 0.5% yeast extract and 1% NaCl) or on LB agar plates at 37uC. S. dysenteriae was routinely cultured in LB broth or on tryptic soy broth agar plates (Becton Dickenson and Company, Sparks, MD) containing 0.01% (wt/vol) Congo red dye (ISC BioExpress, Kaysville, UT) at the indicated temperature. Ironlimited growth conditions were achieved by the addition of 200 mg/ml of deferrated ethylenediamine-N,N'-bis 2-hydroxyphenylacetic acid (EDDHA) to LB. EDDHA was deferrated as detailed previously [53]. Chloramphenicol and Ampicillin were used at a final concentration of 30 mg/mL and 150 mg/mL respectively to select for the presence of reporter or control plasmids. All oligonucleotide primer sequences used in this study are available upon request.

Generation of the Gfp Translation Reporter Plasmid pshuA-gfp
Following isolation from E. coli strain Top10 using a Plasmid Midi Kit (Qiagen, Valencia, CA) according to manufacturer protocols, plasmid pXG-10 [54] was digested with AatII and NheI restriction endonucleases (New England BioLabs, Ipswich, MA) to remove the PLtetO-1 promoter and lacZ fragment. A DNA fragment containing the shuA start codon and 390 upstream nucleotides was amplified from the chromosome of wild-type S. dysenteriae by polymerase chain reaction using primers which contain AatII and NheI endonuclease recognition sites respectively. The amplified product was purified using a QIAQuick gel extraction kit (Qiagen) and then digested with AatII and NheI endonucleases and cloned into the digested pXG-10 plasmid backbone to create pshuA-gfp. The nucleic acid sequence of pshuAgfp was verified by nucleic acid sequencing of both DNA strands.

Generation of the Truncated Wild-type and Mutant
Translational Reporter Plasmids pWT-shuA, pS-shuA and pD-shuA maintain the PLtetO-1 constitutive promoter and remove the lacZ fragment. DNA fragments encoding wild-type and mutant FourU RNA elements were created by the annealing of complementary DNA oligonucleotides generating an insert with nucleic acid sequences at the 59 and 39 ends resembling sites digested with NsiI and NheI endonucleases, respectively. Complementary oligonucleotides were boiled for 10 minutes in 16STE buffer and allowed to cool to room temperatures. After annealing, the oligonucleotides were ligated into the pXG-10 backbone creating the translational fusion between the inserted sequence and the gfp reporter yielding plasmids pWT-shuA, pS-shuA and pD-shuA. Each plasmid generated was verified by nucleic acid sequencing of both strands to confirm insertion of the correct DNA fragments.
Creation of b-galactosidase Reporters pWT-chuA, pS-chuA and pD-chuA 293 nucleotides 59 of the chuA translational start site, containing the chuA 59 utr and the first two codons of the gene were amplified from uropathogenic E. coli UT189 (UPED NC_007946.1) via PCR using primers which introduce NheI and EcoRI restriction sites into the amplified product. Plasmid pBAD2-bgaB and the amplified chuA 59 utr were digested using NheI and EcoRI (Thermo Scientific, Fermentas, St. Leon-Rot, Germany) and ligated together forming the translational fusion and generating plasmid pWT-chuA. Mutations were introduced into the chuA sequence carried on plasmid pWT-chuA using the QuikChangeH mutagenesis kit (Stratagene, La Jolla, USA) according to manufacturer protocols. Specifically, within pWT-chuA the thymine at position 219, relative to the bgaB translational start site, was mutated to a cytosine to create pS-chuA, and an adenine to create pD-chuA. Plasmids were sequence verified by Eurofins (Eurofins, Martinsried, Germany).

b-galactosidase Assays
E. coli DH5a cells carrying the bgaB reporter plasmids pWT-chuA, pS-chuA and pD-chuA were grown overnight in 5 mL Luria-Bertani (LB) medium supplemented with ampicillin (150 mg/mL) at 25uC. 1 ml of the stationary phase culture was used to inoculate a 25 mL LB culture which was grown to an OD 600 of 0.5 prior to the addition of 0.01% arabinose (w/v) to induce transcription from each reporter plasmid. Next, 10 mL of each culture was shifted to pre-warmed flasks at 37uC and 400 mL of sample taken after 30 minutes of incubation. b-Galactosidase assays were performed on each collected sample as described by Rinnenthal et al. [55]. Standard deviations were calculated from three independent experiments.

Western Blot Analysis
All Western blot analyses were performed using whole cell extracts. Bacterial cultures were grown to stationary phase under the indicated growth conditions and the optical density at 600 nm measured using a ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE). A total of 5610 8 bacterial cells from each culture were pelleted and suspended in 200 mL of Laemmli protein dye (Bio-Rad) containing 5% 2-mercaptoethanol and were boiled for 10 minutes. Samples were stored frozen at 220uC until use. ShuA Western blot analysis was carried out as follows: 15 mL of each whole cell protein preparation was separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) using a 7.5% polyacrylamide gel and was then transferred to a PVDF membrane that had been pre-soaked in methanol for 10 minutes and rinsed three times with water. Following protein transfer, the membrane was blocked by incubation overnight at 4uC in a solution of phosphate buffered saline with 0.1% Tween20 (PBST) and 5% milk. After blocking, the membrane was washed 4 times for 5 minutes each time using PBST and then incubated for 1 hour at 4uC with polyclonal a-ShuA IgG (Custom antibody raised in a rabbit -Fisher Scientific, Waltham, MA) diluted 1:1000 in a solution of PBST with 5% milk. Following a series of washes as described above, the membrane was incubated for 1 hour at 4uC with goat anti-rabbit HRP conjugated IgG (Bio-Rad) diluted 1:10,000 in a solution of PBST with 5% milk. Finally, the membrane was rinsed as above, incubated in Immun-Star regents (Bio-Rad) according to manufacturer protocols and imaged using x-ray film.
Gfp Western blot analysis was carried out as detailed for that of ShuA with the following modifications. Following transfer of whole-cell protein samples to the PVDF, as detailed above, the membrane was blocked by incubation overnight at 4uC in a solution of PBST and 10% milk. Next, the membrane was incubated for one hour at 4uC with anti-Gfp monoclonal IgG stabilized antibody preparation (Roche, Indianapolis, IN) diluted 1:1000 in a solution of PBST and 5% milk. The membrane was then washed 3 times for 5 minutes, incubated in a solution of PBST and 10% milk for 10 minutes at 4uC and then incubated for 1 hour at 4uC with goat anti-mouse HRP conjugated IgG (Bio-Rad) diluted 1:20,000 in PBST with 5% milk. Finally, the membrane was washed and the blot was imaged using Immun-Star WesternC reagents (Bio-Rad) and a ChemiDoc XRS+ Imaging System (Bio-Rad). Contrast was altered uniformly over the entire image using Image Lab software (Bio-Rad) to reduce background noise and did not alter relative band intensity or the banding pattern observed.

RNA Extraction
Following growth of the bacterial culture to the indicated growth phase under the specified growth condition and at the indicated temperature, total RNA was extracted using an RNEasyH Mini-kit (Qiagen) according to manufacturer protocols. The purified RNA was then treated with 16 units of amplification grade Dnase I (New England BioLabs) for 1 hour at 37uC. One mL of 100% EtOH, 40 ml of 3 M sodium acetate (pH 5.2) and 100 mL of 1 mM ethylenediaminetetraacetic acid (EDTA) were added to each sample prior to incubation overnight at 280uC. Following precipitation of the RNA by centrifugation at 12,0006g for 15 minutes at 4uC, each sample was washed with cold 75% EtOH, the RNA was pelleted as above and dried. The RNA pellet was resuspended in diethyl polycarbonate (DEPC) treated water and the nucleic acid concentration measured using an ND-1000 spectrophotometer (NanoDrop Technologies). Finally, to ensure the removal of contaminating DNA, each RNA sample was treated using the TURBO DNA-free kit (Ambion, Austin, TX) according to the manufacturer protocols. The lack of DNA in each RNA sample was confirmed by PCR using the purified RNA as template and oligonucleotide primers known to amplify a portion of the shuA open reading frame.
Quantitative Real-Time Polymerase Chain Reaction cDNA was generated from 150 ng of total RNA using the iScript cDNA Synthesis Kit (Bio-Rad) according to manufacturer protocols. Each cDNA sample was diluted 1:10 in water and 5 mL used as template in a 20 mL amplification reaction. Primer concentration and reaction conditions were optimized for each primer set. All quantitative Real-time PCR reactions were carried out using iQ SYBR Green Supermix (Bio-Rad). A six point standard curve was generated for each target during each experimental run to ensure that an acceptable efficiency was achieved in the analysis. All expression values were calculated using the DDCt method, normalized to the level of rrsA measured in each sample and expressed relative to the value obtained in the indicated control sample. All reactions were performed in a Bio-Rad CFX96 Real-Time PCR System. All primer sequences were designed using Beacon Designer 7.5 and are available upon request.

Reverse Transcriptase PCR
A two-step reverse transcriptase PCR was conducted in a PTC-200 DNA Engine Cycler (Bio-Rad) using SuperScript III Reverse Transcriptase (Invitrogen, Grand Island, New York) according to manufacturer protocols followed by an amplification reaction. Using 150 ng of RNA isolated from wild-type S. dysenteriae cultured to stationary phase under iron-limited conditions as template, shuA specific cDNA was generated in the reverse transcription step using primer shu2 that binds within the shuA open reading frame. In the amplification step, specific products were amplified in a 50 mL reaction containing 1 mL template (cDNA), 5 mL of 2.5 mM dNTP, 5 mL of 106 Standard Taq Reaction Buffer, 1 mL of Taq DNA Polymerase, 36 ml of RNase free water, 1 mL of primer shu2 and 1 mL of either primer shu3, primer shu4 or primer shu6 at a final concentration of 0.6 mM of each primer. A ''No-RT control'' in which the isolated RNA was used as template in the amplification step was performed with each primer pair to screen for DNA contamination in the RNA sample. A positive amplification control using primer pair shu2 and shu6 was conducted using S. dysenteriae genomic DNA as template. Amplification products were resolved on a 2% agarose gel with 0.001% ethidium bromide and visualized with a ChemiDoc XRS+ Imaging System (Bio-Rad).

Alignment and in silico Modeling
Nucleic acid sequences for alignment of the promoters and 59 untranslated regions of shuA and chuA were acquired from GenBank. Alignments were conducting using Clone Manager 9 software and were recreated in Adobe Illustrator CS4. Predicted structure of shuA 59 utr, partial and complete, were obtained using Mfold software (http://mfold.rna.albany.edu/?q = mfold/RNA-Folding-Form).

Statistical Analysis
Two tailed, two sample Student's t tests, assuming equal variance, were used throughout our studies to determine significance (P#0.05).

ShuA Production is Influenced by Environmental Temperature
The impact of environmental temperature on ShuA levels was assessed directly by Western blot analysis using a polyclonal anti-ShuA antibody. ShuA levels were measured in whole cell extracts prepared from an equal number of wild-type Shigella dysenteriae (ND100 [56]) cells following growth to stationary phase under iron-limited conditions at 25uC, 30uC and 37uC. Iron-limited growth conditions were used for this study to relieve Fur-mediated repression of shuA transcription [14]. The data demonstrate a positive correlation between environmental temperature and ShuA levels with increased ShuA levels present in S. dysenteriae following growth of the strain at 37uC as compared to those measured following growth of the strain at either of the lower temperatures tested (Figure 1). A strain of S. dysenteriae lacking shuA (DshuA) was used as a negative control in this assay and as expected, shows no specific reactivity with the anti-ShuA antibody ( Figure 1). These data indicate that ShuA levels are modulated in response to environmental temperature with increased ShuA levels achieved at the host-associated temperature of 37uC as compared to the non-host associated temperatures of 25uC and 30uC.

Temperature-dependent Modulation of ShuA Production is Mediated by a Post-transcriptional Regulatory Mechanism
In order to determine which step within the shuA expression pathway is thermoregulated, the effect of environmental temperature on the relative abundance of shuA mRNA molecules was investigated by Quantitative Real-Time PCR (Q-PCR) analysis. shuA transcript levels were measured following growth of wild-type S. dysenteriae to mid-logarithmic or stationary phase at 25uC and 37uC under iron limiting conditions. The temperatures of 37uC and 25uC were chosen for this analysis because they represent the physiologically relevant temperatures encountered by the bacterium during an infection and during transmission, respectively, and because it is at these two temperatures that the greatest difference in ShuA protein levels was observed (Figure 1). Unlike ShuA protein levels, shuA transcript levels are not significantly higher following growth of wild-type S. dysenteriae at 37uC as compared to that following growth of the strain at 25uC (Figure 2). These data indicate temperature does not significantly influence the rate of shuA transcription or the stability of the shuA transcript. Temperature-dependent modulation of ShuA protein levels in the absence of parallel alteration in shuA mRNA levels indicates that the observed thermoregulation is mediated by a post-transcriptional regulatory mechanism.
Nucleic Acid Sequences within the shuA 59 Untranslated Region (utr) Confer Thermo-regulation of ShuA Translation An alteration of translational efficiency is often controlled by nucleotides surrounding or composing the Shine-Dalgarno (SD) sequence within the target mRNA (often within a 59 utr) [44,47,48,49]. To determine the influence of nucleotides within the shuA 59 utr on the observed temperature-dependent modulation of ShuA levels, the existence of the previously predicted shuA 59 utr and its ability to confer thermoregulation was investigated. The existence of the predicted shuA 59 utr was confirmed by reverse transcriptase PCR (RT-PCR) analysis ( Figure 3). These data demonstrate that the shuA 59 utr is between 278 and 342 nucleotides in length, a length that is in good agreement with the 328 nucleotide 59 utr previously predicted by sequence analysis [13].
Next, nucleic acid sequences composing the full-length 59 utr and shuA promoter were cloned into a modified pXG-10 [54] reporter plasmid to generate pshuA-gfp, a plasmid carrying a translational fusion between the first shuA codon and a functional gfp reporter gene. From this plasmid, expression of gfp is controlled by the native shuA promoter and may be influenced by nucleic acid sequences within the shuA 59 utr. The effect of temperature on gfp expression from pshuA-gfp was investigated by a combined approach using both Western blot and Q-PCR analyses following growth of wild-type S. dysenteriae carrying pshuA-gfp under iron limiting conditions to stationary phase at 25uC and 37uC. Western blot analysis using whole cell extracts generated from equivalent numbers of wild-type S. dysenteriae carrying pshuA-gfp demonstrated that Gfp protein levels are higher following growth of the strain at 37uC as compared to those seen following growth of the strain at 25uC ( Figure 4A). Q-PCR analysis demonstrated that temperature does not significantly alter the relative abundance of gfp mRNA present in wild-type S. dysenteriae carrying the pshuA-gfp reporter plasmid ( Figure 4B). The observed temperature-dependent alteration of Gfp protein levels in the absence of a corresponding change in gfp transcript levels from the pshuA-gfp translational reporter plasmid indicate that nucleic acid sequences within the shuA promoter and 59 utr are sufficient to confer post-transcriptional temperature-dependent regulation onto the expression of the gfp reporter gene.  In silico Analysis Predicts the Presence of a FourU RNA Thermometer within the shuA 59 utr Given that nucleic acid sequences composing the shuA promoter and 59 utr are sufficient to confer post-transcriptional thermoregulation, (Figure 4) the predicted secondary structure of this region was evaluated by in silico modeling to identify putative regulatory elements. MFold analysis was used to predict the secondary structure of the entire shuA 59 untranslated region (http://mfold. rna.albany.edu/?q = mfold/RNA-Folding-Form), the most energetically favorable of which is shown in Figure S1. Ten out of twelve structures generated by Mfold included a short hairpin that is formed by canonical and non-canonical base-pairing between four consecutive uracil residues and the predicted Shine-Dalgarno sequence ( Figure 5 and Figure S1). These characteristics are found in FourU RNA thermometers, a relatively new class of cis-acting riboregulators shown to mediate post-transcriptional thermoregulation of target gene expression by occlusion of the ribosomal binding site at non-permissive temperatures [42,45,55,57].

The Nucleotide Sequence Composing the Putative shuA FourU RNA Thermometer is Sufficient to Confer Posttranscriptional Thermoregulation
To determine if the nucleotide sequence comprising the predicted RNA thermometer is sufficient to confer post-transcriptional thermoregulation onto reporter gene expression, a 32 nucleotide long insert containing the putative FourU RNA thermometer ( Figure 5) was cloned between the constitutive PLtetO-1 promoter and the gfp reporter gene on plasmid pXG-10. Such cloning generates a translational fusion between the first codon of shuA and gfp, the transcription of which is driven by the plasmid promoter and protein production is potentially influenced by the putative shuA FourU thermometer cloned into pXG-10. The newly created plasmid, designated pWT-shuA, was introduced into Escherichia coli, and thermoregulation of gfp expression was investigated by Western blot and Q-PCR analyses following growth of the strain to stationary phase at 25uC or 37uC ( Figure 6). Western blot analyses demonstrate that Gfp protein levels are increased following growth of E. coli carrying pWT-shuA to stationary phase at 37uC as compared to those measured following growth of the strain at 25uC ( Figure 6A) while Q-PCR analyses indicate that temperature does not influence the relative abundance of gfp mRNA ( Figure 6B). Together, these data clearly demonstrate that the predicted FourU RNA thermometer contained within the shuA 59 utr is sufficient to confer posttranscriptional thermoregulation onto the production of the Gfp reporter protein.

Site-directed Mutagenesis of the Putative FourU Element Alters Thermoregulation of the gfp Reporter Gene
To further verify the existence of a FourU RNA thermometer within the shuA 59 utr, single nucleotide mutations were introduced into the regulatory element cloned within the pWT-shuA reporter plasmid. The first mutant construct, pS-shuA, contains a thymine For each experimental reaction, primer shu2 was used to generate the shuA specific cDNA product. In each amplification step, the reverse primer shu2 was paired with the indicated forward primer. A control reaction using RNA as template was conducted with each primer set to ensure that the isolated RNA sample did not contain DNA contamination; a (2) indicates RNA was used as template in the amplification reaction while a (+) indicates that cDNA was used as template in the amplification reaction. DNA was used as template in the amplification step to verify that the indicated primer pair was capable of facilitating target amplification. doi:10.1371/journal.pone.0063781.g003 to cytosine mutation to generate a cytosine-guanine base pair within the predicted RNA thermometer ( Figure 7A). This mutation was created to stabilize the closed conformation of the FourU RNA thermometer, a mutation that is predicted to decrease reporter protein production from pS-shuA compared to pWT-shuA when measured at the permissive temperature of 37uC. E. coli was transformed with pS-shuA and Gfp production compared to that from pWT-shuA by Western blot analysis following the growth of each strain at 37uC. Gfp production from pS-shuA following growth of the reporter strain at the permissive temperature of 37uC is inhibited as compared to that from the strain carrying pWT-shuA ( Figure 7C). Q-PCR analysis demonstrates that gfp transcript levels produced from pS-shuA and pWT-shuA are equivalent under the conditions tested, indicating that the observed difference in Gfp levels does not result from altered transcription or transcript stability of the mutated reporter gene ( Figure 7E). Together these data demonstrate that, as expected, mutations that are predicted to stabilize the inhibitory structure within the putative FourU RNA thermometer result in decreased expression of the regulated gene at the permissive temperature of 37uC.
The second mutant construct, pD-shuA, contains a cytosine to adenine mutation which is predicted to abolish a canonical cytosine-guanine base-pair and destabilize the inhibitory structure of the FourU RNA thermometer thus allowing for increased Gfp production at the non-permissive temperature of 25uC ( Figure 7B). E. coli was transformed with pD-shuA and Gfp production compared to that from pWT-shuA by Western blot analysis following the growth of each strain at 25uC. Gfp production from pD-shuA following growth of the reporter strain at the nonpermissive temperature of 25uC is increased as compared to that from the strain carrying pWT-shuA ( Figure 7D). Q-PCR analysis demonstrates that gfp transcript levels produced from pD-shuA and pWT-shuA are equivalent under the conditions tested, indicating that the observed difference in Gfp levels does not result from altered transcription or transcript stability of the mutated reporter gene ( Figure 7F). Together these data demonstrate that, as expected, mutations that are predicted to destabilize the inhibitory structure within the putative FourU RNA thermometer result in increased expression of the regulated gene at the non-permissive temperature of 25uC.
Collectively, these mutagenesis studies demonstrate that thermoregulation conferred by the putative shuA FourU RNA thermometer is mediated by differential stability of the identified inhibitory structure. These data strongly support the conclusion that the shuA 59 utr harbors a functional FourU RNA thermometer.
The 59 utr of chuA, the E. coli Ortholog of shuA, Contains a FourU RNA Thermometer Sufficient to Confer Posttranscriptional Thermoregulation An alignment among the 59 utr of shuA and its ortholog chuA from several strains of pathogenic E. coli shows moderate sequence variation; however, the FourU RNA thermometer and surrounding regions are completely conserved (Figure 8). To determine if the E. coli chuA 59 utr is capable of conferring thermoregulation, this region was amplified from uropathogenic E. coli UT189 (UPEC NC_007946.1) and cloned into plasmid pBAD2-bgaB, a well-established translational reporter system [58]. Specifically, a region containing 293 nucleotides of the chuA 59 utr and the first two codons of the open reading frame was cloned in frame with the bgaB reporter, encoding a heat-stable b-galactosidase, to  . Sequences composing the shuA FourU thermometer are sufficient to confer post-transcriptional thermoregulation onto gfp expression. A) E. coli strain DH5a containing pWT-shuA or pXG-0, as a vector control, was cultured to stationary phase at the temperatures indicated. A Western blot was performed using whole-cell lysates generated from an equal number of cells and anti-Gfp monoclonal antibodies. B) Quantitative Real-time PCR was conducted using RNA isolated from E. coli DH5a carrying pWT-shuA following growth of the strain to stationary phase at 25uC or 37uC. gfp mRNA levels were normalized to rrsA measured in each sample and expressed relative to the amount of gfp transcript measured in the first 25uC sample. All data are representative of three biological replicates and error bars represent one standard deviation. Assuming a confidence interval of 95% (p#0.05), no significant difference exists between the relative levels of gfp transcript measured from pWT-shuA following growth of the strain at 25uC or 37uC. doi:10.1371/journal.pone.0063781.g006 generate the translational reporter plasmid pWT-chuA. Expression of bgaB from pWT-chuA is under control of the pBAD arabinose inducible promoter and is potentially influenced by the putative chuA FourU RNA thermometer contained within the cloned E. coli sequence ( Figure 9A). E. coli DH5a was transformed with pWT-chuA and the thermoregulation of bgaB activity investigated. A temperature shift from 25uC to 37uC resulted in a significant increase in the bgalactosidase activity measured from pWT-chuA ( Figure 9E). These data demonstrate that like that of S. dysenteriae shuA (Figure 6), Figure 7. Mutational analysis demonstrates that the shuA 59 utr contains a functional FourU RNA thermometer. The shuA region cloned into pWT-shuA was mutagenized to further validate the existence of a predicted FourU element within the shuA 5'utr. A) The uracil residue located 19 nucleotides upstream of the gfp translational start site was mutated to a cytosine. This mutation, indicated by the box, is predicted to stabilize the inhibitory structure within the putative FourU RNA thermometer. This mutated sequence was cloned into the gfp translational reporter pXG-10 to generate the stabilized mutant construct designated pS-shuA. B) The cytosine residue 17 nucleotides upstream of the gfp translational start site was mutated to an adenine. This mutation, indicated by the box, is predicted to destabilize the inhibitory structure within the putative FourU RNA thermometer. This mutated sequence was cloned into the gfp translational reporter pXG-10 to generate the destabilized mutant construct designated pD-shuA. Western blot analyses were conducted using monoclonal anti-Gfp antibodies and whole-cell extracts generated from an equal number of E. coli carrying pWT-shuA or pS-shuA cultured to stationary phase in LB at the permissive temperature of 37uC (C), and E. coli carrying pWT-shuA or pD-shuA cultured to stationary phase in LB at the non-permissive temperature of 25uC (D). Quantitative real-time PCR analysis was performed using RNA isolated from E. coli DH5a cells carrying pWT-shuA, pS-shuA and pD-shuA after culturing the strains to stationary phase using the temperatures indicated. gfp transcript levels were normalized to the amount of rrsA in each sample and set relative to the amount of gfp in the first pWT-shuA sample. All data are representative of three biological replicates and error bars represent one standard deviation. Assuming a confidence interval of 95% (p#0.05), no significant difference exists between the relative levels of gfp transcript measured from pWT-shuA and pS-shuA (E) or pD-shuA (F) at the temperatures tested. doi:10.1371/journal.pone.0063781.g007 sequences within the 59 utr of UPEC chuA are sufficient to confer thermoregulation onto the expression of a reporter gene on a translational reporter plasmid. Two control plasmids were utilized in this study. The negative control plasmid, pgyrA, carries the reporter bgaB translationally fused to the 59 utr of the nonthermoregulated E. coli gyrA and is a modification of that used previously [45]. The positive control plasmid, pagsA [58], carries the reporter bgaB translationally fused to a previously characterized FourU RNA thermometer from Salmonella agsA. As expected, a temperature-dependent increase in b-galactosidase activity was also seen from the positive control plasmid pagsA, while no thermoregulation was observed from the negative control plasmid pgyrA ( Figure 9D).
To further characterize the UPEC chuA FourU RNA thermometer, point mutations were introduced to stabilize or destabilize the predicted inhibitory structure cloned within the pWT-chuA reporter plasmid. Non-pathogenic E. coli DH5a was transformed with each mutated reporter plasmid and the influence of the mutation of thermoregulation of the reporter bgaB gene evaluated by b-galactosidase analyses. The first mutant construct, pS-chuA, carries a thymine to cytosine mutation that introduces a C-G pairing within the predicted RNA thermometer ( Figure 9B). As predicted, this stabilizing mutation results in significantly less bgalactosidase activity from the pS-chuA translational reporter at the permissive temperature of 37uC as compared to that measured from pWT-chuA ( Figure 9E). The second mutant construct, pD-chuA, carries a thymine to adenine mutation that destabilizes the predicted chuA FourU RNA thermometer by disrupting a U-A pairing ( Figure 9C). This destabilizing mutation results in significantly more b-galactosidase activity from the pD-chuA translational reporter at the non-permissive temperature of 25uC as compared to that measured from pWT-chuA ( Figure 9E). Together, these data indicate that, like that of S. dysenteriae shuA, the expression of UPEC chuA is regulated in response to environmental temperature by the activity of a FourU RNA thermometer located within the 59 utr of the gene.

Discussion
This study demonstrates that the expression of S. dysenteriae shuA and E. coli chuA, genes essential for the utilization of heme as a source of nutrient iron, is modulated in response to environmental temperature by a post-transcriptional regulatory mechanism. This work is the first to demonstrate temperature-dependent posttranscriptional regulation of shuA and chuA expression and the first to implicate temperature as an environmental factor controlling the production of an iron uptake system in Shigella. Finally, it is demonstrated that the observed post-transcriptional thermoregulation of both shuA and chuA expression is mediated by the activity of a conserved FourU RNA thermometer within the 59 utr of each gene.
The conservation of the FourU RNA thermometer between related pathogenic bacterial species (Figure 8) suggests that it provides a selective advantage by increasing the fitness of each species. Multiple regulatory mechanisms control the expression of both S. dysenteriae shuA and E. coli chuA and this hierarchical regulation likely increases the fitness of each pathogen by limiting the production of the heme binding receptor to environmental conditions mimicking those encountered within the heme containing human body ( Figure 10). This work along with previous studies has demonstrated that the expression of shuA and chuA is regulated by both iron availability and environmental temperature [14,25]. Dual-regulation of shuA and chuA likely increases the overall fitness of S. dysenteriae and E. coli by promoting expression of each gene within the host, where heme is going to be encountered, while efficiently preventing expression of each gene in the non-host environment where heme will not be encountered. Specifically, the iron-limitation and environmental temperature (37uC) encountered within the human-host relieves both the Furmediated inhibition of shuA transcription and FourU RNA thermometer mediated inhibition of shuA translation, respectively Figure 9. A functional FourU RNA thermometer within the 59 utr of E. coli chuA confers thermoregulation. A) Predicted structure of a portion of the chuA mRNA molecule containing the start codon (AUG) and preceding 29 nucleotides. This region contains the putative chuA FourU RNA thermometer. B) The uracil residue located 19 nucleotides upstream of the bgaB translational start site was mutated to a cytosine. This mutation, indicated by the box, is predicted to stabilize the inhibitory structure within the putative chuA FourU RNA thermometer. The mutated sequence was cloned into the pBAD2-bgaB translational reporter to generate the stabilized mutant construct designated pS-chuA. C) The uracil residue located 19 nucleotides upstream of the bgaB translational start site was mutated to an adenine. This mutation, indicated by the box, is predicted to destabilize the inhibitory structure within the putative chuA FourU RNA thermometer. The mutated sequence was cloned into the pBAD2-bgaB translational reporter to generate the destabilized mutant construct designated pD-chuA. The influence of temperature on the expression of bgaB from the positive and negative control plasmids pagsA and pgyrA D) and the translational reporters pWT-chuA, pS-chuA and pD-chuA E) was determined by measuring the b-galactosidase activity produced form E. coli DH5a carrying each. All date are the average of three replicate experiments. Error bars represent one standard deviation and a *indicates a statistically relevant difference between the indicated strains (p#0.05). doi:10.1371/journal.pone.0063781.g009 ( Figure 10). Consequently, host-associated conditions are optimal for shuA and chuA expression and subsequent heme acquisition by S. dysenteriae and pathogenic E. coli. Conversely, the non-host environment encountered during transmission, marked by an increase in available iron and a decrease in temperature, will result in Fur-dependent inhibition of shuA and chuA transcription and FourU RNA thermometer dependent inhibition of translation, respectively ( Figure 10). As a result of this dual regulation, shuA expression will be efficiently inhibited during transmission through non-heme containing environments.
A further advantage that is likely conferred by the regulation of shuA and chuA expression at both the transcriptional and translational level of gene expression is that such regulation likely facilitates rapid adaptation of the pathogen to changing environmental conditions. It is reasonable to assume that upon being shed from an infected human host, S. dysenteriae and pathogenic E. coli experiences an increase in relative iron-availability and a decrease in environmental temperature. Inhibition of both transcription and translation would likely provide an advantage over regulation only at the level of transcription in this scenario, by preventing the translation of existing shuA and chuA mRNA molecules, thus saving the pathogen the energetic cost of producing the heme receptor in an environment in which the protein will provide no advantage to the bacterium.
While shuA is regulated at both the transcriptional and translational level of expression, additional regulatory mechanisms could influence shuA expression. The FourU element is located within the last 30 nucleotides of the fairly well conserved shuA 59 utr that measures over 300 nucleotides (Figure 3). In silico folding analyses predicts a highly structured 59 utr ( Figure S1) that may play additional roles in shuA regulation. The impact of the remaining 59 utr on shuA and chuA expression remains under investigation. It is plausible that additional elements within the 59 utr of chuA and shuA are capable of stabilizing the FourU element and/or regulating expression of each gene in a manner independent of the FourU RNA thermometer.
Only two other FourU thermometers, agsA from Salmonella and icrF from Yersinia species, have been extensively characterized to date [45,55,57,59]. Given the conservation and distribution of the shu locus and identified FourU RNA thermometer among S. dysenteriae and several strains of pathogenic E. coli ( Figure S1) [13], as well as the predicted distribution of FourU RNA thermometers among both Gram-negative and Gram-positive bacterial species [45], the significance of this study reaches beyond S. dysenteriae gene regulation and into the broader fields of enteric pathogenesis and RNA-mediated mechanism of bacterial gene regulation.
Only after multiple FourU RNA thermometers are identified and characterized will a complete understanding of this potentially wide-spread regulatory element be possible. Future studies aimed at characterizing the regulatory function of the shuA FourU RNA thermometer in finer molecular detail as well as the identification and characterization of additional FourU RNA thermometers in Shigella and other bacterial species will not only further the understanding of this newly recognized regulatory element functions but will also reveal the potentially expansive role that Figure 10. Dual regulation controlling the expression of S. dysenteriae shuA and chuA of pathogenic E. coli. The production of S. dysenteriae ShuA and E. coli ChuA is regulated at the levels of transcription and translation by two independent mechanisms. Transcription of the shuA and chuA message is inhibited under high-iron conditions by the iron-dependent transcriptional repressor Fur in a temperature-independent manner. However, under iron-poor conditions Fur mediated repression of shuA/chuA transcription is relieved and mRNA is synthesized. Translation of the shuA/chuA transcript is regulated in response to environmental temperature via a FourU RNA thermometer within the 59 untranslated region of each gene. At a temperature of 25uC, translation of shuA/chuA is blocked by the formation of an inhibitory structure predicted to occlude the ribosomal binding site and prevents ribosomal binding. At a temperature of 37uC the inhibitory structure within the shuA and chuA 59 utr is destabilized and efficient translation of each gene proceeds. This model shows that production of ShuA and ChuA occurs only under iron-poor conditions and at 37uC, conditions which mimic the host environment where S. dysenteriae and pathogenic E. coli will encounter heme as a rich source of nutrient iron. doi:10.1371/journal.pone.0063781.g010 FourU RNA thermometers play in bacterial physiology and pathogenesis. Figure S1 In-silico folding analysis of the full-length shuA 59 utr. The full-length 59 utr and start codon of shuA was submitted to Mfold for modeling of the RNA secondary structure. The predicted FourU RNA thermometer is denoted by a box. (TIF)