Characterization of temperature‐dependent hemin uptake receptors HupA and HvtA in Vibrio vulnificus

Abstract The Gram‐negative pathogen Vibrio vulnificus produces several iron‐sequestration systems including a hemin uptake system in response to iron limitation as a means to acquire this essential element. Strains of this organism are capable of causing serious septicemia in humans and eels, where hemin is abundant and an advantageous source of iron. Vibrio vulnificus hemin uptake systems consist of HupA, a well studied outer membrane protein, and a recently identified HvtA protein receptor. In this study, we confirmed that the expression of the hvtA gene is iron‐regulated in a fur‐dependent manner. When analyzed for virulence in a hemin‐overloaded murine model system, the hupA gene was more important for establishing infection than the hvtA gene. Transcriptional profiling of these genes using strains of two different biotypes, biotype 1 (human pathogen) and biotype 2 (eel pathogen), showed that the expression of the two receptors was also regulated in response to temperature. The expression of hupA was highly induced in elevated temperatures in the human pathogenic strain when tested in iron‐depleted conditions. Conversely, hvtA expression was induced significantly in the eel pathogenic strain at a lower temperature, a condition where the hupA locus was relatively repressed. Our results indicate that although both hupA and hvtA are involved for optimal hemin uptake in V. vulnificus, their expression is dually regulated by the environmental cues of iron concentration and temperature. Together, these data suggest that the virulence genes hupA and hvtA are tightly regulated and strictly induced during iron limitation combined with the physiological temperature of the host organism.


| INTRODUC TI ON
Many Gram-negative bacterial pathogens require iron as an essential element for survival. Iron is known to play significant roles in signaling pathways and is used as a cofactor or prosthetic group for several proteins that are required for proper maintenance of cellular functions (Schaible & Kaufmann, 2004;Wandersman & Delepelaire, 2004). Although iron is abundant within the host environment, it exists predominantly in complex with high-affinity iron-binding proteins such as lactoferrin and transferrin, or is bound to hemin (Fe 3+heme) in red blood cells, making it essentially unavailable (Crosa, 1997;Koster et al., 1991;Wandersman & Delepelaire, 2004;Wilks & Burkhard, 2007). Therefore, pathogens have evolved to express several iron-acquisition systems that utilize low molecular weight high-affinity iron-chelating siderophores and outer membrane receptors to sequester and acquire the metal (Crosa, 1997;Koster et al., 1991).
Vibrio vulnificus is an opportunistic human and marine Gram-negative bacterial pathogen that may cause fever, diarrhea, and necrotizing wound infections with high mortality rates in humans and eels.
It is responsible for almost 95% of all seafood-related deaths around the world (Haq & Dayal, 2005;Morris, 1988;Strom & Paranjpye, 2000). Additionally, V. vulnificus infections can lead to fulminant septicemia in immunocompromised patients where the pathogen invades the bloodstream and causes septic shock (Gulig, Bourdage, & Starks, 2005;Koenig, Mueller, & Rose, 1991;Merkel, Alexander, Zufall, Oliver, & Huet-Hudson, 2001). The utilization of iron by this bacterium has thus been extensively studied to gain insight into its contribution to pathogenesis. It has been shown to express several iron-sequestering systems similar to other Gram-negative species when cultured under iron-limiting conditions (Litwin & Byrne, 1998;Simpson & Oliver, 1983).
In V. vulnificus, iron is acquired through the utilization of two high-affinity small molecular weight siderophores, the catecholic vulnibactin, and a hydroxamate-type molecule, both biosynthesized and secreted in response to iron deprivation that serves as iron scavengers from the extracellular surroundings (Litwin, Rayback, & Skinner, 1996;Okujo et al., 1994;Simpson & Oliver, 1983). Certain outer membrane receptor proteins are also co-expressed that specifically bind to the ferric-siderophore complexes and mediate internalization to the periplasmic space where the complex is finally transported across the inner membrane of the bacteria. Alternatively, the bacterium also expresses hemin uptake systems that utilize outer membrane protein receptors responsible for internalizing iron-bound hemin complex scavenged from the host (Datta & Crosa, 2012;Litwin & Byrne, 1998;Litwin & Quackenbush, 2001;Oh, Lee, Lee, & Choi, 2009). The transport and internalization of both the ferric-siderophore and hemin complexes are energy-dependent processes mediated by a complex of proteins known as the TonB energy-transduction system which is located in the inner membrane of the bacteria (Braun, 1995;Crosa, Mey, & Payne, 2004;Postle & Larsen, 2007;Wright, Simpson, & Oliver, 1981). Vibrio species and Vibrionaceae family members are known to possess two TonB systems (TonB1 and TonB2). Interestingly, V. vulnificus has been shown to contain an additional system, TonB3. (Alice, Naka, & Crosa, 2008;Kuehl & Crosa, 2010;Kustusch, Kuehl, & Crosa, 2011;Stork, Otto, & Crosa, 2007) It has also been demonstrated that the TonB2 (and TonB3) systems consist of a fourth protein TtpC apart from the classic ExbB2, ExbD2, and TonB2 (or ExbB3, ExbD3, and TonB3) proteins (Kuehl & Crosa, 2010;Stork et al., 2007). The expression of many of these iron-regulated virulence factors in V. vulnificus is dependent on the regulatory fur gene, whose product represses transcription of specific genes when an adequate concentration of iron is present (Litwin & Calderwood, 1993;Miyamoto et al., 2009).
Low concentration of iron in host tissues serves as an important signal to direct the expression of virulence factors through the release of the Fur protein from consensus DNA sequences, termed Fur boxes, located in the promoter region of these iron-regulated genes (Escolar, Perez-Martin, & Lorenzo, 1999).
HupA, an outer membrane protein (79.27 kDa), was reported to be a hemin receptor and the expression of the hupA gene that encodes the receptor protein was demonstrated to be transcriptionally regulated (a) in response temperature, (b) by the iron-binding regulatory protein Fur, and (c) a LysR homolog HupR (Litwin & Byrne, 1998;Litwin & Quackenbush, 2001;Oh et al., 2009). We reported the identification of an additional iron-regulated TonBdependent outer membrane hemin receptor HvtA (79.09 kDa), which was shown to facilitate optimum hemin utilization in V. vulnificus using the biotype 1 CMCP6 strain (Datta & Crosa, 2012).
In this report, we further study and compare the regulation of the hupA and hvtA genes in biotypes 1 (CMCP6) and 2 (ATCC33149) V. vulnificus strains and demonstrate that they are dually regulated in response to iron concentration and temperature at the transcriptional level. Further, we hypothesize that apart from the hupA gene, the hvtA gene may also play an essential role in the bacterium's pathogenicity with individual requirements specific to the environment of the host organism (37°C in humans and 25°C in eels) where each gene is upregulated.

| Bacterial strains, plasmids, and growth conditions
Strains and plasmids used in this study are listed in Table 1. Bacteria were routinely grown in trypticase soy broth supplemented with 1% NaCl (TSBS) or on trypticase soy agar supplemented with 1% NaCl (TSAS; V. vulnificus), or in LB broth (E. coli) with appropriate antibiotics: kanamycin (50 μg/ml) and chloramphenicol (30 μg/ ml) for E. coli unless otherwise mentioned. M9 minimal medium (Crosa, 1980) was used for iron-limiting conditions supplemented with 0.2% casamino acids and 5% NaCl with the iron chelator ethylenediamine-di-(o-hydroxyphenylacetic) acid (EDDA) at indicated concentrations. Ferric ammonium citrate (FAC) was added to the medium to obtain iron-rich growth conditions at indicated concentrations. Thiosulfate-citrate-bile-salts-sucrose agar (TCBS; Preiser Scientific, Louisville, KY) was used for selection of V. vulnificus in conjugation experiments. and hvtA (VV21549-1fwd and VV21549-2rev) genes from the ATCC 33149 biotype 2 V. vulnificus strain were designed using the analogous gene sequences in the CMCP6 strain (Table 2).

| Construction of V. vulnificus mutants
Deletion mutants in V. vulnificus CMCP6 strain were generated by allelic exchange using the pDM4 suicide plasmid (Milton, O'Toole, Horstedt, & Wolf-Watz, 1996). Upstream and downstream regions (700-800 bp) flanking the genes were amplified by specific primers, and combined using splicing by overlapping extension (SOE) PCR.
The generated PCR products were cloned into the blunt PCR2.1 vector (Invitrogen, Carlsbad, CA), digested with restriction enzymes, and subcloned into the suicide vector pDM4 also linearized with the same restriction enzymes. The resulting pDM4 derivatives were conjugated into V. vulnificus according to the procedure previously  Milton et al. (1996) pRK2013 Helper plasmid; Km r Figurski and Helinski (1979) TA B L E 1 Strains and plasmids used in this study reported using the helper plasmid pRK2013 (Alice et al., 2008;Figurski & Helinski, 1979).

| Hemin utilization assay
Bioassays were performed to determine whether hemin and hemo- and appearance of the growth halo was monitored after 18 hr.

| Virulence assays
Overnight cultures of V. vulnificus strains were inoculated into 25 ml of TSBS (inoculation ratio 1:100) and grown at 37°C to an OD 600 of ~0.5. The cells were then harvested and washed twice in phosphate-buffered saline (PBS). Next, cells were resuspended to an OD 600 of 1.0 and serially diluted in PBS. Five 4-to 6-week-old CD1 mice (Charles River Laboratories) per dilution were injected intraperitoneally (i.p.) with 100 μl of the strain of interest. Five serial dilutions for each strain were evaluated. Mortality was monitored for 48 hr post-infection, and 50% lethal dose (LD 50 ) calculations were determined by the Reed-Muench method (Reed & Muench, 1938).
For hemin-overloaded models, 100 μl of hemin (10 mg/ml) solution prepared in 10 mM NaOH was injected (i.p.) into the animals two hours before injecting the bacterial strains of interest. All manipu- (CMCP6: ΔhupA, ΔlacZ), respectively, to confirm whether the strains behaved similarly in the host. Briefly, bacterial strains were cultured as described above, and then, equal volumes were mixed and serial dilutions were performed. Intraperitoneal injections of three serial dilutions were put into three to six mice. Animals were checked approximately 9 hr onwards after inoculation, and when they showed signs of sickness (e.g., lethargy, slow movements, and lack of appetite), they were euthanized with CO 2 according to IACUC regulations. Since i.p. injections involve puncture of the skin, skin samples (one square centimeter around the inoculation site) along with spleens, and livers were aseptically extracted and homogenized by using a Seward stomacher laboratory blender in the presence of PBS (1 ml for skin and spleen and 2 ml for liver). Serial dilutions were performed in PBS, and the dilutions were plated on TSAS-5-bromo-4chloro-3-indolyl-β-D-galactopyranoside (TSAS-X-Gal; 0.004%, wt/ vol) plates to determine the CFU of the strains under analysis. CI values were obtained as described previously (Taylor, Miller, Furlong, & Mekalanos, 1987). Statistical analysis was performed using the Student t test with GraphPad Prism 4.0 software. City, CA) to quantify the expressions of hupA and hvtA genes with primers listed in Table 2 for biotype 1 (CMCP6) and biotype 2 (ATCC 33149) strains. The levels of mRNA expression of both genes were calculated in iron-rich and iron-limiting conditions by normalizing to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) expression.

| RNA extraction and transcript analysis
Relative fold changes in transcript levels were then calculated between the two growth conditions.

| TonB specificities of the V. vulnificus hemin receptors
Prior studies demonstrated that hemin and hemoglobin transport are tonB1-and tonB2-dependent, where tonB2 also requires ttpc2 in the V. vulnificus CMCP6 strain (Alice et al., 2008;Datta & Crosa, 2012). This indicates that the TonB systems may have redundant functions in transporting iron sources. To determine whether the hemin and hemoglobin receptors HupA and HvtA exhibited specificities for the TonB systems, a series of receptor mutants were generated in the same strain in combination with the individual TonB systems (VV21614 & VV20360). Hemin and hemoglobin uptake assays were then performed using the combinatorial mutant strains described in Table 3. It was previously demonstrated that both the TonB1 and TonB2 systems facilitate ferrioxamine uptake in V. vulnificus, whereas the transport of aerobactin was only facilitated by the TonB2 system (Kustusch, Kuehl, & Crosa, 2012).
Thus, ferrioxamine and aerobactin were used as controls for the iron utilization assay to ensure the functionality of each TonB system.
We have previously shown that only the HupA receptor is capable of transporting hemoglobin (Datta & Crosa, 2012

| Cloning the hemin receptor genes in V. vulnificus biotype 2 strain
The biotype 2 ATCC33149 strain of V. vulnificus has been shown to be extremely infectious in eels (Tison et al., 1982). Since this biotype can also utilize hemin as a source of iron, it was of interest to determine whether homologs of the hupA and hvtA genes existed in the V. vulnificus biotype 2 ATCC33149 strain as the genomic sequence of this strain is not yet published (Gulig et al., 2010;Tison et al., 1982). PCR was performed using primers designed from the genome sequence of CMCP6 biotype 1 strain (

| Transcription of hupA and hvtA is impacted by temperature
It was reported that the expression of the hupA gene in a human pathogenic V. vulnificus strain was influenced by elevated temperature (Oh et al., 2009). The effect of temperature on hvtA gene expression has not been studied to date. As V. vulnificus is an opportunistic pathogen for both humans and fishes, we sought to determine the expression profiles for both genes in two different strains. Biotype 1 (CMCP6) is a human pathogen, and biotype 2 (ATCC33149) is predominantly an eel pathogen Fouz, Larsen, & Amaro, 2006;Roig & Amaro, 2009). Gene expression profiles were measured and compared after growth at the physiological temperatures of the hosts, 37°C and 25°C, respectively, between iron-rich (M9 minimal media with 10 μg/ml FAC) and iron-limiting (M9 minimal media with 5 μM EDDA) conditions. qRT-PCR analysis revealed the hupA and hvtA genes were upregulated in iron-limiting conditions at both temperatures in both biotypes.
It was also interesting to note that the hupA gene was more upregulated (~375-and 200-fold) at the higher temperature of 37°C (in both biotype 1 and biotype 2 strains, respectively) than the hvtA gene (Figure 4a,b). This observation was consistent at the translational level-while the HupA protein was detected in outer membrane fractions when the biotype 1 CMCP6 strain was cultured at 37°C, HvtA was only detected when overproduced in trans (Datta & Crosa, 2012). Contrastingly, an increase in the expression of hvtA mRNA was observed at the lower temperature (25°C) in both biotypes 1 and 2 by ~400-and ~1,000-fold, respectively. Upregulation of the hvtA gene was significantly higher at 25°C in the biotype 2 than in the biotype 1 strain (~2.5-fold), emphasizing its importance at lower temperatures.

| The hupA and hvtA genes have variable contributions in V. vulnificus mouse virulence
The requirements for the hupA and hvtA genes in biotype 1 V. vulnificus virulence were assessed by comparing the wild-type CMCP6 bacterial strain with the hemin receptor mutants in a murine model system using 4-to 6-week-old CD-1 normal or hemin-overloaded mice. The iron-overloaded mouse model was used because it more closely reproduces the conditions found in humans during a V. vulnificus infection (Starks et al., 2000;Wright et al., 1981). No difference in 50% lethal dose (LD 50 ) was observed between the hemin receptor mutants and the wild-type strain in normal, non-iron-overloaded mice (Table 4). The LD 50 increased 10,000-fold for both the ΔhupA and ΔhupAΔhvtA strains in hemin-overloaded mice relative to the wild-type strain. However, the hvtA mutant strain exhibited only a marginal attenuation (Table 4). These data indicate that the HupA receptor in the biotype 1 CMCP6 strain is the dominant receptor needed for a fully virulent phenotype in a hemin-overloaded mouse model (physiological temperature being 37°C). The biotype 2 strain ATCC 33149 (isolated from diseased Japanese eels and most abundantly found in coastal seawater at lower temperatures) was also tested in both normal mice and hemin-overloaded mice and was found to be avirulent at its highest concentration of 10 8 cells/ ml (data not shown). This is in stark contract to the data presented in Table 4 using the CMCP6 biotype 1 strain. These data suggest that the growth temperature of the bacteria in mice was more important for inducing the hupA gene than the hvtA gene in the human pathogenic strain (Table 4). In addition to lethality studies, Competitive Index were also generated to extract subtle differences between the virulence of the wild-type (ALE-LAC) and hvtA mutant (VSSD58) strains. The ability of the two strains to compete for growth in the iron-overloaded CD-1 mouse was determined by counting the number of viable bacteria isolated from the skin, spleen, and liver of the host. The CI of ΔhvtA to wild-type did not change dramatically from 1.0, indicating that the hvtA mutant strain did not have a growth defect when compared to the wild-type strain. This was likely because the activity from the product of the hupA gene had masked any observable phenotype at this temperature (Figure 5a). To eliminate the effect of the contribution of the hupA gene in iron uptake, another CI analysis was performed to compare the ΔhupAΔlacZ with the ΔhvtAΔhupA strain ( Figure 5b). In this comparison, only marginal differences were seen in the spleen and liver. These were not significantly different than a CI score of 1. These results indicate that the ΔhvtAΔhupA strain did not have a growth defect compared with the ΔhupAΔlacZ strain in these two organs. A small yet significant difference was seen in the skin sample when the ΔhvtAΔhupA strain was compared against the ΔhupAΔlacZ strain and did significantly vary from a CI score of 1 ( Figure 5b). These data indicate that a significant growth defect occurred in the ΔhvtAΔhupA strain compared with the ΔhupAΔlacZ strain in the skin sample. This indicates a role for hvtA in V. vulnificus infection when the bacterium is found at the skin of the murine model.

| D ISCUSS I ON
Vibrio vulnificus is known to cause severe sepsis in patients who carry abnormally high concentrations of iron in their bloodstream.
This organism has developed several iron-acquisition systems allowing both siderophores and hemin uptake. These systems support its evolution as a facultative human and fish pathogen where infections often result in fatality Gulig et al., 2005;Hor, Chang, Chang, Lei, & Ou, 2000;Tison et al., 1982). Commonly, iron complexes are internalized by bacteria by first being bound by outer membrane receptor proteins. The energy supplied for internalization is produced in the inner membrane through the action of proton motive force (PMF). This energy is then harnessed by the TonB systems and transferred to the respective outer membrane receptor to facilitate the internalization of the ferric complexes (Braun, 1995;Crosa et al., 2004;Postle & Larsen, 2007). In this report, we studied and characterized additional outer membrane proteins in V. vulnificus that are necessary for optimal uptake of hemin. This bacterium possesses three distinct TonB systems (TonB1, TonB2, and TonB3) where the TonB3 system was observed to be induced only when bacteria were cultured in human serum (Alice et al., 2008). Here, we demonstrate that HupA and HvtA receptors are dependent on either TonB1 or TonB2, consistent with other iron transport proteins in this organism. The existence of multiple TonB-dependent heminacquisition systems in human pathogens is not uncommon and has been identified in another closely related Gram-negative bacterium V. cholerae (Mey & Payne, 2001). The apparent redundancy in roles appears to emphasize their importance in establishing virulence inside the host where chances of encountering hemin as a source of iron are high (Helms et al., 1984).
Our analysis of the HupA and HvtA proteins identified and sequenced in the eel pathogen V. vulnificus biotype 2 strain (ATCC33149) demonstrates significant homology to the respective proteins in the biotype 1 CMCP6 strain. The HupA protein contains the conserved FRAP motif that has been implicated strongly with receptor function in other iron-sequestering proteins, particularly in hemin uptake systems. In addition, it includes the partial NPNL sequence downstream of the FRAP motif, including a conserved histidine residue between the two motifs that were shown to participate in hemin binding (Bracken et al., 1999). We found that the HvtA protein in the biotype 2 strain contains motifs that are 75% and 50% identical to the FRAP and NPNL boxes, respectively. This observation was consistent with the notion that the NPNL motif is not well conserved in hemin receptors from other Vibrio species (Mey & Payne, 2001).
While the hupA gene in the CMCP6 strain was shown to be monocistronic (Litwin & Byrne, 1998), we found that the hvtA gene is gene. The results also indicated that VV21546 was not part of the operon, whereas VV21550-2 was co-transcribed with the hvtA gene.
This operon exhibits significant homology to the hutR containing operon in V. cholerae, with the functions of the products of genes VV21550-2 remaining unknown.
In this study, we confirmed experimentally that the expression of the hvtA operon in the CMCP6 biotype 1 strain was dependent on iron concentration with Fur regulation as hypothesized by Gulig et al. (2010). A potential binding site for the Fur protein (5′-gtaaatgataattgatgt-3′) was identified in the putative promoter region of the hvtA operon that shares 63% identity with the consensus Fur binding sequence (5′-gataatgataatcattatc-3′; Lavrrar & McIntosh, 2003). The fur mutant strain could not modulate expression levels of examined genes in iron-rich versus iron-limiting conditions. This is consistent with the regulatory patterns of iron-acquisition genes in different Gram-negative bacteria including V. vulnificus which are regulated by Fur in response to this element (Escolar et al., 1999;Pajuelo et al., 2014).
In addition to being regulated in response to iron, transcriptional analysis determined that both the hupA and hvtA genes exhibit temperature-dependent expression in biotype 1 and biotype 2 strains.
This type of temperature-dependent transcriptional regulation was previously reported for the virF gene in Yersiniae, whose product is a transcriptional regulator of the AraC family and plays a vital role in controlling the expression of virulence factors (Falconi, Colonna, Prosseda, Micheli, & Gualerzi, 1998;Konkel & Tilly, 2000;Wattiau & Cornelis, 1994). In addition, recent reports have identified RNA-mediated thermoregulation of iron-acquisition genes in Shigella (Kouse, Righetti, Kortmann, Narberhaus, & Murphy, 2013;Wei, Kouse, & Murphy, 2017). This is particularly significant because V. vulnificus infecting humans (biotype 1) or eels (biotype 2) must regulate gene  (Pajuelo et al., 2014). It is interesting to note that the CECT4999 strain used in the study is a Spanish isolate that is pathogenic for both mice and eels. The ATCC33149 biotype 2 strain however has not been found to exhibit any sign of mice pathogenicity in our experiments or in earlier studies (Biosca, Oliver, & Amaro, 1996). It would be interesting to study the virulence of the hemin receptor mutants of this strain in eels in order to determine their importance. Attempts to generate individual hemin receptor mutants in the ATCC33149 biotype 2 strain have been unsuccessful thus far.
Vibrio vulnificus causes severe septicemia and it is reasonable to postulate that hemin acquisition serves a major role in the pathogenesis of the human host, where the majority of available iron in serum is sequestered by hemin and hemoglobin in red blood cells.
However, it appears that the organism triggers the expression of specific hemin receptors at different temperatures. The significance of the hupA gene as a virulence factor in the murine model at the physiological temperature of 37°C is consistent with the observation that its expression was highly induced at elevated temperatures in iron-limiting conditions. Alternately, it is possible that the hvtA gene has evolved to play a more important role in certain V. vulnificus strains whose primary host is eels, where the growth temperature is closer to 25°C in coastal seawaters. Indeed, hvtA expression was highly upregulated in the biotype 2 strain at 25°C, and was not in-  Mei et al., 2018;Reitman, 2018). Again, this lower temperature found at a potential wound site could help in early-stage infections in acquiring hemin from a host, suggesting a role for hvtA in infections at lower temperatures.
The differential temperature response of these hemin receptor-encoding genes confirms the role of a yet unknown regulatory switch that controls the expression of these genes in response to iron and temperature. Interestingly, Pajuelo et al. reported that a biotype 2 serovar CECT4999 strain showed equal levels of virulence when tested in mice and eels where hemin receptor hupA and vulnibactin receptor vuuA served as the major virulence factors (Pajuelo et al., 2014). The hvtA (named hutR in strain CECT4999) single mutant behaved similarly to the wild-type in either hosts suggesting that hupA and vuuA are host-nonspecific virulence genes, but hvtA is not (Pajuelo et al., 2014).
Bacterial pathogenesis is a complex energy-intensive process involving multiple factors that not only contribute to disease progression but also promote successful replication inside the host organism. Pathogens have evolved to distinguish between environments and ensure that a large number of virulence genes are expressed only when the bacterium is invading host tissues.
Changes in temperature, nutrients, osmolarity, iron, and other ion concentrations serve as cues that promote preferential expression of virulence genes necessary for survival. Our studies indicate that a dual regulation of the hemin receptors exists that responds to both iron concentration and temperature and that this regulation is critical to the pathogenesis of V. vulnificus. Where one of the receptors encoded by the hupA gene is heavily expressed at an elevated temperature, the other, produced from the hvtA gene, is likely more biologically relevant at a lower temperature as observed from its expression profile. These data suggest that the receptors have unique regulatory features associated with them.
Future studies will incorporate the identification and characterization of the regulatory switch in order to unravel the mechanism of gene regulation in this bacterium. Understanding the molecular basis for this dual regulation will provide a platform for understanding virulence gene regulation in other pathogens that alternate between hosts with variable ambient temperatures.

ACK N OWLED G M ENTS
We

CO N FLI C T O F I NTE R E S T S
The authors have no conflicts of interest to disclose.

AUTH O R CO NTR I B UTI O N S
S.D and R.J.K conceptualized the data, curated the data, involved in formal analysis, investigated the data, contributed to methodology, administered the project, involved in resource management, provided software, supervised the data, validated the data, visualized the manuscript, wrote the original draft, and reviewed and edited the manuscript. R.J.K acquired the funding.

E TH I C S S TATEM ENT
All manipulations of mice were approved by the Institutional Animal Care & Use Committee (IACUC) at the Oregon Health Science University, protocol #A802.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data are provided in full in the results' section of this paper apart from the full protein sequences used to create alignments. These sequences are available at the National Center for Biotechnology