Expression of a Yersinia pseudotuberculosis Type VI Secretion System Is Responsive to Envelope Stresses through the OmpR Transcriptional Activator

The Type VI secretion system (T6SS) is a macromolecular complex widespread in Gram-negative bacteria. Although several T6SS are required for virulence towards host models, most are necessary to eliminate competitor bacteria. Other functions, such as resistance to amoeba predation, biofilm formation or adaptation to environmental conditions have also been reported. This multitude of functions is reflected by the large repertoire of regulatory mechanisms shown to control T6SS expression, production or activation. Here, we demonstrate that one T6SS gene cluster encoded within the Yersinia pseudotuberculosis genome, T6SS-4, is regulated by OmpR, the response regulator of the two-component system EnvZ-OmpR. We first identified OmpR in a transposon mutagenesis screen. OmpR does not control the expression of the four other Y. pseudotuberculosis T6SS gene clusters and of an isolated vgrG gene, and responds to osmotic stresses to bind to and activate the T6SS-4 promoter. Finally, we show that T6SS-4 promotes Y. pseudotuberculosis survival in high osmolarity conditions and resistance to deoxycholate.


Introduction
The Type VI secretion system (T6SS) is a macromolecular machine dedicated to the secretion of toxin proteins, widespread in Gram-negative Proteobacteria [1][2][3][4][5]. This system is highly versatile as it can target eukaryotic or prokaryotic cells [6]. The Aeromonas hydrophila and Vibrio cholerae T6SS have been shown to be required for full virulence towards eukaryotic host cells via the transport of protein domains responsible for actin modification [7][8][9][10][11]. Several other T6SS, including those of Pseudomonas aeruginosa, Burkholderia thailendensis, Serratia marcescens V. cholerae, Citrobacter rodentium and enteroaggregative Escherichia coli, are required to eliminate competing bacteria in mixed environments [12][13][14][15][16][17]. The toxins secreted by the P. aeruginosa and S. marcescens T6SS have been recently identified and characterized: the Pseudomonas Tse1, Tse3 and Serratia Rab proteins are translocated into the periplasm of the prey cells where they create lesions in the peptidoglycan layer [18][19][20]. Additional T6SS functions have been reported such as role in stress sensing, biofilm formation or adaptation to environmental conditions, although the mechanistic bases for these functions have not been clearly defined [21][22][23][24].
At a molecular level, the T6SS is assembled via interactions between 13 different components, called Tss proteins [25][26].
Several of these proteins share structural homologies with components of the tail of contractile bacteriophages, including the major tail tube protein (Hcp), the cell puncturing device (VgrG), the sheath (TssB-C) and at least one component of the baseplate (TssE) of bacteriophage T4 [27][28][29][30][31][32]. Mechanistically, it has been proposed that this bacteriophage-like complex assemble a tubular structure similar to the bacteriophage tail tube, wrapped by a contractile sheath-like structure. As evidenced for the bacteriophage, contraction of the T6SS sheath will propel the inner tube to the cell exterior. Indeed, recent cryo-and fluorescence microscopy data demonstrated that the TssB-C proteins form dynamic cytoplasmic structures, oscillating between extended and contracted conformations [33] and that T6SS sheath contraction is correlated with prey cell lysis [17]. Several other T6SS proteins are embedded into the inner or outer membranes where they assemble a trans-envelope spanning complex proposed to anchor the bacteriophage-like structure to the envelope [34][35][36][37][38][39].
Although the architecture of the T6SS seems to be conserved, it is however clear that T6SS have been rerouted to be dedicated to functions important to the need of each individual species. This is also clearly reflected by the regulatory mechanisms identified so far [40][41][42]. T6SS activity is regulated at different levels. T6SS gene transcription is dependent on transcriptional activators, twocomponent systems, histone-like proteins, alternative sigma factors or regulatory RNAs whereas post-translational mechanisms based on the phosphorylation status of a fork-head associated protein modulate T6SS activity [40][41][42][43][44][45][46]. One interesting observation is that T6SS gene clusters can be found in several copies in bacterial genomes. While many species encode a single locus, 2, 3 or more (up to 6) complete T6SS gene clusters can be scattered on the genome [1,5]. This is the case of the Yersinia pseudotuberculosis genome that encodes four complete and two incomplete machineries [5,47]. This bacterium constitutes an interesting model to identify the regulatory mechanisms controlling the expression of these T6SS, potential regulatory cross-talks, the specific function of each T6SS and specificity determinants during assembly.
To gain insights into the regulatory mechanisms underlying expression of the Y. pseudotuberculosis T6SS gene clusters, we constructed transcriptional lacZ and gfp reporter fusions to the T6SS-4 gene promoter, a T6SS locus shared by Y. pseudotuberculosis and Y. pestis. Although the Y. pestis CO92 T6SS-4 gene cluster has been shown to promote phagocytosis and to limit intracellular replication in macrophages, it has no role during rat flea infection or in virulence in murine bubonic plague models [48]. As previously reported, the T6SS-4 promoter was more active at low temperature (28uC vs 37uC) [47][48][49]. By using transposon mutagenesis we identified OmpR as a potential regulator. OmpR binds on the T6SS-4 promoter region in vitro and does not regulate the other T6SS loci or an isolated vgrG gene. OmpR is the response regulator of the EnvZ-OmpR two component systems that is responsive to cell envelope and osmotic stresses [50][51][52]. Interestingly, LptD was also identified in the same screen. LptD is a component of the Lpt machine required for the proper insertion of lipopolysaccharides in the outer leaflet of the outer membrane [53][54][55]. Epistasis experiments showed that LptD acts upstream OmpR suggesting that T6SS-4 expression might be responsive to cell envelope stresses. Indeed, we show that the transcriptional fusion is activated in high osmolarity conditions and in presence of the bile salt deoxycholate, and that T6SS-4 is required for survival after osmotic stress and for increased resistance to deoxycholate.

Yersinia Pseudotuberculosis T6SS-4 Expression is Thermoregulated
The T6SS-4 gene cluster is found in both Y. pestis and Y. pseudotuberculosis its expression has been shown to be thermodependent in both species [47][48][49]. To gain further insights into the regulatory mechanism underlying T6SS-4 expression, we constructed a lacZ fusion to the T6SS-4 promoter at the original locus in strain Y. pseudotuberculosis IP31758. To avoid false positive clones during the transposon mutagenesis screen (e.g., transposon insertion into the lacZ reporter gene), we introduced a second promoter-gfp fusion at the ara locus, yielding the IP31758-41 strain. As previously reported in both Y. pestis and Y. pseudotuberculosis, b-galactosidase activities and GFP fluorescence confirmed that the expression of the T6SS-4 locus is activated at 28uC compared to 37uC (see Fig. 4A).

Transposon Mutagenesis Identifies OmpR and LptD as Candidates for T6SS-4 Regulation
To identify regulators, we screened mini-Tn5 transposon mutants for decreased T6SS-4 expression at 28uC. Tn5 transposons insert randomly in the genomes of Yersinia species [56]. Approximatively 250,000 transposon mutant strains were screened and 128 clones with a white lacZ 2 phenotype were isolated on McConkey plates. These 128 transposon mutants were tested for GFP fluorescence levels to eliminate potential insertion into the lacZ reporter gene. 31 candidates were retained with both decreased lacZ activity and fluorescence levels. The site of mini-Tn5 transposon insertion was determined for 3 of these candidates that displayed stronger decreases. One insertion occurred at close proximity to the T6SS-4 locus (YpsIP31758_3437 to Yp-sIP31758_3420), into the lptD gene (YpsIP31758_3441; gene accession YP_001402396.1). Two independent transposon insertions mapped into the ompR locus (YpsIP31758_3980; gene accession YP_001402928.1) ( Figure 1A).

OmpR Activates Transcription of the Y. Pseudotuberculosis T6SS-4 Locus
OmpR is the response regulator of the two-component EnvZ-OmpR system [57]. To confirm that OmpR is involved in the regulation of the T6SS-4 locus, we cloned the Y. pseudotuberculosis ompR gene under the control of the P BAD promoter into pBAD24 [58]. b-galactosidase activities and fluorescence levels of the transposon mutants carrying the empty pBAD24 or the pBAD24-OmpR vectors were measured. Figure 1B shows that the lptD mutant displayed a 6-fold decrease of b-galactosidase activity and a 7-fold decrease of fluorescence compared to the wild-type strain whereas the two ompR mutant strains (ompR1 and ompR2) displayed a 4-fold decrease of b-galactosidase activity and of fluorescence. The levels of b-galactosidase activity and of fluorescence were complemented by the overproduction of OmpR into the two ompR mutant strains, demonstrating that the OmpR protein produced from pBAD24 was fully functional.
To determine whether LptD acts upstream or downstream OmpR in the regulatory cascade we performed epistasis experiments. In the lptD strain, overproduction of OmpR led to an increased fluorescence and b-galactosidase activity compared to the lptD strain carrying the empty vector, at levels comparable to the WT strain ( Figure 1B). These data suggest that OmpR is acting downstream LptD.
We further tested whether OmpR regulates the expression of the four other Y. pseudotuberculosis T6SS loci (T6SS-1, Yp-sIP31758_0312 to _0339; T6SS-2, YpsIP31758_2511 to _2485; T6SS-3, YpsIP31758_1354 to _1379 and T6SS-5, Yp-sIP31758_0777 to _0805) and of an isolated vgrG gene (Yp-sIP31758_0696). Here again, strains with transcriptional lacZ reporter at the locus and promoter-gfp reporter at the araBAD locus were constructed. The T6SS-3 gene cluster being composed of two divergent operons (YpsIP31758_1361 to _1354 [T6SS-3-rev] and Yps31758_1362 to _1379 [T6SS-3-fwd]), fusions to lacZ and gfp were introduced in both orientations. Figure 2 shows that the bgalactosidase activity and fluorescence levels of these promoters were very low, suggesting these promoters are weakly expressed in rich medium and low temperature (28uC). Aside T6SS-4, overproduction of the OmpR response regulator in these strains did not affect significantly the activity of the promoters.

Purified OmpR Binds to the T6SS-4 Promoter Region
In silico analysis of the T6SS-4 promoter region with Virtual Footprint suggests the existence of OmpR binding motifs ( Figure 3A). To test whether the control by the OmpR protein was direct, we performed electrophoretic mobility shift assays. A recombinant N-terminally 66His-TRX-tagged variant of the Y. pseudotuberculosis OmpR protein was purified to homogeneity by metal-affinity chromatography. The native OmpR Yps protein was obtained by tag proteolysis and gel filtration (see Material and Methods). OmpR being a transcriptional activator associated with a two-component system, its affinity for targets is dependent on its phosphorylation status [59]. Gel shift assays were therefore monitored in presence of acetyl-phosphate to promote OmpR auto-phosphorylation. The purified OmpR Yps protein bound to the promoter region of the ompF gene ( Figure 3B, upper panel, lanes 1-6), a direct target of the OmpR response regulator [60]. This protein-DNA interaction was specific as no OmpR Yps binding was observed on the enteroaggregative E. coli sci-1 promoter ( Figure 3B, upper panel, lane 10) whose expression is regulated by the ferric uptake regulator [61] and independent on OmpR (the sci1-lacZ fusion shares comparable b-galactosidase activities in E. coli WT and DompR strains; data not shown). A shift was observed when the T6SS-4 promoter was used ( Figure 3B, lower panel, lanes 1-6). No shift was observed with the purified Fur protein in both ompF and T6SS-4 promoters ( Figure 3B, lanes 9). The P T6SS-4 -OmpR Yps shift was abolished when a competitor unlabelled DNA corresponding to the ompF promoter was used ( Figure 3B, lower panel, lanes 7-8). Conversely, addition of the unlabelled T6SS-4 promoter decreased OmpR Yps binding on the ompF promoter ( Figure 3B, upper panel, lanes 7-8). Taken together these results demonstrate that OmpR specifically binds to the T6SS-4 promoter region. However, it is worthy to note that the affinity for the purified OmpR protein is higher for the ompF

T6SS-4 Thermoregulation is Independent on OmpR
Recently, Brzostek et al. reported that the temperature-dependent regulation of the Y. enterocolitica invasine inv gene was controlled by OmpR [62]. To test the role of OmpR in the T6SS-4 thermoregulation we compared the effect of the ompR mutations at 28uC and 37uC. Figure 4 shows that similar decreases in T6SS-4 gene cluster expression were observed in the WT ( Figure 4A) and ompR1 mutant ( Figure 4B) strains at these two temperatures (8-fold decrease at 37uC compared to 28uC). These results show that OmpR has no role in the temperature-dependent expression of T6SS-4 and further suggest that T6SS-4 thermoregulation relies on additional regulatory mechanisms.

T6SS-4 is Responsive to Cell Envelope and Osmotic Stresses in an OmpR-dependent Manner
The EnvZ-OmpR two-component system regulates target genes in response to various stresses including cell envelope damages and high osmolarity [50,52,57]. Interestingly, LptD is an outer membrane protein required for LPS transport to the outer membrane [53,55] suggesting that the lptD mutation induces a cell envelope stress. The observation that OmpR acts downstream LptD in the regulatory cascade suggests that the effect of LptD on T6SS-4 expression results from activation of OmpR engendered by a cell envelope stress. We therefore tested whether the expression of the T6SS-4 is modulated by cell envelope stresses engendered by exposure to high osmolarity or to sodium deoxycholate (DOC), a bile salt for 60 min. Figure 5 shows that the b-galactosidase activity and the GFP fluorescence levels of the wild-type strain increased 2.6-and 2-fold respectively in presence of 0.6 M sucrose (,20%) and ,2-fold in presence of 1% DOC. These increased activities of the promoter fusions in presence of osmotic or cell envelope stresses were dependent on OmpR as high osmolarity and bile salts had no effect on T6SS-4 expression in the ompR1 transposon strain (identical results were obtained for the ompR2 transposon strain).

T6SS-4 is Required for Survival after Cell Exposure to an Osmotic Stress and for Resistance to Deoxycholate
The T6SS-4 being activated in high osmolarity conditions or in presence of bile salts we asked whether the Y. pseudotuberculosis T6SSs, and in particular T6SS-4, are required for survival after exposure to various stresses. Interestingly, several T6SSs have been shown implicated in the bacterial adaptation to environmental conditions: the Vibrio anguillarum T6SS regulates the stress response [22], the V. cholerae O1 T6SS is activated at high osmolarity [23] whereas the T6SS inhibits Campylobacter growth at high bile salt concentrations [24].
To test survival upon osmotic stress exposure and resistance to a cell envelope stress, the wild-type Y. pseudotuberculosis strain and its derivatives in which the tssF gene, encoding an essential component of the secretion apparatus [25,26] [YpsIP31758_0800]) were exposed to 20% of sucrose for one hour prior to counting viable cells or spotted onto LB agar plates supplemented with 1% DOC. Figure 6 shows that the survival to high osmolarity exposure (upper graph) and resistance to DOC (lower graph) were affected by the tssF4, lptD and ompR mutations while tssF1, tssF2, tssF3 and tssF5 cells displayed survival behaviours similar to the WT strain. It is worthy to note that the lptD and ompR mutant cells were more severely affected than tssF4 cells. In trans-complementation experiments with pBAD24-OmpR, the survival of ompR and lptD mutant cells was restored while tssF4 mutant cells remain sensitive to high osmolarity exposure and DOC. The tssF4 phenotype was restored by production of a 66His epitope-tagged TssF4 protein.

Discussion
In many species T6SS gene clusters are found in several copies in the genome. Defining how these systems are regulated and if and how they cross-talk with each other is therefore important to better evaluate the contribution of the T6SS in the pathogenesis or fitness in different environments. Pathogenic Yersinia species usually contain 4 to 6 T6SS copies on the genome. In Y. pseudotuberculosis four complete and two incomplete T6SS can be identified [5,47]. To gain insights into the regulatory mechanisms  underlying expression of these systems, we have constructed reporter lacZ and gfp fusions to each of the putative promoter regions of the T6SS gene clusters and have used random mutagenesis to identify putative regulators. In this study, we report data regarding the control of T6SS-4 gene cluster expression. In agreement with previous studies [47][48][49], we showed that expression of the T6SS-4 gene cluster is enhanced at low temperature: a .10-fold increase of the reporter fusions is observed at 28uC compared to 37uC. Using random transposon mutagenesis we identified LptD and OmpR as T6SS-4 regulators. OmpR is the response regulator of the EnvZ-OmpR twocomponent system that responds to a variety of signals including osmotic and cell envelope stresses [50,52]. Once phosphorylated by its cognate sensor kinase EnvZ, OmpR activates transcription of target genes. Indeed, T6SS-4 expression is induced in high osmolarity conditions or in presence of a bile salt, sodium deoxycholate, and in vitro gel shift assays demonstrated that purified phosphorylated OmpR binds to the T6SS-4 promoter. LptD (or Imp or OstA) is an outer membrane b-barrel protein involved in LPS transport and insertion into the outer leaflet of the outer membrane [55]. One hypothesis is that cell envelope defects conferred by the lptD mutation will activate the EnvZ-OmpR twocomponent system. This hypothesis is supported by the observation that overproduction of OmpR into lptD transposon mutant cells rescues T6SS-4 expression, suggesting that OmpR acts downstream LptD in the activation cascade. Our further experiments demonstrated that T6SS-4 is involved in resistance to high osmotic stresses and to bile salts.
While preparing this manuscript, a study reporting T6SS-4 regulation by OmpR was published by Zhang and collaborators [63]. While the authors of this study went in more details in several aspects, our results match with their data: (i) OmpR positively regulates T6SS-4 expression but does not affect expression of the other T6SS gene clusters and (ii) OmpR directly binds to the T6SS-4 promoter region. By using targeted mutagenesis, they conclusively show that the three OmpR binding sequences they identified in the T6SS-4 promoter region are all recognized as only mutations affecting the three boxes abolished OmpR binding and the OmpR-dependent regulation. However, they noted that the OmpR-mediated regulatory mechanism of the Y. pseudotuberculosis T6SS-4 gene cluster is probably more complex than expected. Among the three OmpR binding sites, two (O1 and O3) acts as activating operators while OmpR binding at O2 represses T6SS-4 transcription [63]. In this study, the authors further demonstrated that T6SS-4 is required for acid tolerance by regulating the intracellular pH in Y. pseudotuberculosis [63]. Taken together, the results from both studies demonstrate an OmpRdependent regulation of the Y. pseudotuberculosis T6SS-4 gene cluster and point a role of this system for tolerance to different stresses (cell envelope, acid and osmotic). A role of the T6SS in stress resistance via sensing and activation of the stress response pathway has been demonstrated in V. anguillarum [22]. In V. cholerae O1, the activation of the T6SS has been shown to be dependent on the osmolarity of the milieu [23]. More recently, the Campylobacter T6SS has been shown to be involved to limit the growth when bile salts reach high concentrations [24]. As the role of a secretion system in stress resistance or sensing is not intuitive, further studies should be performed to understand the mechanistic bases for these phenotypes. Our survival experiments under exposure to osmotic stress showed that ompR transposon mutant cells were more severely affected than the tssF4 cells. This suggests that OmpR activates additional genes involved in tolerance to high osmolarity. Osmoregulation of T6SS genes has been demonstrated in V. cholerae. The vas gene cluster of V. cholerae O1 is repressed at low osmolarity through the osmoregulatory repressor OscR [23]. Interestingly, the V. cholerae O1 vas gene cluster is activated at high osmolarity and low temperature [23]. As proposed by Ishikawa and coauthors, these results suggest that the Y. pseudotuberculosis T6SS-4 system might be activated in the environment rather than in the animal host. The observation that T6SS-4 is required for survival in high osmolarity conditions is probably beneficial to Y. pseudotuberculosis in the environment. Although the regulatory mechanisms of the V. cholerae O1 vas and Y. pseudotuberculosis T6SS-4 gene clusters appear similar, several differences can be noted. Notably, OscR represses the vas gene cluster at 37uC in V. cholerae. This is not the case in Y. pseudotuberculosis, in which OmpR has a positive effect on T6SS-4 expression at both 28uC and 37uC. Other regulators are probably involved in the control of T6SS-4 expression. The observation that OmpR has no role in the thermoregulation suggests that at least one additional regulator modulates T6SS-4 expression dependent on the temperature. T6SS-4 gene expression has also been proposed to be under the control of the RovA protein and of quorum sensing [47,64]. One additional candidate is the histonelike H-NS protein. The T6SS-4 gene cluster has a deviating GC content (66% for the gene cluster, 69% for the promoter region compared to the 52% of the genome) that usually characterizes horizontally acquired DNA fragment which are preferred silenced targets of the H-NS protein. Interestingly, the OmpR protein has been shown to antagonize H-NS-mediated silencing of its own gene cluster in Salmonella enterica Typhimurium [65] and a number of studies have pointed regulatory cross-talks between H-NS and OmpR in Gram-negative bacteria [62,66]. However, only three transposon mutant clones with lower expression were sequenced in this study and we can hypothesize that several other putative regulators might be identified by sequencing additional clones. Here, the transposon library was tested for decreased expression at 28uC. A screen to identify clones with higher expression at 37uC will also provide further insights into the control of T6SS-4 expression.

Bacterial Strains, Media and Chemicals
The Y. pseudotuberculosis IP31758 strain [67] was kindly provided by Pr. Elisabeth Carniel (Pasteur Institute, Paris, France). A spontaneous mutant of Y. pseudotuberculosis IP31758 resistant to nalidixic acid was obtained after growth in LB medium supplemented with Nal (5 mg/ml) for 16 hours at 28uC before spreading the liquid culture onto LB agar plates supplemented with Nal (20 mg/ml). This Nal R strain (called RL21758-4) has been used throughout the study. Escherichia coli DH5a (New England Biolabs) was used for cloning procedures, S17-1 lpir or MFDpir [68] for mating experiments with Y. pseudotuberculosis and T7 Iq pLysS (New England Biolabs) for T7 promoter-driven protein production. Strains are listed in Table S1. Unless indicated, cells were grown in Luria broth (LB) at 28uC or 37uC. LB broth and MacConkey agar base were purchased from Difco. Plasmids were maintained by addition of ampicillin (100 mg/mL), chloramphenicol (40 mg/mL) or kanamycin (50 mg/mL). Gene expression was induced by 0.5 mM isopropylb-thio-galactoside (IPTG) or 0.1% arabinose. IPTG, arabinose, sucrose, DOC, X-Gal and ONPG were purchased from Sigma-Aldrich.

Plasmid Construction
Plasmids and oligonucleotides are listed in Tables S1 and S2 respectively.
Suicide plasmids for lacZ fusions. Putative promoter regions were amplified from Y. pseudotuberculosis IP31758 chromo-somal DNA with oligonucleotide pairs introducing restriction sites and PCR products were ligated into the pFUSE vector [69].
Suicide plasmids for gfp fusions. To construct chromosomal promoter-gfp fusions, a vector allowing insertion at the ara locus was designed as previously done for insertion of single-copy lacZ fusions in Y. enterocolitica [70]. Two 1.5-kb fragments ('araHG' and 'araBA') were amplified by PCR from Y. pseudotuberculosis IP31758 chromosomal DNA and ligated into the SacI and SpeI sites of pSR47S [71] with a NotI site between them to yield pRL40. Next, promoterless gfpmut2 was amplified from pUA66 [72] with oligonucleotides introducing NotI, NheI and BglII restriction sites upstream gfpmut2, and NotI downstream, and inserted into the NotI site of pRL40 to yield pRL44, in which the start codon of gfpmut2 is immediately downstream the 'araGH' fragment. The putative promoter PCR fragments were then cloned into the NheI and BglII sites of pRL44. In these constructs, the promoters are fused to gfpmut2 and can be inserted into the Y. pseudotuberculosis genome at the ara locus.
Suicide plasmids for construction of DtssF mutant strains. The Y. pseudotuberculosis DtssF in-frame deletion mutants were made using the sacB + allelic exchange plasmid pRE112 [73]. Two 0.5-kb fragments corresponding to immediately upstream and downstream regions of each tssF genes were PCR-amplified and fused by overlapping PCR, and cloned into the SacI and KpnI sites of pRE112.
For OmpR overproduction, the sequence encoding the ompR gene of Y. pseudotuberculosis was PCR amplified and cloned into pBAD24 [58] by RF cloning [74]. For TssF4 production, the sequence encoding the tssF4 gene of Y. pseudotuberculosis fused to a 66His coding sequence was PCR amplified and cloned between the KpnI and PstI sites of pBAD18-Kan [58]. In these constructs, the ompR and tssF4 genes are under the control of the arabinoseinducible ParaBAD promoter.
For OmpR protein purification, the sequence encoding the ompR gene of Y. pseudotuberculosis was PCR amplified and cloned into the Gateway TM destination pETG-20A vector (kindly provided by Dr. Arie Geerlof, European Molecular Biology Laboratory (EMBL), Hamburg, Germany) according to standard Gateway TM protocols. In this construct, OmpR is fused to an Nterminal hexahistidine-tagged thioredoxin (TRX) followed by a tobacco etch virus (TEV) protease cleavage site.

Strains Construction
Insertion of lacZ fusions using the pFUSE plasmid derivatives. The pFUSE constructs were transferred to Y. pseudotuberculosis RL31758-4 by conjugation using S17-1 lpir as donor. Cultures of donor and recipient strains were cultivated in LB medium (at 37uC for donor strain, 28uC for recipient strain) with agitation until the end of the growth exponential phase. 1 ml of each culture were collected. Cells were harvested by centrifugation and then resuspended in 1 ml of 10 mM MgSO 4 . Mating was performed by mixing 400 ml of donor (Cm R ) and 400 ml of recipient strains (Nal R ) with 2.2 ml of 10 mM MgSO 4 . Cells were recovered by filtration on a 25-mm 0.45-mm pore size membrane. The filter was then placed on LB agar plates and incubated for 16 h at 28uC. Cells were collected and serial dilutions spread onto LB agar plates supplemented with Nal and Cm to select RL31758-4 transconjuguants. The pFUSE plasmid being unable to replicate into Y. pseudotuberculosis it integrated onto the chromosome by homologous recombination at the chromosomal locus. Proper integration of the plasmid was verified by colony PCR.
Insertion of gfp fusions using the pRL44 plasmid derivatives. The pRL44 derivatives were transferred to RL31758-4 by conjugation using S17-1 lpir as donor as described above. Proper integration of pRL44 into the chromosome disrupting the ara locus, plasmid integrants were identified by their Kan R and ara 2 phenotypes on MacConkey-arabinose agar plates. Lost of the plasmid carrying the sacB counter-selectable marker was performed by streaking integrants on LB agar containing 10% sucrose. The Kan s and ara 2 phenotypes of sucrose-resistant clones were verified on LB-Kan plates and MacConkey-arabinose agar plates respectively, and proper insertion of the fusion was verified by colony PCR.
Construction of DtssF mutant strains using the pRE112 plasmid derivatives. The pRE112 derivatives were transferred from S17-1 lpir to Y. pseudotuberculosis RL31758-4 and plasmid integrants were selected on chloramphenicol LB agar plates. Double crossing-over was achieved by selection for sucroseresistant segregants on LB agar plates supplemented with 10% sucrose. Proper in-frame deletion of the tssF genes was verified by colony PCR.

Transposon Mutagenesis and Identification of the Insertion Site
For transposon mutagenesis, plasmid pAJD428 [75] was tranferred by conjugation from MFDpir to the Y. pseudotuberculosis strain carrying lacZ and gfp fusion to the T6SS-4 promoter (RL31758-41). MFDpir being devoid of the dihydrodipicolinate synthase dapA gene, it requires diaminopimelic acid (DAP) to the culture medium [68]. Recipient and donor strains were mixed in 10 mM MgSO 4 as described above. Cells were immobilized on a 25-mm 0.45-mm pore size membrane filter disposed on LB agar plates supplemented with 0.3 mM DAP and incubated for 16 h at 28uC. Cells were collected and serial dilutions spread onto LB agar plates supplemented with Nal, Kan, IPTG and X-Gal (40 mg/ml). Colonies with decreased lac expression were selected and further tested for fluorescence levels. The conditional R6K replication origin of the transposon was used to obtain replicative plasmids. Briefly, candidate transposon mutant strains were grown overnight at 28uC and chromosomal DNA were purified using the Tissue and Blood DNA purification kit (Qiagen). 200 ng of chromosomal DNA were digested by EcoRI. After EcoRI heat-inactivation, fragments were ligated using the T4 DNA ligase (New England Biolabs) and the ligation mixture was electroporated into CC118lpir. Plasmids were purified from Kan R clones and the transposon-chromosome junction was sequenced.
b-galactosidase Activities and Fluorescence Levels b-Galactosidase enzyme activity was measured in permeabilized cells harvested at an OD 600 nm of ,2 as described previously using ortho-nitro-phenyl-b-D-galactopyrannoside (ONPG) as substrate [76]. Individual cultures were assayed in triplicate, and reported values (in Miller units) are averaged from three independent cultures.
Fluorescence levels. Cell cultures were diluted 10-fold into LB and 150 ml were transferred into wells of a black 96-well plate (Greiner). Absorbance at 600 nm and fluorescence (excitation: 485 nm; emission: 530 nm) were measured with a TECAN infinite M200 microplate reader. The relative fluorescence is expressed as the intensity of fluorescence divided by the absorbance at 600 nm, after subtracting the values of a blank sample. Individual cultures were assayed in triplicates and reported values are averaged from three independent cultures.

Cell Survival to Osmotic Stress
Cells were grown at 28uC in LB with shaking to an OD 600 nm ,0.6. 10 3 cells were incubated for 60 min. with or without 0.6 M Sucrose (,20%) at 28uC without shaking and then plated on LB plates. For induction from pBAD vectors, 0.1% of arabinose was added in the medium 60 min. prior to Sucrose addition. After overnight incubation at 28uC, colony forming units (cfu) were counted and cell survival was expressed as % of cfu after sucrose exposure relative to cfu without sucrose treatment.

Cell Survival to Sodium Deoxycholate
Cells were grown at 28uC in LB with shaking to an OD 600 nm ,0.6, harvested and resuspended in LB to an OD 600 nm of 0.5. 10 ml of 10 21 to 10 26 serial dilutions were spotted on LB agar plates supplemented with 1% DOC. For induction from pBAD vectors, 0.1% of arabinose was added in the medium 60 min. prior to cell resuspension, and 0.1% of arabinose was added in LB-DOC plates. Values are reported as the highest dilution for which colonies can be observed after 24 hours of growth at 28uC.

Statistical Analyses
To determine whether two sets of data were significantly different, paired Student's t-tests were performed using the T.TEST function in Excel. p-values obtained from these statistical analyses are reported in the figures (p.0.05 indicated by NS [non significant], p#0.01 indicated by *, p#0.001 by **, and p#0.0001 by ***).
Purification of the Y. pseudotuberculosis OmpR Protein pETG20-OmpR was transformed into the E. coli T7 Iq pLysS expression strain. Cells were grown at 37uC in terrific broth to an OD 600 , 0.9 and ompR expression was induced by addition of 0.5 mM IPTG for 16 hours at 17uC. Cells were harvested and resuspended in Tris-HCl 20 mM (pH8.0), NaCl 150 mM. Lysozyme was added (0.25 mg/ml) and cells were broken by sonication. The soluble proteins were separated from inclusion bodies and cell debris by centrifugation 30 min at 20,0006g. The TRX-66His-TEV-OmpR fusion (42 kDa) was purified using Ni 2+ affinity chromatography (HisTrap 5 ml GE Healthcare on an AKTA FPLC system) and eluted with a step gradient of imidazole. After proteolysis by a Hexahistidine-tagged TEV protease (1:10 (w/w) protease:protein ratio) for 16 hours at 4uC, the TEV protease and contaminants were arrested on a second Ni 2+ affinity column and the native OmpR protein (27 kDa) was collected in the flow through and separated on a preparative Superdex 200 gel filtration column (GE Healthcare) equilibrated in 20 mM Tris-HCl (pH 8.0), 150 mM NaCl. The final concentration of the OmpR solution was 1.5 mg/mL.

Electrophoretic Mobility Shift Assays (EMSA)
Radiolabeled probes were generated by polymerase chain reaction (PCR) using a mix of dNTPs supplemented with [a-32 P]dGTP (5 mCi per PCR in a total volume of 50 ml; Perkin-Elmer), and purified using the Wizard Gel and PCR clean-up kit (Promega). The EMSA protocol with the purified, phosphorylated OmpR protein was adapted from previously published protocols [77,78]. To induce OmpR autophosphorylation, the protein was incubated for 30 minutes with 25 mM acetyl phosphate in Tris-HCl 10 mM (pH7.4), KCl 50 mM, Dithio threitol (DTT) 0.5 mM, MgCl 2 1 mM, Glycerol 4%. Radiolabeled PCR products (4 nM final concentration) were incubated in a final volume of 12 mL of Tris-HCl 10 mM (pH7.4), KCl 50 mM, DTT 0.5 mM, MgCl 2 1 mM, Glycerol 4%, BSA 50 mg/mL, sonicated salmon sperm DNA 50 mg/mL, acetyl phosphate 25 mM in presence of various concentrations of the phosphorylated OmpR protein. The mixtures were incubated for 20 minutes at 25uC and then loaded on a pre-run 8% non denaturing polyacrylamide (Tris-borate) gel. DNA and DNAcomplexes were separated at 80 V in Tris-Borate buffer (45 mM Tris base, 45 mM boric acid, 100 mM MnCl 2 buffer). Gels were fixed in 10% trichloro-acetic acid for 10 minutes, and exposed to Kodak BioMax MR films.