All guinea pig infection procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of School of Basic Medical Sciences, Fudan University (IACUC Animal Project Number 20140226-022-qu) according to national guidelines (Regulations for the Administration of Affairs Concerning Experimental Animals, China).

S. flexneri 2a 301 (Sf301, GenBank accession number AE005674) was kindly provided by Pr. Qi Jin (MOH Key Laboratory of Systems Biology of Pathogens, Institute of Pathogen Biology, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China). The bacterial strains and plasmids used in this study are listed in Table 1. S. flexneri and E. coli were grown in Luria–Bertani medium (LB; Oxoid, Basingstoke, UK) at 37°C. Antibiotics were used at the following concentrations: ampicillin (100 μg/ml), kanamycin (50 μg/ml), tetracycline (10 μg/ml) and gentamicin (50 μg/ml) (Sigma-Aldrich, St. Louis, MO, USA).

S. flexneri phoQ/phoP deletion mutant strain was constructed by one-step inactivation of chromosomal genes using PCR products (Datsenko and Wanner, 2000). First, Sf301 was transformed with pKD46 (Table 1) to express the λ Red recombinase and was selected by ampicillin. The transformants were induced with L-arabinose (1 mM) and made electro-competent cells. A phoPQ homologous recombination fragment with kanamycin resistance gene cassette (kan, 1,394 bp) was amplified from pKD13 with primers PhoPQus-kan-F/PhoPQds-kan-R (containing upstream and downstream regions of Sf301 phoPQ; Table 2). The purified PCR products were then digested with DpnI, suspended in 10 mM Tris (pH 8.0), transferred into Sf301 by electroporation, and grown at 30°C. The phoPQ knockout mutant was screened out by kanamycin, and verified by PCR and sequencing (Sangon Biotech, Shanghai, China). The phoPQ knockout mutant was grown at 43°C to remove pKD46, plasmid removal was confirmed by PCR, and the strain was named ΔphoPQ (Figure S1).

For construction of complementation of phoPQ for ΔphoPQ, the phoPQ operon with its promoter region (2,414 bp) was amplified by PCR with primers ΔphoPQc-F/ΔphoPQc-R (Table 2), which were designed based on the genome of Sf301. The PCR products were ligated with pGEMT (Table 1), and the bacteria with the insertion were selected by ampicillin. After verification by sequencing, the phoPQ complementary plasmid (pphoPQ) was transformed into ΔphoPQ, and the transformants were selected by kanamycin and ampicillin, verified by PCR and sequencing, and named ΔphoPQc. pGEMT was introduced into ΔphoPQ as a vector control, and was named ΔphoPQ (pGEMT).

The construction of S. flexneri icsA, yoaE, yrbL, or rstA deletion mutant strains and their complementation strains used the same method as that of phoPQ with primers listed in Table S4. For construction of the icsA complementary plasmid picsA, the icsA gene with its promoter region (3,642 bp) was amplified from wild-type Sf301 with primers ΔicsAc-F/ΔicsAc-R, and ligated into the pGEMT vector. After verification by sequencing, picsA was then transformed into ΔphoPQ to construct the icsA expression strain ΔphoPQ(picsA), verified by PCR and sequencing.

The invasion ability of strains of S. flexneri was determined by gentamicin protection assay on HeLa cells and Caco-2 cells (Hale and Formal, 1981; Mounier et al., 1992). Cells were grown in 24-well plates at 5% CO₂ [HeLa cells in Dulbecco's Modified Eagle Medium (DMEM) with 10% fetal bovine serum (FBS) and Caco-2 cells in DMEM medium with 20% FBS]. Bacterial strains were inoculated into LB containing 0.3 M NaCl until the OD₆₀₀ reached 1.0. Bacteria were added to semi-confluent HeLa cells or Caco-2 cells at a multiplicity of infection (MOI) of 10. The plates were then centrifuged at 900 g for 10 min. After incubating at 37°C for 5, 15, or 30 min, gentamicin was added to the medium with a final concentration of 100 μg/ml for 60 min at 37°C. The cells were then lysed with 1 ml 0.1% Triton X-100 in PBS for 10 min. The lysates were diluted 1:10 in PBS and plated onto LB agar plates in triplicate, and colonies were counted. The invasion rate was calculated as the number of bacteria recovered from gentamicin-treated cells divided by the total number of inoculated bacteria.

For immunofluorescence assay, HeLa cells grown in 24-well tissue culture plates with coverslips were infected with strains of S. flexneri for 15 min, fixed with 3.7% formaldehyde in PBS for 15 min, and then permeabilized with 0.2% Triton X-100 for 5 min. Bacteria were stained with anti-Shigella O-Ag serum (BD Biosciences, New York, USA) and then with IgG Alexa 488 conjugate (Life Technologies, New York, USA), actin was stained with Texas Red phalloidin (Life Technologies), and nuclei were stained with 4′,6-Diamidino-2-Phenylindole (DAPI, Life Technologies) for 30 min. The coverslips were mounted and observed under a confocal laser scanning microscope (CLSM; Leica TCSSP5, Mannheim, Germany) at × 100 magnification (Chang et al., 2005).

The virulence of S. flexneri was determined by guinea pig keratoconjunctivitis Sereny test (Sereny, 1957). Female guinea pigs (age 6 weeks, about 300 g) were inoculated with 10⁹ Colony-Forming Units (CFU) of bacteria per eye (six guinea pig eyes in each group), and observed at 24, 48, and 72 h. Inoculation with LB served as a negative control. The keratoconjunctivitis induced by the bacteria was scored as follows: −, no disease or mild irritation; +, mild conjunctivitis or late development and/or rapid clearing of symptoms; ++, keratoconjunctivitis without purulence; and +++, fully developed keratoconjunctivitis with purulence. At 72 h post-inoculation, guinea pigs were euthanized with pentobarbital (40 mg/kg) and the eyes were removed and fixed in 4% formalin in PBS (pH 7.2). After hematoxylin and eosin (H&E) staining, the eye sections were examined under a microscope.

Growth curves of the strains were determined by measuring the OD₆₀₀ with an Eppendorf spectrophotometer at 60 min intervals over a period of 14 h. For the low Mg²⁺ growth assay, N medium was used containing 0.1 M Tris-HCl (pH 7.4), 38 mM glycerol, 0.1% (wt/vol) Casamino Acids, 0.37 g/l KCl, 0.087 g/l K₂SO₄, 0.99 g/l (NH₄)₂SO₄ and 0.14 g/l KH₂PO₄ (Barchiesi et al., 2012). Overnight cultures of bacterial strains were inoculated into N medium supplemented with 10 μM or 10 mM MgCl₂ (at 1:50 dilution) and incubated at 37°C with shaking. To assay acid resistance of bacteria, E glucose broth was used containing 0.2 g/l MgSO₄•7H₂O, 2 g/l citric acid, 13.1 g/l K₂HPO₄•3H₂O, 3.5 g/l Na(NH₄) HPO₄•4H₂O, and 0.4% glucose. Overnight cultures were inoculated into E glucose broth at pH 7 or pH 5 (at 1:50 dilution) and incubated at 37°C with shaking (Barchiesi et al., 2012). For the polymyxin B resistance assay, overnight cultures were inoculated into LB, grown with shaking until the OD₆₀₀ reached 0.6, then bacteria were diluted in sterile 0.85% saline to about 5 × 10³ cells per ml and exposed into different concentrations of polymyxin B (5, 10, 20, and 40 μg/ml) for 1 h at 37°C. Surviving bacteria were determined by plating on LB agar plates in triplicate. The survival rate was calculated as the number of bacteria treated with polymyxin B divided by that of the untreated control. All experiments were repeated at least three times.

For microarray analysis, Sf301 and ΔphoPQ were inoculated into LB medium and grown to mid-log phase (6 h) or early-stationary phase (10 h), with three biological replicates. Cells were harvested by centrifugation at 10,000 g for 1 min, and total RNAs were extracted using the RNeasyH Mini Kit (QIAGEN, Hilden, Germany) following the manufacturer's instructions. Agilent custom-specific design GeneChip of Sf301 genomic DNA were used. Each microarray (4^*44k) contained spots with 4168 specific 60-mer oligonucleotides representing the 4168 ORFs of Sf301 in triplicate, carried out by Shanghai Biotechnology Co. Ltd. (Shanghai, China) according to standard protocols provided by Agilent Technologies (Palo Alto, USA). Briefly, the quality and quantity of RNA samples were determined and checked by Agilent 2100 bioanalyzer (Agilent Technologies). The RNA samples were then reverse transcribed to cDNA by MMLV reverse transcriptase (Invitrogen, Carlsbad, USA), followed by transcription with T7 RNA polymerase (New England BioLabs, Beverly, UK) to generate aminoacyl-UTP-labeled cRNA. Amino allyl modified cRNAs were purified and labeled with Cy3 (Cy3 NHS ester, GE Healthcare, Piscataway, NJ). Labeled cRNAs were then fragmented in fragmentation buffer (Agilent Technologies) and mixed with the Gene Expression Hybridization Kit (Agilent Technologies) at 65°C for 17 h with a constant rotation rate of 10 rpm for hybridization. The arrays were scanned by an Agilent DNA Microarray scanner. Microarray data were normalized in the Agilent Feature Extraction software. The ratio of gene expression (ΔphoPQ vs. Sf301) was calculated from the normalized signal intensities. A false discovery rate of 5% (P-value cutoff; 0.05) was used for variance analysis of three biological replicates and an arbitrary threshold of 2.0-fold or 0.5-fold was used for defining significant differences in expression ratios.

For validating the differential expression genes of microarray, qRT-PCR was carried out. Total RNA of bacteria was extracted using the RNeasyH Mini Kit. The extracted RNA was reverse transcribed into cDNA using iScript reverse transcriptase (Bio-Rad, Hercules, CA, USA) with incubation for 5 min at 25°C, followed by 30 min at 42°C and 5 min at 85°C. Subsequently, qRT-PCRs were performed using SYBR green PCR reagents (Premix EX TaqTM, Takara Biotechnology, Dalian, China) in the Mastercycler realplex system (Eppendorf AG, Hamburg, Germany) with amplification conditions of 95°C for 30 s, 40 cycles of 95°C for 5 s and 60°C for 34 s, followed by melting curve analysis. The 16S rRNA methyltransferase coding gene rsmC was used to normalize the transcriptional levels of genes in the qRT-PCRs. All qRT-PCRs were carried out in triplicate with at least three independent RNA samples. The primers (Table S1) were designed based on the genome of Sf301 using Beacon designer software (Premier Biosoft International Ltd., Palo Alto, CA, USA).

For analyzing the interaction of the recombinant PhoP and the promoter regions of putative target genes, EMSA were performed using the DIG Gel Shift Kit (Roche, Basel, Switzerland) according to the manufacturer's instructions. The phoP gene was amplified with primers pET28a-phoP-F/pET28a-phoP-R (Table 2) from the genomic DNA of Sf301 and inserted into the vector pET-28a (+) to obtain the recombinant plasmid pET28a-phoP. The recombinant plasmid was then transformed into Escherichia coli BL21 (DE3). Bacteria were grown to an OD₆₀₀ of 0.6 at 37°C and 0.8 mM IPTG was then added to induce PhoP protein expression for 6 h at 30°C. The expressed His-tagged PhoP protein was purified using the ProBondTM Purification System (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions. PhoP was phosphorylated prior to gel shift reaction by incubating PhoP with 50 mM acetylphosphate for 1 h. The predicted promoter regions of putative target genes were amplified with primers in Table 2 and labeled with digoxigenin using terminal transferase. Each gel shift assay included the probe labeled with digoxigenin plus increasing concentrations of the phosphorylated PhoP (PhoP-P, ranging from 0.16 to 1.6 μM). The coding sequence of virA in Sf301 without PhoP box sequence was designated as a negative control. All samples were incubated at 25°C for 20 min, separated by electrophoresis on 6% non-denaturing polyacrylamide gel, blotted onto a positively charged nylon membrane (Millipore), and detected by an enzyme immunoassay following the manufacturer's instructions.

DNase I footprinting assays were performed according to Wang et al. (2012). The promoter regions of yoaE, mgtA and shf were amplified with primers listed in Table 2, and separately cloned into the pUC18BT vector (Shanghai Biotechnology Corporation, China), which was further used as the template for preparation of fluorescent FAM labeled probes. The probes were purified by the Wizard® SV Gel and PCR Clean-Up System (Promega) and quantified with a NanoDrop 2000C (Thermo Fisher Scientific Waltham, MA, USA). For each assay, 500 ng probes were incubated with different amounts of PhoP in 40 μl binding buffer at 30°C for 30 min. Then 10 μl DNase I (0.01 unit) (Promega) and 100 nmol CaCl₂ were added, incubated at 25°C for 1 min, and the reactions were stopped by adding 140 μl DNase I stopping solution (200 mM unbuffered sodium acetate, 30 mM EDTA and 0.15% SDS). The DNA samples were extracted with phenol/chloroform, dissolved in 30 μl MilliQ water and loaded for carrying out capillary electrophoresis. The data were collected using the GeneScan-500 LIZ dye Size Standard (Applied Biosystems, Foster City, CA, USA).

A promoterless lacZ reporter gene was amplified by PCR with primers lacZ-F/lacZ-R (Table 2) using the pBAD/His/LacZ vector (Invitrogen, Carlsbad, CA, USA) as a template. Purified PCR products were digested with ScaI and NcoI endonucleases (MBI Fermentas, Vilnius, Lithuania) and inserted into the medium-copy-number plasmid pACYC184 (Chang and Cohen, 1978). The bacteria with the insertion were selected using tetracycline, verified by PCR and sequencing, and designated placZ. The promoter-proximal DNA region of phoP, shf, and icsA were amplified by PCR with Pfu DNA polymerase (Takara Biotechnology, Dalian, China) using the primers listed in Table 2. PCR products were digested with AvaI and ScaI endonucleases, and inserted into plasmid placZ. The bacteria with the insertion were selected by tetracycline, verified by PCR and sequencing, and named pphoP::lacZ, pshf::lacZ, and picsA::lacZ. These recombinant plasmids were introduced into Sf301 and the ΔphoPQ strain, respectively. The promoterless placZ was transferred into the bacterial strains as negative control. The reporter bacterial strains were separately grown in LB, or N medium with 10 μM/10 mM MgCl₂, or E glucose broth at pH 7/5.5 or LB with 25 μg/ml polymyxin B. The bacteria were harvested, lysed with 400 μl lysozyme (1 mg/ml) at 37°C for 30 min, and then the β-galactosidase activity in the cellular extracts was measured by the β-Galactosidase Enzyme Assay System (Promega) following the manufacturer's instructions. All experiments were repeated at least three times independently.

Experiments were performed in triplicate and repeated at least three times. The data were analyzed with Student's t-test or one-way factorial analysis of variance in SPSS version 14.0 (Chicago, IL). Differences in means with a P < 0.05 were considered significant.

The complete microarray data set is uploaded in the Gene Expression Omnibus database (http://www.ncbi.nlm.nih.gov/geo/) under accession numbers GPL24308 for the platform design and GSE107365 for the original data set.

To analyze the regulatory mechanism of PhoPQ in Shigella, a S. flexneri phoPQ deletion mutant strain was constructed by using homologous recombination (Datsenko and Wanner, 2000). After transformation of the λ Red recombinase expression plasmid pKD46, the Sf301 was transformed with a fragment of a kanamycin resistance cassette with long flanking regions homologous of phoPQ. The phoPQ knockout mutant was then screened out by kanamycin, verified by PCR and sequencing, pKD46 was removed by growth at 43°C, and the strain named as ΔphoPQ (Figure S1). For construction of the phoPQ complemented strain, the phoPQ operon with its promoter region was amplified from wild-type Sf301, and ligated into the pGEMT vector. The phoPQ expression plasmid was then transformed into ΔphoPQ, selected by growth in the presence of kanamycin and ampicillin, verified by PCR as well as sequencing, and named as ΔphoPQc.

The invasion ability of Sf301, ΔphoPQ and ΔphoPQc were evaluated by the gentamicin protection assay on HeLa cells and Caco-2 cells. Bacterial strains were separately inoculated into HeLa cells or Caco-2 cells at a MOI of 10, incubated for 5, 15, or 30 min, and then gentamicin was added to kill extracellular bacteria. In HeLa cells, invasion rates of ΔphoPQ at 5, 15, or 30 min post-inoculation were reduced by 56-fold, 22-fold, and 15-fold, respectively, compared with those of Sf301. The invasion ability of ΔphoPQc (complementary phoPQ) was recovered to levels of Sf301 (Figure 1A). In Caco-2 cells, invasion rates of ΔphoPQ were similar to those in HeLa cells (Figure 1B). Next, Shigella-infected HeLa cells were observed by confocal immunofluorescence microscopy. HeLa cells grown on coverslips in 24-well tissue culture plates were infected with the strains of S. flexneri at a MOI of 10 and incubated for 15 min. The infected cells were stained with rabbit polyclonal anti-shigella anti-serum, Texas Red-labeled phalloidin, and DAPI (Figure 2). In Sf301 infected cells, more cells displayed membrane ruffling (more than 5 cells showing obvious membrane ruffling in a microscope field of view at ×100 magnification), which indicated actin cytoskeleton changes of HeLa cells, while in ΔphoPQ infected cells no obvious membrane ruffling observed. The membrane ruffling of ΔphoPQc infected cells could be restored to the level of Sf301 infected cells by complementation with pphoPQ (Figures 2A–C).

The Guinea pig Sereny test was used to determine the virulence of Sf301, ΔphoPQ, and ΔphoPQc. Guinea pigs were infected with 10⁹ CFU per eye and observed at different time points. At 24 h post-infection, the guinea pig eyes infected with Sf301 displayed keratoconjunctivitis with or without purulence (+ or ++). At 48 and 72 h post-infection, the inflammation exacerbated with great purulence (+++). The ΔphoPQ-infected guinea pig eyes displayed no obvious inflammation (–) at 24 h post-infection, and a slight keratoconjunctivitis without purulence (+) after 48 h. In ΔphoPQc-infected guinea pig eyes, the inflammation reaction was similar to that of Sf301. Guinea pigs inoculated with PBS were used as a negative control (Figure 3, Table 3). At 72 h post-infection, all guinea pigs were euthanized with pentobarbital and the eyes were removed, stained with hematoxylin and eosin and pathological examinations were carried out. Sf301-infected eyes showed typical inflammatory reactions including luminal debris, epithelial desquamation, neutrophil infiltration and submucosal edema along with severe ulcers (Figures 4A,D). In contrast, ΔphoPQ-infected eyes showed very small numbers of neutrophils in the mucosal lamina or submucosa, and minimal edema in the submucosa (Figures 4B,E). The inflammatory lesions induced by ΔphoPQc recovered to the Sf301 level (Figure 4C).

Extracellular Mg²⁺, pH and antimicrobial peptides have been described as input signals of the PhoP/PhoQ system (Garcia Vescovi et al., 1996; Gunn and Miller, 1996; Bearson et al., 1998; Groisman, 2001; Lejona et al., 2003; Barchiesi et al., 2012). Therefore, the role of PhoPQ in S. flexneri ability to withstand low Mg²⁺, acidic pH, or the presence of polymyxin B was analyzed. Sf301, ΔphoPQ and ΔphoPQc were grown in different conditions and growth curves were determined. In high Mg²⁺ medium (10 mM), the three strains showed no difference in growth pattern (Figure 5A), while in low Mg²⁺ medium (10 μM), ΔphoPQ showed limited growth compared to that of Sf301. At mid-log phase (8 h), Sf301 reached an OD₆₀₀ of 0.68, but ΔphoPQ only reached an OD₆₀₀ of 0.25 and took about 6 h more to reach an OD₆₀₀ of 0.64. This growth deficiency was rescued by complementation with pphoPQ (Figure 5B). In neutral medium (pH 7), growth curves showed no difference among the three strains (Figure 5C). However, in the acidic medium (pH 5), Sf301 and ΔphoPQc showed only limited growth while ΔphoPQ was unable to grow (Figure 5D). At different concentrations of polymyxin B (5, 10, 20, 40 μg/ml), the survival rates of ΔphoPQ were significantly lower than that of Sf301 (P < 0.05), and the resistance defect of ΔphoPQ was restored by complementation with pphoPQ (Figure 5E).

DNA microarray was used to compare the transcriptional profiles of Sf301 and ΔphoPQ at middle-log phase (6 h) or early-stationary phase (10 h) under LB growth conditions. At middle-log phase, 117 differentially expressed genes (DEGs) were identified. Among them, 32 genes were up-regulated and 85 genes were down-regulated in ΔphoPQ (Table 4). At early-stationary phase, 54 DEGs were identified, with 19 genes up-regulated and 35 genes down-regulated in ΔphoPQ (Table S5). The 117 DEGs were involved in metal ion transport (katE, narU, bfr), acid resistance (hdeABCD, gadAB, yhiWX, xasA), lipopolysaccharides (LPS) modification and antibacterial peptide tolerance (rfbU, mdoB, slyB, pagP, msbB2, pmrD), signal transduction (phoPQ, rstA, cstA), bacterial virulence (icsA, virK), respiratory and energy metabolism (hyaABCDEF, appABC) (Table 4). Among them, 44 DEGs were verified by qRT-PCR, with 38 giving a result consistent with that of the DNA microarray (Table 4).

To identify S. flexneri PhoP-regulated genes, the online relational databases (http://genolist.pasteur.fr) was used to search the genes with putative PhoP-binding motif in their promoter regions. A PhoP recognition motif was generated based on conserved pattern PhoP box [5′-(T/G) GTTTA-N5-(T/G) GTTTA- 3′] identified in S. typhimurium and E. coli (Kato et al., 1999; Lejona et al., 2003). The Sf301 genome was searched and 38 putative PhoP recognition motifs were detected (Table 5). Among these genes/operons, phoPQ, mgtA, slyB, rstAB, hdeAB, pagP, yrbL, yoaE, yhiWX, and shf –rfbU-virK-msbB2 were reported as members of the PhoPQ regulon in other bacteria (Kato et al., 1999; Lejona et al., 2003; Minagawa et al., 2003; Zwir et al., 2005), while icsA and ipaH7.8 have not been reported as PhoPQ-regulated genes before (Table 5). The transcriptional levels of these 12 genes or operons showed differential expression both in the microarray and qRT-PCR analyses (Table 4).

To verify those 12 predicted PhoP target genes/operons above in Shigella, EMSAs were performed. The recombinant PhoP-P resulted in a mobility shift of the fragments upstream of 11 genes/operons (phoPQ, mgtA, slyB, icsA, shf –rfbU-virK-msbB2, rstAB, yoaE, hdeAB, yrbL, yhiWX, and pagP) in a concentration-dependent manner, but did not bind to the fragment upstream of ipaH7.8 (Figure 6). As a negative control, the DNA fragment of virA coding sequence (without PhoP box motif) did not form a complex with PhoP under the same conditions (Figure 6).

The PhoP-P binding motif sequences in the promoters of yoaE, shf and mgtA were identified by DNase I footprinting assay. A 25-nt protected sequence in the promoter region of yoaE (−107 to −83 bp) was identified (Figure 7A), and a 35-nt protected sequence was located at −115 to −81 bp in the upstream of the translational start site of shf (Figure 7B). There exist two protected sequences in the mgtA promoter (−184 to −159 bp and −152 to −124 bp) (Figure 7C). All protected sequences were in accordance with PhoP binding consensus motif (Figure 7, Table 5).

Furthermore, the transcriptional levels of genes with PhoP binding activities in low Mg²⁺, acidic pH, or presence of polymyxin B were detected by qRT-PCR. Among those genes, transcriptional levels of phoP, shf and icsA were up-regulated significantly in all three environmental stress conditions (Tables S6–S8). Three LacZ reporter plasmids with the promoter regions of the genes (pphoP::lacZ, pshf ::lacZ, and picsA::lacZ) were then constructed to confirm PhoPQ regulation on those genes expression. After transformation of the plasmids into ΔphoPQ or Sf301 and culture in different mediums, β-galactosidase activity was detected. In low Mg²⁺ (10 μM) medium, the expression of phoP, shf, and icsA in Sf301 was 7.2, 9, and 12.9-fold higher, respectively, than that in ΔphoPQ (Figure 8A), while in high Mg²⁺ (10 mM) medium, the expression of phoP, shf, and icsA in Sf301 was only 1.9, 1.8, and 2.2-fold higher, respectively, than that in ΔphoPQ (Figure 8B). Under acidic pH (pH 5.5) conditions, the expression of phoP, shf, and icsA in Sf301 was 6.4, 6.6, and 7.1-fold higher, respectively, than that in ΔphoPQ (Figure 8C). In contrast, at pH 7, the expression of phoP, shf, and icsA in Sf301 was only 2.2, 2, and 3-fold higher, respectively, than that in ΔphoPQ (Figure 8D). In the presence of polymyxin B (25 μg/ml) in LB medium, the expression of phoP, shf and icsA in Sf301 was 5.2, 4, and 9.2-fold higher, respectively, than that in ΔphoPQ (Figure 8E), while in LB medium only, the expression of phoP, shf and icsA in Sf301 was only 2, 1.9, and 1.5-fold higher, respectively, than that in ΔphoPQ (Figure 8F). It suggested that the expressions of phoP, shf and icsA were regulated by PhoPQ.

As virulence is the key factor in Shigella pathogenesis, we focused on searching for PhoP target genes that are associated with Shigella virulence. Four PhoP-regulated genes (rstA, icsA, yrbL, and yoaE) that may be involved in Shigella virulence were deleted from Sf301, respectively. The virulence of these mutant strains was evaluated by the gentamicin protection assay on HeLa cells and only ΔicsA decreased virulence in HeLa cells compared to Sf301 (Figure S2). IcsA is one of the virulence factors required for Shigella pathogenesis (Bernardini et al., 1989; Brotcke Zumsteg et al., 2014), and its expression is regulated by PhoPQ based on results of the microarray, qRT-PCR, EMSA and β-galactosidase activity assay in our study. The transcriptional level of icsA in ΔphoPQ was significantly reduced both in the microarray (3-fold down-regulated) and qRT-PCR (4.2-fold down-regulated) compared to that of Sf301 (Table 4). A highly conserved motif is found in the promoter region of icsA and PhoP-P results in a mobility shift of the fragments upstream of icsA (Table 5, Figure 6). The β-galactosidase activities of icsA in Sf301 were significantly higher than that in ΔphoPQ (12.9, 7.1, and 9.2-fold higher, respectively) in the environments of low Mg²⁺, acidic pH or presence of polymyxin B (Figures 8A,C,E). As the phoPQ knockout diminished S. flexneri virulence, an icsA expression plasmid (picsA) was introduced into the ΔphoPQ strain [ΔphoPQ(picsA)] to observe whether virulence of ΔphoPQ could recover. A Shigella icsA deletion mutant strain (ΔicsA) and its complementation strain (ΔicsAc) served as controls. The invasion rate of HeLa cells or Caco-2 cells by ΔphoPQ(picsA) was 6.4 and 5.7-fold higher than that of ΔphoPQ, respectively (Figures 9A,B). ΔphoPQ(picsA) resulted in more membrane ruffles indicative of actin cytoskeleton changes in HeLa cells compared to ΔphoPQ (Figure 2F). In the guinea pig Sereny test, guinea pigs inoculated with ΔphoPQ(picsA) showed restored virulence (24 h + or ++ and 48 h ++ or +++) compared to ΔphoPQ (24 h– and 48 h– or +) (Figure 3, Table 3). Guinea pig eyes infected with ΔphoPQ(picsA) showed markedly more inflammatory reactions including epithelial desquamation and neutrophil infiltration in the pathological examination compared to ΔphoPQ-infected eyes (Figure 4I). As a control, the virulence of ΔicsA was decreased both in the gentamicin protection assay (2.7 and 2.2-fold lower, Figures 9A,B) and guinea pig Sereny test (24 h– or + and 48 h +, Figure 3, Table 3), compared to Sf301. It indicated that the virulence of ΔphoPQ(picsA) could be restored partly by complementation with picsA, but it still did not reach the level of Sf301.

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