The Effect of the Potential PhoQ Histidine Kinase Inhibitors on Shigella flexneri Virulence

PhoQ/PhoP is an important two-component system that regulates Shigella virulence. We explored whether the PhoQ/PhoP system is a promising target for new antibiotics against S. flexneri infection. By using a high-throughput screen and enzymatic activity coupled assay, four compounds were found as potential PhoQ inhibitors. These compounds not only inhibited the activity of SF-PhoQc autophosphorylation but also displayed high binding affinities to the SF-PhoQc protein in the Surface Plasmon Resonance response. A S. flexneri cell invasion assay showed that three of these potential PhoQ inhibitors inhibit the invasion of HeLa cells by S. flexneri 9380. In a Mouse Sereny test, mice inoculated with S. flexneri 9380 pre-treated with the potential PhoQ inhibitors 1, 2, 3 or 4 displayed no inflammation, whereas mice inoculated with S. flexneri 9380 alone displayed severe keratoconjunctival inflammation. All four potential PhoQ inhibitors showed no significant cytotoxicity or hemolytic activity. These data suggest that the four potential PhoQ inhibitors inhibited the virulence of S. flexneri and that PhoQ/PhoP is a promising target for the development of drugs against S. flexneri infection.


Introduction
Shigella is a gram-negative facultative intracellular pathogen with enhanced cell invasion, intracellular growth and intercellular spreading capabilities. The bacteria are transmitted fecal-orally and will invade the mucosa of the colon. Infection by only 10 to 100 organisms will cause shigellosis. Because of the overuse of antibiotics, Shigella drug resistance in clinical settings is increasing [1,2,3]. Therefore, new therapeutic targets and drugs are needed to reduce the incidence of shigellosis worldwide. Understanding the regulation of Shigella virulence may lead to the development of new drugs that can inhibit or reduce the virulence of Shigella as well as provide new strategies for treating shigellosis.
PhoQ/PhoP is a two-component system (TCS) that governs virulence, monitors extracellular Mg 2+ , and regulates several cellular activities in many gram-negative species [4]. The PhoQ/ PhoP TCS consists of the transmembrane sensor PhoQ and the cytoplasmic regulator PhoP. PhoQ is a transmembrane histidine kinase with a functional kinase domain that binds ATP. It responds to environmental signals by phosphorylating itself as well as PhoP. PhoP has a functional domain, which when phosphorylated influences virulence by activating a phosphorylation cascade that regulates a series of downstream effecter genes in several bacterial species, including Shigella flexneri, Salmonella enterica, and Escherichia coli [5,6]. In Shigella, a functional phoP gene is important for virulence [7]. It has been proven that PhoP regulates Shigella's susceptibility to polymorphonuclear leucocytes (PMNs) and antimicrobial molecules. A phoP Shigella mutant is highly sensitive to killing by neutrophils [7]. Furthermore, infection of a mouse eye with a wild-type Shigella strain will cause keratoconjunctivitis, whereas infection by a phoP Shigella mutant was resolved more quickly relative to wild type infections [7]. The research of PhoQ/ PhoP TCS in Salmonella showed that mutants in the PhoQ/PhoP system can greatly reduce bacterial virulence and intracellular survival in macrophages [8]. This prompted us to investigate whether PhoQ/PhoP in Shigella would be an appropriate target for the design of novel antibacterial agents.
In the present study, we chose the PhoQ protein of S. flexneri as the target for screening by a chemical library, and four potential PhoQ inhibitors were identified. Both the cell invasion assay and Mouse Sereny test showed that these potential PhoQ inhibitors abate the virulence of S. flexneri. These potential PhoQ inhibitors displayed low cytotoxicity on mammalian cells and had no hemolysis effect. Our data indicate that PhoQ may be a promising target for the development of new antibiotics to treat S. flexneri infection.

Results
Sequence analysis of PhoQ HK and structure-based discovery of potential PhoQ inhibitors in silico Alignment of PhoQ amino acid sequences of the Shigella strains S. flexneri 301, S. dysenteriae 197, S. boydii 227 and S. sonnei 046 show that the proteins share 100% identity. An optimized 3D model of the PhoQ HK domain of Sf301 was constructed based on the crystal structure of PhoQ HK from E .coli, which is highly homologous to S. flexneri (more than 99.8%). The ATP binding site consists of two different cavities connected by a gorge-like channel ( Figure 1). The 3D model was used to identify potential PhoQ inhibitors of S. flexneri PhoQ HK by high-throughput virtual screening (HTVS). Using the drug-selection filter from the SPECS database, 85,000 potential drug-like molecules, constituting an inhouse database (named SPECS_1), were searched for potential binding molecule structures, using the program DOCK4.0 as a primary screening. Then, the most optimal 200 structures were rescored with the programs FlexX and CSCORE. Finally, according to molecular diversity, shape complementarity, and potential for forming hydrogen bonds in the binding pocket of the PhoQ, 100 molecules were manually selected as inhibitor candidates, and 64 compounds were purchased from the SPECS Company for further bioassays.

Effect of the potential PhoQ inhibitors on enzymatic activities of SF-PhoQc
To measure the interaction of the 64 compounds with PhoQ, a prokaryotic expression plasmid containing the Sf301 PhoQ intracellular domain was constructed, the recombinant His-tagged SF-PhoQc was purified and analyzed by SDS-PAGE (Figure S1A), and western blot analysis ( Figure S1B) was performed with a mouse anti-His monoclonal antibody. The molecular mass of the recombinant SF-PhoQc was confirmed by MALDI-MS as 27.914 kDa ( Figure S1C), which is in agreement with the theoretical molecular mass of 27.914 kDa.
The enzymatic activity of SF-PhoQc was determined using the Pyrophosphate Reagent, in which ATP initiates the reaction and ADP generated by ATP is accompanied by the reduction of NADH to NAD + . The rate of NADH reduction to NAD + is detected by a change in absorbance readings at 340 nm. The activity of SF-PhoQc is measured by detecting a decrease in NADH at A 340 . Recombinant SF-PhoQc (25 mmol/L) was added to initiate the reaction, and the A 340 value was recorded by SoftMax vision 4.8 each minute for 30 minutes. The A 340 value reduced from 1.47 to 0.65 in the presence of SF-PhoQc ( Figure 2).
Luminescent Kinase Assay was used to confirm the effects of compounds 1, 2, 3 and 4 on the enzymatic activity of SF-PhoQc. In the Luminescent Kinase Assay, the putative kinase activity of SF-PhoQc protein was measured by quantifying the amount of ATP that remained in solution after the reaction. A direct relationship between the luminescence measured and the amount of ATP was observed, indicating that the purified SF-PhoQc protein possessed ATPase activity. Next, the effect of each compound on the enzymatic activity of SF-PhoQc was measured. The IC 50

Binding affinities of potential PhoQ inhibitors to SF-PhoQc
The binding affinities of compounds 1-4 to the SF-PhoQc were determined by Surface Plasmon Resonance (SPR). All four compounds displayed significant binding affinities to the SF-PhoQc protein in the SPR response ( Figure 5). The data were fitted to a steady-state affinity model for potential PhoQ inhibitors, and a K D value was determined using the Biacore 3000 evaluation software. The dissociation constants of compounds 1-4 to SF-PhoQc were 4.50, 10.6, 7.56 and 9.40 mmol/L, respectively, and they were named as potential PhoQ inhibitors ( Table 1).
Evaluation of potential PhoQ inhibitors on the virulence of S. flexneri and S. typhimurium in vitro At concentration of 200 mmol/L (highest concentration in the present study) the four potential PhoQ inhibitors had no obvious effect on growth of Sf9380. The growth curve of Sf9380 treated with or without potential PhoQ inhibitors at concentration 200 mmol/L were tested. The growth curves of bacteria treated with or without potential PhoQ inhibitors treatment were similar, suggesting that four potential PhoQ inhibitors have no effects on shown was constructed based on the crystal structure of the PhoQ HK of E. coli, which is highly homologous to S. flexneri (more than 99.8%). The ATP binding pocket is fairly large and deep. Two cavities joined by a gorge-like channel create the whole binding pocket. doi:10.1371/journal.pone.0023100.g001 the growth of Sf9380 ( Figure S2). The gentamicin protection assay was used as a cellular model to evaluate the inhibitory effect of the four potential PhoQ inhibitors [9]. The bacteria were treated with each of the four potential PhoQ inhibitors in a serial dilution and grown for 8 hours. Next, the bacteria were inoculated and incubated with HeLa cells for 60 min prior to the addition of gentamicin to kill extracellular bacteria. For testing intracellular growth and spreading to adjacent cell, cells were further incubated for 290 min [9]. The results of the assays were expressed as the number of bacteria recovered from gentamicin-treated cells divided by the number of inoculated bacteria added to the cell. After lysates of HeLa cells were plated, the number of CFU that formed on LB agar were counted. Treated with 12.5 mmo/L of potential PhoQ inhibitor 1, invasion of HeLa cells by Sf9380 was reduced to 1.    Figure 6A). For intracellular growth and spreading to adjacent cell, potential PhoQ inhibitor 1 (12.5 mmo/L) reduced cell invasion of bacterium to 0.061, 1.96 times lower than that of untreated Sf9380 (0.12), potential PhoQ inhibitor 2 (25 mmo/L) was reduced to 0.036, 3.3 times lower than that of untreated Sf9380, PhoQ inhibitor 3 (100 mmo/L) was reduced to 0.046, 2.6 times lower than that of untreated Sf9380, while PhoQ inhibitor 4 (100 mmo/ L) have no obvious effect on the cell intracellular growth (0.12) (shown in Figure 6B).  The Pyrophosphate Reagent detects autophosphorylation. IC 50 represents the concentration of the potential PhoQ inhibitor needed to inhibit 50% of SF-PhoQc's autophosphorylation. Values are means 6 standard deviations from 3 independent wells.
The effects of potential PhoQ inhibitors on cell invasion of Salmonella were determined by gentamicin protection assay. It showed that the cell invasion rate of Salmonella enterica Typhimurium SL1344 treated with potential PhoQ inhibitor 1 (12.5 mmo/L), 2 (25 mmo/L), 3 (100 mmo/L) were 0.89610 22 (6.74 times lower than the control), 1.564610 22 (3.84 times lower than the control) and 0.036610 22 (166 times lower than the control), respectively, while potential PhoQ inhibitor 4 (100 mmo/L) had no obvious effect on the cell invasion (the cell invasion rate was 5.7610 22 ). The cell invasion rate of Salmonella was 6610 22 (shown in Figure 7).

Effects of the potential PhoQ inhibitors on S. flexneri virulence in vivo
A Mouse Sereny test was used to investigate the inhibitory effect of the four potential PhoQ inhibitors on the virulence of S. flexneri Sf9380. The mice were infected with 1610 8 CFUs per eye. Mice inoculated with Sf9380 showed severe keratoconjunctivitis without purulence at 24 hours post-infection, which developed to keratoconjunctivitis with purulence after 48 hours, and the condition was sustained for 96 hours ( Figure 8, Table 2). Mice inoculated with bacteria pre-treated with potential PhoQ inhibitors 1 (12.5 mmol/ L) or potential PhoQ inhibitors 2 (25 mmol/L) displayed a slight conjunctival inflammation at 24 hours after infection, but keratoconjunctivitis was resolved by 48 hours (Figure 8, Table 2). Mice inoculated with bacteria pre-treated with potential PhoQ inhibitor 3 (100 mmol/L) or potential PhoQ inhibitor 4 (100 mmol/L) displayed a slight conjunctival inflammation 24 hours post-infection, with keratoconjunctivitis resolved by 72 hours (Figure 8, Table 2). Mice infected with S. flexneri Sf9380 treated, with the same amount of DMSO (0.1%, v/v) as the potential PhoQ inhibitors, displayed severe keratoconjunctivitis ( Figure 8, Table 2). Mice inoculated with E. coli ATCC 25922 did not develop conjunctival inflammation as a negative control ( Figure 8, Table 2).

Cytotoxicity assay and erythrocyte hemolysis assays of potential PhoQ inhibitors
The cytotoxicity of each potential PhoQ inhibitor towards a Vero cell line was analyzed by the traditional MTT assay. The  CC 50 values of all four compounds were higher than 200 mmol/L, the highest tested concentration of potential PhoQ inhibitors. As a control, an equal concentration of DMSO (0.1%) was added to a well of Vero cells, and no obvious cytotoxicity was observed (Table 1).
To determine whether the four potential PhoQ inhibitors would induce hemolysis of mammalian erythrocytes, the hemolytic activities of each potential PhoQ inhibitor were tested at concentrations of 25 mmol/L and 100 mmol/L. All four potential PhoQ inhibitors showed no hemolytic activity (Table 1).

Discussion
Currently, there is an increase in antibiotic resistance among Shigella isolates [1], and this drug resistance phenomenon is causing complications and difficulties for clinical treatment. Several virulence regulator factors, such as two-component signal systems [10,11,12], quorum sensing systems [11,13,14], type III secretion systems [15], and the assembly of adhesive organelles [16], have been recognized as interesting targets to reduce bacterial infection. Bacterial two-component systems have gained increasing interest as novel antibacterial targets because these systems are required for virulence of pathogenic microorganisms [17,18]. In the present study, we found that the PhoQ/PhoP two-component system of Shigella may be a promising target for developing new antibiotics against S. flexneri infection.
PhoQ/PhoP is a two-component system that governs virulence [4], monitors the extracellular Mg 2+ , and regulates several cellular activities in many gram-negative species. The system also helps bacteria resist antibiotic peptides by regulating lipid A [7,19,20,21,22]. Bivalent cations and antibiotic peptides can competitively bind to the acidic structural domain on the cytoplasmic surface of PhoQ [23]. In addition to the concentration of Mg 2+ or Ca 2+ cations in the cytoplasm, it has been shown that the concentration of antibiotic peptides in the external environment [24], in addition to an acidic environment, will mediate the activation of PhoQ [25]. In Salmonella, PhoQ/PhoP can change the structure of the external cell membrane by regulating the remodeling of lipid A to strengthen a bacterium's resistance to the environment [26]. In Shigella, the PhoQ/PhoP two-component system is necessary for virulence, as demonstrated by an infection of mice with a phoP mutant of Shigella that resulted in milder keratoconjunctivitis than a wild type strain [7].
PhoQ is an attractive target for an antibiotic because it is absent in mammals [10,27]. In this study, we have explored the possibility of using the PhoQ as a potential target by performing a screen for inhibitors. After constructing a 3D model of the PhoQ HK domain of Sf301, 64 compounds were selected as inhibitor candidates based on their molecular diversity, shape complementarities, and potential for forming hydrogen bonds in the binding pocket of PhoQ. To confirm the interaction of the compounds and PhoQ, a prokaryotic expression plasmid containing the Sf301 PhoQ intracellular domain which contains HK domain was constructed, because the main biology activity of PhoQ is depends on its HK domain [28]. To confirm whether these inhibitor candidates targeted the PhoQ HK domain, enzymatic activities of PhoQ were determined in the presence or absence of four compounds. The enzymatic activity of SF-PhoQc was measured using both a Pyrophosphate Reagent and a Luminescent Kinase Assay. The Pyrophosphate Reagent can reflect the reaction of HK and ATP at real time, but not sensitive. The Luminescent Kinase Assay is more sensitive than Pyrophosphate Reagent for kinase reaction but cannot reflect the reaction of HK and ATP at real time. Therefore, in the present study we used two assays to    confirm the results. The different IC 50 values of potential PhoQ inhibitors 1 and 3 determined by the two assays may be the sensitivity difference between the two assays.
By using cell invasion assays, the features of cell invasion process including penetration into epithelial cells and spreading to adjacent cells were tested [29]. The Shigella (Sf9380) were treated with four potential PhoQ inhibitors (100 mmol/L) for 4, 6 or 8 hours, respectively. Compared with cell invasion of the positive control Sf9380 alone, the potential PhoQ inhibitors (100 mmol/L) treated for 8 hours had obvious inhibition effects on the bacteria cell invasion by using gentamicin protection assay, while potential PhoQ inhibitors treated for 4 or 6 hours had no significant inhibition effects on Sf9380 cell invasion ( Figure S3). Therefore, Shigella cell invasion assay and Mouse Sereny test were carried out by the bacteria treated with potential PhoQ inhibitors for 8 hours.
To confirm these four potential PhoQ inhibitors were affecting PhoQ histidine kinase, we made a S. flexneri phoQ/phoP knock-out mutant (Sf301 DphoQ/phoP) and the cell invasion ability was tested. The results indicated that potential PhoQ inhibitors 1, 2, 3 can inhibit HeLa cell invasion ability of Sf301 ( Figure S4) but have no obvious effects on Sf301 phoQ/phoP knock-out mutant ( Figure 6C). It indicated that these potential PhoQ inhibitors can affect PhoQ histidine. The results also suggested that the cell invasion ability of Sf301 DphoQ/phoP decreased significantly compared to the wild type strain. It indicated that phoQ/phoP could regulate the cell invasion of S. flexneri.
The PhoQ of Salmonella is high homology to that of Shigella and in Salmonella PhoP/PhoQ regulates virulence including cell invasion [4,6]. So it was curious for us to evaluate whether these four potential PhoQ inhibitors have similar effect on Salmonella. The results of cell invasion suggested that cell invasion of Salmonella SL1344 was inhibited by the potential PhoQ inhibitors 1, 2, 3 while potential PhoQ inhibitor 4 has no obvious effect on cell invasion which were similar to the results of Shigella (Figure 7). It indicates that the potential PhoQ inhibitors can inhibit the cell invasion of Salmonella.
In the present study, we chose the PhoQ protein, a transmembrane sensor of the PhoQ/PhoP TCS in S. flexneri 2a 301, as the target and found that three potential PhoQ inhibitors can inhibit the bacterial ability to invade HeLa cells. Further, we found out that with phoQ/phoP knocking out, the cell invasion ability of Sf301 DphoQ/phoP decreased significantly, compared to the wild type strain, and no obvious effects of potential PhoQ inhibitors on Sf301 DphoQ/phoP were observed. However, Moss et al has reported that there were no significant differences between wildtype and phoP mutant of S. flexneri serotype 5 strain M90T in HeLa cell invasion [7]. In gram-negative pathogen, there are many cross-talks between two-component systems (TCSs) [30,31,32,33], in which one HK (Histidine Kinase) can regulate several RRs (Response Regulator), and one RR's phosphorylation can be regulated by several HKs to gather several signal pathways of TCSs and induce expression of a battery of downstream genes. The difference between phoQ/phoP knock-out mutant of S. flexneri 2a 301 and phoP knock-out mutant of S. flexneri 5 M90T may due to the cross-talks between TCSs in the bacterium, although in Shigella it remains poorly understood. Involvement of the PhoQ/ PhoP cascade on Shigella virulence across strains, serotypes and species need to be investigated in the future.
In the present study, four potential PhoQ inhibitors, at 200 mmol/L, showed no effect on Shigella growth ( Figure S2). This was expected because the PhoQ/PhoP signaling system does not directly regulate bacterial growth. With increasing knowledge about bacterial virulence, several researchers have found that bacterial virulence genes are essential to mount a harmful infection, but they are usually dispensable for growth of bacteria in vitro. These results indicate that inhibition of microbial virulence without inhibiting their growth may be a promising strategy [34]. In contrast, currently available antibiotics either kill bacteria or prevent their growth. Drugs that block disease without killing the pathogen bacteria [35,36] may cause less selective pressure for the generation of drug resistance [34,36,37]. These alternative drug strategies would presumably induce pathogen resistance at a much slower rate because the targeted non-essential genes or functions are under less selective pressure to mutate. The host will be subjected to intact avirulent bacteria, allowing the host to develop an adequate immune response against the pathogen. This would allow the host to efficiently respond to and eradicate an invader upon re-exposure. Therefore, the strategy to target bacterial virulence factors has become an attractive approach for the development of new therapeutic agents [35].
In addition to novel drug targets, the use of small organic molecules is gaining interest more than genetic-based drugs [38,39]. Small organic molecules that target specific proteins may be used for the prevention or treatment of infections caused by a wide variety of gram-negative bacteria species, including Escherichia coli [17,40], Salmonella typhimurium [18] and Yersinia pseudotuberculosis [41], as well as gram-positive bacteria such as Staphylococcus epidermidis [42]. In this study, four promising potential PhoQ inhibitor candidates were validated using enzymatic activity assays and binding affinities.
In previous studies, some potential PhoQ inhibitors displayed side effects, such as membrane damage or excessive protein binding, which would be an obstacle for their further development [43]. In this study, we found four potential PhoQ inhibitors that reduce the virulence of Shigella that also have low cytotoxicity and hemolysis of mammalian cells at their effective concentrations. We demonstrated that PhoQ/PhoP is a promising target for the development of new drugs against S. flexneri infection and proved that four potential PhoQ inhibitors can inhibit the virulence of Shigella. In future work, we will modify the compound structure to increase the efficacy of the potential PhoQ inhibitors and identify which stage of infection is inhibited by these potential inhibitors which is important to the therapy of shigellosis. Bacterial strains, growth conditions and reagents S. flexneri 2a 301 (Sf301, Genbank Accession number for chromosomes is AE005674) was kindly provided by Pr. Jianguo Xu (Chinese Center for Disease Control and Prevention, Beijing, China). A clinical strain of S. flexneri, Sf9380, was isolated from an epidemic in the Hebei province by Dr. Xuanyi Wang from our laboratory. E. coli ATCC 25922 was kindly provided by Dr. Bijie Hu (Zhongshan Hospital, Shanghai, China). S. enterica typhimurium SL1344 and plasmid pSB890 were kindly provided by Dr Daoguo Zhou (Purdue University, USA). E. coli JM109 and plasmid pQE30 were purchased from Qiagen (Germany). E. coli JM109 was used for both recombinant DNA manipulation and protein production. S. flexneri and E. coli were grown at 37uC in Luria-Bertani medium (LB; Oxoid, Wesel, Germany). All compounds used as inhibitor candidates were purchased from the SPECS Company (Netherlands). Stock solutions (10 mmol/L) of the potential PhoQ inhibitors were prepared in dimethyl sulfoxide (DMSO).

Ethics Statement
3D structure modeling of the PhoQ histidine kinase of S. flexneri Sf301 The PhoQ histidine kinase sequence of S. flexneri 2a strain 301 was retrieved from GenBank (accession number AE005674). To identify potential inhibitors of PhoQ histidine kinase (PhoQ HK) by virtual screening, we obtained the crystal structure of the PhoQ HK domain (PDB entry code: 1ID0) from the Brookhaven Protein Database (PDB http://www.rcsb.org/pdb). The missing residues in the PhoQ HK were added and subjected to energy minimization in Sybyl 6.8 [43] using the steepest descent method up to the gradient tolerance of 0.05 kcal/(mol?Å ) to relieve possible steric clashes and overlaps of side chains. The potential of the 3D structure of PhoQ HK domain was assigned according to the Amber force field with Kollman-united-atom charges encoded in Sybyl 6.8 [43].

Structure-based virtual screening of potential PhoQ HK inhibitors
The ATP-binding pocket formed by residues within a radius of 7 Å around the ATP site of the PhoQ HK catalytic domain was used as the target site for high-throughput virtual screening (HTVS). The optimized 3D model of PhoQ HK was used as the target for virtual screening with the SPECS database (http://www. specs.net). In the first step, 85,000 potential drug-like molecules were selected from the SPECS database, creating an in-house database (named SPECS_1). Next, the SPECS_1 database was searched for potential binding molecule structures using the program DOCK4.0 [44,45]. The most optimal 10,000 structures were subsequently re-scored using the FlexX [46] and CSCORE, a consensus scoring method that integrates five popular scoring functions. Two hundred molecules passed this highly selective filter. Finally, 100 molecules were manually selected as inhibitor candidates, according to their molecular diversity, their shape complementarities, and their potential for forming hydrogen bonds in the binding pocket of the PhoQ HK domain.

Expression and purification of recombinant SF-PhoQc
A DNA segment encoding the cytoplasmic domain of PhoQ HK was amplified by PCR from the Sf301 genomic DNA with primers PhoQc59 (59-CGCGGATCCGTACGAAACCTGAAC-CGATT-39) and PhoQc39 (59-CCCAAGCTTTTATTCATCT-TTCGGCGCAG-39), which have BamHI and HindIII restriction sites, respectively. The amplified DNA segment encodes PhoQ HK amino acids 251-486, which include the phosphoester and catalytic domains of PhoQ HK (SF-PhoQc). The amplified target DNA was purified, digested by BamHI and HindIII (TaKaRa Biotechnology Co., Ltd, China) and cloned into pQE30, creating pQE30-SF-PhoQc. The prokaryotic expression plasmid pQE30-SF-PhoQc was constructed to produce a PhoQc and N-terminal hexahistidine fusion protein. The pQE30-SF-PhoQc plasmid was transformed into E. coli JM109 and cultured overnight at 37uC in LB containing 100 mg/ml of ampicillin. When the bacteria reached OD600 absorbance values of 0.5-0.7, expression of SF-PhoQc was induced by adding 0.8 mmol/L isopropyl-b-Dthiogalacto-pyranoside (IPTG). Incubation continued under rigorous shaking at 30uC for 6 hours. Cells were harvested at 10,0006g for 30 minutes, resuspended in buffer-A (20 mmol/L Tris, pH 8.0, 300 mmol/L NaCl and 10 mmol/L imidazole), and disrupted by sonication on ice. The lysed bacteria were centrifuged to yield a clear supernatant, and the supernatant was loaded onto a column with Ni-NTA resin (Qiagen) pre-equilibrated in buffer A. The column was washed with buffer B (20 mmol/L Tris, pH 8.0, 300 mmol/L NaCl, 30 mmol/L imidazole) and eluted with buffer C (20 mmol/L Tris, pH 8.0, 300 mmol/L NaCl, 200 mmol/L imidazole). The eluted fractions were concentrated and dialyzed against buffer D (20 mmol/L Tris, pH 8.0, 150 mmol/L NaCl) to remove imidazole and stored at 270uC. The purified protein was identified by AutoFlex MALDI-MS (Bruker Daltonics, Inc., Billerica, MA, USA). The operating condition was optimized with standard solution, and the working parameters of the ion source were as follows: lasing light emitter N2, wave length 337 nm; cation with accelerating voltage 20 kV and matrix SA.

Enzymatic activity assays of SF-PhoQc with the potential PhoQ inhibitors
The enzymatic activity of SF-PhoQc was measured using both a Pyrophosphate Reagent (Sigma-Aldrich, MO, USA) and a Luminescent Kinase Assay (Promega, WI, USA). The Pyrophosphate Reagent was performed to confirm the inhibitory activities of the effective potential PhoQ inhibitors in the bioinformatics approach. All reactions were carried out in solid black, flatbottomed 96-well plates. The Pyrophosphate Reagent was performed at 28uC by monitoring the decrease of NADH by a decrease in absorbance readings at 340 nm (e340 = 6180 M 21 ? cm 21 ) in the presence of ATP. The final reaction mixture for the Pyrophosphate Reagent contained 25 mmol/L purified SF-PhoQc, 25 mmol/L KCl, 2.5 mmol/L MgCl 2 , 2 mmol/L ATP, 1 mmol/L phosphoendpyruvate, 0.35 mmol/L NADH, 1.5 U/ml PK, 1.25 U/ml LDH and 50 mmol/L Tris-Cl (pH 8.0). The enzyme activity was determined in the presence of various concentrations of potential PhoQ inhibitors (0-100 mmol/L) to investigate dose-dependent inhibition effects. The A 340 value was recorded by SoftMax vision 4.8 every minute, for 30 minutes. The rate of SF-PhoQc enzymatic activity inhibition by a compound was calculated by the following formula: Rp (Ratio of Phosphorylation) was defined as the ratio of the decrease activity of SF-PhoQc treated with compound vs the activity of SF-PhoQc treated without compounds. Rp was stand for the percentage of decrease activity of SF-PhoQc. IC 50 (the concentration of inhibition of 50%)values of potential PhoQ inhibitors were obtained by fitting the data to a sigmoid dose-response equation using Origin 7.0 software (OriginLab, Northampton, USA).
In the Luminescent Kinase assay, reactions were performed in solid black, flat-bottomed 96-well plates. To determine the IC 50 of potential PhoQ HK inhibitors, each compound underwent 2-fold serial dilutions with DMSO (0.1%) and was added into the reaction mixture. The final reaction mixture contained 4 mg of purified SF-PhoQc, 40 mmol/L Tris-Cl (pH 7.5), 16 mmol/L MgCl 2 , 0.8 mmol/L DTT and different concentration of 100, 50, 25, 12.5, 6.25 or 3.125 mmol/L compound, respectively. After incubation at 30uC for 25 minutes, 3 mmol/L ATP was added, and the reaction was incubated at 30uC for another 25 minutes. Next, an equal volume of Luminescent Kinase Reagent was added to each well, mixed and kept at room temperature for 10 minutes before recording the luminescence reading on a Fluoroskan Ascent FL machine (Thermo Scientific, USA). The rate of phosphorylation inhibition by a compound was calculated with the following formula: The IC 50 value of each compound was determined using the Origin 7.0 software (OriginLab, Northampton, USA).

Binding affinities of potential PhoQ inhibitors to SF-PhoQc
The binding affinities of potential PhoQ inhibitors to SF-PhoQc were determined in vitro using Surface Plasmon Resonance (SPR) technology with the dual flow cell Biacore 3000 instrument (Biacore AB, Uppsala, Sweden). The standard primary amine coupling reaction wizard immobilized SF-PhoQc to the hydrophilic carboxymethylated dextran matrix of the sensor chip, CM5 (Biacore). The SF-PhoQc covalently bound to the matrix was diluted in 10 mmol/L sodium acetate buffer (pH 4.2) to a final concentration of 0.035 mg/ml. Equilibration of the baseline was completed by a continuous flow of HBS-EP running buffer (10 mmol/L HEPES, 150 mmol/L NaCl, 3 mmol/L EDTA, and 0.005% (v/v) surfactant P20, pH 7.4) through the chip for 1-2 hours. All the Biacore data were collected at 25uC with HBS-EP as the running buffer at a constant flow of 30 ml/min. All the sensor grams were processed by using automatic correction for nonspecific bulk refractive index effects. The equilibrium constants (K D values) evaluating the protein-ligand binding affinities were determined by the steady-state affinity fitting analysis of the Biacore data.

Construction of the S. flexneri phoQ/phoP deletion mutant
The plasmid for phoQ/phoP deletion in S. flexneri 2a 301 was constructed. The upstream and downstream of the phoQ/phoP were amplified by using PCR with the primers listed in the Table 3: phoQ/phoP U-F and phoQ/phoP U-R were used to amplify upstream fragment; phoQ/phoP D-F and phoQ/phoP D-R were used to amplify downstream fragment. The amplified upstream fragment of phoQ/phoP was digested with NotI and XbaI restriction endonucleases, ligated into NotI and XbaI-cut pSB890 vector (a suicide vector [47]), resulting in a plasmid name as pSB890up. The amplified downstream fragment of phoQ/phoP was digested with XbaI and SmaI restriction endonucleases and ligated into XbaI and SmaI-cut pSB890up vector, resulting in a plasmid pSB890 phoQ/phoP with the upstream and downstream fragments of phoQ/phoP. The plasmid pSB890 phoQ/phoP was transferred into E. coli SM10lpir cells, then pSB890 phoQ/phoP was introduced into S. flexneri 2a 301 by conjugation. Equal volumes (5 ml) of S. flexneri 2a 301 wild-type recipient and E .coli SM10lpir contain pSB890 phoQ/phoP were mixed, and 100 ml was spotted onto an LB agar plate. After incubation (37uC for 24 h), the bacteria were recovered, resuspended in 1 ml of Luria-Bertani (LB) broth, and cultured on LB agar plate containing ampicillin and streptomycin at 37uC. Individual ampicillin-resistant S. flexneri colonies were inoculated into LB broth, grown overnight at 22uC, and plated onto LB agar supplemented with 5% sucrose. Sucrose-tolerant colonies were screened using PCR with primers 59-CCC-CGCTGGTTTATTTAATGTTTA-39 and 59-ACCCGCCTG-TATATTTCTTGTGTC-39. A Sf301 DphoQ/phoP was screened out and identified by PCR and sequencing.

S. flexneri cell invasion assay
Bacterial ability to invade HeLa cells was tested with a gentamicin protection assay [9]. HeLa cells were grown in sixwell tissue culture plates to semi-confluent monolayers (DMEM, 10% FCS, 5% CO 2 ). Sf9380 or Sf301 DphoQ/phoP was treated with potential PhoQ inhibitors for 8 hours in LB at 37uC (final concentrations of 100,50,25,12.5 or 6.25 mmol/L) prior to infection. The bacteria pre-treated with potential PhoQ inhibitor without washing, were added to semi-confluent HeLa cells at a multiplicity of infection of 100, at the same time the bacteria were diluted and plated onto LB agar plates. The CFUs were counted as initial bacterial number before added into HeLa cells. Then the plates were centrifuged at 9006g for 5 minutes. After incubating at 37uC for 60 minutes (For testing intracellular growth and spreading to adjacent cell, cells were further incubated for 290 min [9].), the cells were washed three times with PBS, and gentamicin was added to the medium to a final concentration of 10 mg/ml for 20 minutes at 37uC. After the incubation, the HeLa cells in each well were lysed in 1 ml of PBS containing 0.1% Triton X-100 for 10 minutes at room temperature. The lysates were diluted and plated onto LB agar plates in triplicate. Colonies grown on LB plates were counted. The results of the assays were expressed as the number of bacteria recovered from gentamicintreated cells divided by the number of inoculated bacteria added to the cell. The cells were inoculated with untreated Sf9380 or Sf301 DphoQ/phoP (treated without DMSO) as a positive control and E. coli ATCC 25922, an avirulent strain treated without DMSO, as a negative control. DMSO is the solvent of the potential PhoQ inhibitor compounds; therefore, Sf9380 was also treated with the same amount of 0.1% DMSO as the other groups to control for any DMSO-specific effects. The cell invasion assay was performed in triplicate for each potential PhoQ inhibitor, and the assay was repeated twice.

Salmonella cell invasion assay
The gentamicin protection assay was used as a cellular model to evaluate the inhibitory effect of the four potential PhoQ inhibitors. Salmonella enterica typhimurium SL1344 was treated with each of the four potential PhoQ inhibitors (final concentrations: potential PhoQ inhibitor 1 was 12.5 mmol/L, 2 was 25 mmol/L, 3 was 100 mmol/L and 4 was 100 mmol/L) and grown for 8 hours. The bacteria pre-treated with potential PhoQ inhibitor without washing, were added to HeLa cells at a multiplicity of infection of 100, at the same time the bacteria were diluted and plated onto LB agar plates. The CFUs were counted as initial bacterial number before added into HeLa cells. Next, the bacteria were inoculated to HeLa cells and incubated with for 60 minutes prior to the addition of gentamicin to kill extracellular bacteria. After the incubation, the HeLa cells in each well were lysed in 1 ml of PBS containing 0.1% Triton X-100 for 10 minutes at room temperature. The lysates were diluted and plated onto LB agar plates in triplicate. Colonies grown on LB plates were counted. The results of the assays were expressed as the number of bacteria recovered from gentamicin-treated cells divided by the number of inoculated bacteria added to the cell. After lysates of HeLa cells were plated, the number of CFU that formed on LB agar were counted. The cells were inoculated with untreated SL1344 (treated without DMSO) as a positive control and E. coli ATCC 25922, an avirulent strain treated without DMSO, as a negative control. SL1344 was treated with the same amount of 0.1% DMSO as the other groups to control for any DMSO-specific effects. The cell invasion assay was performed in triplicate for each potential PhoQ inhibitor, and the assay was repeated twice.

S. flexneri Mouse Sereny test
A Mouse Sereny test was used to evaluate the virulence of S. flexneri Sf9380 [48]. A single red colony of Sf9380 on Congo Red agar (Tryptic soy broth provided by Oxfoid Co. with 1.5% agar and 0.01% Congo red) [48], was inoculated into LB and grown with or without potential PhoQ inhibitors (final concentration 100,50,25,12.5 or 6.25 mmol/L) at 37uC for 8 hours with constant shaking. The female BALB/c mice (about 18 g) were infected with 1610 8 CFUs per eye. The eye of the mouse was inoculated with the bacterium with or without potential PhoQ inhibitor treatment (n = 6 mice in each group). Keratoconjunctivitis in the mice infected with bacteria was observed at 24, 48, 72 and 96 hours after inoculation. In the experiment, mouse eyes inoculated with untreated Sf9380 (treated without DMSO) served as the positive control for keratoconjunctivitis, and mouse eyes inoculated with E. coli ATCC 25922 (treated without DMSO) served as the negative control. Because DMSO is the solvent of the potential PhoQ inhibitor, Sf9380 was treated with the same amount of DMSO (0.1%) as the other groups. The invasiveness of the bacteria was scored: no disease or mild irritation as ''2'', + as mild conjunctivitis or late development and/or rapid clearing of symptoms, ++ as keratoconjunctivitis without purulence, and +++ as fully developed keratoconjunctivitis with purulence [49].

Cytotoxicity assay of the potential PhoQ HK inhibitors
Cytotoxicity of the potential PhoQ inhibitors on cultured Vero cells was measured by using the Cell Proliferation Kit I (MTT) (Roche, Indianapolis, USA). An equal concentration of DMSO (0.1%) in the medium was used as a negative control. The cytotoxicity assay was performed in triplicate for each potential PhoQ inhibitor, and the assay for each potential PhoQ inhibitor was repeated twice. The results were converted to percentage of the control (cells only containing 0.1% DMSO) and CC 50 (concentrations that produce a 50% cytotoxicity effect on Vero cell) of each potential PhoQ inhibitors was calculated using the Origin v7.0 software (OriginLab, Northampton, USA).

Erythrocyte hemolysis assays of potential PhoQ inhibitors
The hemolytic activity of the potential PhoQ inhibitors was determined using human erythrocytes [50]. The erythrocytes were washed three times with sterile saline and resuspended to 5% cell concentration prior to assay. Then, 200 ml erythrocyte suspensions containing 25 mmol/L or 100 mmol/L concentrations of each potential PhoQ inhibitors were added in triplicate to the wells of the 96-well microtiter plates. Cells without potential PhoQ inhibitors treatment or with 1% Triton-100 treatment were used as the zero or 100% hemolysis control, respectively. The cell suspensions were incubated for 1 hour at 37uC and centrifuged for 10 minutes at 12006g. Supernatants (100 ml) were transferred to another sterile plate, and the released hemoglobin was measured absorbance readings at 570 nm. An equal concentration of DMSO (0.1%) in the medium did not affect the integrity of the erythrocyte membrane. The hemolysis assay was repeated twice. The gentamicin protection assay shows the inhibitory effect of the four potential PhoQ inhibitors on Shigella HeLa cells invasion. The S. flexneri Sf9380 were treated with each of the four potential PhoQ inhibitors (100 mmo/L) grown for 4, 6 or 8 hours. Then the bacteria were inoculated with HeLa cells for 60 minutes prior to the addition of gentamicin to kill extracellular bacteria. After the incubation, the HeLa cells in each well were lysed in 1 ml of PBS containing 0.1% Triton X-100. The lysates were diluted and plated onto LB agar plates in triplicate. Colonies grown on LB plates were counted. The results of the assays were expressed as the number of bacteria recovered from gentamicin-treated cells divided by the number of inoculated bacteria added to the cell. Result indicated that compared with cell invasion of the positive control Sf9380 alone, the potential PhoQ inhibitors (100 mmol/L) treated for 8 hours had obvious inhibition effects on the bacteria cell invasion, while potential PhoQ inhibitors treated for 4 or 6 hours had no significant inhibition effects on Shigella cell invasion. (TIF) Figure S4 The effects of potential PhoQ inhibitors on cell invasion of Shigella Sf301. The gentamicin protection The underlined sequences represent the restriction sites. Primers phoQ/phoP U-F and phoQ/phoP U-R were used to amplify upstream fragment of phoQ/phoP; phoQ/ phoP D-F and phoQ/phoP D-R were used to amplify downstream fragment of phoQ/phoP. doi:10.1371/journal.pone.0023100.t003 assay shows the inhibitory effect of the four potential PhoQ inhibitors on S. flexneri 2a 301 HeLa cells invasion. The bacteria were treated with each of the four potential PhoQ inhibitors grown for 8 h. Then the bacteria were inoculated with HeLa cells for 60 minutes prior to the addition of gentamicin to kill extracellular bacteria. After the incubation, the HeLa cells in each well were lysed in 1 ml of PBS containing 0.1% Triton X-100. The lysates were diluted and plated onto LB agar plates in triplicate. Colonies grown on LB plates were counted. The results of the assays were expressed as the number of bacteria recovered from gentamicin-treated cells divided by the number of inoculated bacteria added to the cell. Values are means 6 standard deviations from 6 independent wells. *p,0.01 vs. Sf301. The results indicated that potential PhoQ inhibitors 1, 2, 3 can inhibit HeLa cell invasion ability of Sf301 while potential PhoQ inhibitor 4 had no effect on Sf301 invasion of HeLa cells.