The periplasmic enzyme, AnsB, of Shigella flexneri modulates bacterial adherence to host epithelial cells.

S. flexneri strains, most frequently linked with endemic outbreaks of shigellosis, invade the colonic and rectal epithelium of their host and cause severe tissue damage. Here we have attempted to elucidate the contribution of the periplasmic enzyme, L-asparaginase (AnsB) to the pathogenesis of S. flexneri. Using a reverse genetic approach we found that ansB mutants showed reduced adherence to epithelial cells in vitro and attenuation in two in vivo models of shigellosis, the Caenorhabditis elegans and the murine pulmonary model. To investigate how AnsB affects bacterial adherence, we compared the proteomes of the ansB mutant with its wild type parental strain using two dimensional differential in-gel electrophoresis and identified the outer membrane protein, OmpA as up-regulated in ansB mutant cells. Bacterial OmpA, is a prominent outer membrane protein whose activity has been found to be required for bacterial pathogenesis. Overexpression of OmpA in wild type S. flexneri serotype 3b resulted in decreasing the adherence of this virulent strain, suggesting that the up-regulation of OmpA in ansB mutants contributes to the reduced adherence of this mutant strain. The data presented here is the first report that links the metabolic enzyme AnsB to S. flexneri pathogenesis.


Introduction
Shigellosis, more commonly known as bacillary dysentery, is a diarrhoeal infirmity ensuing from an acute inflammatory reaction in the gastrointestinal (GI) tract caused by the enteric bacterium Shigella. There are four recognized species of the genus Shigella, namely S. flexneri, S. dysenteriae, S. boydii, and S. sonnei [1]. Of the four species, S. flexneri strains are most frequently linked with endemic outbreaks of shigellosis in the developing world. S. flexneri invades the colonic and rectal epithelium of its host and causes severe tissue damage, which manifests in a spectrum of clinical symptoms ranging from watery diarrhea to severe dysentery characterized by fever, abdominal cramping and bloody mucoid stool [2]. Most cases of shigellosis are self-limiting, however in certain individuals, especially malnourished infants, untreated disease can result in several complications which can be fatal [3].
Although a lot has been unraveled about the various steps involved in the pathogenesis of S. flexneri, a complete picture of the bacterial genes that are expressed exclusively during infection of host cells has not been elucidated. A clear understanding of bacterial genes expressed at various host-pathogen interfaces can pave the way to the identification of potential attenuation targets and vaccine candidate antigens. In an attempt to identify immunogenic proteins expressed during infection, Jennison and colleagues [4] used 2-dimensional gel electrophoresis  followed by immunoblotting with sera from convalescent shigellosis patients, to screen the soluble and membrane protein profiles of S. flexneri 2457T. In their study, the chromosomally encoded, metabolic, periplasmic enzyme, L-asparaginase (AnsB) was iden-tified to be immunogenic. This protein is of interest as AnsB has been found to be antigenic in a number of other enteric bacterial species [5][6][7][8], and has not been characterized in S. flexneri, thereby making it a potential attenuation target for vaccine development.
The E. coli L-asparaginase (AnsB) protein, which shares 99.1% amino acid identity with S. flexneri AnsB, is an enzyme that converts asparagine to aspartate and ammonia. AnsB has high affinity for L-asparagine and the expression of ansB is induced by anaerobiosis and positively regulated by the cyclic AMP receptor protein (CRP) and the product of the fnr gene [9].
In this study, a reverse genetic approach was used to characterize AnsB in S. flexneri and determine if this enzyme is required for bacterial pathogenesis. Recombineering, a lambda red-mediated PCR-based approach [10], was used to knockout the ansB gene from a virulent, wild type serotype 3b strain of S. flexneri. The effect of ansB mutation on bacterial physiology and virulence was analyzed using in vitro growth studies, cell culture assays and two in vivo models of shigellosis, namely the murine pulmonary model and the C. elegans infection model, a new potential animal model of shigellosis [11]. The results of this study indicate that the activity of AnsB is required during the initial stages of bacterial pathogenesis, as in vitro bacterial adherence assays clearly showed that ansB mutants had decreased adherence to host cells. To gain further insight into how AnsB affects bacterial adherence, we screened the proteomes of the ansB mutant and wild type S. flexneri strains and found that mutant cells up-regulate the expression of the prominent outer membrane protein, OmpA. These findings highlight the importance of AnsB in the pathogenesis of S. flexneri, paving the way to designing new therapeutic or vaccine strategies.

Materials and Methods
All mice experiments were performed under the protocol approved by the Australian National University Experimentation Ethics Committee.

Bacterial strains and growth conditions
All S. flexneri strains used were derivatives of the highly virulent S. flexneri 3b strain, SFL1520, obtained from the International Centre for Diarrhoeal Diseases Research Bangladesh (ICDDRB). Bacterial cultures were grown routinely in Luria Bertani (LB) broth or on LB agar plates. Where appropriate, antibiotic selection was with ampicillin (100 mg/mL) and/or kanamycin (50 mg/mL). For growth studies under nutrient stress, Minimal Essential media (MM) was used. E. coli strains were grown and maintained at 37uC, while S. flexneri cultures were grown at 30uC to maintain the virulence plasmid.

Recombinant DNA methods
Restriction endonucleases (New England Biolabs), alkaline phosphatase (Fermentas) and T4 DNA ligase (Promega) were used in accordance with the manufacturer's specifications. Plasmids were maintained in the E. coli strain JM109 [12] and isolated using the alkaline lysis method [13] or the QIAprep Spin MiniPrep kit (Qiagen). All primers were synthesized by Sigma-Aldrich. All sequencing was performed at the Biomolecular Resource Facility, Australian National University. PCR was carried out using the high fidelity PfuUltraII polymerase (Stratagene) for cloning and sequencing or iTaq (Scientifix) for screening.

Construction of mutant strains
Recombineering was used to generate knockouts in SFL1520. A linear knockout template consisting of the kanamycin resistance gene flanked by ,1000 bp regions showing homology to the upstream and downstream regions of each of the target genes was generated using PCR. Electrocompetent cells of the wild-type SFL1520, transformed with either of the helper plasmids pKD46 or pKM208 [10], were prepared. Freshly prepared electrocompetent cells were transformed with 1-2 mg of each purified knockout template. Transformed cells were recovered in Super Optimal Broth with Catabolite repression (SOC) [14] for 4 hours to allow for homologous recombination. Recovered cells were plated on LB agar containing 50 mg/mL kanamycin and incubated at 30uC for up to 3 days. Resulting colonies were routinely restreaked onto fresh LB agar plates containing 50 mg/mL kanamycin to eliminate false positives. Colonies were then screened for successful gene disruption by colony PCR and sequencing, after which the presence of the virulence plasmid in both mutant strains was confirmed using PCR to amplify two virulence plasmid indicator genes, apyI and virG. These genes were chosen as they lie on opposite sides of the 220 kb virulence plasmid.

Growth studies
Overnight cultures of strains were diluted 1:20 in LB as well as MM and sampled at 30 minute intervals. Growth studies were carried out at 30uC and 37uC under aerobic conditions.

SDS PAGE and Western Blot
The secretion of type 3 effector proteins was induced using congo red (final concentration 10 mM) for 30 minutes. Bacterial cells were harvested by centrifugation and the supernatants were passed through 0.45 mm filters to eliminate cellular debris. Secretory proteins were precipitated using 25% trichloroacetic acid (TCA) in acetone. Protein precipitates were washed twice using acetone and pellets were allowed to dry completely, after which they were resuspended in solubilization buffer (9 M urea, 4% CHAPS, 1% DTT, 1% ampholytes, 35 mM tris base). Protein concentration was estimated using the BCA kit (Pierce) according to the manufacturer's instructions and 20 mg of proteins in the supernatant samples were loaded onto 12% (v/v) SDS-PAGE gels for Western blots. The gels were stained with Coomassie Brilliant blue R250 to ensure equal loading of samples. Once samples were equalized the separated proteins were transferred onto Hybond-P PVDF membranes (Millipore) and blocked in 5% skimmed milk in Phosphate Buffered Saline (PBS) at 4uC overnight. Following blocking, the membranes were washed thrice in PBS containing 0.05% (v/v) Tween-20 (PBST) (Sigma). Membranes were then incubated with the primary antibodies (Anti-IpaB and Anti-IpaD antisera generated in-house in mice) for 2 hours. Unbound antibodies were washed off using three washes with PBST as outlined above, following which the membranes were incubated with the secondary antibodies (Anti-mouse IgG; Sigma) for 1 hour. Membranes were washed thrice as done previously. The binding of antibodies was then detected by chemiluminescence using SuperSignal West Pico Chemiluminescent Substrate (Pierce) as described by the manufacturer. Chemiluminescence was then detected using the Fisher Biotec chemiluminescence system.

Reverse transcription polymerase chain reaction (RT-PCR) and Quantitative Real-time Reverse transcription polymerase chain reaction (qRT-PCR)
For both RT-PCR and qRT-PCR, RNA was isolated using Trizol reagent (Invitorgen) according to the manufacturer's instructions. Briefly, overnight cultures of S. flexneri strains, (SFL1520, SFL2283 or SFL2443) were diluted 1:50 and grown to log phase at 37uC. 1610 9 CFU were harvested and treated with 1 mL Trizol reagent followed by incubation at room temperature for 5 minutes. Samples were treated with 200 ml chloroform to separate RNA from DNA and proteins. RNA in the aqueous phase was washed with chloroform and precipitated using 2 volumes of isopropanol. RNA pellets were collected by centrifugation and washed using 1 mL freshly prepared 75% ethanol. RNA pellets were air-dried after ethanol wash and suspended in 50 ml of nuclease-free water. The isolated RNA was cleaned up using the Qiagen RNeasy Kit; this step was performed to remove any residual DNA and organic salts. Isolated RNA was treated with Turbo DNase (Ambion) to minimise any contaminating genomic DNA. cDNA was made from 200 ng of RNA using Superscript II (Invitrogen) according to the manufacturer's instructions. qRT-PCR was performed on the cDNA samples using the power SYBR Green RT-PCR kit (Applied Biosystems) according to the manufacturer's instructions except that primers were used at a final concentration of 0.4 mM and the final reaction volume was reduced to 10 ml. Expression of hisG was used as a control to normalize the expression of all genes studied. All Real-Time RT-PCRs were performed in triplicate with a no reverse transcriptase (NRT) and a no template control (NTC) set up for each run. Reactions were run in a Rotor-Gene Q Real-Time cycler (Qiagen).

Measurement of asparaginase activity
Asparaginase activity of SFL1520 and SFL2283 was measured using the ammonia assay kit (Sigma) with log phase culture supernatants, according to the manufacturer's instructions. Bacterial cells expressing L-asparaginase would break down asparagine with the release of ammonia, thus increased ammonia levels correlate with increased hydrolysis of asparagine. Briefly, fresh log phase cells were harvested and washed once in PBS, and 3610 8 colony forming units (CFU) of bacteria was resuspended in 1 ml PBS with 5 mM asparagine. Aliquots were taken after 30, 60, 90 and 180 minutes of incubation at 37uC and centrifuged to collect bacterial cells. The ammonia concentration in the cell-free supernatants was measured using the ammonia assay kit, according to the manufacturer's guidelines.
In vitro adherence assay BHK cells in 6-well plates were infected with 4610 8 CFU of each bacterial strain. Bacterial cells were allowed to infect the monolayer for 90 minutes at 37uC, 5% CO 2 . Unbound bacterial cells were washed off the BHK monolayer using PBS, after which the mammalian cells were lysed using 0.05% (v/v) Triton-X in PBS. Appropriate dilutions of the cell lysates were plated on LB agar plates carrying the appropriate antibiotics to determine the number of adherent bacterial cells.

Coverslip adherence assay
The coverslip adherence assay was performed based on the protocol developed by Cravioto et al, [15]. Briefly, overnight cultures of all S. flexneri strains grown at 30uC were diluted and grown to log phase at 37uC (OD 600 = 0.6-0.8). 4610 8 CFU of each strain was used to infect a confluent layer of BHK cells grown on coverslips placed within wells of 6-well culture plates. The plates were incubated for 90 minutes at 37uC to allow for infection. The coverslips were then washed thrice in PBS to remove any unbound bacterial cells. Samples were fixed using fresh 70% (v/v) methanol and stained for 30 minutes using 10% (v/v) Giemsa stain. Excess stain was washed off. Coverslips were mounted on glass slides and examined using an oil immersion lens.

C. elegans strains and growth conditions
The C. elegans wild type N2 strain [16] was used in this study. Nematodes were maintained at 22uC on modified nematode growth medium (NGM) agar medium seeded with E. coli OP50 [16]. A synchronous population of C. elegans was established by isolating eggs from adult worms through bleaching [17]. After hatching, L1 worms were transferred to NGM agar plates seeded with E. coli OP50 and allowed to grow at 22uC until they reached the L4 stage.

C. elegans bacterial accumulation
All bacterial strains used for the accumulation assay were grown overnight at 37uC on NG agar medium to stimulate expression of virulence plasmid-encoded genes [11]. Plates were cooled to room temperature before they were inoculated with 50-100 nematodes (L4s). The worms were allowed incubated at 22uC for 24 hours after which 10 worms were picked and washed thoroughly using sterile S-basal with 1 mM sodium azide. Worms were then treated with 200 mg/mL gentamycin for 3 hours to remove surfacebound bacterial cells, after which they were washed thoroughly to remove any residual antibiotic. Washed worms were suspended in S-basal +0.1% Triton-X and lysed by mechanical disruption using glass beads as described by [18]. Appropriate dilutions of the lysates were plated onto LB agar with the appropriate antibiotics, to obtain gut bacterial counts. In order to visualize bacterial accumulation within nematode guts, worms were fed S. flexneri strains tagged with GFP + . Following 24 hours of infection, bacterial fluorescence was observed using the EVOS digital inverted microscope (AMG) and quantified using the Infinite M1000 Pro (Tecan).

C. elegans survival assays
Overnight cultures of S. flexneri strains maintained at 30uC were diluted 1: 50 in LB media and grown to log phase at 37uC. A synchronized population of L4 stage C. elegans worms was treated with 200 mg/mL gentamycin for 3 hours, after which they were washed thoroughly with S-basal to remove any residual antibiotic. Approximately 20 washed L4s were transferred into each well of a 24-well plate containing 100 mL of the appropriate log-phase bacterial culture to be tested. The volume of solution in each well was adjusted to 500 mL with S-basal and plates were incubated at 22uC. The number of live worms in each well was scored every 12 hours and the percentage survival was calculated. Worms that showed no pharyngeal pumping and remained immobile on tapping the plate were considered dead. A minimum of three replicates for each test strain was set up per trial.
Mouse pulmonary model of shigellosis [6][7][8] week old female Balb/c mice weighing approximately 20-25 g (Animal Resource Centre, WA) were inoculated intranasally with a sub-lethal dose (2610 7 CFU/10 mL) of either the pathogenic wild type SFL1520 strain or SFL2283 (DansB). The mice were lightly anesthetized with isoflurane prior to inoculation. 20 mice were euthanized 24 hours post infection and their lungs were extracted and homogenized in PBS with 0.05% Triton-X. Appropriate dilutions of the homogenates were plated on LB agar with antibiotics, where appropriate, to obtain the live counts of intracellular bacteria per lung.

Isolation of total bacterial protein for two-dimensional gel electrophoresis
Overnight cultures of SFL1520 (wild type parent) and SFL2283 (DansB) grown in LB at 30uC were diluted 1/100 and grown to log phase at 37uC. Bacterial cells were collected and washed thrice using cell wash solution (10 mM Tris pH 8.0, 5 mM magnesium acetate). Bacterial pellets were resuspended in 500 ml of cell lysis buffer containing 7 M urea, 2 M thiourea, 30 mM tris-base, 4%(w/v) CHAPS and Complete Protease Inhibitor (Roche) and incubated on ice for 10 minutes. Subsequently, the suspensions were sonicated 10 times for 15 s with 30 s intervals, on ice. Cellular debris was sedimented by centrifugation. Supernatants were treated with 10 volumes of ice-cold acetone and incubated at 220uC overnight to precipitate proteins. The protein pellets were washed thrice using ice-cold acetone, following which they were dried. Acetone-free proteins were solubilized in 200 ml of cell lysis buffer and pH was adjusted to 8.5. For 2-dimensional (2-D) gel electrophoresis, total protein from wild type and ansB mutant strains were isolated in triplicate over three independent experiments. Protein concentration was determined using the BCA kit (Pierce) and 500 mg of each protein sample was labeled with fluorescent dyes Cy3 or Cy5 (GE Healthcare); an internal standard, consisting of a mixture of 250 mg of each sample, was labeled with Cy2. Samples were labeled with the CyDye DIGE Fluors (minimal dyes) for Ettan DIGE kit (GE Healthcare), according to the manufacturer's instructions, with minor modifications as described by Mathesius et al, 2001. Briefly, 500 mg of each protein sample was labeled with 400 pmol amine reactive cyanine freshly dissolved in anhydrous dimethyl formamide. CyDye labeling was carried out on ice and in the dark for 30 minutes. Each reaction was terminated using 10 nmol lysine to eliminate any unbound dye. Labeled samples were treated with DTT (20 mg/ml) and Bio-ampholytes (50 ml/ml). To identify differences in the proteomes of SFL1520 and SFL2283, 500 mg of total protein from the wild type parent (SFL1520) labeled with Cy3 or Cy5 was combined with 500 mg of oppositely labeled proteins from the DansB mutant strain. Each wild type-mutant pair was then mixed with 500 mg of the combined protein preparations of all protein samples that were labeled with Cy2 as an internal control.

2-D electrophoresis
2-D electrophoresis was performed in darkness to maintain the stability of the Cy dyes. Immobiline pH 3-10 NL Drystrips (24 cm, GE Healthcare) were used for the first dimension isoelectric focusing (IEF). The strips were rehydrated overnight in rehydration solution containing, 8 M urea, 0.5% (w/v) CHAPS, 0.2% (w/v) DTT, 0.52% (w/v) bio-ampholytes and 0.6% (w/v) bromophenol blue. Wild type-mutant pairs with internal controls were loaded onto rehydrated Immobiline strips and IEF was carried out in a Multiphor II electrophoresis system (GE Healthcare) at 20uC for a total of 35,000 volt hours [19].
For separation of proteins across the second dimension, self-cast 12.5% SDS-PAGE gels were cast using the EttanDALTsix system (GE Healthcare). The gels were cast using low fluorescence glass plates, which are compatible with visualizing the CyDyes. Focused first dimension strips were equilibrated as described by [19] and placed on the second dimension gels. SDS-PAGE was carried out at 10uC in SDS running buffer (25 mM Tris, pH 8.0, 192 mM glycine and 0.1% SDS) at 600 V, 10 mA, and 2.5 W per gel for the first hour; 600 V, 40 mA, and 13 W per gel until the bromophenol blue front reached the bottom of the gel.

Gel imaging and image analysis
After the second dimension electrophoresis, DIGE-labeled proteins were visualized using a Typhoon Trio laser scanner (GE Healthcare). Gels were scanned with the specific excitation wavelengths of Cy3 (532-nm laser and a 580-nm band pass 30 emission filter), Cy5 (633-nm laser and a 670-nm band pass 30 emission filter) and Cy2 (488-nm laser and a 580-nm band pass 40 emission filter). Spot detection and analysis was carried out using DeCyder Version 5 (GE Healthcare) software followed by careful manual confirmation and rematching of matching errors. Statistics and identification of differentially expressed spots were carried out in the DeCyder DIA and BVA modules (one-way ANOVA).

In-gel trypsin digestion of protein spots and mass spectrometry
After analysis of gel images, identified spots of interest were excised from the 2D-DIGE gels using an Ettan spot picker (GE Healthcare). The spot picker was calibrated and tested several times prior to spot excision. In-gel trypsin digest was carried out as described by Mathesius et al, 2001 with a few modifications. Briefly, excised protein spots were washed four times in acetonitrile: 50 mM ammonium bicarbonate (50:50, v/v). Spots were then dried in 100% acetonitrile for 30 minutes following which they were air-dried to eliminate all acetonitrile. Gel pieces were then rehydrated with a trypsin solution (20 units; Promega) and incubated for 2 hours at 4uC followed by overnight incubation at 37uC. Peptides were then extracted from the gel pieces using an extraction buffer consisting of acetonitrile: water: trifluroacetuic acid (TFA) (50%:50%:1%, v/v) followed by gentle sonication in a sonic water bath for 40 minutes. This extraction step was performed twice using a reduced volume of extraction buffer and 20 minutes sonication. Peptides were collected and dried completely to remove all traces of TFA. Dried peptides were resuspended in 20 ml of acetonitrile:water:formic acid (10%:89.9%:0.1%, v/v) of which 6 ml were loaded onto the mass spectrometer. The peptides were identified at the Mass Spectrometry Facility, The Australian National University, on an Agilent 6530 Q-TOF LC/MS (Santa Clara, CA, USA) with a ChipCube ion source interface (Agilent Technologies, Inc., Palo Alto, CA) containing a liquid chromatographic chip (ProtID-Chip-150(II), separation: 150 mm675 mm, enrichment: 4 mm 40 nL, packed with 5 mm Zorbax 300SB-C18 particles) The LC separation system included a binary capillary pump operated at a flow rate of 4 mL/min, used for loading the samples, and a nanoflow gradient pump using a linear gradient from 8 to 38% mobile phase B in 47 min at a flow rate of 300 nL/min. The column was then washed with 90% mobile phase B for 5 min. Mobile phase A 0.1% formic acid and mobile phase B was 90% acetonitrile/water containing 0.1% formic acid.
The spray from the chip was subjected to positive polarity electrospray ionisation (ESI) using the following settings: gas flow rate 4 L/min, gas temperature 300uC, capillary voltage 1900 V, fragmentor 175 V, skimmer 65 V and octopole RF peak 750 V.

Results
Insertional inactivation of ansB in the wild type S. flexneri serotype 3b strain The ansB gene was successfully disrupted in wild type S. flexneri serotype 3b (SFL1520) using recombineering to produce DansB (SFL2283). Excessive handling and manipulation of S. flexneri strains often results in the loss of the virulence plasmid, therefore the presence of the virulence plasmid in S. flexneri serotype 3b (SFL1520) and DansB (SFL2283) was confirmed by PCR before every virulence assay performed in this study as described in the methods.
S. flexneri virulence-plasmid encoded invasion protein antigens (Ipa) are important components of the type three secretion needle complex, required for virulence. In order to determine whether DansB (SFL2283) efficiently expressed and secreted virulenceplasmid encoded proteins, we investigated the expression and secretory levels of IpaB and IpaD, two key components of the type three secretory apparatus, using western immunoblots. Western immunoblots showed no differences in the expression and secretion of IpaB and IpaD in DansB (SFL2283) and the wild type serotype 3b strains (SFL1520) (Figure S1), thus suggesting that expression and secretion of key virulence plasmid-encoded proteins remain unaffected by the ansB mutation.
Prior to carrying out any physiological studies, we confirmed that the ansB gene was expressed in wild type serotype 3b (SFL1520) when grown in vitro under nutrient rich and starved conditions, using reverse transcriptase PCR (RT-PCR) (data not shown). AnsB is an identified L-asparaginase in E. coli; to determine if this function was conserved in S. flexneri, we used an ammonia assay to measure levels of ammonia released when bacterial cells were treated with asparagine. Ammonia assays showed that wild type S. flexneri serotype 3b (carrying intact ansB) released significantly more ammonia than the DansB strain ( Figure 1). These results suggest that AnsB in S. flexneri functions as an L-asparaginase.

The activity of AnsB is not essential for bacterial growth in vitro
Once we established that the ansB gene was expressed and functional in S. flexneri serotype 3b when grown in vitro, we carried out in vitro growth studies to determine whether the metabolic activity of AnsB was required for bacterial growth. No mutationrelated growth retardation was observed when strains were grown under nutrient rich (LB media) conditions at both 37uC and 30uC ( Figure S2.A). To determine if ansB expression was required for growth under stress conditions, growth studies were carried out under reduced nutrient conditions using minimal essential salts media (MM) ( Figure S2.B). No significant (p.0.05, unpaired t-test) growth retardation was exhibited by the ansB mutant strain, suggesting that, although this metabolic gene is expressed, its activity is not vital for bacterial growth in vitro.
AnsB activity is required for S. flexneri adherence to baby hamster kidney (BHK) cells in vitro adherence assays were performed to investigate whether the activity of AnsB contributed to bacterial adherence to epithelial cells. The DansB (SFL2283) strain showed a significant drop in the number of adherent bacteria when compared with the wild type parental strain (SFL1520) (Figure 2.A). This decrease in adherence was restored upon complementation of the ansB mutation in SFL2309 (SFL2283 + plasmid expressing ansB).
In order to visualize the differences in adherence patterns of the mutant and wild type strains, a coverslip adherence assay was performed. Giemsa staining of infected BHK monolayers also showed a decrease in the number of adherent bacteria in monolayers infected with DansB (SFL2283) (Figure 2.C) compared with wild type S. flexneri 3b (SFL1520) (Figure 2.B) and the ansB complemented strain SFL2309 (Figure 2.D). The number of adherent bacteria per BHK cell was determined in order to validate the microscopic findings. DansB (SFL2283) cells showed a significant drop in the number of adherent bacterial cells when compared with SFL1520 and SFL2309 (Figure 2.E).

AnsB activity is required for S. flexneri-mediated killing and bacterial accumulation in the intestinal lumen of C. elegans
To determine if the reduced adherence observed in vitro in the DansB (SFL2283) strain had an effect on bacterial virulence; preliminary in vivo studies were carried out in C. elegans. Kesika and colleagues have previously reported that virulent S. flexneri serotype 2b kills C. elegans in liquid culture [18]. We performed liquid killing assays to compare the killing rates of DansB (SFL2283) and wild type S. flexneri serotype 3b. C. elegans fed wild type S. flexneri 3b (SFL1520) died much faster (with 50% killing (TD 50 ) = 4261 h) than worms maintained on DansB (SFL2283) (,50% killing in 48 h) (Figure 3.A).
Burton E.A. et al [11] have demonstrated that virulent S. flexneri strains accumulate in the intestinal lumen and kill C. elegans, while avirulent S. flexneri strains are digested by the nematodes. To determine whether the survival of nematodes fed the DansB strain correlated with low intraluminal bacterial numbers, C. elegans bacterial accumulation assays were performed. Results of bacterial accumulation assays correlate with nematode killing, as the wild type S. flexneri 3b strain (SFL1520) accumulated within the gut of the worms, while DansB (SFL2283) cells were digested by the nematodes (Figure 3.B). Complementation of the ansB mutation restored the intestinal accumulation phenotype of the wild type strain.
In order to visualize bacterial accumulation, nematodes were fed GFP + tagged S. flexneri strains. Consistent with the bacterial accumulation assays, worms fed wild type S. flexneri 3b (SFL1520) showed accumulation of GFP + tagged bacterial cells within their intestinal lumen (Figure 3.C), while those fed the DansB (SFL2283) strain failed to show GFP + accumulation (Figure 3.D). GFP + accumulation assays could not be carried out using the complemented strain as the plasmid expressing GFP + was not compatible with the plasmid used for the complementation of ansB mutation.
The enhanced survival of nematodes fed DansB (SFL2283) and digestion of wild type S. flexneri 3b following mutation of ansB suggests that this metabolic enzyme is required for bacterial virulence.

AnsB activity is required for S. flexneri infection in the murine pulmonary model of shigellosis
Since preliminary in vivo studies in C. elegans indicated that the ansB mutant showed reduced virulence, the virulence of this strain was also assessed using the murine pulmonary model, an established animal model of shigellosis [21][22][23]. Virulent strains of S. flexneri, when inoculated intranasally, colonize the lungs of infected mice and cause pneumonia. The lung counts obtained from mice inoculated with the DansB strain 24 hours post infection were significantly lower than the wild type strain (p = 0.0035, Figure 1. AnsB in S. flexneri hydrolyses asparagine. S. flexneri parental strain SFL1520 or isogenic mutant DansB (SFL2283) at 3610 8 colony forming units (CFU) per ml were incubated in phosphate buffered saline (PBS) with 5 mM asparagine. Ammonia production was measured after 60 minutes incubation at 37uC. Hydrolysis of asparagine was measured by quantifying ammonia production. There is a significant drop in the levels of released ammonia by ansB mutant cells (p = 0.0006, unpaired t-test). Error bars represent the standard error of means obtained from three independent experimental measurements. doi:10.1371/journal.pone.0094954.g001 unpaired t tests) (Figure 4), further demonstrating that AnsB activity is required for the virulence of S. flexneri.

2-Dimensional gel electrophoresis identifies changes in the S. flexneri proteome caused by ansB mutations
To gain insight into why the DansB strain shows reduced adherence to epithelial cells, we used two dimensional differential in-gel electrophoresis (DIGE) to compare the proteomes of DansB (SFL2283) and wild type S. flexneri serotype 3b (SFL1520) strains. Our proteomic studies identified 17 differentially expressed proteins in DansB (SFL2283). Using liquid chromatography mass spectrometry (LC-MS) we identified 6 (out of 17) proteins as upregulated in the DansB (SFL2283) strain (RplL, Udp, Mdh, OmpA, GroEL and DnaK) (Table S1). Quantitative real-time reverse transcription polymerase chain reaction (qRT-PCR) was performed to confirm the results obtained in the DIGE analysis ( Figure 5). We analysed the transcript levels of proteins identified in the DIGE analysis, in three independent biological replicates, using the bacterial hisG gene as a control gene to normalize the expression of all genes tested. Similar to our DIGE analysis, the expression of rplL, udp, ompA, groEL and dnaK was up-regulated in the DansB (SFL2283) strain. However, no significant differences in the transcript levels of mdh were observed ( Figure 5).
Overexpression of the outer membrane protein OmpA in wild type S. flexneri serotype 3b decreases bacterial adherence to BHK cells and virulence in C. elegans Since the ansB mutant exhibited decreased adherence and upregulation of OmpA, a prominent outer membrane protein, we decided to determine whether the up-regulation of OmpA in DansB was responsible for the observed virulence phenotypes. To investigate whether the up-regulation of OmpA had an effect on bacterial virulence, we overexpressed OmpA in wild type 3b (SFL1520) by introducing a plasmid carrying ompA into SFL1520, to generate SFL2443 (SFL1520 + ompA). qRT-PCR was used to confirm that there was an increase in ompA transcripts in SFL2443 compared with the wild type strain ( Figure 6). SFL2443 was then used to perform cell culture adherence assays and in vivo C. elegans virulence assays in order to assess whether the decreased adherence shown by the ansB mutant strain was a result of the up-regulation of outer membrane protein, OmpA.
Adherence assays performed using SFL2443 (SFL1520 + ompA) clearly indicate that the overexpression of OmpA in wild type S. flexneri serotype 3b significantly decreased the adherence of wild type cells (Figure 7.A). Similarly, in vivo bacterial accumulation assays in C. elegans showed that the overexpression of OmpA in wild type 3b, reduced bacterial accumulation in the nematodes to levels similar to DansB (SFL2283) (Figure 7.B). Results of both the BHK adherence assays and C. elegans bacterial accumulation assays suggest that the up-regulation of the outer membrane protein OmpA is responsible for the decreased virulence of the DansB strain.
ansB regulates the expression of ompA in S. flexneri Since DIGE analysis indicated that OmpA levels were elevated in ansB mutant cells and virulence studies showed that the upregulation of OmpA in wild type cells lead to decreased bacterial adherence, we decided to determine if the presence of AnsB does in fact affect the expression of ompA. In order to determine whether the presence of ansB had an effect on the expression of ompA in S. flexneri, qRT-PCR was performed to compare the levels of ompA transcripts in the wild type 3b strain, ansB mutant and the ansB mutant strain complemented with a plasmid copy of ansB. Using this approach, we show that on complementation of the ansB mutation, the levels of ompA transcripts are restored to wild type levels ( Figure 8). These results suggest that ansB does in fact play a role in the regulation of ompA expression.

Discussion
Here we characterize AnsB, a metabolic enzyme identified to be immunogenic in S. flexneri [4] and found that the activity of this enzyme is required for the virulence of S. flexneri. Using in vitro bacterial adherence assays and two in vivo models of shigellosis, we found a marked decrease in the adherence and virulence of the ansB mutant strain (SFL2283), clearly suggesting that AnsB is involved in the initial stages of S. flexneri pathogenesis.
The L-asparaginase activity of AnsB has previously been found to enhance bacterial colonization of mammalian cells by increased hydrolysis of asparagine in the case of Campylobacter jejuni [24][25][26] and Helicobacter pylori [6,27]. However, the role of this enzyme in S. flexneri virulence has not been elucidated. The results of ammonia release assays suggest that AnsB in S. flexneri functions as an Lasparaginase, as is the case in E. coli [9] and Salmonella enterica [28]. This reaction features in several metabolic pathways, including nitrogen, cyanoamino acid, alanine and aspartate metabolism [29].
The involvement of L-asparaginase activity in several metabolic pathways suggests that mutating ansB could lead to metabolic impairments, which in turn could decrease overall bacterial fitness. However, our in vitro growth studies found no differences in the growth rates of mutant and wild type cells, which suggests that although ansB is expressed and functional in wild type S. flexneri cells grown in vitro, its activity is not required for bacterial growth in vitro under nutrient rich or nutrient stressed conditions ( Figure S2). The results of our growth studies are not surprising as AnsB in E. coli, is essential for growth under low oxygen, poor carbon source conditions by providing an alternative electron acceptor. Therefore although this gene is expressed in S. flexneri cells grown aerobically, the activity of this enzyme may only be critical under anaerobiosis [30]. Therefore further growth studies need to be performed under anaerobic conditions in order to determine whether ansB mutant cells show reduced bacterial fitness.
Successful establishment of bacterial infection requires adherence to host tissue. Members of the family Enterobacteriaceae use a plethora of strategies to adhere to host tissues, ranging from the use of pili to the secretion of highly specialized adhesion molecules elegans liquid killing assays. 20 synchronized adult nematodes were treated with log phase cultures of S. flexneri 3b (SFL1520-black) or DansB (SFL2283-grey) strains grown at 37uC to express virulence factors. Infected nematodes were monitored for 48 hours and survival was scored every 12 hours. Worms treated with the DansB strain show a significant increase (p,0.0001, Logrank test) in survival rates compared to nematodes infected with wild type S. flexneri 3b. Error bars represent standard error of mean across at least 3 independent biological repeats. B: Bacterial accumulations assays; young adult hermaphrodite nematodes were fed SFL1520, SFL2283 or ansB complemented strain SFL2309 (SFL2283 + ansB) for 24 hours. 20 nematodes were picked and mechanically disrupted to release internalized bacteria. Diluted lysates were plated on LB agar plates carrying appropriate antibiotics, and colonies were scored in order to quantify S. flexneri cells associated with each nematode. There is a significant decrease in the accumulation of SFL2283 compared to SFL1520 and SFL2309 (p,0.01, One-way ANOVA). Error bars represent standard error of mean across 3 independent biological repeats. Young adult hermaphrodite nematodes infected as above with GFP + -tagged SFL1520 (C) and SFL2283 (D). Wild type S. flexneri 3b accumulates in the nematode gut lumen and DansB (SFL2283) strain fails to accumulate in the worm gut lumen. Scale bar = 400 mm. doi:10.1371/journal.pone.0094954.g003 [31]. The molecular mechanisms used by S. flexneri to adhere to host cells are relatively unknown. Previous studies have shown that the bacterial Type III secretory Ipa proteins, especially IpaB, facilitate adherence to mammalian tissue [32][33][34]. In this study, western immunoblots confirmed that IpaB levels in DansB and wild type cells are comparable ( Figure S1.B), which suggests that there may be an IpaB-independent mechanism involved in S. flexneri adhesion to host cells.
DIGE analysis was performed to identify changes in the S. flexneri proteome in response to mutating ansB. Comparative analysis of proteomes of the ansB mutant and wild type S. flexneri strains showed that this mutation exerts pleiotropic effects on the expression of several S. flexneri proteins (Table S1), including proteins involved in stress responses (DnaK and GroEL), translation (RplL), metabolism (Udp and Mdh) and an outer membrane protein (OmpA). OmpA is a prominent outer membrane protein found in Gram negative bacteria, required to maintain the structural integrity of cellular outer membranes [35,36] and the role of this protein in Gram negative bacterial pathogenesis has been extensively studied [37][38][39]. Furthermore, studies in E. coli [40] and enterohemorrhagic E. coli (EHEC)  . Reverse transcription quantitative PCR (qRT-PCR) analysis to confirm two dimensional differential in-gel electrophoresis (DIGE) analysis. Transcript levels of proteins identified by DIGE analysis, rplL, udp, mdh, groEL, dnaK and ompA, were measured in DansB (SFL2283) and wild type S. flexneri serotype 3b (SFL1520) and expressed as a ratio of the ansB mutation-induced levels and the basal wild type levels. Data represent the means of three independent biological repeats, each replicate measured in triplicate and normalized to the control gene hisG. Asterisks indicate statistically significant differences calculated using unpaired Student's t-test (p,0.05). doi:10.1371/journal.pone.0094954.g005 Figure 6. qRT-PCR analysis confirms the up-regulation of ompA in SFL2443 (wild type + ompA). qRT-PCR was used to compare the mRNA transcript levels of ompA in SFL1520 (wild type S. flexneri serotype 3b), SFL2283 (DansB) and SFL2443 (SFL1520 + a plasmid carrying ompA to account for the up-regulation of OmpA seen in the DansB mutant). Results represent the means of three biological repeats, each measured in triplicate and normalized to the control gene hisG.
Error bars indicate the standard errors of the means. OmpA mRNA levels in SFL2443 are significantly higher than SFL1520 (p,0.001, unpaired ttest) and similar to the up-regulated levels seen in SFL2283 (p.0.05, unpaired t-test). doi:10.1371/journal.pone.0094954.g006 [41,42] have found that OmpA plays a key role in the initial stages of bacterial adhesion and invasion. In this study we show that AnsB is required for the adherence of S. flexneri, and ansB mutant cells up-regulate the expression of OmpA, therefore the effect of overexpressing OmpA in S. flexneri was investigated.
ansB mutants showed an up-regulation of OmpA, it is generally believed that the overexpression of membrane proteins affects the integrity of cell membranes and cell viability [43]. Therefore the overexpression of OmpA would affect cellular protein homeostasis, which in turn would have an effect on the overall integrity of the cellular envelope. Therefore to determine the effects of elevated OmpA levels on bacterial adherence, we overexpressed OmpA in the wild type strain and found that the overexpression of this outer membrane protein decreased both bacterial adherence to BHK cells and accumulation in C. elegans (Figure 7.A and B). These results suggest that the up-regulation of OmpA in ansB mutant cells could disrupt the bacterial cell envelope integrity and therefore may be one of the factors that contribute to the decreased adherence of these mutant cells.
qRT-PCR analysis of ompA expression in the ansB complement showed that complementation of the ansB mutation resulted in decreased ompA expression. This result suggests that the presence of ansB directly or indirectly affects ompA expression. The expression of OmpA has been known to be environmentally responsive. In E. coli, OmpA levels are elevated in response to nitrogen shortage, to facilitate an increased uptake of peptides to alleviate the nitrogen shortage [44,45]. One of the identified functions of AnsB is to provide nitrogen and carbon sources from the hydrolysis of exogenous asparagine [46]. Nitrogen depletion increases the expression of ansB in yeast [47,48], suggesting that AnsB plays a role in nitrogen metabolism. ansB mutant cells would thus have lower nitrogen levels compared with wild type cells; Figure 7. Overexpression of OmpA decreases S. flexneri adherence in vitro and virulence in C. elegans. A: Results of in vitro adherence assays for SFL2443 (SFL1520 + a plasmid expressing ompA to account for the up-regulating of OmpA seen in the DansB mutant), wild type S. flexneri 3b (SFL1520), and DansB (SFL2283) strains plotted as percentage adherence, a ratio of the number of adherent and intracellular bacteria to the total number of bacterial cells in the infecting inoculum (y-axis). The adherence of SFL1520 is significantly decreased as a result of the overexpression of OmpA in, SFL2443 (p,0.005, unpaired t-test). Results are the means of three independent blind repeats with standard errors (error bars).B: Results of in vivo C. elegans bacterial accumulation assays for SFL2443, SFL1520, SFL2283 strains. The accumulation of SFL1520 is significantly higher than SFL2443 (p,0.001, unpaired t-test). Results are the means of three independent blind repeats with standard errors (error bars). doi:10.1371/journal.pone.0094954.g007 Figure 8. qRT-PCR analysis confirms that the presence of ansB regulates the expression of ompA in S. flexneri. qRT-PCR was used to compare the mRNA transcript levels of ompA in SFL1520 (wild type S. flexneri serotype 3b), SFL2283 (DansB) and SFL2309 (SFL2283 + a plasmid carrying ansB to complement the ansB mutation). Results represent the means of three biological repeats each measured in triplicate and normalized to the control gene hisG. Error bars indicate the standard errors of the means. A significant increase in ompA expression is observed in SFL2285 (p,0.0001) when compared with SFL1520 and SFL2309. No significant differences were observed in the expression of ompA between the wild type parental strain (SFL1520) and the ansB complement (SFL2309) (p.0.05, unpaired t-test), suggesting that complementation of the ansB mutation in SFL2309, restores the expression of this gene to wild type levels. doi:10.1371/journal.pone.0094954.g008 therefore we propose that the expression of OmpA in these mutant cells may be increased to alleviate this nitrogen shortage.
The exact mechanism involved in the AnsB-mediated regulation of OmpA remains unclear at this stage nonetheless our finding provided clear evidence suggesting that the presence of AnsB is required for the adherence of S. flexneri and for regulating OmpA levels.

Conclusions
Here we have shown that the periplasmic enzyme, AnsB, is required for the virulence of S. flexneri. in vitro cell culture assays indicate that this enzyme is essential for bacterial adherence. Differential in-gel electrophoresis showed that ansB mutants exert pleiotropic effects on the expression of a number of S. flexneri genes, including a prominent bacterial outer membrane protein, OmpA that is known to be required for adherence to host cells. We also show that the up-regulation of OmpA, could be one of the factors responsible for the defective adherence of ansB mutant cells. This is the first report in S. flexneri where the function of AnsB has been found to extend beyond its canonical metabolic role. The requirement of AnsB for the virulence of S. flexneri makes this gene an attractive candidate for designing new therapeutic measures and preventive approaches to curb the spread of shigellosis. Figure S1 ansB mutation does not affect the expression and secretion of invasion protein antigens-IpaB and IpaD. Secreted proteins were isolated from the S. flexneri serotype 3b parental strain (SFL1520), DansB (SFL2283) and a S. flexneri strain cured of the virulence plasmid (SFL1223) [20] used as a negative control. Proteins were separated using SDS-PAGE, following which they were electroblotted onto PVDF membranes and probed with anti-IpaB (A) and anti-IpaD (B) primary antibodies. The levels of secreted IpaB (,62 kDa) and IpaD (,34 kDa) produced by all strains was visualized using a chemiluminescence reader. No differences were observed in the secretory levels of both IpaB and IpaD between the wild type and ansB mutant strains. (TIFF) Figure S2 AnsB activity is not required for S. flexneri growth in vitro. The growth curves of wild type S. flexneri serotype 3b (SFL1520) (black) and DansB strain (SFL2283) (grey), plotted as optical density readings at 600 nm (OD 600 ) (y-axis) versus time (x-axis). A: Growth in Luria-Bertani (LB) broth at 37uC (bold lines) and at 30uC (dashed lines). B: Growth in minimal essential salts media (MM) at 37uC (bold lines) and at 30uC (dashed lines). No significant difference in the growth patterns of SFL1520 and SFL2283 were detected under all four conditions examined (p.0.05, unpaired t-test). Error bars represent standard error of means obtained from three independent biological repeats. (TIFF)

Author Contributions
Conceived and designed the experiments: DG NV. Performed the experiments: DG. Analyzed the data: DG NV UM CB. Contributed reagents/materials/analysis tools: DG NV UM CB. Wrote the paper: DG NV UM CB.