The Burkholderia pseudomallei Proteins BapA and BapC Are Secreted TTSS3 Effectors and BapB Levels Modulate Expression of BopE

Many Gram-negative pathogens use a type III secretion system (TTSS) for the injection of bacterial effector proteins into host cells. The injected effector proteins play direct roles in modulation of host cell pathways for bacterial benefit. Burkholderia pseudomallei, the causative agent of melioidosis, expresses three different TTSSs. One of these systems, the TTSS3, is essential for escape from host endosomes and therefore intracellular survival and replication. Here we have characterized three putative TTSS3 proteins; namely BapA, BapB and BapC. By employing a tetracysteine (TC)-FlAsH™ labelling technique to monitor the secretion of TC-tagged fusion proteins, BapA and BapC were shown to be secreted during in vitro growth in a TTSS3-dependant manner, suggesting a role as TTSS3 effectors. Furthermore, we constructed B. pseudomallei bapA, bapB and bapC mutants and used the well-characterized TTSS3 effector BopE as a marker of secretion to show that BapA, BapB and BapC are not essential for the secretion process. However, BopE transcription and secretion were significantly increased in the bapB mutant, suggesting that BapB levels modulate BopE expression. In a BALB/c mouse model of acute melioidosis, the bapA, bapB and bapC mutants showed a minor reduction of in vivo fitness. Thus, this study defines BapA and BapC as novel TTSS3 effectors, BapB as a regulator of BopE production, and all three as necessary for full B. pseudomallei in vivo fitness.

Introduction vitro in a TTSS3-dependent manner [22] was BopC. Disruption of bopC decreased B. pseudomallei invasion of human lung epithelial A549 cells in vitro, suggesting a specific role for BopC in host cell invasion [22]. Recently, a proteomic analysis of two B. pseudomallei TTSS-hypersecreting mutant strains (bipD and bsaP mutants) identified 26 proteins as likely TTSS3-dependent effectors [32]. The proteins identified included BopE, BopA and BopC, as well as BapA. Furthermore, BapA, BopA and BprD were confirmed as being secreted in a TTSS3-dependent manner in wild-type B. pseudomallei strain, using an epitope tagging strategy.
In this study, we characterized the putative TTSS3 proteins, BapA, BapB (BPSS1527) and BapC, with respect to their possible functions as bacterial effectors, their roles in controlling expression of other TTSS3 components and their involvement in B. pseudomallei pathogenesis. A tetracysteine (TC)-FlAsH™ labelling technique was used to monitor the secretion of TCtagged fusion proteins in vitro. These analyses confirmed that BapA was secreted in a TTSS3-dependent manner and identified BapC as a novel TTSS3 effector. B. pseudomallei bapA, bapB and bapC mutants were constructed and tested for a range of in vitro and in vivo phenotypes. The essential role of each protein in TTSS3 function was analyzed by using the well-characterized TTSS3 effector BopE as a marker of TTSS secretion activity. All mutants were able to secrete BopE, indicating that BapA, BapB and BapC are not essential for TTSS3 function. However, transcription of bopE and secretion of the BopE protein were increased in the bapB mutant, indicating that BapB is likely to play a role in modulating bopE expression. Finally, all three proteins were shown to have a minor role in B. pseudomallei in vivo fitness in BALB/c mice.

Materials and Methods
Bacterial strains and cell culture Bacterial strains and plasmids are described in S1 Table. The B. pseudomallei wild-type strain K96243 [33] (kindly provided by Dr. Brenda Govan, James Cook University, Townsville, Australia) was used as the parent strain for mutagenesis. Escherichia coli DH5α was primarily used for growth and amplification of plasmids, and strain S17-1/λpir was used for mobilization of DNA (pDM4 or pBHR1 constructs) into B. pseudomallei by conjugation [34]. Bacterial culture media were purchased from Oxoid (Hampshire, UK) and solidified by adding 1.5% (wt/vol) agar as required. Antibiotics were purchased from Sigma-Aldrich (St. Louis, MO, USA). Unless indicated otherwise, E. coli was cultured in Lysogeny broth (LB); ampicillin (Amp; 100 μg/ml), chloramphenicol (Cm; 20 μg/ml), kanamycin (Kan; 50 μg/ml) or tetracycline (Tet; 10 μg/ml) were added when required. B. pseudomallei was cultured in LB supplemented with gentamicin, (Gen; 8 μg/ml), Cm (50 or 100 μg/ml), Kan (1 mg/ml) or Tet (25 μg/ml) as required. All bacterial cultures were grown at 37°C, with broth cultures shaken at 200 rpm. The murine RAW 264.7 macrophage-like cell line and RAW 264.7 cells stably expressing GFP-LC3 were maintained in antibiotic-free medium as described previously [9,31]. The human respiratory epithelial A549 cell line was obtained from the American Type Culture Collection (Manassas, VA, USA) and maintained at 37°C in 5% CO 2  Double-crossover, allelic exchange mutagenesis of bapA, bapB and bapC B. pseudomallei bapA, bapB, and bapC mutant strains were generated by double-crossover allelic exchange using the λpir-dependent vector pDM4 which contains the sacB gene for counterselection [35]. Strains and plasmids used are listed in S1 Table. Specific primer pairs (S2 Table) were used to amplify separate sequences upstream and downstream of the target genes, and these were cloned into pDM4. For bapA mutagenesis, the primers MC5532 and MC5533 were used to amplify an upstream fragment of bapA, and the primers MC5516 and MC5517 were used to amplify a downstream fragment encompassing the entire bapB and bapC genes. Both fragments were cloned into SpeI/XbaI-digested pDM4. For bapB mutagenesis, the primer pairs JT6156/JT6157 and JT6175/JT6176 were used to amplify the upstream and downstream fragments respectively, and both fragments cloned into SphI/SpeI-digested pDM4. For bapC mutagenesis, the primer pairs JT6319/JT6320 and JT6179/JT6180 were used to amplify the upstream and downstream fragments, which were then cloned sequentially into XmaI/SacIdigested then SalI-digested pDM4. The tetracycline resistance gene tetA(C), recovered from pUTminiTn5Tc [36], was then ligated into the central BglII site of pDM4 in order to generate the mutagenesis constructs pDM4::bapA::tetA(C), pDM4::bapB::tetA(C) and pDM4::bapC:: tetA(C). These constructs were introduced by transformation into the conjugative donor strain E. coli S17-1/λpir, and then mobilized into B. pseudomallei by conjugation [9]. Each conjugation reaction was plated onto LB agar containing 8 μg/ml Gen and 25 μg/ml Tet, and plates incubated at 37°C for up to 2 days. Tet resistant transconjugants were patched onto LB agar supplemented with 20% (wt/vol) sucrose in order to select for double-crossover mutants based on sacB counter-selection as described previously [37]. Putative double-crossover mutant strains were confirmed by PCR and DNA sequencing (data not shown) and correct strains designated ΔbapA, ΔbapB and ΔbapC.

Complementation using pBHR1
To confirm that identified phenotypes were due specifically to inactivation of the bap genes, each mutant was complemented with the intact gene cloned into the B. pseudomallei replicating plasmid pBHR1 as described previously [9,28]. Strains and plasmids used are listed in S1 Generation of tetracysteine (TC)-tagged bopE, bapA, bapB and bapC Initially a TC-tagged version of bopE was constructed in a modified mini-Tn7 vector (pUC18Tmini-Tn7T::tetA(C)::P glmS2 ) that contained a tetracycline gene for selection and the constitutively active B. pseudomallei glmS2 promoter directing transcription of the cloned bopE gene. Strains and plasmids are listed in S1 Table and primers in S2 Table. The primers JT6929 and JT6930 were used to amplify an 813 bp fragment containing the full-length bopE which was then ligated into XmaI/SpeI-digested pUC18Tmini-Tn7T::tetA(C)::P glmS2 . The modified TC tag, encoding proline (CCG) and glycine (GGC) as spacers between four cysteine sequences (TGC) and a stop codon (TAA) [38], was generated by annealing the primers JT6931 and JT6932 and then cloning this fragment in-frame at the 3' end of the bopE gene, generating pUC18Tmini-Tn7T::tetA(C)::P glmS2 ::bopETC. Initial testing indicated that the expression of TC-tagged BopE from this construct was weak (data not shown), so the TC-tagged bopE was transferred into the multi-copy plasmid pBHR1. The primers JT7125 and JT7126 were used to amplify an approximately 1,200 bp fragment containing the glmS2 promoter, bopE, the TC tag, and the two terminator sequences (T 1 and T 0 ) from pUC18Tmini-Tn7T::tetA(C)::P glmS2 :: bopETC, and this fragment was then cloned into AclI/NcoI-digested pBHR1. The resultant recombinant plasmid was introduced into E. coli S17-1/λpir and then transferred into the B. pseudomallei wild-type strain by conjugation. Transconjugants were selected on LB agar containing 8 μg/ml Gen and 1 mg/ml Kan and verified by PCR and nucleotide sequencing (data not shown). One clone containing the correct plasmid was designated B. pseudomallei [bopETC] and used for further experiments.
Constructs expressing TC-tagged bapA, bapB and bapC were generated by cloning the TC tag into each of the corresponding complementation constructs. In each case, the TC tag was generated by annealing the primers JT7147 and JT7241 prior to ligation into AclI-digested pBHR1::bapA, pBHR1::bapB or pBHR1::bapC. A single correct plasmid representing each construct was then introduced into E. coli S17-1/λpir and then mobilized into B. pseudomallei by conjugation. Single pBHR1::bapATC, pBHR1::bapBTC and pBHR1::bapCTC containing strains were identified and designated K96243 [bapATC], [bapBTC] and [bapCTC], respectively. In addition, each of the pBHR1::P glmS2 ::bopETC, pBHR1::bapATC, pBHR1::bapBTC and pBHR1::bapCTC constructs was transferred into a B. pseudomallei bsaS mutant [30]  Precipitation of proteins using sodium deoxycholate/trichloroacetic acid (DOC/TCA) and protein sample preparation For precipitation of proteins, DOC was used, in combination with TCA, to act as a co-precipitant to enhance protein precipitation as described previously [39,40] with some modifications. In brief, culture supernatant samples collected from bacterial cells grown to the required growth phase were filtered through a 0.45-μM syringe filter (Pall Life Science, USA) and subsequently incubated with 0.02% (wt/vol, final concentration) DOC at room temperature for 15 min. Total proteins were precipitated by incubation with 10% (wt/vol, final concentration) TCA overnight at 4°C, followed by centrifugation (10,000 x g, 20 min, 4°C). The pellet was washed twice with ice-chilled methanol prior to air-drying to eliminate any methanol residue. Laemmli sample buffer (1X) containing β-mercaptoethanol (BME) was used to resuspend the protein pellet. The samples were then denatured by boiling at 99°C for 10 min and cooled to room temperature before determination of total protein in the samples using the 2-D Quant Kit (GE Healthcare, NSW, Australia) according to the manufacturer's instructions.

TC-FlAsH™-based fluorescence labelling
TC-tagged proteins were visualized using FlAsH-labelling as described previously [38,41] with some modifications. For visualization of TC-tagged proteins from live cells, B. pseudomallei was grown to exponential phase, harvested by centrifugation (4,293 x g, 5 min, room temperature) and washed with sterile phosphate-buffered saline, pH 7.4 (PBS) prior to resuspension in 50 μl of PBS. Stock FlAsH reagent, from the TC-FlAsH™ II In-Cell Tetracysteine Tag Detection Kit (Life Technologies™, USA), was added to the bacterial cell suspension to obtain a final concentration of 5 μM prior to incubation at 37°C with shaking at 1,000 rpm, for 1 h. Labeled samples were washed once with PBS to remove unbound FlAsH reagent prior to resuspension in 1X Laemmli sample buffer containing 10 mM (final concentration) dithiothreitol (DTT) and heated at 99°C for 10 min. Labeled proteins were separated by SDS-PAGE, and green fluorescence emission visualized immediately at 520 +/-10 nm upon excitation at 488 nm. For visualization of secreted TC-tagged proteins in vitro, DOC/TCA precipitated supernatant samples were labeled with FlAsH reagent according to the manufacturer's instructions with the following minor modifications. Precipitated proteins were labeled with the FlAsH reagent (20 μM, final concentration), using 2-mercaptoethanol as the reducing agent, and incubated at 70°C with shaking for 10 min. Samples were cooled to room temperature, polypeptides separated by SDS-PAGE and fluorescently labeled samples visualized as described above.

Western immunoblotting
Following SDS-PAGE, proteins were transferred to PVDF membranes (Merck Millipore, USA) and the membranes probed by immunoblotting with rabbit anti-BopE 78-261 antiserum [27] as described previously [30]. Antibody binding was analyzed using Amersham ECL Western Blotting Detection Reagent (GE Healthcare, NSW, Australia), and chemiluminescent signals detected with X-ray film (Kodak, NY, USA). BopE expression in each sample was normalized for sample loading by densitometric analysis of Coomassie Blue stained whole cell lysates and the BopE expression reported as relative to the expression levels in the wild-type strain (foldchange relative to wild-type expression).

Purification of RNA and quantitative real-time RT-PCR (qRT-PCR)
Total bacterial RNA was extracted using TRIzol 1 (Life Technologies™, USA), and cDNA was produced using reverse transcriptase and random hexamers as described previously [42]. qRT-PCR was conducted with the Mastercycler 1 ep realplex PCR system (Eppendorf South Pacific, Australia) using FastStart Universal SYBR Green Master (Rox) as the master mix (Roche Diagnostics, Australia). The primer pairs JT7472/JT7473 and JT7474/JT7475 (S2 Table) were designed for the amplification of bopE and the internal control rpoA (BPSL3187), respectively, using the Primer3 primer design software (http://simgene.com/Primer3). These primer pairs amplified PCR products of 119 and 90 bp, respectively. Each qRT-PCR mixture was prepared by addition of 5 μl of 1:10 dilution of cDNA to 15 μl of PCR master mix containing 10 μl of FastStart Universal SYBR Green Master (Rox), 0.2 μl of each primer (100 μM) and 4.6 μl of DEPC-treated nuclease-free water (Ambion). The genomic DNA of B. pseudomallei wild-type strain K96243 and RT-negative samples were used as positive and negative controls, respectively. The cycling parameters for amplification were as follows: 1 cycle of 95°C for 2 min, 40 cycles of 95°C for 15 sec and 60°C for 1 min. All qRT-PCR experiments were carried out in technical triplicate on biological duplicates.

In vivo competitive growth assays
To determine the ability of each strain to grow in vivo, relative growth assays in BALB/c mice were carried out as described previously [9]. In brief, each of the mutant and the wild-type strains was grown to an OD 600 of 0.2 (corresponding to 2 x 10 8 CFU/mL). Equal volumes of the mutant and the wild-type cultures were combined and 10 μL of the 10 −2 dilution of the combined mixture (corresponding to 2 x 10 4 CFU) was inoculated intranasally into three to five, 6-to 8-week-old, female BALB/c mice. The infection was allowed to proceed for 20 h, and mice were then euthanised in an ethically approved manner (Monash University Animal Ethics Committee approval #SOBS/M/2008/2). Numbers of mutant and wild-type bacteria in input cultures, or in output cultures recovered from mouse spleens, were determined by patching recovered bacteria onto LB plates with or without Tet (25 μg/ml). The competitive index (CI) was defined as the ratio of mutant to wild-type bacteria in the output pool divided by the ratio of mutant to wild-type bacteria in the input pool. The statistical significance of a reduction in CI for each mutant was determined by a one sided z-test [43,44] and the difference in CI between mutant and complemented strains assessed by Mann-Whitney U test.

Statistical analyses
Two-way ANOVA was used to analyze intracellular survival assays, and Student's unpaired ttest was used to compare means between each mutant and the wild-type. A P value of less than 0.05 was accepted as indicating a statistically significant difference between samples.

Results
The B. pseudomallei BapA and BapC proteins are secreted by the TTSS3 The bapA, bapB and bapC genes are located within the TTSS3 locus, but the functions of the encoded proteins are unknown. To investigate if these genes might encode secreted TTSS3 effectors, each of the genes was fused with a sequence encoding a TC tag and recombinant C-terminal tagged proteins visualized by labelling with the membrane-permeant 4',5'-bis (bis1,3,2-dithioarsolan-2-yl) fluorescein (FlAsH) reagent. As a control, the gene encoding the well characterized effector BopE was also fused to the same TC tag. Firstly, to identify whether the FlAsH reagent was able to access and bind to the TC-tagged proteins in the recombinant B. pseudomallei [bopETC], [bapATC], [bapBTC] and [bapCTC] strains, mid-exponential phase cultures of these recombinant strains were incubated with the FlAsH compound and then proteins in whole cell lysate samples separated by SDS-PAGE ( Fig 1A); as a control for sample loading the total proteins in each sample were visualized by PAGE and Coomassie Blue staining ( Fig  1B). A fluorescent signal at 33 kDa, predicted to correspond to TC-tagged BopE, was observed in the B. pseudomallei [bopETC] sample, but not in the B. pseudomallei [pBHR1] sample ( Fig  1A). In samples derived from B. pseudomallei [bapATC], a strongly fluorescent signal with a mobility corresponding to approximately 120 kDa was observed (Fig 1A). Although the size of BapA is predicted to be 88 kDa, as no fluorescent signals were observed at this position in any other samples including the B. pseudomallei [pBHR1] control sample, we concluded that this protein corresponds to TC-tagged BapA. A fluorescent signal corresponding to the predicted size of BapC (23 kDa) was observed in samples derived from the B. pseudomallei [bapCTC] strain (Fig 1A). A fluorescent signal of approximately 25 kDa (Fig 1A) was observed in samples derived from both the [bopETC] and [pBHR1] strains, indicating that some endogenous B. pseudomallei proteins can form a fluorescent complex with the FlAsH compound, as has been observed previously with this technique [45][46][47]. These results indicate that the TC-FlAsH™ labelling technique can be used for labelling proteins in B. pseudomallei. No fluorescent signals corresponding to TC-tagged BapB were observed in the B. pseudomallei [bapBTC] samples despite analysis under a wide range of labelling and protein separation conditions (data not shown), suggesting that either BapB is not expressed under the conditions analyzed, or the TC tag was not accessible for FlAsH labelling in the bapBTC strain.
To determine whether BapA and BapC are TTSS3 secreted, we analyzed culture supernatants of the B. pseudomallei wild-type strain and a TTSS mutant (bsaS) strain [30]. Strains were grown to mid-exponential phase, supernatant samples collected, proteins precipitated with DOC/TCA and then labeled with the FlAsH reagent (Fig 2A). Total proteins were also visualized by PAGE and Coomassie Blue staining as a control for loading ( Fig 2B). As expected, a fluorescent signal corresponding to TC-tagged BopE was detected in the DOC/ TCA precipitated supernatant samples of the B. pseudomallei [bopETC] strain (Fig 2A), but not in the TTSS3-deficient ΔbsaS[bopETC] strain, confirming the TTSS3-dependent secretion of BopE. In the supernatant samples from B. pseudomallei [bapATC] cultures, a fluorescent signal of approximately 120 kDa was observed (Fig 2A) as previously identified for TC-tagged BapA in whole cell lysates (Fig 1A). Similarly, a fluorescently labeled protein of 23 kDa, corresponding to TC-tagged BapC, was identified in the supernatant samples from the B. pseudomallei [bapCTC] strain (Fig 2A). Importantly, no fluorescently labeled proteins were detected in samples derived from strains ΔbsaS [bapATC] or ΔbsaS[bapCTC] (Fig 2A). These data strongly suggest that BapA and BapC are secreted in a TTSS3-dependent manner.

BapB plays a role in modulating BopE secretion
To determine whether BapA, BapB or BapC are essential for the function or regulation of the TTSS3 we used the secretion of BopE as a measure of TTSS3 activity. The secretion of BopE was assessed with BopE-specific antiserum in each of the B. pseudomallei ΔbapA, ΔbapB and ΔbapC strains, and the B. pseudomallei ΔbopE::pDM4 strain as a negative control. The levels of BopE were assessed in whole cell lysates and culture supernatants collected from strains grown to early-, mid-or late-exponential growth phase (Fig 3).
BopE secretion at levels similar to, or higher than, that observed in the wild-type strain was observed in certain growth phases for all strains except ΔbopE::pDM4, indicating that BapA, BapB and BapC are not essential TTSS3 secretion apparatus components. In early-exponential phase culture supernatants (Fig 3A), the level of secreted BopE was indistinguishable in the wild-type, ΔbapA and ΔbapC strains, but was substantially increased (30-fold; P < 0.05) in the ΔbapB strain. Similarly, a 30-fold increase in BopE was also observed in total culture samples derived from the ΔbapB strain (Fig 3B). Increased BopE production was also observed in total culture samples derived from the ΔbapC strain (10-fold; P < 0.05) (Fig 3B). Analysis of midexponential phase cultures indicated that again the ΔbapB strain exhibited a significant  -(A, B), mid-(C, D) and late-exponential (E, F) growth phases. Supernatant or total culture samples were separated by SDS-PAGE and transferred to a membrane prior to western immunoblotting using anti-BopE antiserum. The relative expression levels, determined by densitometric analysis of the western blots and normalized against the wild-type expression, are presented as the mean ± SEM of three biological replicates. *P < 0.05, ***P < 0.0001. A single representative western blot is shown above the quantified averaged data for comparison. A vertical black line on the western blot indicates unrelated intervening lanes have been removed. The samples were derived from the ΔbapA, ΔbapB, ΔbapC, ΔbopE or the the wild-type (WT) strain. increase in secretion of BopE (2-fold, P < 0.05). However, no increase was observed in the level of BopE present in the total cultures of the ΔbapB strain. In late-exponential phase, the ΔbapB strain again showed an increase in secreted BopE (4-fold, P < 0.05) (Fig 3E), but no difference was observed in total cultures (Fig 3F). Thus, the ΔbapB strain showed a consistent increase in secreted BopE production at all growth phases tested. The ΔbapA strain showed a minor decrease in BopE production during mid-exponential phase, both in secreted samples (1.5-fold, P < 0.05) (Fig 3C) and total cultures (2-fold, P < 0.05) (Fig 3D) and this strain exhibited a slight increase (2-fold; P < 0.05) in secreted BopE, during late-exponential phase.

The transcription of bopE is increased in the ΔbapB strain
Given the substantial increase in BopE secretion observed in the ΔbapB strain across all growth phases, and the increase in total BopE production observed in early-exponential phase total culture samples, we investigated the transcription of bopE using quantitative RT-PCR to determine if these changes might be due to increased expression of bopE. The transcription of bopE was assessed in early-, mid-and late-exponential phase cultures of the B. pseudomallei ΔbapA, ΔbapB and wild-type strains. At early-exponential phase (Fig 4A), the level of expression of bopE in the wild-type and ΔbapA strains was indistinguishable, but in the B. pseudomallei ΔbapB strain it was increased approximately 5-fold (P < 0.05). At both mid-exponential (3-fold; P < 0.0001) and late-exponential (3-fold; P < 0.0001) growth phases, bopE expression was also significantly increased in the ΔbapB strain ( Fig 4B). Thus, bopE expression was increased between 3-to 5-fold in the B. pseudomallei ΔbapB strain during early-, mid-and late-exponential growth phases. The expression of bopE was reduced in the ΔbapA strain at both mid-exponential (3-fold; P < 0.001) and late-exponential (4-fold; P < 0.001) growth phases.

BapA, BapB and BapC play a minor role in the in vivo fitness of B. pseudomallei
To examine the importance of bapA, bapB and bapC for in vivo growth, growth assays were conducted in BALB/c mice and competitive indices (CI) determined. The CI of the B. pseudomallei ΔbapA, ΔbapB and ΔbapC strains was significantly reduced compared to the wild-type strain (0.38 ± 0.36, 0.46 ± 0.21 and 0.56 ± 0.23, respectively; Fig 5), while the growth rates of the mutant and wild-type strains were indistinguishable in LB in vitro (data not shown). Thus, these data suggest that BapA, BapB and BapC play a minor role in in vivo growth of B. pseudomallei.
To verify that the reduced in vivo growth of the mutant strains was due specifically to inactivation of the bap genes, each of the mutants was complemented with an intact copy of the relevant gene cloned into the plasmid pBHR1 [34]. The complemented strains and the appropriate controls (S1 Table) were tested for in vivo growth using competitive growth assays in BALB/c mice. The B. pseudomallei ΔbapB[bapB] complemented strain was first used in a competitive growth assay with the wild-type strain. The average CI of the B. pseudomallei ΔbapB strain was 0.46 ± 0.21 (Fig 5) and the average CI of the B. pseudomallei ΔbapB[pBHR1] vector only strain was not significantly different at 0.51 ± 0.20 (Mann Whitney U test; P > 0.05; Fig 5). However, the average CI of the B. pseudomallei ΔbapB[bapB] complemented strain was 0.86 ± 0.32, indicating significant, albeit partial, restoration of the wild-type in vivo fitness phenotype (Mann Whitney U test; P = 0.01).
As the bapB complementation construct did not fully restore wild-type in vivo fitness we assessed the stability of the pBHR1 constructs in the absence of selection for 12 h and 20 h.   strains was even lower, with no retention of either complementation plasmid observed following 20 h growth in the absence of antibiotic selection (data not shown). Accordingly, we did not test the complemented strains further.
BapA, BapB and BapC are not essential for bacterial invasion, intracellular survival and replication, escape from host phagosomes, actin-mediated motility or multinucleated giant cell (MNGC) formation B. pseudomallei TTSS3 mutants display a range of altered phenotypes, including reduced invasion, intracellular survival and replication, endosomal escape, MNGC formation and intercellular migration [16,19,20,26,27,48]. Therefore, we analyzed the bapA, bapB and bapC mutants with regard to their possible role(s) in bacterial pathogenesis using a range of in vitro phenotypic assays, specific for previously known TTSS3 functions [9,20,23,26,27,31,49]. Each of the B. pseudomallei ΔbapA, ΔbapB and ΔbapC strains was able to invade human lung epithelial A549 cells, as efficiently as the wild-type strain (data not shown). In addition, all the bap mutant strains retained the ability to survive and multiply within the murine macrophage-like cell line RAW 264.7, escape from host phagosomes, induce MNGC formation and form actin tails for actin-based motility (data not shown), indicating that the bap genes are not required for any of these functions.

BapA and BapC are TTSS3-secreted proteins
The B. pseudomallei TTSS3 is a critical virulence factor, with the secreted effector proteins directly responsible for the intracellular effects of the TTSS and likely to play key roles in modulating host cell functions [50]. In order to identify novel effector proteins, fluorescent-tagging of putative effectors [51,52] was employed in this study. This technique has been used with success to investigate secretion of TTSS effectors in S. flexneri, Enteropathogenic E. coli (EPEC) and S. enterica serovar Typhimurium [38,41,53]. We fused the BapA, BapB and BapC proteins with a C-terminal TC tag and the secretion of the fusion proteins assessed by FlAsH labelling. As a positive control, the well-characterized TTSS3 effector BopE [27] was TC-tagged and confirmed as being secreted in a TTSS3-dependent manner. Our optimized labeling conditions were then used to investigate the secretion of the TC-tagged BapA, BapB and BapC proteins and show that BapA and BapC were secreted in a TTSS3-dependent manner. To our knowledge, this is the first time FlAsH labelling has been used to investigate the secretion of B. pseudomallei TTSS molecules. Very recently, BapA was shown to be a bone fide secreted TTSS3 effector, by initial identification of secreted proteins in a B. pseudomallei hypersecretion mutant and then by using a c-Myc epitope tagging strategy [32]; our data confirm this result. Furthermore, we identified BapC as a novel TTSS3-secreted protein. Interestingly, BapC was not identified as a T3SS3 effector via proteomic analysis of B. pseudomallei hypersecretion mutants [32]. It is possible that BapC secretion is not significantly increased in the bsaP or bipD mutants used in that study or that even in hypersecreting strains the total amount of secreted BapC is below detection levels. However, with the sensitive detection of FlAsH labeling, our analysis was able to identify BapC as a bone fide TTSS3 effector using wild-type B. pseudomallei. In our hands, the TC-tagged BapB could not be identified either in total cultures or supernatant samples. As we have shown by RT-PCR that bapB is expressed during in vitro growth (data not shown), the lack of labelling suggests that the TC tag in BapB was not accessible for FlAsH compound binding. Based on bioinformatics analyses, BapB is a predicted acyl carrier protein (ACP) containing a pantetheine 4' phosphate (PP) group. The PP prosthetic group has high conformational flexibility and functions in the transfer of acyl molecules during fatty acid biosynthesis [54,55]. Given that the distance between the two pairs of cysteines in the TC fusion protein has to be accurately matched to the spacing of the bi-arsenic molecules of the FlAsH compound [56], such conformational flexibility may inhibit the TC-FlAsH interaction [46].

BapB affects secretion of BopE
To test whether loss of BapA, BapB or BapC affected the activity of the TTSS3, production and secretion of the known effector BopE were used as markers of TTSS3 activity. All mutant strains were able to secrete BopE, showing that BapA, BapB and BapC are not essential for TTSS3 function. However, the level of secreted BopE was significantly increased in the B. pseudomallei ΔbapB strain at all three growth phases tested, and total production of BopE was also significantly higher in the ΔbapB strain at early-exponential growth phase. Furthermore, there was significantly increased transcription of the bopE gene in the ΔbapB strain at all growth phases tested. These data suggest that BapB levels negatively affect the transcription of bopE. A similar scenario has been found in an AraC-family TTSS regulator BsaN (BPSS1546), which in association with its chaperone BicA (BPSS1533), functions by controlling the expression of not only several TTSS3 components, including BopA, BopE, BapA, BapB and BapC, but also T6SS molecules involved in facilitating bacterial survival within the host, thereby promoting bacterial virulence [57].
As stated above, BapB is a predicted ACP that displays most similarity to the TTSS ACP IacP of S. enterica serovar Typhimurium (27% amino acid identity). A Salmonella iacP mutant demonstrated reduced invasion of non-phagocytic cells, altered host actin rearrangement, a significant decrease in secretion of the effector proteins SopA, SopB and SopC and increased secretion of the flagella subunit FljB [58]. By contrast, the ΔbapB strain showed no difference in invasion of A549 cells (data not shown) and increased effector secretion (Fig 3); thus, IacP and BapB are unlikely to have similar functions. While BapB shows highest amino acid identity to IacP, other features of BapB, such as its acidic pI and small size, suggest it may function as a TTSS3 chaperone. In general, TTSS chaperones are essential not only for preventing the degradation and/or misfolding of bacterial effector proteins prior to secretion into the host, but also play important roles in preventing undesirable interactions of the effectors with other TTSS components [59][60][61][62]. TTSS chaperones can broadly be divided into different groups (class IA, IB, II and III) according to their substrate specificities [63]. Phylogenetic analysis of BapB and other TTSS chaperone proteins from Yersinia, Shigella and Salmonella spp. (S1 Fig) indicated that BapB clustered with the class III chaperone FliT. Among the class III chaperones, FliT functions as a negative regulator of the flagella biosynthesis operon of Salmonella [64][65][66]. Moreover, some TTSS chaperones, such as Salmonella SicA and Shigella IpgC play a role in regulating the transcription of effector genes [63,67,68,69]. Taken together, these data suggest a possible function of BapB as a TTSS3 chaperone with a negative regulatory function. Although BapB was not identified by a large scale bioinformatics genomic screen to identify putative TTSS chaperone-effector pairs in bacterial genomes, this may be because that screen applied a molecular weight filter of between 12 and 20 KDa and BapB is slightly smaller than this at 10 KDa [70].
The secretion of BopE in the ΔbapA strain was similar to that of the wild-type strain at early-exponential growth phase, but showed a slight decrease and then increase at mid and late-exponential growth phase, respectively. Surprisingly, transcription of bopE in the ΔbapA strain was slightly decreased at both mid and late-exponential growth phases. These data clearly show that BapA is not required for the TTSS3 function, but indicate that loss of BapA may play a minor or indirect role in regulation of TTSS activity.
BapC is predicted to contain a TTSS-associated lytic transglycosylase (LT) domain. In general, proteins with LT activity function by cleaving the β-1,4 glycosidic bond between N-acetylmuramoyl and N-acetylglucosaminyl residues of bacterial peptidoglycan (PG) for the recycling of PG, cell division and insertion of either flagella or secretion system components including the TTSS [71][72][73][74] into the membrane. Alignment of the amino acid sequence of BapC with other TTSS proteins that contain LT domains indicated that the two TTSS-associated LTs, HpaH and Hpa2, from the plant pathogens Xanthomonas campestris pv. vesicatoria and X. oryzae pv. oryzae, respectively, displayed the highest identity (approximately 39%; S2 Fig). HpaH plays a role in promoting TTSS assembly and secretion of other TTSS proteins, at least in part, by remodeling PG [75]. However, the secretion of HpaH, as a TTSS effector, has not yet been verified. The X. oryzae Hpa2 is likely to function as part of the translocon complex and together with another translocon component HrpF interacts with the host cell membrane [76,77]. Both HpaH and Hpa2 are required for bacterial virulence [75][76][77]. BapC also shows identity (approximately 36%) with the putative TTSS effectors, IagB and IpgF from Salmonella and Shigella respectively. Both these proteins share the three conserved motifs of the LT domain with BapC (S2 Fig) and cleave PG. However, IagB and IpgF play roles in bacterial invasion, but not in virulence [78][79][80]. Thus, the role of BapC in B. pseudomallei TTSS3 function appears different from each of these related TTSS proteins.
Taken together, these data suggest that BapA, BapB and BapC act to modulate the efficient function of the TTSS3 needle-like apparatus. None of the proteins is essential for TTSS function, but production of BapA, BapB and BapC appears necessary for full B. pseudomallei in vivo fitness in the BALB/c mouse model. BapA and BapC are clearly TTSS3 secreted proteins and BapC may play a role in facilitating TTSS needle protrusion and expansion through the PG. BapB may act both as a chaperone and a negative regulator that prevents premature secretion of certain TTSS3 effectors. pseudomallei BapC, X. campestris pv. vesicatoria HpaH, X. oryzae pv. oryzae Hpa2, Salmonella IagB and Shigella IpgF. The three conserved motifs of the LT domains are indicated by the boxes. Regions 1 and 3 are predicted to form α-helices and region 2 is predicted to form a β-sheet. Residue E in each of the α-helix regions (boxed) is typically the catalytic glutamate residue responsible for LT domain cleavage of β-1,4 glycosidic bonds of bacterial peptidoglycan. ':' or '.' designate amino acids with strongly and weakly conserved properties, respectively. ' Ã ' designates identical amino acids. (TIF) S1