Pyridoxal phosphate synthases PdxS/PdxT are required for Actinobacillus pleuropneumoniae viability, stress tolerance and virulence

Pyridoxal 5’-phosphate (PLP) is an essential cofactor for numerous enzymes involved in a diversity of cellular processes in living organisms. Previous analysis of the Actinobacillus pleuropneumoniae S-8 genome sequence revealed the presence of pdxS and pdxT genes, which are implicated in deoxyxylulose 5-phosphate (DXP)-independent pathway of PLP biosynthesis; however, little is known about their roles in A. pleuropneumoniae pathogenicity. Our data demonstrated that A. pleuropneumoniae could synthesize PLP by PdxS and PdxT enzymes. Disruption of the pdxS and pdxT genes rendered the pathogen auxotrophic for PLP, and the defective growth as a result of these mutants was chemically compensated by the addition of PLP, suggesting the importance of PLP production for A. pleuropneumoniae growth and viability. Additionally, the pdxS and pdxT deletion mutants displayed morphological defects as indicated by irregular and aberrant shapes in the absence of PLP. The reduced growth of the pdxS and pdxT deletion mutants under osmotic and oxidative stress conditions suggests that the PLP synthases PdxS/PdxT are associated with the stress tolerance of A. pleuropneumoniae. Furthermore, disruption of the PLP biosynthesis pathway led to reduced colonization and attenuated virulence of A. pleuropneumoniae in the BALB/c mouse model. The data presented in this study reveal the critical role of PLP synthases PdxS/PdxT in viability, stress tolerance, and virulence of A. pleuropneumoniae.


Introduction
Vitamin B6 is an essential cofactor in a multitude of cellular enzymatic reactions and is required in the metabolism of carbohydrates, amino acids, and fatty acids [1]. Pyridoxal 5'phosphate (PLP) is the biochemically active form of vitamin B6 [2]. To date, studies have identified more than 160 enzymes that are functionally dependent on PLP [3,4]. Vitamin B6 plays an important role in living organisms, and humans and animals must obtain it from food. By contrast, bacteria, fungi, and plants are able to synthesize PLP via de novo pathways. PLOS  virulence of A. pleuropneumoniae remain unclear. In the present study, we identified and characterized the function of enzymes PdxS and PdxT in the vitamin B6 biosynthesis pathway in A. pleuropneumoniae. We constructed knockouts in the pdxS and pdxT genes, respectively and investigated the role of PdxS and PdxT in viability, stress tolerance, and virulence of A. pleuropneumoniae.

Ethics statement
Animal experiments were approved by Animal Ethics Committee of Harbin Veterinary Research Institute of the Chinese Academy of Agricultural Sciences (CAAS) and carried out in strict accordance with animal ethics guidelines and approved protocols (Heilongjiang-SYXK-2011-022). All efforts were made to minimize animal suffering.

Bacterial strains and growth conditions
The bacterial strains, plasmids, and primers used in this study are described in S1 and S2 Tables. The A. pleuropneumoniae strains were cultured in a brain heart infusion (BHI) medium supplemented with 10 μg/mL nicotinamide dinucleotide (NAD) (Sigma-Aldrich, USA). For the culture of A. pleuropneumoniae transconjugants (single crossovers), BHI medium was supplemented with 10 μg/mL of NAD and 7 μg/mL of chloramphenicol. E. coli β2155 was grown in Luria-Bertani (LB) medium supplemented with 1 mM diaminopimelic acid (DAP) (Sigma-Aldrich, USA). The chemically defined medium (CDM) was prepared as previously described [15], without the addition of NH 4 Cl. All strains were routinely grown at 37˚C.

Protein expression and purification
The coding sequences of A. pleuropneumoniae pdxS and pdxT genes were PCR-amplified from S-8 genomic DNA using specific primers SF/SR and TF/TR (S1 Table). The digested PCR products were ligated with NdeI/XhoI-digested pET22b(+) (Novagen). The recombinant plasmids were confirmed by sequencing and used to transform into E. coli BL21 (DE3). The expression of the each target protein was induced for 18 h at 16˚C with 0.5 mM isopropyl 1-thio-β-D-galactopyranoside (IPTG) in LB broth containing 50 μg/ml ampicillin. The His6tag fusion proteins were loaded onto a Ni Sepharose 6 Fast Flow column (GE Healthcare, United States) and purified as previously described [16]. The recombinant protein concentrations were determined using a bicinchoninic acid (BCA) protein assay kit (Beyotime, China).

Glutaminase activity assay
Glutaminase activity was measured as described previously [17]. Samples of 8 μM PdxS, 8 μM PdxT, or 8 μM mixture of both proteins was incubated with 10 mM glutamine, 6 units of bovine glutamate dehydrogenase, and 0.5 mM 3-acetylpyridine adenine dinucleotide (APAD; Sigma) in a total reaction volume of 300 μl of 50 mM Tris-Cl (pH 8.0). The reaction mixture was incubated at 37˚C. The absorbance of each sample was read at a wavelength of 363 nm (OD 363 ).

PLP formation assay
PLP formation assay was performed using a procedure as previously described [17]. Samples of 8 μM PdxS, 8 μM PdxT, or 8 μM mixture of both proteins was incubated with 0.5 mM ribose 5-phosphate, 1mM DL-glyceraldehyde 3-phosphate, 10 mM glutamine or 10 mM ammonium sulfate was added to a final volume of 300μl of 50 mM Tris-HCl (pH 8.0). The reaction mixture was incubated at 37˚C. The absorbance of each sample was read at a wavelength of 414 nm (OD 414 ).

Construction of gene deletion mutants and complemented strains
The primers used for the construction of the deletion mutants S-8ΔpdxS and S-8ΔpdxT are listed in S1 Table. Primers SUF/SUR and SDF/SDR were used to amplify the two segments flanking with the pdxS gene. Using single-overlap extension PCR (SOE PCR), the fragment with a 487 bp internal deletion in the pdxS gene (from nt 54 to 539) was generated, and cloned into the conjugative vector pEMOC2 [18] to construct plasmid pEM4pdxS. Using E. coli β2155 and a single-step transconjugation system [19,20], plasmid pEM4pdxS was applied to introduce the pdxS mutation into the S-8 strain. After two homologous recombination steps, the A. pleuropneumoniae S-8ΔpdxS mutant was verified by sequencing and PCR analyses using primers SCF/SCR and RTSF/RTSR (Fig 2A).
The deletion of pdxT gene in A. pleuropneumoniae S-8 was performed using the same procedure as described above. The primers TUF/TUR, and TDF/TDR were used to generate a 491 bp internal deletion in the pdxT gene (from nt 7 to 497), and the S-8ΔpdxT mutant was confirmed by PCR analyses with the primers SCF/TCR and RTTF/RTTR ( Fig 2B).
The promoter of PdxS and PdxT was predicted using the BPROM program (http://www. softberry.com/). The pLpdxS plasmid was constructed by cloning the 1573-bp PCR product amplified with the primers SCF/SCR (S1 Table), which contained the entire pdxS open reading frame (ORF) and the upstream region to include the native promoter, into plasmid pLS88 [21]. Using SOE PCR, the 1264-bp fragment containing the upstream promoter and entire pdxT ORF was generated, and ligated into the plasmid pLS88 to yield plasmid pLpdxT plasmid (Fig 2A and 2B). The plasmids pLpdxS and pLpdxT were electroporated into S-8ΔpdxS and S-8ΔpdxT, respectively for trans complementation. The complemented strains were selected on BHI agar containing 20 μg/mL of kanamycin, and were verified by PCR using the primers SCF/SCR or TCF/TCR.

RNA isolation and qRT-PCR
For RNA isolation, A. pleuropneumoniae strains were grown to mid-logarithmic phase in 3 ml of BHI medium. The cultures were harvested by centrifugation at 10,000 g at 4˚C. Total RNA was extracted using RNeasy kit (Qiagen) and complementary DNA (cDNA) was synthesized using the PrimeScript RT reagent kit (TaKaRa, Japan) according to the manufacturer's instructions. The primers used for analysis of gene expression are listed in S2 Table. The cDNA samples were amplified using SYBR Green I (TakaRa). Real-time polymerase chain reactions were performed in a MicroAmp Optical 96-well reaction plate using qTOWER 2.2 system (Analytikjena, Germany). The quantitative qRT-PCR experiments were performed in triplicate, with recF as an internal control [22].

In vitro growth assays
In vitro growth assays were conducted as previously described [23]. The A. pleuropneumoniae wild-type (WT) S-8, S-8ΔpdxS mutant, S-8ΔpdxT mutant and complemented strains S-8ΔpdxScomp and S-8ΔpdxTcomp were grown in 3 ml of BHI medium supplemented with 10 μM PLP for 18 h, then washed three times with CDM medium to remove residual PLP and diluted to optical densities at 600 nm (OD 600 ) of 0.1. The fresh cultures in 10 ml of CDM medium were respectively supplemented with 10 μM PLP (Sigma), 1 mM ammonium sulfate or not supplemented any chemicals. The cultures were then incubated at 37˚C. The OD 600

In vitro stress assays
In vitro stress assays were performed as described previously [14]. The A. pleuropneumoniae WT S-8, S-8ΔpdxS mutant, S-8ΔpdxT mutant and complemented strains S-8ΔpdxScomp and S-8ΔpdxTcomp were grown to OD 600 0.8 in BHI medium supplemented with 10 μM PLP. We inoculated bacteria at different time points according to the growth curves to try to get all strains to achieve the OD600 of 0.8 at the same time. In addition, we inoculated bacteria at an interval of 0.5 h to make sure that. Cells (10 8 CFU/mL) of each strain from the broth cultures were washed three times with un-supplemented BHI medium, and harvested by centrifugation at 4000 g for 10 min. To test the tolerance of the cells to oxidative stress, cells of each strain were resuspended in 1 ml of BHI medium containing 5 mM H 2 O 2 and incubated in the presence or absence of PLP for 45 min at 37˚C. To test the tolerance of the cells to osmotic stress, cells of each strain were resuspended in 1 ml of BHI medium containing 0.4 M NaCl and incubated in the presence or absence of PLP for 45 min at 37˚C. Cells of each strain were resuspended in BHI medium in the presence or absence of PLP as controls and incubated for 45 min at 37˚C. All of the cultures of each stress assay were washed three times with BHI medium to remove residual H 2 O 2 or NaCl, and serially diluted and cultured on BHI agar plates supplemented with 10 μM PLP. The cell count was determined after 24 h of incubation. Values of stress resistance were calculated as [(stressed sample CFU/ ml)/ (control sample CFU/ ml)]×100.

Transmission electron microscopy
The A. pleuropneumoniae WT S-8, S-8ΔpdxS mutant, S-8ΔpdxT mutant and complemented strains S-8ΔpdxScomp and S-8ΔpdxTcomp were cultivated in BHI medium supplemented with 10 μM PLP at 37˚C for 16 h, then washed three times with un-supplemented BHI medium to remove residual PLP and diluted to optical densities at 600 nm (OD 600 ) of 0.05. The fresh cultures were grown to mid-exponential growth phase in 10 ml of BHI medium with or without 10 μM PLP. The cells were washed three times with PBS and fixed for 2 h in 2% osmic acid. Dehydration was performed in upgraded ethanol, and then the samples were embedded in SPI-Pon 812 resin (emicron) for two days. Resin-soaked sample blocks were polymerized at 70˚C, and samples were counterstained with uranyl acetate and lead citrate. Electron micrographs were recorded with a transmission electron microscope (JEM-1200EX, JEOL, Japan).

Mouse in vivo experiments
BALB/c mouse has been acknowledged as an appropriate model for A. pleuropneumoniae infection [24,25]. Specific-pathogen-free, six-week-old female BALB/c mice (Beijing Vital River Laboratory Animal Co., Ltd.) were used for virulence evaluation of A. pleuropneumoniae WT S-8, S-8ΔpdxS mutant, and S-8ΔpdxT mutant strains. Briefly, all A. pleuropneumoniae strains were cultured in BHI medium supplemented with 10 μM PLP at 37˚C, and harvested during the mid-exponential phase and washed three times with sterile PBS. A total of 160 mice were randomly divided into 16 groups (n = 10/group). For each strain, five experimental groups were inoculated intraperitoneally with 100 μL of PBS containing varying concentrations of bacterial suspension (10 5 -10 9 CFU). Non-infected mice in the control group were inoculated with 100 μL of sterile PBS (pH 7.4). Animal suffering was reduced with buprenorphine (0.05 mg/kg), given subcutaneously for analgesia every 12 hours during the first 72 hours after infection. The health status and the weight of the mice were monitored twice daily for a 14-day period and humane endpoints used to determine if the mice met criteria to be euthanized [26]. These criteria included weight loss >10-15%, lethargy, inability to stand, anorexia or flocked together for more than 6 hours. Mice meeting criteria were euthanized by cervical dislocation under isoflurane anesthesia. The 50% lethal dose (LD 50 ) of A. pleuropneumoniae WT strain S-8, S-8ΔpdxS mutant, and S-8ΔpdxT mutant strains were calculated as previously described by Reed-Muench [27].

Enumeration of bacterial load in organs
A total of 15 specific-pathogen-free, six week-old female BALB/c mice were randomly divided into 3 groups (n = 5), and each group was intraperitoneally administered with 1.0×10 7 CFU of WT S-8 strain, S-8ΔpdxS mutant, and S-8ΔpdxT mutant strains respectively. Three (3) days post-infection, mice from each group were humanely euthanized and the organs of lung, liver, and kidney were removed aseptically. Samples were weighed, and homogenized using a tissue homogenizer (100 mg weight/ml of PBS). Viable counts in serial dilutions of homogenates were determined following culture on BHI agar plates for 24 h at 37˚C.

Statistical analysis
All statistical analyses were performed using GraphPad Prism version 5.01 (GraphPad Software Inc., USA). The data are expressed as the means +/-standard deviation. P-values less than 0.05 were considered statistically significant.

PdxS and PdxT are involved in PLP synthesis in A. pleuropneumoniae
The ORFs of pdxS and pdxT genes are 888-bp and 576-bp in length, respectively, and encode the putative pyridoxal 5'-phosphate synthase subunits PdxS and PdxT. Searching the respective conserved domains in NCBI database revealed that PdxS has SOR_SNZ domain which represents a family of pyridoxal 5'-phosphate synthase and PdxT has SNO domain which represents glutamine amidotransferase family. Both PdxS and PdxT shared a high identity in the amino acid sequences with their orthologous sequences in Bacillus subtilis (84 and 55%, respectively) and Streptococcus pneumoniae (64 and 41%, respectively). In this study, to determine the activities of PdxS and PdxT in vitamin B6 formation in vitro, A. pleuropneumoniae pdxS and pdxT genes were overexpressed in E. coli BL21 (DE3) and the recombinant His6-tag fusion proteins were purified using Ni 2+ -affinity chromatography. SDS-PAGE analysis showed that the recombinant proteins, with the molecular weight of approximately 35.3 kD and 23.4 kD (S1 Fig), respectively, were obtained when the cells were induced with 0.5 mM IPTG at 16˚C for 18 h.
Glutaminase activity of PdxT was only detected in the presence of its partner protein PdxS. As shown in Fig 1B, varying the concentration of glutamine in the assay produced typical Michaelis-Menten kinetics, the catalytic constants K M and k cat were 1.72±0.14 mM and 6.12±0.14 min -1 respectively. Furthermore, the ribose 5-phosphate and glyceraldehyde 3-phosphate were served as substrates of PLP synthase PdxS and PdxT, and tested for their ability to form PLP. The kinetic constants were determined as K M = 159.8±13.5 μM, k cat = 0.020±0.0005 min -1 for ribose 5-phosphate (Fig 1C), and K M = 175.7±23.0 μM, k cat = 0.023 ±0.0011 min -1 for glyceraldehyde 3-phosphate, respectively (Fig 1D).
A. pleuropneumoniae S-8ΔpdxS and S-8ΔpdxT mutants are auxotrophic for PLP To demonstrate the function of PLP biosynthesis in A. pleuropneumoniae, pdxS and pdxTknockout mutants, S-8ΔpdxS and S-8ΔpdxT, respectively, were generated using the doublecrossover homologous recombination approach and confirmed by PCR and sequencing (Fig  2). The results of qRT-PCR showed that the transcription levels of the downstream genes were unaffected, confirming that the mutations in S-8ΔpdxS and S-8ΔpdxT strains were nonpolar (S2 Fig). Complementation of the S-8ΔpdxS and S-8ΔpdxT were achieved using the plasmid pLpdxS and pLpdxT, respectively (Fig 2). The complemented mutants were designated S-8ΔpdxScomp and S-8ΔpdxTcomp. The results of qRT-PCR showed that the transcriptions of pdxS in S-8ΔpdxS and pdxT in S-8ΔpdxT were virtually undetectable and partially restored in the complemented strains.
The growth properties of the A. pleuropneumoniae WT S-8, S-8ΔpdxS, S-8ΔpdxT, S-8ΔpdxScomp, and S-8ΔpdxTcomp strains were investigated in vitro. Upon cultivation in the absence of PLP, both S-8ΔpdxS and S-8ΔpdxT mutants exhibited severe growth defects (Fig 3A). It is suggested that PLP is essential for A. pleuropneumoniae growth in vitro. We therefore analyzed the growth of S-8ΔpdxS and S-8ΔpdxT mutants in the presence PLP supplementation, the defective growth of these mutants could be recovered, fundamentally similar to growth curves of the WT strain S-8 ( Fig 3B). The growth curves of the S-8, S-8ΔpdxScomp, and S-8ΔpdxTcomp strains were similar in the absence and presence of PLP. These data suggest critical roles for PdxS and PdxT proteins in A. pleuropneumoniae PLP biosynthesis.

The PLP synthesis activity of PdxS is not dependent on PdxT in the presence of ammonium
The growth properties of the A. pleuropneumoniae WT S-8, S-8ΔpdxS mutant, S-8ΔpdxT mutant and complemented strains S-8ΔpdxScomp and S-8ΔpdxTcomp were also analyzed in the presence of ammonium supplementation in vitro. With ammonium sulfate supplementation, the defective growth rate of S-8ΔpdxT mutant was partially restored, but the defective growth of the S-8ΔpdxS mutant remained unchanged (Fig 4A). We then examined PLP synthesis activity of PdxS using ammonium substitute for glutamine in the reaction. Unlike using glutamine as substrate, the individual PdxS could produce PLP in the addition of ammonium (Fig 4B and 4C). The PLP synthase had a specific activity of 996 ± 54 pmol min -1 mg -1 using glutamine as substrate, and of 508 ± 38 pmol min -1 mg -1 using ammonium sulphate as substrate.
A. pleuropneumoniae S-8ΔpdxS and S-8ΔpdxT mutants are sensitive to NaCl and H 2 O 2 The viability of the WT S-8, S-8ΔpdxS mutant, S-8ΔpdxT mutant and complemented strains S-8ΔpdxScomp and S-8ΔpdxTcomp were investigated when exposed to oxidative stress and osmotic stress conditions. As shown in S3A Fig, in the absence of PLP and when the cells were exposed to H 2 O 2 -induced oxidative stress, the viable counts of S-8ΔpdxS and S-8ΔpdxT were much lower than that of the S8 cells. The survival rates of the S-8ΔpdxS mutant and S-8ΔpdxT mutant were 18.19% and 34.73%, respectively, which were significantly lower (P < 0.01) than that of WT S-8 (75.68%) (Fig 5A). To exclude the influence of growth defects of S-8ΔpdxS and S-8ΔpdxT in the absence of PLP, the cells of each strain were also exposed to oxidative stress in the presence of PLP supplementation. The viable counts of S-8ΔpdxS and S-8ΔpdxT were still lower than that of the S8 cells, although increased compared to the viable counts of S-8ΔpdxS and S-8ΔpdxT exposed to oxidative stress in the absence of PLP (S3B Fig). As shown in Fig 5B, in the presence of PLP supplementation, 74.86% of WT S-8 survived. However, S-8ΔpdxS and S-8ΔpdxT exhibited 28.90% (P < 0.01) and 47.16% (P < 0.01) survival rates, respectively. Similar results were obtained when the cells were exposed to osmotic stress. The viable counts of S-8ΔpdxS and S-8ΔpdxT were lower than that of the S8 cells (S3C and S3D Fig). In the absence of PLP, WT S-8 cells under osmotic stress exhibited the 85.89% survival rate, whereas the S-8ΔpdxS mutant and the S-8ΔpdxT mutant respectively exhibited 32.08% (P < 0.01) and 52.13% (P < 0.05) survival rate (Fig 5C). In the presence of PLP supplementation, the survival rate of the S-8ΔpdxS mutant was 38.58%, which was significantly lower (P < 0.01) than that of WT S-8 (86.09%); and the survival rate of S-8ΔpdxT mutant was 59.16%, which was also significantly lower (P < 0.05) than that of WT S-8 (Fig 5D). These results suggest that PLP synthases PdxS/PdxT are linked to resistance to osmotic stress and oxidative stress.

Loss of pdxS and pdxT leads to morphological defects of A. pleuropneumoniae
The effects of PLP on the morphologies of WT S-8, S-8ΔpdxS mutant, S-8ΔpdxT mutant and complemented strains S-8ΔpdxScomp and S-8ΔpdxTcomp were observed using transmission electron microscopy. These data showed that the morphologies of the five strains were similar in the presence of PLP supplementation, and were consistent with the normal morphology of coccobacilli ( S4 Fig). However, in the absence of PLP, the S-8ΔpdxS mutant and S-8ΔpdxT mutant exhibited irregular and aberrant shapes, showing holes and deep craters on their surface. S-8ΔpdxS also appeared partially lysed or swollen (Fig 6). Approximately 49% of S-8ΔpdxS and 33.67% of S-8ΔpdxT cells were abnormal cells, whereas S-8, S-8ΔpdxScomp and S-8ΔpdxTcomp strains exhibited only 2.3%, 7.7% and 6% of abnormal cells, respectively ( Fig  6D). These data indicate that PLP production is required for maintaining normal cellular morphology of A. pleuropneumoniae. Roles of PdxS and PdxT in Actinobacillus pleuropneumoniae viability, stress tolerance and virulence Loss of pdxS attenuates the virulence of A. pleuropneumoniae in the BALB/c mouse model To address whether pdxS or pdxT deletion affected the virulence of A. pleuropneumoniae, BALB/c mice were inoculated intraperitoneally with WT S-8, S-8ΔpdxS mutant, and S-8ΔpdxT mutant strains at various doses, and LD50 values were determined. The LD 50 value of the S-8ΔpdxS mutant was 2.51×10 8 CFU, higher than 5.01×10 6 CFU for WT S-8. The virulence of the S-8ΔpdxS mutant was 50.1-fold attenuated compared to that of WT S-8. Conversely, the LD 50 value of the S-8ΔpdxT mutant (7.76×10 7 CFU) was found to be only 15.5-fold lower than WT S-8. These results suggested that the loss of both pdx enzymes attenuates the pathogenicity of A. pleuropneumoniae, with a greater impact observed with the deletion of pdxS.
The capacity of WT S-8, S-8ΔpdxS mutant, and S-8ΔpdxT mutant strains to colonize mice was then tested. The A. pleuropneumoniae load in tissues of systemically infected mice was determined by culturing the lungs, livers, and kidneys homogenates 3 days post-infection. As shown in Fig 7, the viable counts of the S-8ΔpdxS mutant in the lungs, livers and kidneys were significantly lower compared with those of WT S-8 (P<0.01). Similarly, significant differences (P < 0.05) in bacterial loads were also found between the WT S-8-inoculated mice and the S-8ΔpdxT-inoculated mice in lungs. In addition, the bacterial loads in the livers and kidneys of the S-8ΔpdxT-inoculated mice were decreased compared to the WT S-8-inoculated mice, although these differences were not statistically significant (Fig 7). Taken together, the results Roles of PdxS and PdxT in Actinobacillus pleuropneumoniae viability, stress tolerance and virulence showed that the S-8ΔpdxS mutant of A. pleuropneumoniae displayed an impaired colonization ability in BALB/c mice.

Discussion
In the A. pleuropneumoniae genome, only the adjacent pdxS and pdxT genes are annotated as belonging to the pdx family. Our results in the present study demonstrated that A. pleuropneumoniae synthesizes PLP via the DXP-independent pathway through the utilization of the two enzymes: the synthase subunit PdxS and the glutaminase subunit PdxT. In this study, we observed that respective disruption of the pdxS and pdxT genes rendered A. pleuropneumoniae auxotrophic for PLP (Fig 3), suggesting the importance of PLP biosynthesis in the growth of A. pleuropneumoniae. In addition, the results of the PLP synthase activity showed that both PdxS and PdxT proteins were required for PLP production in vitro in the presence of glutamine ( Fig  4B and 4C). These findings are consistent with the previous observations in other bacteria.
In this study, the supplementation of exogenous ammonium in the growth broth complemented the defective growth rate of the S-8ΔpdxT mutant, but did not influence the defective growth rate of the S-8ΔpdxS mutant (Fig 4A). Similar results have also been reported in the pdxT deletion mutant in S. pneumoniae and B. subtilis whose defective growth rate could also be compensated in the presence of ammonium [9,28], suggesting that the preliminary function of PdxT protein in PLP synthesis is producing ammonium from glutamine. Therefore, PdxS can produce PLP in the absence of PdxT when the glutamine is replaced by ammonium in the reaction (Fig 4C). These observations in this assay are in agreement with the proposed two-step model for PLP formation in previous studies [29,30]. In this model, PdxT depends on PdxS to hydrolyze glutamine into ammonium, and PdxS utilizes ammonium for generating PLP by incorporation with other chemicals.
Intriguingly, we found that the PLP synthases PdxS/PdxT were implicated in the stress responses of A. pleuropneumoniae. To the best of our knowledge, this is the first study to show the relation between the PLP synthases PdxS/PdxT and stress response in bacteria. From the data presented here, it appeared that S-8ΔpdxS and S-8ΔpdxT mutants showed different responses in vitro to stress conditions compared with WT S-8. We found that pdxS and pdxT gene deletion caused increased susceptibility of A. pleuropneumoniae to H 2 O 2 -induced oxidative stress and NaCl-induced osmotic stress (Fig 5). Though these findings were not described in bacteria, PLP formation has been linked to stress tolerance in plants [31][32][33]. In Arabidopsis thaliana, disruption of PLP synthases lead to reduced plant size, slowed root growth and Roles of PdxS and PdxT in Actinobacillus pleuropneumoniae viability, stress tolerance and virulence hypersensitivity to NaCl [31,32]. In addition, PLP production in A. thaliana has also been associated with defense against cellular oxidative stress since A. thaliana mutants deficient in PLP synthase were shown to be highly sensitive to oxidative stresses [33]. Combining our findings presented here and previous work by others, it is highly likely that PLP synthases PdxS/ PdxT play a vital role in osmotic stress and oxidative stress responses in A. pleuropneumoniae. However, the supplementation of exogenous PLP did not reverse the stress responses of the S-8ΔpdxS and S-8ΔpdxT mutants. We speculate that PdxS/PdxT may participate in other critical cellular processes involved in the biosynthesis of important enzymes associated with stress responses; however further investigations are required for confirmation.
Another objective of this study was to clarify whether PLP synthases PdxS/PdxT are essential for A. pleuropneumoniae pathogenicity in a mouse model. The results of this study indicated that the virulence and colonization of the S-8ΔpdxS and S-8ΔpdxT mutants were attenuated (Fig 7), although there were differences regarding the magnitude of attenuation, suggesting that the loss of the PLP biosynthesis pathway influences the pathogenicity of A. pleuropneumoniae in BALB/c mice. These findings are correlated with observations presented by other previous studies made in M. tuberculosis, H. pylori and S. pneumoniae [8,9,23]. We subsequently analyzed the probable reasons for attenuated virulence. It is likely that the PLP concentration in mice is insufficient for optimal growth of these A. pleuropneumoniae mutants in vivo. The PLP concentration in plasma of mice has been shown to fluctuate between 100 and 200 nM [34]. However, the addition of 100 nM of PLP to the broth did not significantly compensate the defective growth of the S-8ΔpdxS and S-8ΔpdxT mutants in vitro in our pilot experiments (data not shown), suggesting that the levels of PLP in mice may be too low to support the optimal growth of these mutants, which further affected the colonization and virulence of the S-8ΔpdxS and S-8ΔpdxT mutants. Another probability is the involvement of PLP in the biosynthesis of surface-exposed structures of A. pleuropneumoniae. Indeed, it has been shown that PLP biosynthesis is essential for the synthesis of surface-exposed structures involved in adhesion and colonization in some pathogenic bacteria, such as lipopolysaccharide (LPS) and flagella [23,35,36]. In this study, the S-8ΔpdxS and S-8ΔpdxT mutants exhibited aberrant shapes in the absence of PLP (Fig 6). However, whether this change directly affected the surface-exposed structures of A. pleuropneumoniae or then consequently affected the adhesion and colonization of A. pleuropneumoniae, needs further investigation.
Additionally, it is worth noting that the PLP biosynthesis pathway is present in various pathogenic bacteria, but is absent in humans and mammals [37,38]. Therefore, the de novo pathway of PLP biosynthesis is of particular interest as novel potential drug targets for the therapy of bacterial infections. The data presented in this study demonstrated that disruption of PdxS/PdxT pathway of PLP biosynthesis inhibited the normal growth of A. pleuropneumoniae. This finding offers the new possibility of prevention and treatment of A. pleuropneumoniae infection which ravages the swine industry. In conclusion, the present study revealed the critical role of PLP synthases PdxS/PdxT in viability, stress tolerance, and virulence related to A. pleuropneumoniae. In addition, bacteria are also able to synthesize PLP by a salvage pathway [39], in which PLP is synthesized from other B6 vitamers. Future studies should include attempts to determine whether this pathway is present in A. pleuropneumoniae and its function in A. pleuropneumoniae pathogenicity.