PtrA Is Functionally Intertwined with GacS in Regulating the Biocontrol Activity of Pseudomonas chlororaphis PA23

In vitro inhibition of the fungal pathogen Sclerotinia sclerotiorum by Pseudomonas chlororaphis PA23 is reliant upon a LysR-type transcriptional regulator (LTTR) called PtrA. In the current study, we show that Sclerotinia stem rot and leaf infection are significantly increased in canola plants inoculated with the ptrA-mutant compared to the wild type, establishing PtrA as an essential regulator of PA23 biocontrol. LTTRs typically regulate targets that are upstream of and divergently transcribed from the LTTR locus. We identified a short chain dehydrogenase (scd) gene immediately upstream of ptrA. Characterization of a scd mutant revealed that it is phenotypically identical to the wild type. Moreover, scd transcript abundance was unchanged in the ptrA mutant. These findings indicate that PtrA regulation does not involve scd, rather this LTTR controls genes located elsewhere on the chromosome. Employing a combination of complementation and transcriptional analysis we investigated whether connections exist between PtrA and other regulators of biocontrol. Besides ptrA, gacS was the only gene able to partially rescue the wild-type phenotype, establishing a connection between PtrA and the sensor kinase GacS. Transcriptomic analysis revealed decreased expression of biosynthetic (phzA, prnA) and regulatory genes (phzI, phzR, rpoS, gacA, rsmX, rsmZ, retS) in the ptrA mutant; conversely, rsmE, and rsmY were markedly upregulated. The transcript abundance of ptrA was nine-fold higher in the mutant background indicating that this LTTR negatively autoregulates itself. In summary, PtrA is an essential regulator of genes required for PA23 biocontrol that is functionally intertwined with GacS.


INTRODUCTION
Public concern over the use of chemical pesticides together with the potential for acquiring resistance to these compounds has led to renewed interest in alternative strategies for management of diseases affecting plants. Pseudomonas chlororaphis strain PA23 is a soybean root-tip isolate that demonstrates excellent antifungal (AF) activity against Sclerotinia sclerotiorum (Lib.) de Bary (Savchuk and Fernando, 2004). Strain PA23 produces an arsenal of compounds including the diffusible antibiotics phenazine 1-carboxylic acid (PCA), 2hydroxyphenazine (2-OH-PHZ), and pyrrolnitrin (PRN) together with degradative enzymes (Poritsanos et al., 2006;Zhang et al., 2006;Selin et al., 2010). PRN is the primary antibiotic responsible for biocontrol; conversely, phenazines (PHZ) are not essential for fungal suppression, but do play a role in biofilm formation (Selin et al., 2010). Despite the fact that many biocontrol agents perform well in the greenhouse, they often exhibit reduced efficacy under field conditions (Cook, 1993;Walsh et al., 2001;Haas and Keel, 2003). Variable expression of genes and gene products required for biocontrol likely contributes to poor performance in the field. It is essential, therefore, that molecular mechanisms underlying biocontrol are well understood so that production of the pathogen-suppressive factors can be optimized in the environment.
In both pathogenic and biocontrol pseudomonads, expression of secondary metabolites is controlled by a multi-tiered network of regulation. Situated at the top of this hierarchy is the GacS/GacA two-component signal transduction system, comprised of the sensor kinase GacS and its cognate response regulator GacA (Heeb and Haas, 2001). For many biocontrol strains, including PA23, a mutation in gacS or gacA leads to a loss of fungal antagonism (Heeb and Haas, 2001;Poritsanos et al., 2006). A second system, called Rsm, functions in concert with Gac and consists of a combination of RsmAlike repressor proteins and untranslated regulatory RNAs. The repressor proteins act at the post-transcriptional level by binding to and blocking the ribosome-binding site (RBS) of target mRNA (Lapouge et al., 2008). The regulatory RNAs antagonize repression by titrating out the RsmA-like proteins, rendering RBSs accessible to the translational machinery (Lapouge et al., 2008). Several additional regulatory components govern expression of PA23 AF metabolites including the PhzR/PhzI quorum-sensing (QS) system , the stationary phase sigma factor RpoS (Manuel et al., 2012), a regulator of RpoS called PsrA , and the stringent response (Manuel et al., 2012). Cross-regulation between the regulators themselves adds to the increasingly complex nature of this regulatory hierarchy (Manuel et al., 2012;Selin et al., 2012Selin et al., , 2014. We have recently identified a novel regulator in PA23 called PtrA (Pseudomonas transcriptional regulator) (Klaponski et al., 2014). The phenotype of a ptrA mutant is very similar to that of a gac-deficient strain exhibiting a complete loss of AF activity (Poritsanos et al., 2006;Selin et al., 2014). PtrA belongs to the LysR-type transcriptional regulator (LTTR) family, which is the most abundant class of transcriptional regulators found among prokaryotes (Schell, 1993). LTTRs act as either activators or repressors and are known to control a diverse range of metabolic functions including cell invasion and virulence, QS, oxidative stress, and amino acid metabolism (Cao et al., 2001;Sperandio et al., 2002;Kim et al., 2004;Heroven and Dersch, 2006;Byrne et al., 2007;Kovaleva and Gelfand, 2007;Hernández-Lucas et al., 2008;Maddocks and Oyston, 2008). Preliminary proteomic and phenotypic analysis of a ptrA mutant revealed 59 differentially expressed proteins together with decreased PHZ and PRN production, consistent with the loss of AF activity (Klaponski et al., 2014).
LTTRs frequently control expression of genes that are upstream of and divergently transcribed from the lttr locus (Schell, 1993;Maddocks and Oyston, 2008). Immediately upstream of ptrA lies a gene encoding a short-chain dehydrogenase, designated scd. At present, it is not known what role scd plays in PA23 biological control. Moreover connections between PtrA and other members of the regulatory network have not been investigated. The aim of the current study was to conduct greenhouse studies to establish whether PtrA is required for PA23-mediated control of Sclerotinia stem rot. We also generated an scd mutant and determined its role in PA23 fungal suppression. Finally, a combination of complementation and transcriptional analysis was used to explore interactions between PtrA and other regulators of PA23 biocontrol.

PCR
Polymerase Chain Reaction (PCR) was performed as follows. Each reaction contained 2.5 µL of both forward and reverse primers (12 µM), 1 µL of template DNA, 10 µL of 10× Taq Buffer (without Mg added), 1.5 µL of MgSO 4 , 1 µL of Taq polymerase (Thermo Fisher Scientific, Carlsbad, USA) and nuclease-free water to a final volume of 100 µl. PCR reaction conditions included an initial denaturation at 98 • C for 2 min, followed by 30 cycles of 98 • C for 30 s, 55 • C for 30 s, and 68 • C for 1 min/kb, followed by a final extension of 68 • C for 5 min.

Nucleic Acid Manipulation
Cloning, purification, electrophoresis, and other manipulations of nucleic acid fragments and constructs were performed using standard techniques (Sambrook et al., 1989).

Generation of a scd Mutant
To generate PA23scd, a 542-bp internal fragment of the scd gene was PCR amplified from PA23 genomic DNA using primers scd-pKNOCK FWD and scd-pKNOCK REV. The amplicon was gel purified and digested with BamHI and XhoI and cloned into the same sites of pKNOCK-Tc. The pKNOCK-scd plasmid was then mobilized into PA23 through triparental mating with the donor strain E. coli DH5αλpir containing pKNOCK-scd and the helper strain DH5α (pRK600). Pseudomonas Isolation Agar (PIA, Difco) supplemented with Tc (150 µg/mL) was used to screen for transconjugants. Insertion of the plasmid into scd was Frontiers in Microbiology | www.frontiersin.org confirmed by PCR using primers tet FWD and new ptrA TL start FWD followed by sequencing of the amplicon.

Plasmid Construction
For complementation analysis, pUCP22-retS and pUCP22-ladS were generated as follows. retS was PCR amplified from PA23 genomic DNA using primers retS-F2 and retS-R2. The 2.9-kb amplicon was digested with BamHI & HindIII before cloning into the same sites of pUCP22. The ladS gene was amplified using primers ladS-F-BamHI and ladS-R-HindIII. The 2.8-kb ladS-containing fragment was subject to digestion with BamHI & HindIII and cloned into pUCP22 digested with the same enzymes. pUCP22-retS and pUCP22-ladS were verified through sequence analysis.

Antifungal Assays
Radial diffusion assays to assess fungal inhibition in vitro were performed according to previously described methods (Poritsanos et al., 2006). Five replicates were analyzed for each strain and assays were repeated three times.

Greenhouse Assays
Strains PA23 (pUCP22) suspended in sterile distilled water with 0.02% Tween 20 and kept in a growth chamber (24/16 • C, 16-h photoperiod). Twentyfour hours after bacterial inoculation, plants were sprayed with ascospores of S. sclerotiorum (8 × 10 4 spores mL −1 ) suspended in water containing 0.02% Tween 20. Pathogen control plants were inoculated with ascopores only, while healthy control plants were sprayed with water (0.02% Tween 20). After pathogen inoculation, plants were incubated in a humidity chamber for 72 h, after which they were placed back in the growth chamber. Fourteen days after ascospore inoculation, symptom development on the stem and leaves was scored according to Selin et al. (2010). Ten plants were used for each treatment and the plant studies were repeated two times.

Pyrrolnitrin Analysis
Production of the antibiotic PRN was quantified according to the methods outlined by Selin et al. (2010). Briefly, 20 mL cultures of PA23 and its derivatives were grown for 5 days in M9 minimal media (1 mM MgSO 4 ; 0.2% glucose) and PRN was extracted with an equal volume of ethyl acetate. Before extraction, toluene was added to each sample as an internal control. Toluene and PRN UV absorption maxima were recorded at 225 nm with a Varian 335 diode array detector. PRN peaks were detected at 4.7 min. Samples were analyzed in duplicate.

Autoinducer Analysis
The production of AHL was analyzed by spotting 10 µL of an overnight culture onto LB agar plates seeded with C. violaceum CV026. This strain is able to detect exogenous AHLs with carbon chain length structures ranging from C4 to C8, resulting in a purple halo surrounding the colonies. The diameter of the purple zones was measured at 24 h.

Semi-Quantitative Reverse Transcriptase PCR
To monitor expression of genes involved in biocontrol, semiquantitative reverse transcriptase (RT) PCR was used. PA23 and its derivatives were grown to early stationary phase and total RNA was extracted using a RNeasy Mini Kit (QIAGEN, Valencia, USA). Residual genomic DNA was removed by treatment with TURBO RNAase-free DNAse I (Ambion, Carlsbad, USA) during the RNA isolation procedure. RNA concentrations were measured at 260 and 280 nm and only RNA samples with A 260 /A 280 between 1.8 and 2.0 were used in subsequent steps. cDNA was generated by reverse transcription using the Maxima First Strand cDNA Synthesis Kit (ThermoScientific, Rockford, USA) and random hexamer primers in a 20 µL total reaction volume. The following conditions were employed: initial heating at 25 • C for 10 min, 50 • C for 15 min for reverse transcription and 85 • C for 5 min for enzyme denaturation. Sequences for the genes of interest from PA23 were obtained from GenBank. The primer sequences are listed in Table 1. PCR was performed using a CFX96 Connect TM Real-Time PCR Detection System (Bio-rad, Hercules, USA) and SsoFast TM EvaGreen R Supermix (Bio-rad). The final 10-µL volume mixture in each well contained 0.4 µL of both forward and reverse primers (12 µM), 1 µL of 1:20 diluted cDNA, 5 µL of SsoFast TM EvaGreen R Supermix and 3.4 µL of nuclease-free water. PCR reaction conditions included an initial denaturation at 98 • C for 2 min, followed by 39 cycles of 98 • C for 5 s, 60 • C for 30 s, and 60 • C for 5 s. Melt-curve analysis was performed to evaluate the formation of primer dimers and other artifacts to validate results. Each reaction was performed in triplicate and experiments were repeated three times with three biological replicates. Relative gene expression was calculated using the ∆∆Ct method as described by Livak and Schmittgen (2001) using rpoB as the reference gene and the CFX Manager TM software (Bio-rad).

Statistical Analysis
All statistical analysis was performed using unpaired Students's t-test.

PtrA is Essential for PA23 Biocontrol of S. sclerotiorum in the Greenhouse
The wild-type PA23, PA23-443 and the complemented PA23-443 (ptrA-pUCP22) were tested for their ability to protect canola from stem rot disease caused by S. sclerotiorum. Two parameters were evaluated, namely incidence of leaf infection and stem rot disease severity. As illustrated in Figure 1, the ptrA mutant showed a significant reduction in its ability to control fungal infection of leaves and stems and reduce overall disease severity. Compared to the disease control, the ptrA mutant mediated a modest decrease in leaf infection ( Figure 1A) and no difference in disease severity ( Figure 1B). Addition of ptrA in trans restored the ability of PA23-443 to prevent both leaf infection and stem rot to wild-type levels (Figure 1).

scd, Which Lies Upstream of PtrA, Does Not Appear to Be Involved in PA23 AF Activity
A 115-bp intergenic region separates ptrA and an upstream gene encoding a short-chain dehydrogenase, designated scd. To determine whether this allele is involved in PtrA regulation, an scd insertional mutant was generated. Unlike the ptrA mutant that is devoid of AF activity, the scd mutant showed near wildtype fungal suppression (Figure 2). Due to the plasmid insertion, only the first 100 nt (of 600) of the scd open reading frame remain intact (data not shown). It seems highly unlikely that a functional truncated Scd is being produced, but we cannot rule out this possibility entirely.

gacS is Able to Partially Complement the PtrA Mutant
To reveal interactions between PtrA and other members of the regulatory hierarchy overseeing PA23 biocontrol, plasmid-borne copies of regulatory genes constitutively expressed from the lac promoter were transformed into the ptrA mutant. Genes that are able to fully or partially complement the mutant are predicted to lie downstream of PtrA in the regulatory cascade. We began our characterization of the PA23-443 transformants by analyzing AF activity. As expected, providing ptrA in trans restored AF activity close to wild-type levels (Figure 2). The only other gene that resulted in partial complementation of fungal suppression was gacS (Figure 2). PA23 produces the diffusible antibiotics PHZ and PRN. Antibiotic analysis revealed that strain PA23-443 synthesized markedly lower levels of both compounds, whereas no difference in antibiotic production was observed for the scd mutant ( Table 2). Addition of ptrA in trans led to partial and full restoration of PHZ and PRN production, respectively. Consistent with the AF analysis, plasmid-borne gacS resulted in partial complementation of antibiotic synthesis in the ptrA mutant ( Table 2). Our protease activity profiles closely mirrored what was observed for the antibiotics. The ptrA mutant was devoid of protease production, whereas the scd mutant showed wildtype activity (Figure 3). Addition of ptrA and gacS in trans lead to full and partial rescue of protease activity in the ptrA mutant background, respectively (Figure 3).

PtrA Regulates AHL Signal Production
The AHL signal generated by the PhzRI QS system activates the CVO26 biosensor resulting in a white to purple color change due to the production of the QS-controlled pigment violacein. While CVO26 cells form a purple halo around colonies of PA23 and the scd mutant, this is not observed for PA23-443 indicating that the latter is AHL deficient. Addition of ptrA, gacA, and gacS in trans rescued AHL production in the ptrA mutant to varying degrees (Figure 4).
downregulated, indicating positive regulation by PtrA. Next, we examined the orphan sensor kinase-encoding retS and ladS genes. A decrease in retS activity was observed; whereas ladS transcription remained at wild-type levels. The gene showing the most dramatic change in transcript abundance was ptrA. In the PA23-443 background, ptrA expression increased nine fold indicating that this LTTR is subject to negative autoregulation. No change in scd transcription was observed further supporting that PtrA regulation is not mediated through this divergently transcribed gene. Quantitative RT-PCR was used to monitor FIGURE 5 | Biocontrol gene expression in PA23-443 compared to wild type. qRT-PCR fold change in gene expression in the ptrA mutant vs. wild type was determined using rpoB as the reference gene. The level of gene expression in the PA23 wild type was normalized to 1.0 (indicated by the dotted line). Gene expression that differs significantly from wild type is indicated with an asterisk (*p < 0.05; **p < 0.01; ***p < 0.0001).
expression of three genes immediately upstream of scd, encoding a membrane protein (AIC18438.1; primers Up1), a hypothetical protein (AIC18437.1; primers Up2) as well as nhaA (AIC18435.1; primers Up3). No differences in transcript abundance were detected in the ptrA mutant compared to the wild type (data not shown).

Phylogenetic Analysis of PtrA Homologs
Phylogenetic analysis revealed that PtrA was conserved amongst pseudomonads, including both biocontrol and pathogenic strains. Homologs from well-known biocontrol strains P. chlororaphis O6 and P. protegens Pf-5 were found to cluster with PtrA (Figure 6). A similar pattern was observed when a phylogenetic tree was constructed using genomic sequences (Supplementary Figure 1). As closely related homologs of PtrA are found in several biocontrol strains, they are expected to play a similar role in regulating genes responsible for the production of antifungal compounds. A list of the PtrA homologs used for the phylogenetic analysis can be found in Supplementary Tables 1, 2.

DISCUSSION
LTTRs, which represent the largest family of prokaryotic transcriptional regulators, frequently regulate divergently transcribed genes; however, targets can be located elsewhere on the chromosome (Schell, 1993;Maddocks and Oyston, 2008). Several pieces of evidence suggest that PtrA fits into this second paradigm, functioning as a global transcriptional regulator. First, 59 differentially regulated proteins distributed across 16 different COG categories were identified in the ptrA mutant (Klaponski et al., 2014) and genes encoding these proteins are scattered about the chromosome (Loewen et al., 2014). Second, insertional inactivation of the divergently transcribed scd gene upstream of ptrA resulted in no observable phenotype. AF activity, antibiotic and AHL production, and protease activity were similar to wild type (Figures 2-4; Table 2). Moreover, expression of scd was unchanged in the ptrA mutant ( Figure 5). These findings indicate that scd is not involved in PA23 biocontrol and does not appear to be linked to PtrA. Finally, our qRT-PCR results showed altered expression of several biosynthetic and regulatory genes involved in PA23 biocontrol (Figure 5). For the most part, gene expression profiles corresponded well with the ptrA mutant phenotype. For example, phzA and prnA were both significantly downregulated (Figure 5). Negligible expression of both genes is consistent with the loss of orange pigmentation, fungal suppression (Figure 2) and antibiotic production ( Table 2) exhibited by the ptrA mutant. In strain PA23, the Phz QS system regulates expression of antibiotics and degradative enzymes and so it is necessary for biocontrol . Transcriptional profiling revealed that phzI and phzR were significantly decreased in the ptrA mutant compared to the wild type. Reduced phzI expression coincides with the loss of AHL signal production in this background (Figure 4). While rpoS expression was markedly down in the ptrA-deficient strain, psrA transcription was elevated, albeit modestly ( Figure 5). Because RpoS is a negative regulator of PA23 biocontrol (Manuel et al., 2012) the decrease in rpoS transcription was unexpected as the ptrA mutant is no longer capable of fungal suppression. It is important to note, however, that cross regulation occurs between the PhzRI QS system and RpoS , which may obscure interpretation of findings. We also explored genes FIGURE 6 | Molecular phylogenetic analysis of PtrA homologs by the Maximum Likelihood method. The evolutionary history was inferred using the Maximum Likelihood method based on the Le Gascuel Method (Le and Gascuel, 1993). This analysis involved 41 amino acid sequences including PtrA. The tree with the highest log likelihood is shown. The percentage of trees in which the associated taxa clustered together is indicated next to the branches.
belonging to the Gac-Rsm regulatory network. In P. protegens strain CHA0, this regulatory network functions as follows. Upon binding to an unknown signal, the sensor kinase GacS undergoes autophosphorylation and phosphotransfer to the response regulator GacA (Lapouge et al., 2008). Phosphorylated GacA can then activate expression of sRNAs, including RsmXYZ, that titrate out the translational repressors RsmA and RsmE allowing expression of target genes (Lapouge et al., 2008). While all of these components have been identified in strain PA23, the way in which they function has not yet been explored. In the closely related P. chlororaphis 30-84, RsmE was found to repress production of PHZ and AHL signaling molecules; however, RsmA exerted no regulatory effect over these compounds (Wang et al., 2013). Moreover, of the three small RNA molecules, only constitutively expressed rsmZ was able to rescue a gacA mutant for PHZ and AHL production; rsmX exhibited no effect and elevated levels of rsmY were reportedly lethal (Wang et al., 2013). Collectively, these findings suggest that in terms of PHZ and AHL production in P. chlororaphis 30-84, RsmE is the primary repressor protein and RsmZ is the small RNA responsible for lifting repression. If we assume that this circuitry functions in a similar manner in strain PA23, the elevated rsmE expression and reduced rsmZ transcription are consistent with the loss of AF activity exhibited by the ptrA mutant.
Even though gacS is not under PtrA transcriptional control, a regulatory link clearly exists between the two. Our phenotypic assays showed either full or partial complementation of the ptrA mutant by gacS when provided in trans. We hypothesized that PtrA might be controlling expression of regulatory elements that impact signal transduction through the Gac system and that overexpression of gacS is able to overcome this effect. In P. aeruginosa PAO1, the orphan sensor kinases RetS and LadS modulate the Gac-Rsm circuitry (Ventre et al., 2006;Goodman et al., 2009). RetS blocks GacS autophosphorylation and subsequent activation of GacA, while LadS exerts a positive effect on Gac regulation (Ventre et al., 2006;Goodman et al., 2009). To explore whether RetS and LadS are in some way linked to PtrA, we attempted to complement the ptrA mutant by providing retS and ladS in trans. No changes in fungal suppression, protease activity, antibiotic and AHL production were observed in the ptrA mutant harboring plasmid-borne copies of these genes (Figures 2-4; Table 2). We did, however, discover a two-fold decrease in retS transcription in PA23-443 ( Figure 5). Because RetS functions to antagonize the Gac system, decreased retS expression is expected to have a positive impact on biocontrol. Taken together, these findings do not support a role for RetS and LadS in the PtrA -GacS interaction.
Phylogenetic analysis illustrates that a wide range of Pseudomonas species harbor a PtrA homolog (Figure 6). For both symbiotic and pathogenic pseudomonads, secreted products play a significant role in the biocontrol and virulence properties of these organisms. In addition, regulatory factors overseeing their expression are in many cases conserved. For example, the Gac two-component system and QS positively regulate exoproducts that play a key role in symbiotic and pathogenic interactions (Heeb and Haas, 2001;Bassler, 2002). It is not surprising, therefore, to find that PtrA is highly conserved amongst Pseudomonas species. To the best of our knowledge, this transcriptional regulator has not been characterized in other pseudomonads.
In summary, we have shown that PtrA is essential for biocontrol of S. sclerotiorum stem rot of canola. PtrA appears to function as a global regulator controlling expression of unlinked genes across the chromosome. Future studies will be directed at analyzing the PtrA transcriptome on a global scale so we can better comprehend how this LTTR is controlling expression of biocontrol factors in PA23. Due to the conserved nature of this regulator, we hypothesize that PtrA governs expression of secondary metabolites in biocontrol strains and pathogenic pseudomonads alike. Elucidating the mechanisms underlying PtrA regulation will have far-reaching implications for our understanding of how these bacteria interact with other members of their environment including prokaryotic and eukaryotic organisms.

AUTHOR CONTRIBUTIONS
NS, NK, WF, MB, and Td conceived and designed the study. NS and NK drafted the manuscript with input from Td. NS, NK, and RR performed the phenotypic characterization of the ptrA mutant; CS was responsible for the greenhouse analysis. All authors read and approved the final manuscript.