The two component sensor kinase system involving GacS and GacA is conserved in plant-associated pseudomonads and regulates production of many of the biocontrol active components. The system involves activation of phosphorylation of GacS by an as yet unknown signal followed by phospho-transfer to the GacS regulator. Phosphorylated GacA activates changes expression from genes encoding small regulatory RNAs, such as RsmX, RsmY and RsmZ, which compete with translational repressors, RsmA and RsmB (Brencic et al., 2009). The GacS/GacA regulon is extensive, for example encompassing about 10% of the genes in P. fluorescens Pf-5 (Hassan et al., 2010).

The genome of P. chlororaphis O6 (Loper et al., 2012) possesses genes potentially encoding GacS, GacA, and the proteins, RsmA and RsmB interacting with rRNAs. However, findings with P. chlororaphis O6 and other pseudo-monads reveal that the GacS-regulated traits differ between strains. For example, a gacS mutant of P. chlororaphis O6 has increased swimming and swarming motilities (Kim et al., 2014a), whereas lack of GacS in P. fluorescens Pf-5 has no effect on swimming, and decreases swarming activity, due to mainly loss in production of a surfactant (Hassan et al., 2010).

This study employed a proteomic approach for better understanding the genes regulated by GacS in P. chlororaphis O6, and we compared proteomes from the wild type and gacS mutant strains. Proteins that were down regulated in the gacS mutant were identified and their potential functions deduced. The role of two of these proteins, one involved in pyrrolnitrin production and the second in tryptophan biosynthesis, was explored using mutants in the GacS-regulated genes, prnA and trpE. The abilities of the prnA and trpE mutants to colonize roots and induce systemic resistance were investigated.

Bacteria were stored at −70°C in 25% glycerol. The gacS mutant and the gacS-complemented strain of P. chlororaphis O6 were constructed previously (Kang et al., 2004) and grown at 28°C with shaking at 200 rpm in King’s medium B broth. A prnA mutant of P. chlororaphis O6 was constructed previously (Park et al., 2011). Cultures of Escherichia coli DH5α were grown at 37°C on Luria-Bertani (LB) broth.

Extracts were prepared from the wild type, the gacS mutant and the complemented mutant by sonication of cells grown to stationary phase in LB and harvested by centrifugation. Proteins were extracted by washing cells twice in ice-cold phosphate buffered saline before suspension in sample buffer containing 7 M urea, 2 M thiourea, 4% (w/v) 3-[(3-cholamidopropyl dimethyammonio-1-propanesulfonate, 1% (w/v) dithiothreitol and 2% (v/v) pharmalyte and 1 mM benzamidine. After sonication for 10 seconds and incubation for one h at room temperature, the mixture was vortexed and centrifuged at 15,000 × g for one h at 15°C to obtain the soluble fraction used in gel analysis. Changes in protein profiles between the wild type, the gacS mutant, and the complemented gacS mutant were detected by two-dimensional gel electrophoresis using procedures described previously (Oh et al., 2013b). Protein spot intensities of wild type and gacS mutant on 2-D PAGE analyses were assessed by Student's t test. Three independent two dimensional protein analyses were performed, and significantly up- or down-regulated protein spots in the gacS mutant were selected and identified using Q-TOF analysis (Oh et al., 2013b).

Twelve protein spots were observed in the gacS mutant at lesser intensities than in the gels from the wild type strain (Fig. 2A). Based on a cut-off value of twofold change, peptide identification revealed they were a catalase/peroxidase with homology to KatG from P. fluorescens Pf-01 (Table 1), a tryptophan halogenase (PrnA), catalyzing the first step in pyrrolnitrin synthesis, a single-strand DNA binding protein (Ssb), a serine protease (PspB), a recombination associated protein (RdgC), and a potential secretin (CpaC), involved in pilus synthesis, and an outer membrane protein (OprF) (Table 1). Proteins showing fold changes between two- and four- fold were: a protein with a LysM domain associated with binding peptidoglycan (LysM), the anthranilate/para-aminobenzoate synthase component I (TrpE), involved in tryptophan synthesis from chorismate, glucose 1-phosphate thymidilate transferase (RmlA) required for rhamnose synthesis, a protein functioning in isoprenoid biosynthesis (ElbP), and glutathione peroxidase (Gpx) (Table 1).

The study was continued with a transcript analysis to determine the extents to which changes protein level in the gacS mutant correlated with altered gene expression. RNA accumulation from the genes predicted to encode the GacS-regulated proteins was evaluated by endpoint RTPCR. Cells grown in LB broth were harvested in stationary phase (OD_{600 nm} = 2.4). Total RNA was isolated using the Trizol method following the user's manual (GIBCO BRL, Rockville, MD, USA). RT-PCR was performed using the QuantiTect SYBR Green reverse transcription-PCR kit (Qiagen Inc., Valencia, USA). A reaction mixture of 25 μl was incubated at 50ºC for 30 min for reverse transcription, followed by RT-PCR with each primer set (Supplementary Table 1). A Rotor-Gene 2000 Real Time Cycler machine (Corbett Research Inc., Australia) was used for 35 cycles with denaturation at 94ºC for 15 seconds, annealing at 55°C for 30 seconds, and extension at 72ºC for 30 seconds. PCR reactions were stop after 20 cycles and PCR products were loaded on 2% agarose gel. The results provided are typical of three independent studies. Findings shown in Fig. 2 are from stationary phase cells and comparable loading between samples was demonstrated from the consistency of the PCR products for the 16S rRNA genes (Fig. 2).

Transcripts in the gacS mutant compared to the wild type were most reduced for katG, prnA, rdgC, cpaC and trpE. For other genes, expression was impaired (psbB, ssb and rmlA) by the gacS mutation. However, for gpx, lysM, and oprF transcript levels were near wild type. These findings suggest that other mechanisms other than direct effects on transcription for control of protein levels may be operating.

We have prior evidence that genes, such as prnA and trpE, showing changes in transcript levels in the gacS mutant, were part of the RpoS regulon in P. chlororaphis O6 (Oh et al., 2013a; Park et al., 2011). These findings agree with control of pyrronitrin production by quorum sensing as shown in another P. chlororaphis isolate PA23 (Selin et al., 2012) and specifically by mutation in gacA in P. chlororaphis isolate 30-84 (Wang et al., 2013). Indeed, extraction and assay showed pyrrolinitrin formation was eliminated in an rpoS mutant (Park et al., 2011), as later confirmed for another isolate by Selin et al. (2012). We believe that the observations with the gacS mutant occurred because rpoS transcription and RpoS protein abundance are controlled by GacS (Kang et al. 2004: Oh et al., 2013a). Two other proteins also were regulated similarly, the catalase/peroxidase KatG and a potential glutathione peroxidase, Gpx. Both of these proteins have significant roles in cellular protection against oxidative stress. Identical changes in catalase/peroxidase and superoxide dismutase isozyme patterns were observed in rpoS and gacS mutants (Oh et al., 2013b). A recent report indicated that P. chlororaphis 30-84 expression from rpoS is under GacA control (Wang et al., 2013).

It is interesting that two of the other GacS-regulated proteins have roles in DNA repair: the single strand DNA binding protein Ssb could be associated with repair of DNA breaks caused by hydrogen peroxide (Ananthaswamy and Eisenstark, 1977) and the protein, RdgC, is proposed to act with RecA to aid in DNA repair when there is damage in replication forks or by double strand breaks (Briggs et al., 2010).

Two proteins down-regulated in the absence of GacS, OprF and LysM, were associated with cell wall functions. OprF is a major outer membrane protein with multiple potential roles that includes the formation of outer membrane vesicles (Wessel et al., 2012). Studies with other pseudomonads also linked oprF expression with quorum sensing and Gac regulation. Crespo and Valverde (2009) observed mutations in the oprF gene sequence in P. fluorescens CHA0 that affected activity of the repressor proteins RsmA/E which are involved in GacA regulation. Additionally an oprF mutant in P. aeruginosa showed reduced production of the quorum-sensing signals, acyl homoserine lactones (Fito-Boncompte et al., 2011). The second GacS-regulated protein has a LysM domain associated with peptidoglycan binding (Bateman and Bycroft, 2000). It is possible that changes in the peptidoglycan binding proteins in the gacS mutant were involved in the elongated growth of these cells observed by atomic force microscopy for cells grown in a biofilm (Anderson et al., 2005) and SEM analysis (Kim et al., 2014b). Additional cell surface changes also are implicated due to the control by GacS in isolate P. chlororaphis O6 of the enzyme, RmlA, involved in generating rhamnose (Zuccotti et al., 2001), and a potential secretin, CpaC, involved in pilus formation (Bitter, 2003). We note changes in colony surface morphology and in biofilm formation for the gacS mutant compared to the wild type strain (Anderson et al., 2005; Kim et al., 2014b).

To characterize the role of the GacS-regulated genes, trpE and prnA mutants were constructed. A trpE mutant was made by homologous marker exchange mutagenesis. A PCR product from the trpE gene was generated using genomic DNA and two specific primers based on the genome sequence of P. chlororaphis O6 (Loper et al., 2012); forward (5′-ATG ATC CGC GAA GAA TTC CT-3′) and reverse (5′-TCA GTC CGG GGT TTG CTC GG-3′) in polymerase chain reactions (PCR). The PCR fragment (about 1.5 kb) was cloned into cloning vector pGEM-7Z. The trpE sequence in the pGEM-7Z plasmid was disrupted by the insertion of a 0.9 kb KpnI fragment containing the kanamycin resistance gene from the plasmid pRL648 (Elhai and Wolk, 1988). The chromosomal trpE gene in P. chlororaphis O6 was mutated using the exchange vector pCPP54 (Tc^R) and pRK2073 helper plasmid strain as previously described (Miller et al., 1997). The mutants were selected on LB agar containing 5% sucrose based on their sensitivity to tetracycline and resistance to kanamycin. A prnA mutant of P. chlororaphis O6 was constructed previously (Park et al., 2011).

The effects of mutations in prnA and trpE mutants on the requirement for tryptophan for growth was studied (Fig. 3) using cells pre-grown on LB agar plates for 2 days at room temperature. The cells were washed several times by re-suspension in sterile distilled water, and the suspension adjusted to OD_600nm = 0.1 before being applied to M9 minimal medium agar plates with or without a supplement of 5 mM tryptophan (Sigma Co., MO, USA) to assess growth. As anticipated, the trpE mutation rendered the mutant unable to grow on minimal medium without the addition of 5 mM tryptophan (Fig. 3). The prnA mutant grew well on the minimal medium with or without the tryptophan (Fig. 3).

Induction of systemic resistance by colonization of roots with the mutants was measured in tobacco cv. Xanthi seedlings using a challenge of the soft rot bacterium, Pectobacterium carotovorum SCCI (Han et al., 2006). Briefly, sterilized seeds were placed on 1 ml of 0.5% (w/v) MS agar supplemented with 3% sucrose contained in the wells of a 12-well microtiter plate (SPL Inc., Korea). After growth for three weeks, the seedlings were inoculated with suspensions of cells of the wild type, the gacS mutant, the complemented gacS mutant, the trpE mutant and the prnA mutant. The cells for the inocula were grown to an OD_600nm = 2.0 in LB broth, pelleted by centrifugation and suspended in sterile 50 mM potassium phosphate buffer (pH 7.5) to an OD_600nm = 0.2. The inocula, 10 μl, of bacterial suspensions containing 1 × 10⁸ colony forming units (cfu)/ml, were applied to the seedling roots. Sterile 50 mM potassium phosphate buffer was applied to control plants. One week after bacterial treatments, tobacco plants were challenged with P. carotovorum subsp. carotovora SCCI by pipetting 2 μl of pathogen inoculum onto a leaf (1× 10⁸ cfu/ml) as described by Han et al. (2006). One to two days after pathogen challenge, the extent of soft-rot was rated visually and the disease severity assessed. The trpE and prnA mutants displayed less ability than the wild type or the complemented gacS mutant to protect the plants against the pathogen (Fig. 4). The extent of protection from the prnA and trpE mutants was low and similar to that displayed by the gacS mutant.

The efficacy of root colonization of the wild type, the trpE mutant and prnA mutant was compared on the tobacco cv. Xanthi seedlings using methods described previously with three week old seedlings raised in the well plates (Han et al., 2006). Roots were excised at defined times after inoculation, and the fresh weight was measured before transferred into 10 ml of sterile distilled water. After vigorous vortexing for 1 min, serial dilutions of these washings were plated onto LB-agar plates containing antibiotics appropriate to the strain. Colonies were scored after incubation at 28°C for two days. Studies were repeated two times with three plants/each treatment. Root colonization in terms of cfu/g fresh weight roots was calculated for each time point. Data were statistically analyzed by ANOVA with the IBM SPSS Statistics version 21 (IBM Corp., Somers, New York, USA). The lack of induced resistance for the trpE mutant correlated with loss of root colonization which declined with time over a 15-d trial (Fig. 5). In contrast both the wild type and the prnA mutant colonized the root surfaces with increases in cfu/root during the first 10 d of the 15 d assessment period (Fig. 5).

The failure of the trpE mutant to colonize the plant roots suggested that the supply of tryptophan in the tobacco root exudates was inadequate to maintain wild type level of growth at the root surface. The ability of the prnA mutant to colonize but not to induce systemic resistance suggested that pyrronitrin was active as an effector. This suggestion indicated that pyrrolnitrin had a dual role for the biocontrol-active pseudomonad, both as a direct anti-microbial compound as well as an effector of induced resistance (Park et al., 2011). A similar role was determined for the phenazines produced by this bacterial isolate (Kang et al., 2007). However, our findings of impaired colonization with the trpE mutant emphasized that the formation of pyrrolnitrin in the rhizosphere would be influenced by the supply of tryptophan in the plant root exudates.

In summary, the proteomics analysis of control by GacS confirmed its role as a key regulator of proteins with anticipated roles not only in the formation of antimicrobials but also in oxidative stress, cell signaling, secretion and cell surface properties. Production of KatG and Gpx, involved in protection against oxidative stress, overlapped between the rpoS mutant and the gacS mutant. The newly identified GacS-regulated proteins in this paper, Ssb, PspB, RdgC, LysM, RmlA, and ElbP, indicate further the diverse role of this regulatory system in a rhizosphere-competent pseudomonad with biocontrol potential.