A Highly-Conserved Single-Stranded DNA-Binding Protein in Xanthomonas Functions as a Harpin-Like Protein to Trigger Plant Immunity

Harpins are produced by Gram-negative phytopathogenic bacteria and typically elicit hypersensitive response (HR) in non-host plants. The characterization of harpins in Xanthomonas species is largely unexplored. Here we demonstrate that Xanthomonas produce a highly conserved single-stranded DNA-binding protein (SSBX) that elicits HR in tobacco as by harpin Hpa1. SSBX, like Hpa1, is an acidic, glycine-rich, heat-stable protein that lacks cysteine residues. SSBX-triggered HR in tobacco, as by Hpa1, is characterized by the oxidative burst, the expression of HR markers (HIN1, HSR203J), pathogenesis-related genes, and callose deposition. Both SSBX- and Hpa1-induced HRs can be inhibited by general metabolism inhibitors actinomycin D, cycloheximide, and lanthanum chloride. Furthermore, those HRs activate the expression of BAK1 and BIK1 genes that are essential for induction of mitogen-activated protein kinase (MAPK) and salicylic acid pathways. Once applied to plants, SSBX induces resistance to the fungal pathogen Alternaria alternata and enhances plant growth. When ssbX was deleted in X. oryzae pv. oryzicola, the causal agent of bacterial leaf streak in rice, the resulting ssbXoc mutant was reduced in virulence and bacterial growth in planta, but retained its ability to trigger HR in tobacco. Interestingly, ssbXoc contains an imperfect PIP-box (plant-inducible promoter) and the expression of ssbXoc is regulated by HrpX, which belongs to the AraC family of transcriptional activators. Immunoblotting evidence showed that SSBx secretion requires a functional type-III secretion system as Hpa1 does. This is the first report demonstrating that Xanthomonas produce a highly-conserved SSBX that functions as a harpin-like protein for plant immunity.

PTI is a form of basal defense or nonhost-mediated resistance in plants. PTI and effector-triggered immunity (ETI) activate similar signaling pathways and defense responses in plants. However, ETI generally activates a more prolonged, robust resistance than PTI [2]. Signal transduction pathways associated with PTI and ETI include mitogen-activated protein kinase (MAPK) cascades, calcium fluxes, and the activation of reactive oxygen species (ROS). Furthermore, both ETI and PTI are associated with modulations in hormonal signaling pathways including those associated with production of salicylic acid (SA) for systemic acquired resistance (SAR), jasmonic acid (JA) for induced systemic resistance (ISR) and ethylene (Eth) [17]. Unlike ETI, PTImodulated signaling requires BAK1, which is a BRI1-ASSOCI-ATED KINASE 1 that regulates plant signaling by functioning as an adaptor for multiple RLKs [2,[17][18][19][20]. For example, the FLS2/ BAK1 complex phosphorylates BIK1 (Botrytis-induced kinase 1) for signal transduction to the MAPK cascade [21]. The latter may then activate the expression of WRKY transcription factors that regulate SA-, JA-or Eth-dependent genes by binding the W-box [22]. However, it is unclear whether the proteins mentioned above are also involved in harpin-triggered immunity.
The genus Xanthomonas contains 307 species or pathovars that infect at least 124 monocotyledonous and 268 dicotyledonous plants and causes enormous agricultural losses [46]. Despite the huge host range of Xanthomonas, few species in this genus are known to cause disease on tobacco, suggesting that tobacco may sense a conserved molecule in Xanthomonas and potentially initiate plant immunity. In this report, we present evidence that a highlyconserved single-stranded DNA-binding protein (SSB X ) in Xanthomonas is regulated by HrpX, secreted via the T3SS, required for full virulence in planta, and elicits HR in nonhost plants. These novel results indicate that SSB X functions as a harpin-like protein and modulates plant immunity in tobacco.
Determination of SSB X -elicited HR in planta. To investigate whether ssb Xoc triggers HR in tobacco, full-length ssb Xoc (552 bp) was amplified by PCR with the primer pairs ssb X -F/ssb X -R (Table S2) using the genomic DNA of strain RS105 as template. The amplified product was then cloned into PVX vector pgR107 [54] at ClaI and SalI sites, resulting in pPVXssb X (Table S1). The hpa1 [8] and bax [55] genes were also cloned into pgR107, generating pPVXhpa1 and pPVXbax (Table S1), which were used as controls. These constructs (along with the empty vector) were transferred into A. tumefaciens GV3101, resulting in strains SSB X , Hpa1, Bax and PVX, respectively. Suspensions of A. tumefaciens strains were adjusted to OD 600 = 0.5 and infiltrated into N. benthamiana with needleless syringes. HR symptoms were scored 48 hours post inoculation (hpi). Three independent experiments were performed and similar results were yielded. Representative results from one of these experiments are presented here.

SSB Protein Expression and Purification
SSB Xoc homologues were amplified from X. oryzae pv. oryzicola RS105, X. oryzae pv. oryzae PXO99 A , X. campestris pv. vesicatoria 85-10, X. axonopodis pv. citri 306, X. campestris pv. campestris 8004, R. solanacearum ZJ3721 and E. amylovora 0065, E. coli BL21(DE3), P. syringae pv. tomato DC3000, and P. fluorescens (Table S1). Each ssb homologue was amplified by PCR from corresponding genomic DNAs using the primers listed in Table S2. PCR products were then cloned into pET41a (+) resulting in constructs designated pSSB Xoc, pSSB Xoo , pSSB Xcv , pSSB Xac , pSSB Xcc , pSSB Rs , pSSB Ea , pSSB Ec , pSSB Pst , and pSSB Pf respectively (Table S1). These constructs were transformed into E. coli BL21 (DE3) (Table S1) as recommended in the Novagen pET System manual (Novagen, USA). Proteins were expressed as recommended by Novagen. Briefly, a single colony of each recombinant strain was inoculated to 2 ml LB broth containing Km. After incubation at 37uC for 12 h, 2 ml of culture was transferred into 200 ml of fresh LB liquid containing 1.0 mM isopropyl b-D-thiogalactopyranoside (IPTG, Sigma) and incubated for 4 h at 37uC. Cells were harvested by centrifugation, and pellets (1 g) were resuspended in 5 ml PBS buffer (137 mM NaCl, 2.7 mM KCl, 10 mM Na 2 HPO 4 and 2 mM KH 2 PO 4 , pH 7.4); the solution also contained 20% glycerol, 5 U/ml DNaseI and 5 ml PMSF (phenylmethanesulfonylfluoride). Bacterial cells were lysed by sonication (20 kHz, 10 min). After centrifugation (15,0006g) for 15 min at 4uC, the supernatants were purified using a GSTrap TM FF column as recommended by the manufacturer (Purification Manual, GE Healthcare, Germany). The purified proteins were digested by thrombin to remove the GST-tag, and the purified proteins were quantified with the Easy Protein Quantitative Kit (TransGen Biotech, China) and a NANODROP 1000 Spectrophotometer (Thermo). Purified proteins were then used for HR induction assays in tobacco. The Hpa1 protein purified and the empty vector preparation (EVP) by the same procedure was used as a positive and negative control, respectively.
HR assays. Purified proteins were tested for their ability to elicit HR on N. benthamiana or N. tabacum cv. Xanthi by infiltration into plant tissues using needleless syringes. Plant responses were observed 48 hpi for the HR. All plants were grown in growth chambers at 25uC with a 12 h photoperiod. Experiments were repeated at least three times.
To measure minimal HR-eliciting concentrations, purified SSB Xoc and others were diluted in PBS buffer at the following concentrations: 50, 25, 10, 5.0, 2.5, 1.0, 0.5, 0.1, 0.05 and 0.01 mM, while the purified Hpa1 diluted in PBS at the same concentrations above was used as positive control. All concentrations of the tested SSB proteins and Hpa1 were infiltrated into tobacco leaves and photographed 48 hpi.
To characterize biochemical activity, purified SSB Xoc (1 mM) and Hpa1 (1 mM) were heat-treated at 100uC for 10 min and incubated with protease K (0.5 U/ml) at 37uC for 10 min, respectively. To investigate potential susceptibility to eukaryotic metabolic inhibitors, SSB Xoc (1 mM) and Hpa1 (1 mM) were mixed with 1 mM LaCl 3 , 0.71 mM actinomycin D, and 0.1 mM cycloheximide, respectively. Treated and untreated SSB Xoc , Hpa1, and EVP were infiltrated into tobacco leaves. Three independent biological experiments were performed and yielded similar results. Representative results from one of these experiments are presented.
To determine whether the HR induced by SSB Xoc was dependent on SA accumulation in planta, purified SSB Xoc (1.0 and 5.0 mM) was infiltrated into wild-type and NahG tobacco, respectively. Purified Hpa1 (1.0 and 5.0 mM), wild-type strain RS105 (OD 600 = 0.5), and EVP were used as controls. Three independent biological experiments were performed and yielded similar results. Representative results from one experiment are presented in this report.
DNA laddering assays. Genomic DNA of cv. Xanthi leaves infiltrated with purified SSB Xoc (1 mM), Hpa1 (0.5 mM) and PBS buffer was isolated at 3, 6, 12, 24, 36 and 48 hpi, respectively. DNase-free RNase A was then used to digest existing RNA. Genomic DNA (10 mg) from each sample was subjected to electrophoresis in 2% agarose gels for at least 10 h under low voltage. Three independent experiments were performed and similar results were yielded. Representative results from one experiment are shown here.
H 2 O 2 assays. Tobacco leaves (N. benthamiana) were infiltrated with purified SSB Xoc (1 mM), Hpa1 (0.5 mM), and EVP, respectively, using needleless syringes. Eight hours later, treated leaves were collected and incubated in diaminobenzidine (DAB) for 8 h at 25uC, and then boiled in 95% ethanol for 10 min to remove the dye [56]. After 4 h further incubation in ethanol, leaves were fully bleached and brown precipitates were observed, indicating H 2 O 2 accumulation and the production of ROS. Epidermal peels were performed at the injection sites 0 and 8 hpi with purified SSB Xoc (1 mM), Hpa1 (0.5 mM), and EVP; these were then stained with DAB for 8 h at 25uC, and visualized with an OLYMPUS IX71 microscope. Three independent experiments were performed and similar results were yielded. Representative results from one experiment are presented here.
Callose deposition assays. To observe callose deposition, tobacco leaves (N. benthamiana) were infiltrated with purified SSB Xoc (1 mM), Hpa1 (0.5 mM), or EVP with a needleless syringe. After infiltration (0 and 8 h), leaf epidermal peels of the infiltrated area were removed and incubated with aniline blue (0.1% in 0.15% K 2 HPO 4 , pH 8.2) for 0.5 h. Fluorescence (400 nm excitation and 510 nm emission) and bright-field images were obtained with an OLYMPUS IX71 microscope. Three independent experiments were performed and similar results were yielded. Representative results from one experiment are displayed here.
Northern blot assays. Total RNA of N. benthamiana leaves infiltrated with SSB Xoc (1 mM), Hpa1 (0.5 mM) and EVP, was extracted 8 hpi. Prior to electrophoresis, RNA samples were treated with RNase-free DNaseI (TaKaRa, China) to remove potential traces of genomic DNA. RNA samples (30 mg) were then separated by electrophoresis in 1% agarose gels. Biotin-labeled DNA probes were prepared with the BrightStar Psoralen-Biotin Labeling kit (Ambion, USA) as recommended by the manufacturer. The primers for DNA probes are listed in Table S2. RNA was transferred to Hybond N+ membranes (Amersham Pharmacia Biotech, USA), hybridized with specific probes (Table S2) at 42uC using Northern Max (Ambion, USA), and detected using Bright-Star BioDetect (Ambion) according to the manufacturer's instruction.

Assays for Plant Growth Promotion and Disease Prevention by SSB Xoc
To detect potential plant growth promoting activity of SSB Xoc , seeds of tobacco cv. Xanthi and Arabidopsis. thaliana (Col-0) were treated with SSB Xoc (1 mM), Hpa1 (0.5 mM), EVP, and sterile water (DDW) for 8 h at 4uC. Treated seeds were placed on MS agar medium and measured for root length and fresh weight two weeks after treatment.
The potential effect of SSB Xoc in enhancing plant disease resistance was investigated on tobacco inoculated with A. alternata strain TBA28A (Table S1), the causal agent of brown spot disease. Ten plants of two-month-old tobacco were spray-inoculated with SSB Xoc (1 mM in 0.5% Tween 20 solution) or Hpa1 (0.5 mM); plants were sprayed twice in three-day intervals. EVP was used as a negative control. Three days after the second spray, plants were inoculated with A. alternata TBA28A fresh disc. Infection was measured as the diameter of necrotic brown spots by statistical analysis.
Construction of ssb Xoc deletion mutants. Experiments were designed to generate nonpolar deletion mutants of ssb X in X. oryzae pv. oryzicola RS105 and RDhpa1, a hpa1 deletion mutant [45]. The DNA sequences flanking ssb X were amplified from RS105 genomic DNA using primer pairs ssbI-F/ssbI-R and ssbII-F/ssbII-R (Table S2), cloned into pMD18-T (Takara, China), and verified by sequencing. After digestion with BamHI/XhoI and XhoI/PstI, the two fragments were cloned into the suicide vector pKMS1 [57] at BamHIand PstI sites, resulting in pKDssb X (Table  S1). This construct was introduced into the wild-type RS105 and the hpa1 deletion mutant RDhpa1, and the isolation of ssb X deletion mutants was performed as described previously [57]. The ssb Xoc deletion mutant RDssb X and the double mutant RDhpa1Dssb X were verified by PCR using primers ssbI-F/ssbII-R (Table S2) and by Southern blot analysis using ssb X as a probe.
Bacterial pathogenicity and HR assays. Pathogenicity assays were performed as described previously [8]. X. oryzae pv. oryzicola derivatives were examined for their ability to cause disease symptoms in rice IR24 (Oryza sativa ssp. indica) or to trigger a HR in tobacco cv. Xanthi. Rice adult plants (two-months-old) were inoculated by leaf-needling and fully-expanded tobacco leaves were infiltrated by needleless syringes with bacterial suspensions (,3610 8 cfu/ml). Lesion lengths in rice were scored 14 days postinoculation (dpi) and the HR in tobacco 2 dpi. All plants were maintained in growth chambers at 25uC with a 12 h photoperiod. Experiments were repeated at least three times.
Measurement of bacterial growth in rice. Bacterial cell suspensions (3610 8 cfu/ml; OD 600 = 0.3) were infiltrated into recently expanded leaves of two-week-old rice IR24 with needleless syringes at three spots per leaf. Three 0.8 cm diameter leaf discs were harvested with a cork borer from each infiltrated area. The leaf discs were surface sterilized with 70% ethanol first and then with 30% hypochlorite, macerated with a sterile mortar and pestle in 1 ml of distilled water, serial dilutions were plated in triplicate on NA with appropriate antibiotics. Plates were incubated at 28uC for 3-4 days until single colonies could be counted. Bacterial numbers (cfu/cm 2 ) were calculated, and standard deviations were determined using colony counts from three triplicate spots in each of three samples per time point per inoculum. Experiments were repeated at least three times.
Promoter activity assays and quantitative real-time PCR. To construct a transcriptional fusion between the ssb X promoter and glucuronidase (GUS), the promoter region (21 to 2350 bp) upstream of ssb X was amplified from the genomic DNA of X. oryzae pv. oryzicola RS105 with the primer pair pssb-F/pssb-R (Table S2). This PCR product was then fused with the promoterless gusA gene, which was obtained with primers gusA-F/gusA-R (Table S2). The ssb X -gusA fusion was then cloned into pUFR034 [58] at the EcoRI site, resulting in pPIPAGUS (Table S1). In another experiment, a mutation was introduced into the PIP-box of the ssb X promoter using primers mpssb-F/pssb-R (Table S2) and fused with gusA, resulting in pPIPBGUS (Table S1).
For quantitative real-time PCR analysis (qRT-PCR), the bacteria were cultured as described for the GUS activity assay in this report or cultured in rice suspension cells as described by Li and her colleagues [44]. Total RNA was extracted using Trizol reagent according to the manufacturer's instructions (Invitrogen, USA). Total RNA was quantified by measuring the OD 260 / OD 280 , and the quality was examined by gel electrophoresis. Before synthesis of the first stranded, total RNA was treated with RNase-free DNaseI (TaKaRa, China) to remove genomic DNA. Removal of contaminating DNA was confirmed by using extracted RNA as a template to amplify selected target genes using the primers listed in Table S2. cDNA synthesis and PCR were conducted with AMV (TaKaRa) and Ex-Taq DNA polymerase (TaKaRa, China) using the primers listed in Table S2. Semiquantitative RT-PCR was performed using the following program: step 1, 95uC for 3 min; step 2, 95uC for 20 s; step 3, 55uC for 30 s; step 4, 72uC for 40 s; 35 cycles of steps 2-4; and step 5, 72uC for 10 min. Quantitative real-time PCR was performed on an Applied Biosystems 7500 Real-Time thermocycler using SYBR Premix Ex Taq TM (Takara, China). Conditions for quantitative RT-PCR were as follows: denaturation at 95uC for 30 s and 41 cycles for 95uC, 5 s; 60uC, 34 s. The results were analyzed using Applied Biosystems 7300 System SDS software and the RQ study application. Expression of the 16S rRNA gene was used as an internal standard to verify the absence of significant variation in cDNA levels. The comparative-threshold method by log 2 value was used to calculate the relative mRNA level with respect to the corresponding transcript of ssb Xoc in the wild-type RS105 and the hrpG and hrpX mutants (Table S1) cultured in NA medium and rice suspension cells, respectively. All the real-time quantitative RT-PCRs were performed in triplicate.
SSB Xoc secretion assays. To determine whether the secretion of SSB Xoc was dependent on the T3SS, X. oryzae pv. oryzicola strains, containing Ssb Xoc -c-myc or hpa1-c-myc fusion (as a positive control) (Table S1), were pre-incubated in NB medium to logarithmic phase. Bacterial cells were harvested and adjusted to OD 600 = 2.0 with sterilize water and washed twice. Then, 40 ml of bacterial suspension was poured into 1 ml of the hrp-inducing medium XOM3 [49] and incubated at 28uC for 8 h. Cell and supernatant fractions were separated by centrifugation, and the protein in the supernatant fraction was precipitated with 12.5% trichloroacetic acid. Proteins were separated on 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gels and transferred to membranes for immunoblotting using anti-c-Myc primary antibody (Genescript, China). Primary antibodies were recognized by anti-rabbit secondary antibodies (Genescript, China) and visualized on autoradiographs with the Western-Light Chemiluminescence System (Transgene, Beijing, China). Three independent experiments were performed and yielded similar results. Representative results from one of these experiments are presented here.

Results
ssb Xoc Encodes a Single-stranded DNA-binding Protein Eliciting HR in Tobacco Mutagenesis of hrpG or hrpX in X. oryzae pv. oryzicola abolishes the elicitation of HR in tobacco and pathogenicity in rice [8]. Thus, we assumed that the expression of HR-eliciting genes, including hpa1, are also regulated by HrpG and HrpX [60]. Using cDNA microarrays of X. oryzae pv. oryzicola strain RS105 and the hrpG & hrpX mutants (unpublished), we discovered that the expression of XOC_1514, which encodes a single-stranded DNAbinding protein (AEQ95695.1) [61], was positively regulated by HrpG and HrpX in pathogen-infected rice cells (Fig. 1). This protein, which was designated SSB Xoc , is rich in glycine (20% of the total amino acids) but lacks cysteine residues (Table S3, Fig. S1); these are characteristics typical of the harpin protein family. To confirm this, we expressed ssb Xoc in PVX vector pgR107, which is typically used to screen HR elicitors in tobacco [54]. SSB Xoc triggered HR in N. benthamiana that was similar to Hpa1 [8] and Bax [55] (Fig. 2A), suggesting that SSB Xoc functions as a harpin in X. oryzae pv. oryzicola.
Previously, we reported that the minimum concentration of Hpa1 for HR induction is 0.1 mM [8]. To determine the concentration of SSB Xoc required for HR induction, we overexpressed the protein in E. coli BL21 (DE3) ( Table S1). Purified SSB Xoc was infiltrated into tobacco at concentrations ranging from 0.01 to 50 mM. The minimum concentration of SSB Xoc needed for HR induction in tobacco cv. Xanthi was 1.0 mM, approximately 10-fold higher than the minimum effective concentration of Hpa1 (Fig. 2B). oryzae pv. oryzicola wild-type RS105 and mutants RDhrpG and RDhrpX. Strains were grown in NB or rice suspension cells, and designated as (2) and (+), respectively. The ratios (shown in units of log 2 ) reflect ssb Xoc transcript levels between different strains in two different growth conditions. 1. +RS105/2RS105; 2. 2RDhrpG/2RS105; 3. 2RDhrpX/ 2RS105; 4. +RDhrpG/+RS105; 5. +RDhrpX/+RS105. Data represent the means 6 standard deviations (SD) from three replicates. doi:10.1371/journal.pone.0056240.g001 Nucleotide and protein searches using the NBCI database (http://blast.ncbi.nlm.nih.gov/Blast.cgi) indicate that SSB Xoc homologues exist in other bacteria. Protein sequence alignment of SSB Xoc with homologues from other Gram-negative bacteria indicated that the differences of SSB proteins between Xanthomonas and other prokaryotic bacteria mainly existed in the glycine-rich regions (see rectangle marked with dashes, Fig. S1). A phylogenetic analysis showed that SSB proteins could be classified into one of three groups (Fig. S2). Group I contained SSB proteins from closely related Xanthomonas species, group II SSB homologues from Xyllela fastidiosa, R. solanacearum, Thermus aquaticus, P. aeruginosa, and P. syringae pv. tomato, and group III from Candidatus Liberibacter asiaticus, P. fluorescens, E. amylovora, Dickeya dadantii, Escherichia coli and Shigellia dysenteriae (Fig. S2). Percentages of glycine-rich amino acids of SSB X in X. oryzae pv. oryzicola RS105 strain and other bacteria are also shown in Table S3.
Electrophoretic mobility shift assays (EMSA) demonstrated that SSB Xoc from X. oryzae pv. oryzicola, as the representative of these proteins in Xanthomonas bound randomly synthesized singlestranded DNAs (DNA1 and DNA2, Fig. S3). This is consistent with that the single-stranded DNA-binding protein is for ssDNA protection from nucleolytic digestion in bacterial cell viability [62], implying that SSB X of Xanthomonas rather than other plant pathogenic bacteria is coevolved to be recognized as a potential HR-elicitor by plants.
Previous reports indicate that harpin proteins are heat-stable and protease-sensitive [6,7,34,41,42]. To investigate these characteristic for SSB X. , purified protein (1 mM) was incubated in a water bath at 100uC for 10 min and with protease K (0.5 U/ml) at 37uC for 10 min, while Hpa1 was used as positive control. Heat-or protease-treated SSB Xoc was then inoculated into tobacco cv. Xanthi. At 48 hpi, heat-treated SSB Xoc still triggered HR in tobacco ( Fig. 2D panel 8), but protease-treated SSB Xoc did not (Fig. 2D, panel 9).

SSB Xoc -elicited HR is a form of Programmed Cell Death
The next experiments were designed to determine whether SSB Xoc , like Hpa1, is toxic to plant cells or not and SSB Xoc leads to a metabolic response that triggers HR. Three metabolic inhibitors were used: actinomycin D (inhibits eukaryotic RNA polymerase II), cycloheximide (targets 80S ribosomes), and LaCl 3 (a calcium channel blocker). These inhibitors were incubated with purified SSB Xoc (see Methods) and then assayed for HR induction in tobacco. All three inhibitors prevented the SSB Xoc -elicited HR in tobacco when co-infiltrated with the purified SSB Xoc (Fig. 2D, panels [12][13][14]. These results indicated that the SSB Xoc -elicited HR is an active process and requires de novo gene expression, protein synthesis and calcium flux across membranes. Thus, SSB Xoc acts as an elicitor, like Hpa1, of HR but is not directly toxic to plant cells. It is well-documented that harpin-elicited HR is a form of programmed cell death (PCD), which is accompanied by DNA laddering [63]. To determine whether the SSB Xoc -elicited HR is a form of PCD, DNA laddering experiments were performed. Total genomic DNA from SSB Xoc -infiltrated tobacco leaves were extracted at different time points after infiltration and analyzed on 2% agarose gels. As shown in Fig. 2E, DNA ladders were clearly observed in SSB Xoc -inoculated leaves at 24 hpi, 12 h later than that in Hpa1-inoculated leaves. Thus, SSB Xoc , like Hpa1 (Fig. 2E), elicits PCD that is characterized by DNA laddering.
We then investigated whether SSB Xoc -elicited HR occurs with the activation of known HR marker genes including HIN1 (harpininduced 1) [64], HSR203J [65], and the SA-dependent marker, PR1a [66]; the JA-dependent gene, PR1b [67], was also conducted. The expression of these genes was evaluated in tobacco leaves infiltrated with SSB Xoc , Hpa1, and EVP 6 hpi. All four genes were induced in response to SSB Xoc ; however, Hpa1 did not induce the expression of PR1b (Fig. 2F). Transcripts started to accumulate 6 hpi with SSB Xoc and Hpa1 and were expressed at high levels up to 24 h (data not shown). These findings indicate that SSB Xoc -elicited HR was accompanied by the expression of HR markers and plant defense genes.

SSB Xoc -elicited HR is Dependent on SA Accumulation
It has been reported that HR induction by harpins requires SA accumulation [28,68]. To investigate whether this is valid for SSB Xoc -elicited HR, we utilized transgenic tobacco expressing NahG; this line produces salicylate hydroxylase which degrades SA and blocks its accumulation [69]. Purified SSB X and Hpa1 produced a HR in wild-type tobacco (Fig. 3B), but not in the NahG line (Fig. 3A). Thus, SSB Xoc -induced HR relies on SA accumulation in planta, which is the case for other harpins. It is noteworthy that the wild-type RS105 of X. oryzae pv. oryzicola still elicited HR in SA-deficient tobacco (Fig. 3A). Thus, in addition to SSB Xoc and Hpa1, other unidentified HR-elicitor(s) exist(s) in X. oryzae pv. oryzicola to trigger HR on tobacco in SA-independence.

SSB Xoc Activates Plant Basal Defense
The oxidative burst, which involves the generation of ROS in response to microbial elicitors, occurs quite quickly in resistant plant cells [70,71]. Thus we investigated whether the oxidative burst is generated in SSB Xoc -treated tobacco cells. At 8 hpi, DAB staining resulted in necrotic brown spots indicative of H 2 O 2 production in both SSB Xoc -and Hpa1-infiltrated leaves (Fig. 3C).
Along with the oxidative burst, plants often mobilize multiple forms of basal defense to inhibit pathogen ingress, including callose deposition in cell walls [72]. To determine whether SSB Xoc elicits callose deposition in tobacco, epidermal peels from SSB Xocinfiltrated tissue were stained with aniline blue and examined by fluorescence microscopy. Both SSB Xoc -and Hpa1-infiltrated leaves showed evidence of callose deposition beginning at 8 hpi (Fig. 3D). Thus, SSB Xoc , like Hpa1, functions as an elicitor of basal defense responses and stimulates callose deposition.
The oxidative burst and callose deposition in tobacco infiltrated with SSB Xoc prompted us to speculate that SSB Xoc may function as a PAMP and activate the expression of genes involved in PTI signaling pathways. Our results indicate that BAK1, BIK1 and MAP3K (Fig. 3E), and NPR1, EIN2, COl1 and PR4 (Fig. 3F) genes are activated in response to SSB Xoc or Hpa1. The data show that SSB Xoc triggers a cascade of events similar to those triggered by Flg22 [12,71,73,74], which leads to the oxidative burst and callose deposition and activates the expression of PR genes. These results support our presumption that SSB Xoc acts as a PAMP like Hpa1.

SSB Xoc Induces Plant Disease Resistance and Promotes Plant Growth
Tobacco infiltrated with SSB Xoc shows elevated expression of SA-and JA-dependent genes, along with the oxidative burst and callose deposition (Fig. 2F; Fig. 3C, D). Therefore, we hypothesized that SSB Xoc may stimulate induced resistance to pathogen infection. For this, we inoculated a fungal pathogen, A. alternata TBA28A (Table S1), causal agent of tobacco brown spot disease, to fully-expanded tobacco leaves that were previously sprayinoculated twice with SSB Xoc (1 mM) in three-day intervals. The necrotic areas in tobacco leaves treated with SSB Xoc were significantly smaller (P = 0.01, t test) than those observed on leaves inoculated with EVP (Fig. 4). Like SSB Xoc , we found that Hpa1 (0.5 mM) also induced similar resistance to A. alternata. The data suggest that both SSB X and Hpa1 induce SAR against pathogen infection.
The application of a harpin, HrpN from E. amylovora, enhances plant growth, particularly because the Eth-dependent genes are activated [30,75]. The activation of Eth-dependent genes, e.g. EIN2 and PR4 (Fig. 3F), led us to determine whether SSB Xoc promotes plant growth or not. For this, germinating seeds of Arabidopsis thaliana Col-0 and tobacco cv. Xanthi were soaked in a solution containing 1 mM SSB Xoc for 8 h and then transferred to MS medium for 14 days. The results showed that root lengths of SSB Xoc -treated Arabidopsis and tobacco plants were nearly 2-fold longer than plants treated with EVP or water. There was no significant difference in root lengths between SSB Xoc and Hpa1 treatments (P = 0.01, t test; Fig. 5). In addition, fresh weight of SSB Xoc -treated plants, like those treated with Hpa1, was nearly three times more than those treated with EVP or water (Fig. 5B). Thus, SSB Xoc promotes plant growth, possibly through the activation of the Eth signaling pathway (Fig. 3F).

ssb Xoc is Required for Full Virulence and Bacterial Growth in Rice
To investigate the potential contribution of SSB X to virulence, ssb Xoc was deleted both in X. oryzae pv. oryzicola RS105 and the hpa1 deletion mutant, RDhpa1 (Table S1). The RDssb X Dhpa1 double mutant (Table S1) was constructed using a two-step integration procedure [57]. Inoculation studies were conducted by inoculating one half of a rice leaf with wild-type RS105 and the remaining half with one of the following: ssb Xoc deletion mutant RDssb X , hpa1 mutant RDhpa1 [44], the double mutant RDssb X Dhpa1 (Table S1), the complemented strain CRDssb X and the T3SS mutant RDhrcV [76]. Symptoms in RDssb X -inoculated leaves were reduced relative to the wild-type strain, but were not as attenuated as RDhpa1mediated symptoms (Fig. 6A). Lesion lengths in RDssb X -inoculated leaves were significantly smaller than those induced by the wildtype RS105 (P = 0.01, t test) but larger than those induced by RDhpa1 (Fig. 6B). The double mutant RDssb X Dhpa1 did not lose pathogenicity in rice (Fig. 6A), but lesions were significantly smaller than those induced by the wild-type and single mutants (RDssb X and RDhpa1) (Fig. 6B). As expected, the T3SS mutant, RDhrcV, produced no obvious disease symptoms in rice (Fig. 6A,  B). Disease lesion lengths for the complemented strain CRDssb X were equivalent to those induced by the wild-type RS105 (Fig. 6A,  B), indicating that the mutant could be complemented for symptoms with the ssb Xoc gene in trans.
To determine whether ssb xoc contributes to growth of X. oryzae pv. oryzicola in rice, we compared the population dynamics of the wild-type RS105, RDssb X , CRDssb X , RDhpa1, RDssb X Dhpa1, and RDhrcV. The populations of RDssb X were significantly lower than the wild-type RS105 beginning 2 dpi, but higher than the population of RDhpa1 and RDssb X Dhpa1 (Fig. 6C). Growth of RDssb X was restored to wild-type levels when ssb Xoc was present in trans (Fig. 6C). These results indicated that ssb Xoc , like hpa1, contributes to bacterial growth in planta, although the effect was not as pronounced as seen with the T3SS mutant (RDhrcV) or RDhpa1. Furthermore, mutations in ssb Xoc and hpa1 did not abolish the ability of the pathogen to elicit HR in tobacco (data not shown), implying that other HR-elicitor(s) exist in X. oryzae pv. oryzicola.

SSB X Secretion is Dependent on a Functional TTSS
The T3SS deficient mutant RDhrcV did not trigger HR in tobacco implies that HR elicitors, including SSB Xoc and Hpa1, may be secreted via the T3SS. Bioinformatics analysis of SSB Xoc did not show obvious T3SS secretion signals that are commonly found in T3SS effector proteins [77,78]; thus it was not clear whether SSB Xoc secretion required a functional T3SS. We used immunoblotting and SSB Xoc tagged with a c-Myc epitope to Figure 3. SSB Xoc from X. oryzae pv. oryzicola may function as a PAMP and activates PTI in tobacco. SSB Xoc -triggered HR depends on the accumulation of salicylic acid (SA). X. oryzae pv. oryzicola strain RS105 (OD 600 = 0.5), SSB Xoc (1.0 and 5.0 mM), Hpa1 (1.0 and 5.0 mM), and EVP were inoculated into a NahG tobacco leaves (A) or wild-type tobacco cv. Xanthi (B). Photographs were taken 48 hpi. (C) SSB Xoc -triggered HR is accompanied by the oxidative burst. The production of H 2 O 2 was evaluated in tobacco leaves by staining with 3,39-diaminobenzidine tetrahydrochloride (DAB). The reaction mixture contained 200 ml of 0.5 mM DAB in 50 mM Tris acetate buffer (pH 6.0) with purified SSB Xoc (1 mM) or Hpa1 (0.5 mM). Fully-expanded tobacco leaves were infiltrated with needleless syringe, incubated at room temperature for 0 and 8 h, and decolorized in 80% (v/v) ethanol for 10 min at 70uC. Leaves were examined with an OLYMPUS IX71 microscope. PBS buffer was included as a negative control. (D) SSB Xoc elicits callose deposition in tobacco cell walls. Callose deposition in tobacco leaves was observed using fluorescence microscopy (OLYMPUS IX71) and staining with aniline blue at 0 and 8 hpi. Purified SSB Xoc (1 mM) or Hpa1 (0.5 mM) was infiltrated into tobacco leaves (N. benthamiana) using needleless syringes, and EVP was used as a negative control. Inoculated epidermal peels were incubated with aniline blue for 0.5 h. Fluorescence images were captured using a 400 nm exposure for 510 nm absorbed light (UV), and bright-field (BF) images were captured using general bright light. Panels (E) and (F) show Northern blot analysis of PTI signaling pathways in tobacco treated with SSB Xoc , Hpa1, and PBS buffer (control). (E) The expression of BAK1, BIK1 and MAP3K; (F) The expression of NPR1, EIN2, COl1, and PR4. Purified SSB Xoc (1 mM) or Hpa1 (0.5 mM) was infiltrated into tobacco leaves using needleless syringe, and PBS buffer was used as a negative control. RNA was extracted 8 hpi and 10 mg aliquots were separated on 1% agarose gels and transferred to nylon membranes. Blots were hybridized with digoxigenin-labeled cDNA probes of the indicated genes. doi:10.1371/journal.pone.0056240.g003 Figure 4. SSB Xoc induces resistance to tobacco brown spot disease caused by A. alternata. Fully-expanded tobacco leaves (cv. Xanthi) were sprayed twice in three-day intervals with purified SSB Xoc (1 mM), Hpa1 (0.5 mM) and EVP (negative control). Three days after the second application of SSB Xoc , leaves were inoculated with A. alternata strain TBA28A. Diameters of brown spot lesions were measured and photographed 14 dpi. Lesion size (diameter) are shown 6 SD of triplicate measurements. Different letters above columns indicate significant differences at P = 0.01 using the Student's t test. doi:10.1371/journal.pone.0056240.g004 explore whether SSB xoc secretion was T3SS-dependent. The construct for expressing c-Myc-tagged SSB Xoc was transferred into the wild-type and mutants defective in hrcV (encodes an inner membrane component of the T3SS), hrcC (encodes an outer membrane component), hrpE (encodes protein subunits of the Hrp pilus), hpaB and hpaP (encode exit control proteins for T3SE secretion) [8]. These mutants were designated RDhrcV, RDhrcC, RDhrpE, RDhpaB and RDhpaP (Table S1), respectively. RDssb X with the empty vector pUFR034 was used as a negative control. When the wild-type RS105 and mutants RDhrcC, RDhrcV, RDhrpE, were incubated in a hrp-inducing medium XOM3 [50] and examined by immunoblotting, the SSB Xoc -c-Myc protein, like Hpa1-c-Myc protein (as a positive control), was found in the supernatants (SN) of the wild-type and RDhrpE, RDhpaB, and RDhpaP but absent from the SN fraction of RDhrcV, RDhrcC and RDhrpE (Fig. 6D, E). These results indicate that a functional T3SS is needed for secretion of SSB Xoc and Hpa1. The SSB Xoc -c-Myc protein was detected in the SNs and total extracts (TE) of RDhpaB and RDhpaP, indicating that HpaB and HpaP are not required for secretion of SSB Xoc and Hpa1 (Fig. 6D, E). Moreover, the secretion of Hpa1 was not impaired by the mutagenesis in ssb Xoc , and vice versa (Fig. 6D, E). ssb X is positively regulated by HrpX. The down-regulated expression of ssb Xoc in the RDhrpX and RDhrpG mutants indicates that ssb Xoc is positively regulated by HrpX and HrpG (Fig. 1). To investigate this further, we used promoter prediction software (HUhttp://www.fruitfly.org/seq_tools/promoter.html) to analyze the ssb Xoc promoter region in the X. oryzae pv. oryzicola BLS256 genome [61]. This analysis revealed an imperfect PIP-box (TTCGC-N 19 -TTCGT) upstream of the ssb Xoc start codon (Fig. 7A), suggesting that ssb Xoc may be regulated by HrpX [79]. To determine whether ssb Xoc expression depends on the putative PIP-box and HrpX, we constructed a recombinant plasmid pPIPAGUS, which contains the ssb Xoc promoter region fused to a promoter-less gusA in pUFR034, resulting in pPIPAGUS (Fig. 7A, Table S1). A mutated ssb Xoc promoter (first two TT nucleotides replaced with AA, see Fig. 7A) was also fused to gusA in pUFR034, generating pPIPBGUS (Fig. 7A, Table S1). Plasmids pPIPAGUS and pPIPBGUS were transformed into the wild-type RS105 and mutants RDhrpX and RDhrpG, incubated in the hrpinducing medium XOM3, and GUS activities were measured. GUS activity of pPIPAGUS was significantly lower in the hrpG and hrpX mutants than in the wild-type strain (P = 0.01, t test). GUS activity of the mutated ssb Xoc transcriptional fusion (pPIPBGUS) was similar in the wild-type, RDhrpG, and RDhrpX strains (P = 0.01, t test) (Fig. 7B). We also used real-time PCR to evaluate expression levels of ssb Xoc in strains RS105, RDhrpG, and RDhrpX. The expression of ssb Xoc was higher in the hrp-inducing medium XOM3, but significantly reduced in the nutrient-rich NB, regardless of the genetic background (Fig. 7C). ssb Xoc expression levels were consistently lower in RDhrpG and RDhrpX than the wild-type (Fig. 7C). Collectively, results indicate that ssb X transcription is positively regulated by HrpG and HrpX and suggest that ssb Xoc contains a PIP-box that is likely regulated by HrpX.
Our results suggest that the HR-eliciting SSB X protein is highly conserved in Xanthomonas species (Fig. S1, Fig. 2C), leading us to investigate whether PIP-box promoters drive ssb X expression in other xanthomonads or not. Interestingly, PIP-box promoters were identified upstream of the ssb coding sequences in X. oryzae pv. oryzae PXO99 A , X. campestris pv. vesicatoria 85-10, X. axonopodis pv. citri 306, and X. campestris pv. campestris 8004 (data not shown), implying that ssb expression in these bacteria is also induced in planta. Thus, we performed real-time PCR to investigate the expression levels of ssb X in various Xanthomonas species grown in the hrp-inducing media (see Methods). ssb X expression was significantly lesions lengths were used to assess the virulence of X. oryzae pv. oryzicola RS105 and selected mutants. One half of a rice leaf (IR24, two-months old) was inoculated with wild-type RS105, and the remaining half was inoculated with one of the following deletion mutants: ssb Xoc deletion mutant RDssb X , hpa1 mutant RDhpa1, the double mutant RDssb X Dhpa1, the complemented mutant CRDssb X , and the T3SS mutant RDhrcV. Ten leaves were inoculated with each strain (OD 600 = 0.3; approximately 3610 8 cfu/ml) by leaf-needling, and the assay was conducted in triplicate. Bacterial leaf streak symptoms were photographed 14 dpi, and representative symptoms are shown (A). The average lesion lengths formed by the wild-type and mutants were measured 14 dpi (B), and data represent means 6 SD from three replicates. Different letters in each data column indicate significant differences at P = 0.01 (t test). (C) Bacterial growth assays in planta. Strains (OD 600 = 0.3) were infiltrated into leaves of rice seedlings (IR24, two-weeks old) with blunt-end plastic syringes, and the cfu/cm 2 of tissue was evaluated as described in Methods. Data represent means 6 SD from three replications. (D) and (E) demonstrated the secretion of SSB Xoc (D) and Hpa1 (E) are dependent on a functional T3SS of X. oryzae pv. oryzicola. This experiment utilized X. oryzae pv. oryzicola RS105 and strains containing mutations in the following genes: hrcV (RDhrcV), hrcC (RDhrcC), hrpE (RDhrpE), hpaB (RDhpaB), hpaP (RDhpaP), hpa1 (RDhpa1) and ssb Xoc (RDssb X ) to express ssb Xoc -c-myc or hpa1-c-myc fusion (as a positive control). After incubation (8 h) in hrpinducing medium XOM3, total cell extracts (TEs) and culture supernatants (SNs) were analyzed by SDS-PAGE and immunoblotted with an anti-c-Myc antibody. The immunoblotting assay was conducted twice, and similar results were obtained each time. For the detection of SSB Xoc , the strain RDssb X with the empty vector pUFR034 was used as a negative control (D). doi:10.1371/journal.pone.0056240.g006 higher in hrp-inducing media than in the nutrient-rich media, which suggests that these genes have a functional PIP-box and are regulated by HrpX.

Discussion
In this report, we demonstrate that single-stranded DNAbinding proteins from Xanthomonas elicit HR in tobacco. This activity was not demonstrated with SSB proteins obtained from other prokayrotes, so it may be a unique feature of Xanthomonas. Like Hpa1, SSB Xoc contributes both to bacterial growth and virulence in rice (Fig. 6) and also triggers programmed cell death (Fig. 2, 3). This is the first report that Xanthomonas produce a highlyconserved SSB X protein that functions as a harpin-like protein.
Furthermore, we showed that SSB X binds nonspecifically to singlestranded DNAs (Fig. S3), perhaps a potential role in protecting ssDNA from nucleases [62]. Unlike some PAMPs with a very narrow distribution, such as Ax21 in X. oryzae pv. oryzae [80], SSBx may be widely distributed in Xanthomonas species. It has been proposed that PAMPs are conserved throughout classes or genera of microbes and contribute to general microbial fitness [81].
Harpins are generally highly constrained structures that are difficult for plant pathogenic bacteria to alter because they have evolved to help bacteria avoid recognition in plants. The first identified harpin, HrpN from E. amylovora [7], has been identified in related pathogens including Pantoea stewartii subsp. stewartii [82] and D. dadantii [83]. Another harpin, HrpZ, first identified in P. syringae pathovars [6,84], but was later shown to be present in nonpathogenic pseudomonads including P. putida and P. fluorescens [85]. The harpin HrpW, which contains harpin and pectate-lyase domains, is widely conserved across genera and has been identified in E. amylovora [41], P. syringae [42], D. dadantii [86], R. solanacearum [39], and X. campestris pv. campestris [46]. Interestingly, there is no HrpW homologue in the genomes of X. oryzae pv. oryzae, X. oryzae pv. oryzicola, X. campestris pv. vesicatoria, or X axonopodis pv. citri [46,61,87]. Thus, some harpins may have a more narrow distribution. For example, R. solanacearum contains an SSB protein but this does not elicit HR in tobacco (Fig. 2D), possibly because SSB Rs lacks the conserved glycine-rich region of SSB X (Fig. S1) Plasmids pPIPAGUS and pPIPBGUS were transferred into the wild-type RS105 and mutants RDhrpG and RDhrpX. The recombinant strains were then grown in hrp-inducing medium XOM3 for 16 h. GUS activity was determined by measuring the OD at 415 nm using r-nitrophenyl-â-D-glucuronide as a substrate. Data represent the mean 6 SD of triplicate measurements. The different letters above each horizontal column indicate significant differences at P = 0.01 (t test). (C) Expression of ssb Xoc in hrp-inducing and nutrient-rich media. Real-time quantitative RT-PCR was used to compare relative expression of ssb Xoc in X. oryzae pv. oryzicola strains RS105, RDhrpG, and RDhrpX. RNA was isolated from strains grown in a nutrient-rich medium (NB) and the hrp-inducing medium (XOM3) for 16 h. The relative mRNA levels of ssb Xoc in the hrpG and hrpX mutants were calculated with respect to the wild-type strain. Values given are the means 6 SD of triplicate measurements from a representative experiment, and similar results were obtained in two other experiments. Different letters above horizontal columns represent significant differences at P = 0.01 using the Student's t test. (D) Real-time RT-PCR evaluation of ssb X expression in Xanthomonas species. Strains were grown at 28uC for 16 h in NB or one of the following hrp-inducing media: XOM3 for X. oryzae pv. oryzicola RS105 and X. oryzae pv. oryzae PXO99 A (Xiao et al. 2007), XVM2 for X. axonopodis pv. citri 306 & X. campestris pv. vesicatoria 85-10, and MMX for X. campestris pv. campestris 8004 (see methods). Relative mRNA quantitative of ssb X was calculated with respect to the levels observed for wild-type strains grown in NB. Genes encoding 16S rRNA were used as internal controls. Data represent means 6 SD of triplicate measurements (P = 0.01, t test). doi:10.1371/journal.pone.0056240.g007 and the pathogen causes bacterial wilt in tobacco. It is also interesting to recall that flg15, a truncated version of flagellinderived flg22, does not act as an elicitor in Arabidopsis or N. benthamiana, while it is fully active in tomato [88]. In Xanthomonas, only Hpa1 [8,43] and SSB X (Fig. 2; this study) have been identified as harpins that elicit HRs in tobacco. However, the hpa1-ssb Xoc double mutant RDhpa1Dssb X still elicited HR in the wild-type and SA-deficient (NahG) tobacco plants (data not all shown), implying that Xanthomonas produce other elicitors that trigger a SA-independent HR. This double mutant is a valuable resource for identifying other harpin(s) that exist in Xanthomonas.
Mutations in the T3SS of Gram-negative phytopathogenic bacteria often disrupt the ability of the pathogen to elicit HR in tobacco, presumably because harpin proteins are secreted via the T3SS [6][7][8][41][42][43]. In Hrp group II, the cis-acting PIP-box promoter is followed by a 210 box, and collectively these elements indicate regulation via HrpX [76,79,89]. However, the PIP-box in ssb X genes is not followed by the 210 box, possibly because proteins in addition to HrpX are involved in regulation [44,60]. At the amino acid level, harpin proteins possess T3SS signals that are characterized by at least 20% Pro and Ser residues in the first 50 amino acids at the N-terminus; this protein signature is required for secretion through the T3SS [78,90]. Intriguingly, SSB X protein does not have the T3SS signal (Fig. S1), but is secreted, as Hpa1, through the T3SS independently of HpaB and HpaP (Fig. 6D,  6E). Thus, the compelling topics for future studies include understanding of how ssb X is regulated by HrpX and the mechanism of SSB x secretion.
Hpa1 and SSB X function for HR induction in tobacco (Fig. 2) and are required for full virulence in rice (Fig. 6); these dual functions are difficult to reconcile. Harpins may activate defense by entering into plant membranes and modulating ion channels or may be recognized by unidentified receptor(s) [24]. Recent reports show that the key a-helical domain in harpins, including HpaG of X. campestris pv. glycines that is orthologous to Hpa1, is required for amyloidogenesis [23,27]. However, the Hpa1 a-helix does not show any similarity to SSB X proteins (data not shown). We speculate that the conserved domain of SSB X proteins (Fig. S1) is required for HR induction, and experiments to investigate this hypothesis are underway in our lab. The fact that harpins do not elicit HR in host plants suggests that the recognition of harpins differs in host and non-host plants. It is also important to mention that harpins function in the translocation of T3SEs, and this function is required for a full level of virulence in host plants [91]. Whether SSB X plays a role in T3SE translocation or not, remains unknown; however, the mutation in ssb Xoc did not impaired the secretion of Hpa1 (Fig. 6E) and the ssb Xoc -hpa1 double mutant still caused disease in rice (Fig. 6A, B, C), so at least some effectors were delivered to rice cells inoculated with the double mutant (Fig. 6). These findings are reminiscent of those reported for P. syringae pv. tabaci flagellin mutants, which were abrogated in elicitor activity and displayed reduced virulence due to impaired motility [92,93].
Plant immune responses triggered by harpins are often associated with HR and SAR [6,7,24,28]. In the present study, we show that SSB x induces PCD (Fig. 2E), the oxidative burst (Fig. 3C), the expression of HR and SAR marker genes ( Fig. 2F and 3F), and callose deposition (Fig. 3D), which stimulate plant defense. SSB x -induced HR, like Hpa1-induced, could be blocked by eukaryotic metabolic inhibitors (Fig. 2D). It will be interesting to determine whether SSB proteins from diverse genera can elicit HR, and such studies will help us understand how pathogens recruit molecules that are instrumental for bacterial fitness and redeploy them as agents for plant defense.
Although plant-associated microbes can potentially produce many molecules with conserved signatures, only a few PAMPs have been identified, and most of these trigger a similar set of responses. In the current study, we evaluated the expression of PTI signature genes, e.g. BAK1, BIK1 and MAP3K [19,73,94]. The activation of these genes by Hpa1 and SSB Xoc (Fig. 3) further supports the contention that PTI is a variant of ETI [18,81]. Recently, Ax21 of X. oryzae pv. oryzae, which is perceived by Xa21 in rice [80], was shown to be recognized by FLS2 in Arabidopsis [95]. Thus it remains possible that Hpa1 and SSB Xoc , like Ax21, may recruit FLS2 in a receptor complex together with other receptors and adaptors that modulate PTI. The identity of receptors for Hpa1 and SSB Xoc and whether these PAMPs interact with FLS2 remains unclear.
There is abundant evidence in the literature showing that harpins display pleiotropic effects both on HR & SAR and also impact plant growth [28,29,96]. Our results also showed that Hpa1 and SSB Xoc enhanced growth of Arabidopsis and tobacco (Fig. 5); this was correlated with increased expression of EIN2 and PR4 genes (Fig. 3F) that are essential for Eth-signaling [75]. This is consistent with the contention that Eth-signaling regulates the accumulation of the FLS2 receptor and is required for the oxidative burst leading to PTI [97]; thus, Hpa1-and SSB Xocmediated plant immunity may also require Eth-signaling. Eth-and SA-signaling may be regulated by WRKY transcription factors that are phosphorylated by the MAPK cascade [98,99]. Nevertheless, SSB x may possibly have pleiotropic effects in plants.
Based on the results of this study, we propose a working model for SSB X function that is also applicable to Hpa1 (Fig. 8). The ssb x gene in Xanthomonas is regulated by HrpG and HrpX; the latter protein potentially binds to the PIP-box promoter and activates transcription. Although the mechanistic basis of secretion is not Figure 8. Working model of SSB X function. SSB Xoc , like Hpa1, which are regulated at the transcriptional level by HrpG and HrpX, are potentially secreted into the plant apoplast via the T3SS. The grey circle and black triangle indicate SSB X and Hpa1 proteins that may be secreted through the Hrp pilus (encoded by hrpE), but not translocated through the translocon as other T3SEs (different white shapes), and then are possibly recognized by an unidentified receptor (question mark) which associates with BAK1. This interaction may result in phosphorylation of BIK1 and subsequent phosphotransfer to the MAPK cascade to activate the expression of genes involving in SA-, JA-and Eth-signaling pathways that lead to induced resistance (SAR and/or ISR) accompanied by callose deposition on cell walls and enhanced plant growth. MAPK signaling regulates NADPH oxidase-dependent oxidative burst in the early stages of plant defense. doi:10.1371/journal.pone.0056240.g008 totally understood, SSB x secretion depends on the functional T3SS, but does not depend on the presence of HpaB and HpaP. We hypothesize that, besides protecting ssDNAs from nucleases in bacterial cells, SSB x , possibly like Hpa1, is secreted through the T3SS, but not translocated into plant cells, and perceived in plant apoplast where it is recognized by an unknown receptor, possibly a plasma membrane-localized PAMP receptor-like kinase (RLKs) that recruits other proteins, like BAK1, and activates downstream signal transduction cascades for HR induction (Fig. 8). We speculate that signaling leads to expression of Eth-dependent genes for plant growth and SA-or JA-dependent genes for plant defense. These hypotheses are the subject of ongoing experiments in our laboratory by undertaking the investigation of an unknown SSB Xinteracting protein in plants. Figure S1 Comparison of single-stranded DNA-binding proteins in Xanthomonas species and other prokaryotes by multiple sequence alignment. The sequences within the black dashed-line rectangle represent conserved region in Xanthomonas but variable in other prokaryotes. Protein accession numbers are indicated. The abbreviations are as follows: Xoc, X. oryzae pv. oryzicola; Xoo, X. oryzae pv. oryzae; Xcv, X. campestris pv. vesicatoria; Xcc, X. campestris pv. campestris; Rs, Ralstonia solanacearum; Ya, Yersinia aldovae; Ea, Erwinia amylovora; Pst, Pseudomonas syringae pv. tomato; Ec, Eschericha coli, and Xf, Xylella fastidiosa. (TIFF) Figure S2 Phylogenetic analysis of SSB proteins in various bacterial species. A neighbor-joining bootstrap tree was derived from the amino acid sequences of SSB proteins using the Vector NTI Align program (http://www.invitrogen.com). Protein accession numbers are indicated after the bacterial species or strain designation. Based on phylogenetic analysis, SSB proteins were classified into one of three groups (I, II and III) for HR induction in nohost tobacco. (TIFF) Figure S3 SSB Xoc binds to single-stranded DNAs in electrophoretic mobility shift assays (EMSA). Randomly synthesized DNA1 and DNA2 (Table S2) were labeled with the Biotin 39 End DNA Labeling Kit (Thermo, USA). EMSA was performed using protocols supplied with the LightShift Chemoluminescent EMSA Kit (Thermo, USA). Five mg of purified SSB Xoc protein was mixed with 20 ml of the binding buffer and 20 fmol of biotin-labeled DNA1 (left panel) or DNA2 (right panel); in competition assays (lanes marked with*), labeled DNA was mixed with a 200-fold molar excess of unlabeled DNA1 or DNA2. The mixtures were incubated at room temperature for 20 min. Samples were then loaded on 5% polyacrylamide gels in 0.5X TBE buffer (pH 8.3). Gels were transferred to Hybond N+ membranes (Amersham, Pharmacia), and signals were detected by chemoluminescence according to the manufacturer's instructions. The experiment was repeated twice and similar results were obtained. Lanes that are labeled (2) do not contain SSB x ; lanes labeled (+) contain SSB Xoc and DNA. The middle lane in each panel clearly shows the retardation of DNA mobility due to SSB Xoc binding. (TIFF)