The Xanthomonas campestris pv. vesicatoria Type-3 Effector XopB Inhibits Plant Defence Responses by Interfering with ROS Production

The bacterial pathogen Xanthomonas campestris pv. vesicatoria 85–10 (Xcv) translocates about 30 type-3 effector proteins (T3Es) into pepper plants (Capsicum annuum) to suppress plant immune responses. Among them is XopB which interferes with PTI, ETI and sugar-mediated defence responses, but the underlying molecular mechanisms and direct targets are unknown so far. Here, we examined the XopB-mediated suppression of plant defence responses in more detail. Infection of susceptible pepper plants with Xcv lacking xopB resulted in delayed symptom development compared to Xcv wild type infection concomitant with an increased formation of salicylic acid (SA) and expression of pathogenesis-related (PR) genes. Expression of xopB in Arabidopsis thaliana promoted the growth of the virulent Pseudomonas syringae pv. tomato (Pst) DC3000 strain. This was paralleled by a decreased SA-pool and a lower induction of SA-dependent PR gene expression. The expression pattern of early flg22-responsive marker genes indicated that MAPK signalling was not altered in the presence of XopB. However, XopB inhibited the flg22-triggered burst of reactive oxygen species (ROS). Consequently, the transcript accumulation of AtOXI1, a ROS-dependent marker gene, was reduced in xopB-expressing Arabidopsis plants as well as callose deposition. The lower ROS production correlated with a low level of basal and flg22-triggered expression of apoplastic peroxidases and the NADPH oxidase RBOHD. Conversely, deletion of xopB in Xcv caused a higher production of ROS in leaves of susceptible pepper plants. Together our results demonstrate that XopB modulates ROS responses and might thereby compromise plant defence.


Introduction
Plants respond to bacterial pathogens with a vast array of defence responses. Recognition of the invading pathogen is a prerequisite for the activation of the defence system and the successful inhibition of bacterial propagation. Plants have employed a two-tier defence system to combat microbial invaders [1][2][3]. The first layer of defence is triggered by the recognition of pathogen associated molecular patterns (PAMPs) by surface-localised pattern recognition receptors leading to PAMP-triggered immunity (PTI). The second layer referred to as effectortriggered immunity (ETI) is activated by intracellular receptors that detect the presence and/ or activity of pathogen-derived effector molecules that have been secreted into the host cell. Generally, both perception systems elicit the activation of partially overlapping signalling cascades and defence responses [4]. However, responses are usually stronger and prolonged during ETI and often culminate in a kind of programmed cell death called hypersensitive response (HR) [5]. While PTI is sufficient to prevent multiplication of a wide range of non-adapted invaders, ETI is thought to be effective against adapted pathogens [2,3,5].
Most of our current knowledge of signalling events during PTI derives from the model plant Arabidopsis thaliana and relates to the perception of the PAMP flagellin [3,6,7]. The leucine rich repeat receptor kinase FLAGELLIN SENSING 2 (FLS2) recognises flg22, a conserved epitope present in the N-terminus of bacterial flagellin [8]. Binding of flg22 induces the instantaneous dimerization of the FLS2 receptor with the receptor-like kinase BRI1-associated kinase1 (BAK1) which is required to initiate signalling [9][10][11]. The subsequent phosphorylation of both proteins triggers downstream events such as the rapid influx of Ca 2+ ions, the production of reactive oxygen species (ROS) and the activation of mitogen-activated protein kinase (MAPK) cascades [9,11].
The FLS2-BAK1 complex interacts with another, cytoplasmic receptor-like kinase termed Botrytis induced kinase 1 (BIK1) that becomes phosphorylated by BAK1 upon PAMP stimulation and is subsequently released from the complex to activate downstream processes. Recent work by Kadota et al. [12] Li et al. [13] uncovered that BIK1 directly interacts and phosphorylates the NADPH oxidase Respiratory burst oxidase homolog D (RBOHD). Beside this, RBOHD becomes activated by Ca 2+ which directly binds to the EF-hand domains present in their N-terminal regions as well as by phosphorylation by Ca 2+ -dependent protein kinases (CDPKs) [14,15]. In response to flg22 treatment, RBOHD is the major enzyme for production of reactive oxygen species (ROS) [15][16][17]. Besides, the apoplastic class III peroxidases AtPRX33 and AtPRX34 are important for PAMP-mediated ROS production [18][19][20]. The rapid and transient production of ROS like superoxide anions and the more stable hydrogen peroxide (H 2 O 2 ) is one of earliest responses starting a few minutes after PAMP treatment [21,22]. ROS can have a direct, toxic effect on the bacterial pathogens or act to confine microbial growth by cross linking plant cell wall proteins or by stimulating the production of phytoalexins. In addition, ROS function as signalling molecules by inducing defence gene expression and are involved in the redox control of proteins [21,22]. Pharmacological experiments revealed that the ROS burst is dependent on the Ca 2+ influx [7,23]. The Ca 2+ burst is a PTI hallmark [24,25] and is an important stimulus for many downstream responses. Thus, the activation of MAPKs and the subsequent reprogramming of gene expression that are observed upon PAMP stimulation are at least partially dependent on the Ca 2+ burst [7,23,26,27]. Four CDPKs, namely CPK4, CPK5, CPK6 and CPK11, are also involved in transcriptional reprogramming upon flg22 stimulation [28]. MAPK-and CDPK-dependent signalling pathways can either act synergistically or independently to control expression of flg22-responsive genes [28]. Late responses upon PAMP treatment are amongst others the production salicylic acid (SA), the accumulation of antimicrobial secondary metabolites, the expression of pathogenesis-related (PR) proteins as well as the cell wall fortification by callose depositions [6].
Many Gram-negative pathogenic bacteria are capable to subvert PTI and/or ETI responses, by directly injecting type-3 effector proteins (T3Es) into the host cells using a type-3 secretion system (T3SS) [29][30][31]. The understanding of the in planta activities of T3Es, the elucidation of their modes of action and the identification of targeted host processes are major goals of pathophysiological research and is key to discover novel host-related defence components. The compatible interactions between Pseudomonas syringae pv. tomato (Pst) DC3000 and A. thaliana as well as between Xanthomonas campestris pv. vesicatoria (Xcv) (re-annotated as X. euvesicatoria) and tomato (Solanum lycopersicum) or pepper (Capsicum annuum) plants represent current model systems to unravel T3E functions [32,33]. Both bacteria inject at least 28 functional T3Es into host cells [34,35]. Although the functional analysis of T3Es is often hindered by their overlapping properties and functional redundancy, for some of them the mode of action has been uncovered (summarized in [29,30]). These studies revealed that some T3Es mimic eukaryotic proteins in structure and function, exhibit enzymatic activities and act for instance as proteases, E3-ubiquitin ligases, phosphotransferases or acetyltransferases [29,30]. Others T3Es function as transcriptional regulators or disrupt cellular structures to suppress plant defence. However, for many of these bacterial effectors their enzymatic activity and host targets remain to be unravelled.
We aim to understand the in planta function of the Xcv effector protein XopB. XopB has been identified as a T3SS-dependently secreted protein [36,37]. It does not belong to the conserved core group of T3Es and is absent in most Xanthomonas strains [35]. XopB shows high sequence similarity to known avirulence proteins such as HopD1 from Pst and AvrPphD of P. syringae pv. phaseolicola and to so far not investigated predicted T3E proteins of X. fuscans or X. citri. The protein consists of 613 amino acids, and does not possess any conserved domains or known linear motifs. In an early study, deletion of xopB in Xcv did not significantly alter symptom development and bacterial growth in susceptible pepper plants [36]. In contrast, Schulze et al. [37] found that infection with an Xcv ΔxopB strain led to less severe disease symptoms in susceptible plants than infection with the Xcv wild type. Ectopic expression of xopB in transgenic tobacco and tomato plants caused severe phenotypic alteration with malformed leaves and elicited cell death particularly in young and meristematic tissues [38]. Expression of xopB in yeast inhibited cell proliferation while transient expression in N. benthamiana resulted in the appearance of cell death symptoms [37,39]. XopB plays a role in PTI and ETI, as it suppresses HR-like cell death induced by several avirulence factors, but suppression of PTI and ETI might be caused by different mechanisms [37]. In PTI, XopB suppressed the flg22-mediated activation of the AtNHL10 promoter, but did not influence the flg22-triggered phosphorylation of MAPKs. Together with its localization in Golgi vesicles and in the cytoplasm the authors suggested that XopB may inhibit PTI by interfering with vesicle transport processes [37].
In our previous work, we showed that XopB suppresses the induction of cell wall-bound invertase (cw-Inv) which is an important enzyme to fuel the sugar-enhanced defence responses [38]. In plants, an increased cw-Inv activity and/ or transcript accumulation has been observed in response to infection with different groups of pathogens including fungi, oomycetes, virus or bacteria [40,41]. Soluble sugars generated by cw-Inv activity provide energy to feed plant defence and act as signalling molecules to regulate gene expression and photosynthesis [41,42]. Moreover, cw-Inv activity is stimulated by PAMPs such as chitosan or a Fusarium oxysporum lycopersici elicitor preparation [43].
In this study, we analysed the role of XopB in PTI-related plant defence responses in more detail. To this end we generated transgenic A. thaliana plants expressing xopB under control of an ethanol-inducible promoter. Expression of xopB in transgenic A. thaliana plants led to a reduced accumulation of SA and a lower increase in PR marker gene expression as well as to an increased growth of the virulent Pst DC3000 strain indicating that XopB functions as a virulence factor. We show that XopB interferes with PTI responses by supressing the ROS burst after flg22-treatment in both N. benthamiana and A. thaliana plants. XopB contributes also in host plants to the suppression of defence responses most likely by interfering with the generation of ROS.

Results
XopB contributes to disease symptom development and alters SA content in susceptible pepper plants Infection of susceptible pepper plants with an Xcv ΔxopB deletion strain did not affect the timing of disease appearance and bacterial virulence in an early study [36]. However, a more recent study reported that deletion of xopB led to significantly reduced disease symptoms in infected pepper plants, while the bacterial growth in planta was not altered [37]. We also investigated the contribution of XopB to disease symptom development. Leaves inoculated with Xcv wild type developed severe necrotic lesions 5 days post infection (dpi), while those infected with the Xcv ΔxopB deletion strain displayed less severe signs of necrosis at this time point (Fig 1A). This delayed symptom development could be complemented by ectopic expression of xopB under control of its own promoter ( Fig 1A). The complemented strain Xcv ΔxopB + xopB caused even stronger disease symptoms than the Xcv wild type strain which might be due to higher expression of xopB from the plasmid. No signs of disease were observed in Mock-infected control plants.
Next, we wanted to know whether XopB influences the accumulation of salicylic acid (SA) which is a plant hormone important for activation of plant defence [44]. Pepper plants were infected with Xcv wild type, Xcv ΔxopB and Xcv ΔxopB + xopB and the contents of free and glycosylated SA (SAG) were measured in leaf samples taken before, and 1 and 3 dpi. In Xcv wild type-infected leaves, an accumulation of free SA was visible at 1 dpi while SAG levels were clearly increased only 3 dpi ( Fig 1B). Infection with the Xcv xopB deletion strain resulted in an about 2-fold higher accumulation in the amounts of free and conjugated SA as compared to Xcv wild type-infected leaves. In contrast, infection with the complementation strain Xcv ΔxopB + xopB induced an increase in SA-levels similar to Xcv wild type. These results suggest that the presence of XopB causes a suppression of plant defence responses.
This conclusion was further confirmed by measuring transcript abundance of the defence genes CaPR1b1 (pathogenesisrelated protein 1b1) and CaPRQ (chitinase) by quantitative realtime RT-PCR (qPCR). Expression of both genes is regulated by SA while CaPRQ is also stimulated by sugars [45,46] and is suppressed by XopB as shown by northern blotting in previous work [38].
In Xcv ΔxopB-infected leaves, the mRNA levels for CaPR1b1 and CaPRQ were 9-fold and 4-fold higher, respectively, compared to Xcv wild type-infected leaves (Fig 1C), consistent with the elevated SA pool. Infection with the xopB complementation strain induced expression of both PR genes to a similar extent as Xcv wild type ( Fig 1C).
Together, these findings demonstrate that XopB suppresses SA accumulation and SAdependent defence gene expression in pepper plants during Xcv infection.

Inducible expression of xopB causes cell death in transgenic Arabidopsis plants
In order to investigate the function of XopB in planta in more depth we generated transgenic A. thaliana expressing xopB under the control of an ethanol-inducible promoter [47]. To this end, the open reading frame was cloned and inserted into the Bin19-derived vector p35S::alcR as described [38]   Leaves of five week old susceptible pepper plants were inoculated with Xcv wild type (Xcv WT), Xcv ΔxopB, or a xopB deletion strain complemented with xopB (Xcv ΔxopB + xopB) under control of its own promotor at a concentration of 10 9 cfu ml -1 . All strains harbour the plasmid pRBB1-MCS5. As control, leaves were infiltrated with 10 mM MgCl 2 . A) Formation of disease symptoms in infected pepper leaves. Only the lower halves of the leaves were infiltrated. Pictures were taken 5 days post infection (dpi). Similar results were observed at least three times. B) Levels of free (free SA) and conjugated SA (SAG) in leaves, before, 1 and 3 days after infection with the Xcv strains indicated. Values represent the means +/-SE Approximately 6-week old, soil-grown plants were watered with 10 ml 1% EtOH and expression of xopB was verified by RT-PCR and western blotting using a XopB-specific antibody (Fig 2A and 2B). Both the xopB-specific transcript and the protein were already detectable 4h after ethanol induction. The xopB-specific mRNA was not detectable after 48h, while the corresponding protein was still present at this time point (Fig 2A and 2B).
After induction of xopB expression by ethanol the transgenic plants stopped growing and leaves developed first chlorotic spots after about three days. In particular, older leaves residing directly above the soil were affected at this time point (Fig 2C). In the following hours also upper leaves developed chlorosis. Subsequently, chlorotic leaves became necrotic (day 4) and died back (day 7). Appearance of cell death symptoms started from the petioles and spread over the midrib into the leaf lamina ( Fig 2C). In contrast, no phenotypic differences between wild type and transgenic plants were observed before treatment with ethanol.
These results show that the T3E protein XopB causes severe phenotypic changes leading to cell death when ectopically expressed in A. thaliana suggesting that XopB may interfere with plant metabolism [38] and / or signalling.

Expression of xopB in Arabidopsis supports in planta growth of Pst DC3000
In order to investigate whether XopB has an impact on bacterial virulence, leaves of wild type and xopB-expressing plants were infected with Pst DC3000 18 h after induction of xopB expression. Compared to Pst DC3000-infected wild type leaves, which displayed only mild chlorosis, leaves of infected xopB-expressing plants were already severely damaged 3 dpi and died earlier than Mockinfected control leaves (Fig 3). Next, the in planta growth of Pst DC3000 was determined in order to analyse if the visible increase in disease symptom development caused by xopB expression was accompanied by an altered bacterial titre. Two and three days after infection, the xopB-expressing lines contained a more than one order of magnitude higher number of bacteria ( Fig 4A) than wild type plants, as expected from the observed accelerated disease symptom development (Fig 3).
In addition to Pst DC3000, wild type and transgenic plants were infected with the TTSSdeficient mutant Pst ΔTLR1 to investigate whether basal defence is significantly dampened by XopB. The in planta growth of the Pst ΔTLR1 strain, however, was not altered in the xopBexpressing lines as compared to wild type plants ( Fig 4B). This suggests that the heterologous expression of XopB is not sufficient to promote the in planta growth of the TTSS-deficient Pst strain, but rather acts synergistically with T3Es translocated by Pst DC3000 to suppress plant defence responses and to support bacterial propagation in planta.

XopB modulates SA accumulation and SA-dependent gene expression in Arabidopsis
Since XopB suppresses SA-responses in Xcv-infected pepper plants we tested whether this was also the case in the Pst DC3000-infected Arabidopsis plants. To this end, wild type and transgenic plants were infected with Pst DC3000 18 h after ethanol-watering to induce xopB of four different samples. Statistically significant differences to Xcv wild type-infected leaves were determined using two-tailed t-test and are indicated by asterisks (***p<0.001; *p<0.05). The experiment was repeated with similar results. C) Transcript levels of CaPR1b1 and CaPRQ were quantified by qPCR from samples taken 3dpi with the different Xcv strains. Values were normalized to CaEF1alpha and displayed relative to the expression level of Xcv wild type-infected leaves which were set to one. Values are means +/-SD of three independent samples each from a pool of two plants. Significant differences to Xcv wild type-infected leaves were calculated using t-test and are indicated by asterisks (***p<0.001; *p<0.05). Similar results were obtained in three independent experiments. doi:10.1371/journal.pone.0159107.g001 expression. Leaf samples were taken before infection and 1 dpi and subjected to HPLC analysis. In response to Pst DC3000, Arabidopsis wild type leaves accumulated substantial amounts of free and conjugated SA ( Fig 5A). The increase in free and glycosylated SA contents was about half of that in both xopB-expression lines ( Fig 5A) indicating that the presence of XopB compromised their accumulation. Before infection, meaning 18 h after induction of xopB expression, XopB did not significantly influence the amounts of SA and SAG. However, there was a significant increase in the content of SAG in Mock-infected xopB-expressing plants as compared to the Mock-infected wild type plants ( Fig 5A) which may indicate increased stress level due to expression of xopB and wounding. In addition to the SA levels the transcript amounts of the SA-dependent PR genes AtPR1 and AtPR3 were determined by qPCR. Consistent with the strong SA/SAG accumulation in wild type plants, there was a significantly increased expression of both genes after Pst DC3000 infection ( Fig 5B). Relative to Mock-infected wild type plants, mRNA levels of AtPR1 and AtPR3 were 8-and 2.3-fold higher in Pst DC 3000-infected plants, respectively. Transcript abundance of AtPR1 and AtPR3 was clearly reduced in xopB-expressing lines before infection ( Fig 5B) suggesting that the heterologous expression of XopB suppresses plant defence responses. Similar to the SAG pool, an increase in the transcript amounts of AtPR1 and AtPR3 was detected in the Mock-infiltrated transgenic plants, while infection with Pst DC3000 caused only a slight further increase in mRNA levels of AtPR1 and AtPR3 and their absolute levels were significantly lower than in Pst DC3000-infected wild type plants ( Fig 5B).

XopB differentially modulates expression flg22-responsive marker genes
Since it has been shown that XopB suppresses PTI when transiently expressed in Arabidopsis protoplasts [37], we analysed whether PTI is also attenuated in xopB-expressing A. thaliana plants. Expression of flg22-responsive markers genes was quantified, including AtFRK1 and WRKY22 as MAPK-specific targets and AtPHI1 as a CDPK-dependently regulated gene [26,28]. In addition, AtNHL10 was chosen as a gene controlled synergistically by both pathways [26,28]. The fls2 mutant served as a control, because it does not respond to flg22 [8].
Expression of AtNHL10 was 3.8-fold induced by flg22 treatment in wild type plants, while its expression was not flg22-inducible in the fls2 mutant (Fig 6, S1 Table). Similarly to the fls2 mutant, AtNHL10 mRNA level was not or only slightly increased in response to the flg22 treatment in both xopB-expressing lines. Remarkably, expression of AtNHL10 was significantly diminished in water-treated transgenic and fls2 plants as compared to wild type plants (Fig 6). These results support the findings that XopB suppresses both the basal and the flg22-mediated activation of the NHL10 promoter [37].
The mRNA levels of AtWRKY22 were also lower in water-treated transgenic plants as compared to the wild type, but in contrast to AtNLH10, AtWRKY22 was strongly induced in response to flg22 treatment in wild type as well as in xopB-expressing plants (Fig 6). Thus, flg22 treatment elicited an about 15-fold increase in AtWRKY22 transcript amounts in wild type plants and a 14-and 25-fold induction in xopB-expressing lines 10 and 12, respectively (S1 Table). As expected, there was no flg22-mediated increase in AtWRKY22 mRNA amounts in the fls2 plants (Fig 6, S1 Table). AtFRK1 was similarly expressed in non-stimulated xopBexpressing and wild type plants. Flg22 treatment resulted in a significant increase in AtFRK1 transcript abundance in wild type and both transgenic plants, but not in the fls2 mutant. As for AtWRKY22, the flg22-mediated increase in AtFRK1 expression was even higher in the xopB line 12 as compared to the wild type (Fig 6, S1 Table). These results suggest that the flg22-elicited induction of the MAPK-dependent genes AtWRKY22 and AtFRK1 was not affected or even stimulated by XopB. In contrast, the flg22-induced transcriptional activation of AtNHL10 was hampered by XopB. Hence, it is tempting to speculate that in the xopB-expressing plants the CDPK-dependent rather than the MAPK-dependent signalling pathway is influenced. In fact, expression of the CDPK-dependent gene AtPHI1 was not increased by flg22 treatment in Expression of xopB in A. thaliana supports the in planta growth of the Pst DC3000 strain, but not of the Pst ΔTLR1 strain. A. thaliana wild type and transgenic EtOH::xopB plants (lines 10 and 12) were watered with 10 ml 1% EtOH 18 h before infiltration with (A) the Pst DC3000 strain or (B) the TTSS-deficient mutant strain (Pst ΔTLR1) to monitor the bacterial growth in planta. Pst DC3000 and Pst ΔTLR1 were infiltrated using a needle-less syringe with a bacterial density of 5*10 5 cfu ml -1 or 10 6 cfu ml -1 , respectively. Bacterial titres were determined at indicated time points from 3 biological replicates (3 pools out of 6 plants). Values represent means +/-SE. Statistically significant differences between the wild type and the EtOH:: xopB lines 10 and 12, respectively, were determined using a two-tailed t-test assuming normal distribution and are indicated by asterisks. Similar results were obtained in three independent experiments. the xopB-expressing lines, but was 2.6-fold increased in wild type plants (Fig 6). Moreover, AtPHI1 expression was not altered by XopB in water-treated plants, but was decreased in the fls2 mutant after flg22 treatment. thaliana wild type and xopB-expressing plants (lines 10 and 12) were watered with 10 ml 1% EtOH 18 h before infiltration with the virulent Pst strain DC3000 at a bacterial density of 2.5*10 6 cfu ml -1 or with 10mM MgCl 2 . A) Contents of free (free SA) and conjugated SA (SAG) were quantified with HPLC before (0 dpi) and 1 day after infection (1 dpi). Values represent the means +/-SE of two different experiments each with four independent replicates. Statistically significant differences in the SA/ SAG contents between the wild type and the EtOH::xopB lines 10 and 12, respectively, were determined using a two-tailed t-test assuming normal distribution and are indicated by asterisks (*p<0.05). B) Total RNA was extracted before (0 dpi) and 1 dpi and reverse transcribed into cDNA. Abundances of AtPR1-and AtPR3-specific transcripts were determined by qPCR, normalized to AtTUB4 and displayed relative to the expression level in wild type plants. Values are means +/-SD of three independent replicates each measured in triplicates. Statistically significant differences compared to wild type plants were determined using two-tailed t-test and are indicated by asterisks (*p<0.05). The experiment was repeated twice with similar results.

XopB interferes with flg22-mediated ROS burst and inhibits callose deposition
The results so far suggested that XopB might inhibit the CDPK-dependent branch of flg22responsive gene expression, but not the MAPK-dependent one. This supports findings by Schulze et al. [37] that XopB did not alter the phosphorylation of MAPKs. Several publications showed that the calcium response is upstream of two separate signalling pathways; one leading to activation of MAPK and the other to ROS production [7,23,48]. Since MAPK signalling appeared not to be addressed by XopB, we investigated whether it affects the flg22-mediated ROS burst.
To this end, leaf discs from xopB-expressing and A. thaliana wild type plants were stimulated with flg22 and the time-dependent accumulation of ROS was measured. As expected, flg22 elicited a strong, but transient increase in ROS in wild type plants (Fig 7). Interestingly, leaf discs from xopB-expressing lines 10 and 12 showed a 70% and 80% lower ROS production compared to wild type plants, respectively (Fig 7) indicating that XopB suppresses the flg22-triggered ROS generation. To verify this finding, we measured the ROS burst after flg22 treatment in N. benthamiana leaves transiently expressing xopB or an empty vector. Like in Arabidopsis, expression of xopB also abolished the flg22-induced ROS burst in this experimental system (S2 Fig). To investigate whether the altered ROS production in xopB-expressing lines is reflected by altered expression of ROS-related genes we determined the transcript abundance of AtRBOHD, AtPRX33, AtPRX34 and of OXIDATIVE SIGNAL-INDUCIBLE1 (AtOXI1) 30 min after flg22 treatment. Control plants were treated with water and the fls2 mutant was again included.
Expression of AtOXI1 which was shown to be transcriptionally activated by H 2 O 2 [49] was similar in wild type and transgenic water-treated plants, but was ca. 2-fold higher in fls2 controls compared to wild type. As expected, flg22 did not increase expression of AtOXI1 in the fls2 mutant, but elicited an about 9-fold increase in mRNA abundance in wild type plants (Fig  8). In comparison, there was only a 3-fold induction of AtOXI1 by flg22 in the xopB-expressing plants (Fig 8), in accordance with the diminished ROS production in these lines. The mRNA abundance of AtRBOHD was already 4-to 6-times lower in the water-treated xopB-expressing lines compared to wild type controls (Fig 8), indicating an interference of XopB with the redox homeostasis in the absence of PAMPs. Upon flg22 stimulation AtRBOHD mRNA level increased 2.7-fold in wild type plants, but no significant induction was detectable in the fls2 mutant. Although the fold-change compared to the control situation was similar in xopBexpressing lines and wild type (2.6-and 4.0-fold in lines 10 and 12, respectively), the absolute AtRBOHD transcript amounts were lower in both xopB-expressing lines upon flg22 treatment compared to those in wild type in the absence of the PAMP. The mRNA abundance of AtPRX33 and AtPRX34 were also 2-to 4-times lower in water-treated transgenic plants compared to wild type plants supporting the hypothesis that XopB may constitutively affect expression of ROS producing enzymes. Expression of both peroxidases was hardly induced by flg22 in the xopB-expressing lines as well as in the fls2 mutant. The amount of AtPRX33 mRNA was clearly increased in flg22-treated wild type plants, while expression of AtPRX34 was not significantly up-regulated in wild type plants (Fig 8). A. thaliana wild type (WT), xopB-expressing lines 10, 12 and fls2 mutant plants were watered with 10 ml 1% EtOH. After 18 h leaves were infiltrated with 1 μM flg22 or deionized water as a control. Samples were taken after 30min. Total RNA was isolated and reverse transcribed into cDNA. Transcript accumulation of AtOXI1, AtRBOHD, AtPRX33, and AtPRX34 was measured by qPCR. Expression of AtTUB4 was used to normalize the expression of each sample. Expression levels are shown relative to the water-treated wild type. Values are means +/-SD of three independent replicates. Statistical differences (p<0.05) were determined (a) between water-treated wild type and the transgenic or fls2 plants and (b) between water-and the flg22-treated plants, respectively, using two-tailed t-test assuming normal distribution. Similar results were obtained in two independent experiments. doi:10.1371/journal.pone.0159107.g008 A well-known late ROS-dependent response to flg22 treatment is the fortification of cell walls by callose deposition [6,16,21]. A strong accumulation of callose was detected in wild type plants by aniline blue staining 18 h upon flg22 stimulation (Fig 9). In contrast, callose deposition was significantly reduced in the xopB-expressing lines 10 and 12 after flg22 treatment (Fig 9).
Together these data suggest that XopB suppresses the basal and PAMP-mediated expression level of peroxidases and AtRBOHD. This may lead to a lower ROS production upon PAMP stimulation and disturb downstream signalling pathways thatfor example control callose deposition.

Deletion of xopB in Xcv causes higher production of H 2 O 2 in susceptible pepper plants
In order to elucidate whether XopB also inhibits the ROS production during the Xcv-pepper interaction, leaves of susceptible pepper plants were infiltrated with Xcv wild type, Xcv ΔxopB, Xcv ΔxopB + xopB and MgCl 2 (Mock) and stained with 3,3-diaminobenzidine-tetrahydrochloride (DAB) at 3 dpi to monitor H 2 O 2 accumulation [50]. Compared to mock-inoculated plants, a higher amount of brownish precipitates was visible in pepper leaves after Xcv wild type infection indicating an increased pathogen-mediated H 2 O 2 accumulation (Fig 10A). The accumulation of H 2 O 2 was even stronger, when plants were infiltrated with the Xcv ΔxopB strain (Fig 10A). Infection with the complemented strain Xcv ΔxopB + xopB caused a colouration similar to the Xcv wild type strain (Fig 10A). Quantification of the images revealed an approximately 50% higher DAB intensity in Xcv ΔxopB-infiltrated pepper leaves compared to leaves infected with the Xcv wild type or the xopB complemented strain (Fig 10B). These results corroborate the finding of XopB-mediated inhibition of ROS production.

Discussion
Transgenic plants have been proven to be a versatile tool to study T3E function and to identify host target processes [51]. For example, expression of the Pst DC3000 T3E AvrPto in transgenic A. thaliana plants provided the first evidence that PTI is suppressed by this effector [52]. Subsequently, numerous further studies contributed to elucidate its molecular mechanisms by using transgenic plants and showed that AvrPto interacts with the kinase domain of different PRRs including FLS2 and thereby inhibits downstream signalling events (e.g. [53,54]).
In order to analyse the role of the Xcv effector protein XopB in more detail, we generated transgenic Arabidopsis plants expressing xopB in an inducible manner. Expression of xopB in Arabidopsis caused severe phenotypic alteration leading eventually to cell death. This is in accordance with earlier observations showing that XopB caused detrimental effects when expressed in plants [37][38][39]. Although the underlying molecular mechanisms are not clear, these results led us to conclude that this effector modulates cellular processes that finally provoke cell death.
Our recent work suggested that the T3E protein XopB from Xcv inhibits the pathogen-stimulated increase of cw-Inv expression and activity in pepper plants [38]. Cw-Inv catalyses the cleavage of the transport sugar sucrose into glucose and fructose and is a key enzyme for supplying sink-organs with carbohydrates [55]. Its expression and activity was shown to be increased in source leaves in response to infection with various pathogens (see [41]). The induction of cw-Inv is thought be crucial to feed an increased metabolic demand of defence responses [56,57]. A fast increase in hexose amounts appeared to be important to mount an effective and successful plant defence. This observation was supported by the finding that heterologous expression of a yeast invertase in either the apoplast or the vacuole of transgenic tobacco plants resulted in high amounts of hexoses and an increased resistance against potato virus Y [56]. Moreover these plants exhibited strongly increased levels of SA and SAG and a higher expression of SA-regulated defence genes [56]. This indicated that an increased invertase activity and the local accumulation of soluble sugars was a prerequisite for the accumulation of SA and an effective defence. Here, we showed that the T3E XopB is involved in modulation of SA/ SAG levels and of SA-dependent genes such as PR1 and PR3. This was seen in the interaction between A. thaliana and Pst as well as between C. annuum and Xcv. Thus, infection of susceptible pepper plants with an Xcv xopB deletion strain caused a higher accumulation of SA/SAG and SA-dependent PR transcripts compared to infection with an Xcv wild type strain. This is in line with the increased cw-Inv expression and activity found after infection of pepper leaves with Xcv ΔxopB [38]. In xopB-overexpressing Arabidopsis plants, however, both the increase in the amount of SA/SAG as well as in PR gene expression was clearly lower in response to Pst DC3000 infection than in wild type plants indicating that XopB suppresses defence responses during infection with Pst DC3000. The latter findings were accompanied by an increased bacterial growth and an accelerated development of disease symptoms suggesting that XopB acts in concert with the T3Es secreted by Pst DC3000 to suppress plant defence. Pst DC 3000 harbours the T3E HopD1 which is a homolog of XopB. Infection of Arabidopsis plants with an Pst hopD1 mutant led to a slight, but significant reduction in bacterial  Leaves of pepper plants were infiltrated either with Xcv wild type (Xcv WT), a xopB deletion strain (Xcv ΔxopB) or a xopB deletion strain complemented by xopB (Xcv ΔxopB + xopB) at a concentration of 10 9 cfu ml -1 . All strains harbour the pBBR1-MCS5 vector. As a control, leaves were infiltrated with 10 mM MgCl 2 . Two leaves of two pepper plants were infiltrated using a needle-less syringe. Three days post infection six 4 cm 2 leaf discs were taken and stained for ROS using DAB solution. After chlorophyll clearance six randomly chosen microscopic images per leaf disc were documented. A) A representative image for each scenario is shown. B) Pixel intensities of microscopic images were determined by means of the ImageJ software. Values shown are means +/-SE of 18 images and were presented relative to the mean value of pepper plants infiltrated with Xcv wild type strain which was set to one. Statistically significant differences to Xcv wild type-infected leaves are indicated by asterisks (***p<0.001).
doi:10.1371/journal.pone.0159107.g010 growth, but growth of a DC3000 strain was not altered in transgenic plants expressing HopD1 ectopically [58]. However, even though HopD1 and XopB show sequence homology their mode of action might by different since further work by Block et al. [58] revealed that HopD1 reduces ETI responses by targeting the ER-localised transcription factor NTL9, but did not affect PTI, while XopB affects both PTI and ETI [37].
Generally, individual bacterial strains express multiple effectors with apparently distinct and redundant activities which are collectively essential to support their life style, but their effector repertoire shows a high degree of variability [59]. XopB and homologous T3Es from other bacterial strains do not belong to the core group of conserved effector proteins [35], but are rather specific T3Es exploited by some species. Hence, it might conceivable that XopB amend the effector repertoire of Pst DC3000 and thereby enhancing its bacterial virulence.
Although the SA-mediated defence responses were activated upon infection of pepper plants with Xcv ΔxopB, the bacterial growth of this strain was not altered [36,37]. Functional redundancy with other Xcv T3E could be the reason why bacterial growth is not significantly affected. Accordingly, Schulze et al. [37] suggested that XopB and XopS fulfil redundant functions based on studies with a double knock out strain. Also a P. syringae pv. phaseolicola avrPphD mutant (AvrPphD is another homolog of XopB) showed no effect on bacterial growth indicating that it is functionally redundant with other T3E [60]. Infection with a single xopB deletion Xcv strain caused milder disease symptoms ( [37]; this study) suggesting that Xcv virulence is weakened but obviously not sufficiently enough to affect bacterial growth.
Here, we further analysed how XopB affects PAMP-triggered defence responses. This was stimulated by work of Tsuda et al. [4] who showed that SA levels increase in response to flg22-treatment and that SA is a major component of the PTI signalling cascade. Moreover, Schulze et al. [37] reported that XopB supresses PTI responses like the flg22-mediated activation of the NHL10 promoter, a well-established marker for PTI-signalling studies [26,28]. Expression of xopB also decreased basal activity of pNHL10 in Arabidopsis protoplasts [37]. Similarly, both basal and flg22-stimulated expression of NHL10 was reduced in transgenic A. thaliana plants expressing xopB. The flg22-mediated increase in the expression of AtPHI1 was also abolished in the xopB-expressing plants. In contrast, the flg22-stimulated induction of AtWRKY22 and AtFRK1 was not significantly altered or even higher in xopB-expressing lines as compared to wild type plants. MAPK-dependent as well as CDPK-dependent signalling pathways contribute to the transcriptional re-programming during plant immune responses and both pathways control different groups of target genes [28]. Genes such as AtFRK1 and AtWRKY22 are MAPK-specific, while AtPHI1 is controlled by CDPKs. Other targets like AtNHL10 are regulated by both CDPK-and MAPK-dependent signalling pathways. The expression pattern of flg22-responsive marker in the xopB-expressing Arabidopsis plants genes led us to conclude that XopB alters CDPK-but not MAPK-dependent signalling pathways. This assumption is in accordance with the finding that phosphorylation of MAPKs was not influenced upon flg22 treatment by expression of xopB in Arabidopsis protoplasts [37].
Several lines of evidence indicate that PAMP-activation of MAPK may occur independent of the PAMP-triggered ROS production at least during the early stage of PTI (see [61,62]). For instance, the bik1 mutant is compromised in AtRBOHD-dependent ROS production and callose deposition, but not in flg22-induced MAPK activation [63,64]. The flg22-mediated ROS burst is largely controlled by AtRBOHD [16,17] and was shown to be dependent on the transient influx of calcium ions from the apoplast [7,23]. AtRBOHD activity itself is regulated by calcium binding and by phosphorylation [14,17]. A number of CDPKs have been identified that phosphorylate AtRBOHD [65,66]. Recent work has shown, however, that AtRBOHD becomes phosphorylated by BIK1 upon PAMP stimulation in a calcium-independent manner [12,13]. Although we were not able to directly measure the impact of XopB on Ca 2+ influx, due to transgene silencing after crossing the xopB-lines with the aequorin reporter lines, our results clearly show that expression of XopB strongly suppressed the flg22-stimulated ROS burst and subsequent signalling processes in Arabidopsis. Hence, the transcript abundance of AtOXI1, a kinase that is transcriptionally activated by H 2 O 2 [49] was induced to a lower extent in the xopB-expressing Arabidopsis plants than in wild type. Interestingly, AtRBOHD transcript abundance increased after flg22-treatment in both wild type and xopB-expressing Arabidopsis lines, but the absolute levels were lower in the transgenic plants compared to wild type plants.
In addition, the expression of the two apoplastic peroxidases, AtPRX33 and AtPRX34, was down-regulated in xopB-expressing control plants compared to wild type and their expression did not increase in response to flg22. Only expression of PRX33, but not of PRX34, was significantly increased 30 min after flg22 treatment in wild type (Col-0) plants. Results published by Daudi et al. [20] showed an at least 10-fold up-regulation for both genes 2 h after flg22 treatment. Even though these peroxidases were shown to account for about half of the flg22-triggered ROS production [19,20], their expression might be strongly stimulated at later time points after PAMP treatment. Hence, Kadota et al. [15] proposed that the immediate ROS burst after PAMP treatment is entirely dependent on AtRBOHD which then triggers a secondary, late PRX33/34-dependent ROS production.
Transcript accumulation of AtRBOHD as well as of AtPRX33 and AtPRX34 are positively regulated by H 2 O 2 . Moreover, AtRBOHD expression was shown to be dependent on the peroxidase-mediated H 2 O 2 production [19]. Torres et al. [67] postulated, that NADPH oxidases like AtRBOHD have to be activated by an NADPH oxidase-independent source of ROS, which might be peroxidases. Therefore, it seems conceivable that XopB decreases expression of the apoplastic peroxidases AtPRX33 and AtPRX34 which accounts for a reduced basal H 2 O 2 production and subsequently for a decreased AtRBOHD expression. The lower absolute level of AtRBOHD may then cause the reduced ROS production upon flg22 stimulation observed in the xopB-expressing lines.
The prx33 and prx34 as well as the rbohd mutant are impaired in PAMP stimulated callose deposition [16,20] supporting the idea that callose deposition is ROS-dependent. Accordingly, a lower callose deposition was observed in the xopB-expressing Arabidopsis plants.
Several PAMP-associated genes were found to be down-regulated in un-challenged prx33 and prx34 T-DNA insertion lines which led the authors to suggest that a low level of ROS production is required to "preprime" basal defence [20]. Likewise, the observed down-regulation of AtWRKY22 and AtNHL10 in un-treated xopB-expressing plants might be also related to the low expression of both peroxidases. Moreover, Mammarella et al. [68] reported that transgenic A. thaliana plants in which expression of AtPRX33 and AtPRX34 (and probably other peroxidases) was knocked-down by antisense expression of a cell wall peroxidase cDNA clone from French bean [18] were severely impaired in the induction of SA-responsive genes, such as AtPR1.
Remarkably, these plants were not more susceptible to infection with a Pst hrcC-mutant than wild type plants, although these plants are compromised for a variety of PTI-responses. Similarly, expression of xopB caused a down-regulation of the SA-responsive genes AtPR1 and AtPR3 and a dampened PTI-response, but did not alter the in planta growth of a T3SS-deficient Pst ΔTLR1 strain. Since effects caused by XopB expression in A. thaliana strongly suggest that it supresses PTI responses this finding was unexpected and requires further investigations. Expression of XopB in A. thaliana inhibited the flg22-induced ROS burst and down-stream signalling events, but the MAPK-dependent signalling was not inhibited or even weakly activated. Together with other, parallel operating PAMP-stimulated signalling pathways this may provide sufficient activation of defence responses to combat propagation of the Pst ΔTLR1 strain in xopB-expressing lines.
Infection of susceptible pepper plants with a Xcv xopB-deletion strain led to a higher ROS accumulation compared to infection with the Xcv wild type strain indicating that XopB interferes with ROS accumulation also in the host plants. In pepper plants, CaPO2 was identified as an apoplastic peroxidase pivotal for the oxidative burst [69]. Silencing of CaPO2 in pepper compromised PR gene expression and caused enhanced susceptibility to virulent Xanthomonas strains [69]. Moreover, infections of Arabidopsis lines overexpressing CaPO2 with virulent Pst DC3000 resulted in enhanced ROS production [69] confirming the important role of this enzyme. The effect was accompanied by a reduced growth of Pst DC3000 and an elevated mRNA level of PR genes, including SA-dependent ones [69]. Expression of xopB in A. thaliana plants caused similar molecular and biochemical changes as observed in prx33/34 mutants, while infection of pepper plants with an Xcv xopB deletion strain led to phenotypes comparable to plants expressing CaPO2 suggesting that XopB may interfere with expression and/ or activity of peroxidases.
How can the observed effects by XopB be linked to its localisation in vesicle-like structures described by Schulze et al. [37]? Vesicle trafficking emerged as an important means of plant defence contributing to the correct localisation of PAMP receptors [70]. In more detail, endocytosis was recently shown to be involved in PAMP-stimulated ROS production [70,71] and regulation of AtRBOHD activity [72]. XopB altered the ROS production during PAMP-treatment and Xcv infection, but appeared not to affect MAPK signalling. Other T3E proteins have been described which also impair the ROS burst, but show normal MAPK signalling such as AvrAC and AvrPphB [63,64]. AvrAC is an uridyl transfrase that reduces BIK activity, while AvrPphB is a cysteine protease cleaving BIK1 and other PBS1-like kinases. Host proteins targeted by XopB are so far not known, but need to be identified in future work to unravel the molecular and cellular control mechanisms exerted by XopB. Our current data suggest that XopB inhibits peroxidase mediated ROS-production. How this is related to its proposed inhibitory effect on vesicle transport (or endocytotic processes) needs to be learned from further work.

Generation of transgenic Arabidopsis plants
Stable transformation of Arabidopsis thaliana Col-0 plants of the EtOH::xopB construct [38] was performed by Agrobacterium-mediated gene transfer as described previously [73].
For selection of transgenic lines, T1 seeds were surface-sterilized and sown onto Murashige-Skoog medium supplemented with 50 μg ml -1 kanamycin. Expression of the transgene was confirmed by northern blotting. Numbers of T-DNA insertions were determined by southern blotting.

Plant material and growth conditions
A. thaliana plants (T3 or T4 generation of transgenic lines) were sown on soil and kept for 2-3 days at 4°C in darkness to synchronize germination. Subsequently, seedlings were grown for two weeks under short day conditions (8 h of light, 22°C; 16 h darkness, 20°C) in phytotrons (CLF, Wertingen, Germany) before transfer into individual pots. Further cultivation was done in a plant climate chamber (Plant Master PGR 3045, CLF Wertingen, Germany) under short day conditions with approximately 100 μmol quanta m -2 s -1 light and 60% relative humidity. Expression of xopB was induced by watering individual pots with 10 ml 1% (v/v) ethanol and was confirmed by western blotting using a XopB-specific antibody [38] in every experiment.

Bacterial strains and infection experiments
Bacterial strains used in this study are described in S2 Table. Pseudomonas syringae pv. tomato (Pst) strain DC3000 Pst and Xanthomonas campestris pv. vesicatoria (Xcv) 85-10 strains were cultivated at 28°C on nutrient-yeast-glycerol (NYG) medium supplemented with appropriate antibiotics. Antibiotics were added to the media at following final concentrations: kanamycin, 50 μg ml -1 ; rifampicin, 100 μg ml -1 and gentamycin 15 μg ml -1 . Infection of five week old A. thaliana leaves with Pst DC3000 was carried out as follows: Pst strains were plated from DMSO cryo-stocks on NYG plates supplemented with appropriate antibiotics and incubated at 28°C for two to three days. Bacterial cells were further cultivated in liquid media, washed and adjusted with sterile 10 mM MgCl 2 to desired cell densities. For in planta growth studies, virulent and TTSS-deficient Pst strains were syringe-infiltrated with a density of 5 Ã 10 5 cfu ml -1 or 10 6 cfu ml -1 , respectively. Bacterial titres of Pst in planta were determined as described in Sonnewald et al. [38]. For SA quantifications and qPCR studies, Pst DC3000 was syringe-infiltrated with a bacterial density of 2.5 Ã 10 6 cfu ml -1 . Xcv infection of pepper leaves was performed as described previously [38].

Flg22-treatment
Five week old A. thaliana wild type or transgenic plants were infiltrated with 1 μM flg22 peptide or with deionized water. For analysis of gene expression leaf material was harvested 30 or 60 min after infiltration, frozen in liquid nitrogen and stored at -80°C until RNA extraction.

Isolation of genomic DNA and Southern blotting
Genomic DNA was isolated from 5 g leaf material of 5 week old Arabidopsis plants. Restriction enzyme digestion with EcoRI and southern blot analysis was performed as described previously [74].

RNA isolation and quantitative RT-PCR
Isolation of total RNA was performed as described in [75]. Copy DNA was usually synthesized from 2 μg total DNaseI-treated RNA using Oligo dT30 primers. Quantitative real time PCR (qPCR) was essentially performed as described by [76]. For each primer pair the efficiency was determined which was between 96% -113%. A melting analysis was performed at the end of each run to ensure that unique products were formed. Relative expression values of target genes were calculated using the Pfaffl method [77]. Fold changes in gene expression were calculated according to Livak and Schmittgen [78]. The values were normalized to elongation factor 1 alpha (EF1 alpha) from pepper and tubulin 4 (TUB4) from Arabidopsis, respectively. For semi-quantitative RT-PCR, cDNA was diluted 1:10 and subjected to standard PCR protocol. Primers used for qPCR and semi-quantitative PCR analyses are listed in S3 Table. Western blotting Leaf discs (0.5 cm 2 ) were homogenized and processed as described previously [38]. Proteins were detected by using an anti-XopB antibody by means of chemiluminescence.

Luminol-based detection of ROS
Leaf discs (0.25 cm 2 ) were incubated overnight at 22°C in the dark in a 96-well micro titer plate filled with 0.2% EtOH (to induce expression in transgenic A. thaliana plants) or deionized water (for N. benthamiana). The solution was then substituted by 200 mM sodium phosphate buffer, pH 8.0 and a luminol assay solution consisting of 0.2 mM luminol (Sigma, A8511) and 0.01 mg ml -1 type VI horseradish peroxidase (Sigma, P8375). The micro titre plates were placed in the plate reader (Mithras LB 940, Berthold, Bad Wildbad, Germany) and background signals were detected with 1 sec integration time per well for 5 min. After manual addition of 1 μM flg22 the luminescence was recorded with 1 sec integration time for 45 min. For each genotype the mean background value was calculated and subtracted from flg22-triggered luminescence values.

Detection of callose deposition
A. thaliana Col-0 wild type and transgenic plants were watered with 10 ml 1% EtOH 18 h before treatment with 1 μM flg22 or deionized water. Two leaves of three independent plants were infiltrated for each treatment. After 18 h leaves were de-stained in 96% EtOH and further processed as described by [81]. Epifluorescence microscopy was conducted with a Leica DMR microscope using UV excitation and a DAPI emission filter. Six randomly chosen areas per leaf were documented and callose depositions per mm 2 were counted.

Transient expression of xopB in N. benthamiana
For transient expression of xopB, the ORF was amplified with the primers 5'-xopB (5'aggatcctctagaatgaaggcagagctcacacgatccca-3') and 3'-xopB (5'-gtcgaccggctcaggcgcgggttggtg-3') and inserted into expression vector pBinAR under the control of the 35S promoter [82] via BamHI and SalI restriction sites (restrictions sites in primer sequences are underlined). For infiltration of N. benthaniama leaves, Agrobacterium tumefaciens C58C1 strains were infiltrated into the abaxial leaf site of 4-to 6-week-old plants. A. tumefaciens strains harbouring xopB, the silencing suppressor p19 [83] or the empty vector were cultivated overnight at 28°C in the presence of appropriate antibiotics. The cells were harvested by centrifugation, re-suspended in infiltration buffer (10 mM MgCl 2 , 10 mM MES pH 5.6 and 100 μM acetosyringone), adjusted to oD 600 of 0.4. Cell suspensions containing either the xopB expression construct or the empty vector were mixed with p19 (1:1 ratio). Samples for assaying ROS production and protein detection were taken 48h after Agro-infiltration.

Histochemical detection of ROS
Leaf discs (4 cm 2 ) of infected pepper plants were placed into a solution containing 1 mg ml -1 3.3´-diaminobenzidine-tetrahydrochloride, pH 3.8 (DAB, Applichem, A0596). After incubation in the dark for 18 h the DAB solution was replaced by 96% EtOH and boiled at 80°C until the chlorophyll was completely removed. Leaf discs were mounted in 40% glycerol and observed under a Leica DMR microscope using bright field. For quantification of the pixel intensities, 6 randomly chosen areas per leaf discs were recorded resulting in 18 images per condition. Quantification with ImageJ was essentially conducted as described by Rodríguez-Herva et al. [84]. Images with a 200-fold magnification were used. Pixel intensities of samples taken Xcv wildtype infected pepper leaves served as a calibrator and were arbitrary set to one.
Supporting Information S1 Fig. Characterisation of transgenic xopB-expressing Arabidopsis thaliana lines. A) Schematic representation of the T-DNA harbouring the EtOH::xopB expression cassette. The transcriptional regulator AlcR from Aspergillus nidulans is expressed under control of the cauliflower mosaic virus 35S promoter (35S) and the nopaline synthase terminator (NOS) from Agrobacterium tumefaciens, while expression of the type III effector xopB from Xanthomonas campestris pv. vesicatoria (cloned in 5´-3´orientation) is driven by a modified promoter of the alcohol dehydrogenase AlcA from Aspergillus nidulans and 35S terminator (35S T). Numbers indicate base pairs number of the corresponding elements. The right border and the kanamycin resistance cassette of the T-DNA are not shown. The depicted T-DNA construct has three EcoRI sites within the T-DNA sequence and one additional EcoRI site at base pair position 463 of xopB. B) Selection of transgenic lines by northern blotting. Total RNA was extracted from different transgenic lines upon floating of detached leaves with 0.2% EtOH in the dark overnight. Twenty μg of total RNA was separated in a formaldehyde-containing agarose gel and blotted onto nitrocellulose membrane. Nine xopB expressing transgenic T1 lines from A. thaliana could be identified using a xopB specific [32] P-labelled probe. After stripping the same membrane was incubated with a probe specific for the small subunit of RubisCO  Table. Fold induction (flg22-vs. H 2 O-treatment) of target gene expression in wild type, xopBexpressing Arabidopsis lines (10, 12) and fls2 mutant. Values were calculated according to [78] and are the mean ± SD of three independent replicates. Significant differences (pvalue 0.05) compared to wild type are indicated by Ã and corresponding values are highlighted in bold. (DOCX) S2