Genome sequences of 100 strains belonging to the P. syringae complex were downloaded from GenBank and the presence of HopAO1 and HopAO2 was analyzed by blastn using Geneious software v7.1.13¹ and the corresponding amino acid sequences of P. savastanoi pv. savastanoi NCPPB 3335 as template. Positive hits were determined using the “Grade” value, a percentage calculated by Geneious combining the query coverage, e-value and identity values for each hit with weights 0.5, 0.25, and 0.25, respectively. Genomes yielding a Grade value ≥92% were considered to encode the corresponding protein.

The evolutionary history of HopAO1 and HopAO2 were inferred by using the maximum likelihood method (Jones et al., 1992). The trees with the highest log likelihood (-1584,3692), which are drawn to scale, with branch lengths measured in the number of substitutions per site, are shown (Figures 1A,B). All positions containing gaps and missing data were eliminated. Evolutionary analyses were conducted in MEGA7 (Kumar et al., 2016). The accession number of the protein sequences used for the construction of the phylogenetic tree and the abbreviated names used for pathovar designations are presented in Supplementary Table S1.

Bacterial strains and plasmids used are described in Supplementary Tables S2, S3, respectively. Primers used in this study for various purposes are described in Supplementary Table S4. All Pseudomonas strains were grown in King’s B (KB) medium (King et al., 1954), Lysogeny Broth (LB) medium (Bertani, 1951), Hrp-inducing medium (HIM) (Huynh et al., 1989) or Super Optimal Broth (SOB) (Hanahan, 1983) at 28°C. Escherichia coli and Pseudomonas strains were grown in LB medium at 37°C or 28°C, respectively. When required, the medium was supplemented with ampicillin (Ap), 100 μg/mL; gentamicin (Gm), 10 μg/mL; kanamycin (Km), 10 or 50 μg/mL; tetracycline (Tc), 10 μg/mL; or spectinomycin (Sp), 10 μg/mL.

Construction of the ΔhopAO1 (AER-0000610) mutant from P. savastanoi pv. savastanoi NCPPB 3335 was performed by marker exchange mutagenesis as previously described (Matas et al., 2014), and the correct exchange of the hopAO1 gene by the Km resistance cassette was determined by Southern blot analysis.

Basic DNA and molecular techniques were performed following standard methods (Sambrook and Russell, 2001). Genomic DNA was extracted using the Jet Flex Extraction Kit (Genomed; Löhne, Germany). Plasmid DNA for cloning purposes was extracted using GenElute^TM Plasmid Miniprep Kit (Sigma–Aldrich, USA).

The translocation assays were performed as described by Matas et al. (2014). This method is based on the construction of fusion proteins between identified T3E and the calmodulin-dependent Cya reporter domain. Electrocompetent P. savastanoi pv. savastanoi NCPPB 3335 and NCPPB 3335- T3 cells were transformated with Cya fusions. Sp^R transformants were tested by PCR using a forward primer designed specifically and a reverse primer annealing to the cya gene (Supplementary Table S4). Cya assays were performed in N. tabacum var. Newdel plants, as previously described (Schechter et al., 2004). P. savastanoi pv. savastanoi NCPPB 3335 and NCPPB 3335-T3 transformants carrying plasmids expressing T3E-Cya fusions were scraped off of the LB plates, washed twice and resuspended to an optical density at 600 nm (OD₆₀₀) of 0.5 (approximately 10⁸ CFU/mL) in 5 mM morpholinoethanesulfonic acid (pH 5.5) and 100 μM isopropyl β-D-1-thiogalactopyranoside (IPTG). Cell suspensions were injected into the fully expanded upper plant leaves using a 1 mL syringe. Leaf disks were collected at 6 h post-inoculation (hpi) with a 10 mm inner-diameter cork borer, frozen in liquid nitrogen, ground to a powder and suspended in 250 μl of 0.1 M HCl. The samples were incubated at -20°C overnight, and cyclic AMP (cAMP) levels were determined using a 1:100 dilution of the samples and the Correlate-EIA cAMP immunoassay kit according to the manufacturer’s directions (Assay Designs, Inc., Ann Arbor, MI, USA).

The Cya activity of the Cya fusion protein expressed in E. coli XL1Blue from pCPP3234 derivatives were assayed as previously reported (Schechter et al., 2004). Bacterial cells were grown in 5 mL of LB medium containing 100 μM IPTG to an OD₆₀₀ of 0.6 to 0.8. The culture was centrifuged, and the pellet was washed and resuspended in sonication buffer (20 mM tris-HCl [pH 8.0] and 10 mM MgCl2). The bacteria were sonicated with a microtip for 2 min, and the cellular debris was pelleted by centrifugation. Cya activity was determined in the presence or absence of bovine calmodulin (Calbiochem, Farmstadt, Germany) by using 5 μl of each lysate (Sory and Cornelis, 1994).

Pure cultures of the wild-type P. savastanoi pv. savastanoi NCPPB 3335 and its DhrpL mutant were grown overnight in KB medium at 28°C. The cells were diluted in fresh KB medium and incubated with shaking at 28°C to an OD₆₀₀ of 0.5. The sample was split into two. One half was pelleted, and frozen (for non-inducing condition) and the other half was pelleted, washed twice with 10 mM MgCl2 and resuspended in the same volume of HIM (Huynh et al., 1989) supplemented with 5 mM mannitol and 0.0006% ferric citrate. After 6 h of incubation, the cells were pelleted and processed for RNA isolation using TriPure Isolation Reagent (Roche Applied Science, Mannheim, Germany) according to the manufacturer’s instructions, except that the TriPure was preheated at 65°C, the lysis step was performed at 65°C, and 1–bromo–3–chloropropane (BCP; Molecular Research Center, Cincinnati, OH, USA) was used instead of chloroform. Total RNA was treated with the RNAeasy kit (QIAGEN) as detailed by the manufacturer. The RNA concentration was determined spectrophotometrically, and its integrity was assessed by agarose gel electrophoresis. DNA-free total RNA was retrotranscribed to cDNA using the cDNA Reverse Transcription kit (Applied Biosystems, Foster City, CA, USA) and random hexamers. The primer efficiency tests, quantitative real-time PCRs (qRT-PCRs) and confirmation of the specificity of the amplification reactions were performed. The relative transcript abundance was calculated using the ΔΔ cycle-threshold (Ct) method (Livak and Schmittgen, 2001). Transcriptional data were normalized to the housekeeping gene gyrA using the Roche LightCycler 480 Software and are presented as the fold change in expression compared to the expression of each gene in the wild-type strain. The relative expression ratio was calculated as the difference in quantitative PCR threshold cycles (ΔCt = Ctgene of interest-CtgyrA). One PCR cycle represents a twofold difference in template abundance; therefore, fold-change values were calculated as 2-ΔΔCt as previously described (Pfaffl, 2001; Rotenberg et al., 2006). qRT-PCRs were performed in triplicate.

His-tagged P. savastanoi pv. savastanoi NCPPB 3335 HopAO1 and HopAO2 were generated using plasmids pENTR-hopAO1 and pDEST42 as entry and destination vectors, respectively. Site-directed mutagenesis was carried out on pENTR-hopAO1 using the QuickChange II Site-Directed Mutagenesis Kit (Stratagene, Santa Clara, CA, USA) following the supplier’s instructions. E. coli BL21 (DE3) cultures containing plasmids expressing wild-type or both mutant proteins, HopAO1 or HopAO2, were grown at 24°C to an OD₆₀₀ of 0.5 and induced with 0.4 mM IPTG for 4 h Cells were harvested by centrifugation at 5000 g for 15 min, and the pellets were frozen overnight at -20°C. Native and mutant HopAO1 and HopAO2 proteins were purified according to protocol 12 of the QIAexpressionist handbook (Qiagen). The same protocol was used for protein extraction from E. coli BL21 (DE3), and the resulting fraction was used as the negative control.

Phosphatase activity was performed using the sensolyte fluorescein diphosphate (FDP) protein phosphatase assay kit (AnaSpec, Fremont, CA, USA) according to the manufacturer’s instructions. This kit provides a fluorogenic assay for measuring the activity of protein phosphatases that convert the fluorescein diphosphate, tetraammonium salt (FDP) into fluorescein, which has a high extinction coefficient and emission quantum yield, thus providing high assay sensitivity. The phosphatase assay was performed by mixing 50 μl of a protein containing sample with 50 μl of FDP reaction solution. The reaction was incubated at 25°C for 30 min, and 50 μl of stop solution was added to stop the reaction. The fluorescence signal was measured using excitation/emission = 485 nm/535 nm. As a negative control, samples without protein (distilled water) were used. The phosphatase activity for the control sublines was set at 100%, and the respective values for the experiment subline were calculated with respect to that value.

The N. benthamiana and N. tabacum (var. Newdel) plants used in this study were 5–7 and 4–6 weeks old, respectively. The plants were grown with a 16-h light/8-h dark photoperiod with day/night temperatures of 26°C/22°C. N. tabacum var. Newdel plants were used for the HR assays. The leaves were infiltrated with bacterial suspensions (10⁷ and 10⁸ CFU/mL) of P. syringae pv. tomato DC3000D28E or its derivatives harboring plasmids expressing each of the different P. savastanoi pv. savastanoi T3E (Supplementary Table S3). The generated symptoms, scored 48 h after inoculation, were captured with a high-resolution digital camera (Nikon DXM 1200, Nikon Corporation, Tokyo, Japan).

For competition assays, the N. benthamiana leaves were inoculated with mixed suspensions containing equal CFU (approximately 10⁴ CFU/mL) of P. syringae pv. tomato DC3000D28E and each of its transformants carrying pCPP5040 derivatives expressing P. savastanoi pv. savastanoi T3E (Supplementary Table S3). Input and output pools, assayed 1 and 6 h after inoculation, respectively, were plated onto LB-Sp and LB-Sp-Gm to select for DC3000D28E and the transformants with pCPP5040 derivatives, respectively. A competitive index (CI) was calculated by dividing the output ratio (CFU transformant: CFU DC3000D28E) by the input ratio (CFU transformant: CFU DC3000D28E). Competition indices of the transformant strains expressing P. savastanoi pv. savastanoi T3E versus DC3000D28E were normalized with respect to the CI obtained for DC3000D28E (pCPP5040). The CIs presented represent the mean of three replicates demonstrating typical results from three independent experiments. The results were statistically analyzed using Student’s t-test.

Olive plants derived from seed germinated in vitro (originally collected from an ‘Arbequina’ plant) were micropropagated, rooted, and maintained as previously described (Rodríguez-Moreno et al., 2008, 2009). To analyze the pathogenicity of the P. savastanoi pv. savastanoi NCPPB 3335 hopAO1 mutant in 1-year-old olive explants (woody plants), micropropagated olive plants were transferred into soil and maintained in a glasshouse at 27°C with a relative humidity of 58% under natural daylight. The plant stems were wounded and infected with approximately 10⁶ CFU of P. savastanoi pv. savastanoi strains. Morphological changes scored at 28 (in young explants) or at 90 days (in 1-year explants) post-inoculation, were captured with a high-resolution digital camera (Nikon DXM 1200).

Reactive oxygen species production was observed after 3-3′-diaminobenzidine (DAB) staining (Thordal-Christensen et al., 1997) 4 h after inoculation with P. fluorescens Pf55 [pLN18] derivatives. Bacterial suspensions (10⁸ CFU/mL) were inoculated into N. tabacum var. Xanthi leaves using a blunt syringe. Small pieces of tobacco leaves cut from around the injection area were placed into a syringe and stained by vacuum infiltration of a freshly prepared 1 mg/mL solution of DAB (Sigma–Aldrich D-8001, Sigma–Aldrich, Inc., St. Louis, MO, USA) in 8 mM HCl, pH 3.8. Chlorophyll was removed by submerging the leaves into a solution of ethanol/lactic acid/glycerol [3:1:1 (vol/vol/vol)] at 60°C and stored overnight at room temperature on water-soaked filter paper. At least ten biological replicates from each specimen were mounted on slides in a 50% glycerol (vol/vol) solution and observed with a Nikon Eclipse E800 microscope (Nikon Corporation, Tokyo, Japan) under bright field. DAB staining produces an intensely brown precipitate at the sites of ROS production, which were next to the infection zone.

Callose deposition samples were developed 12 h after inoculation and stained as previously described (Guo et al., 2009). Chlorophyll was removed in 95% (vol/vol) ethanol from small pieces of tobacco leaves, which were cut from around the injection area, and staining was performed in a 0.02% (wt/vol) solution of aniline blue (Sigma–Aldrich #415049, Sigma–Aldrich, Inc., St. Louis, MO, USA) in 150 mM potassium phosphate, pH 9, for 1 h in the dark. At least ten biological replicates from each specimen were mounted in 50% (vol/vol) glycerol on glass slides. Observations were conducted under UV-light excitation using the filter UV-2^a (EX 330- 380, DM 400; BA 420) on a Nikon Eclipse E 800 microscope (Nikon Corporation, Tokyo, Japan).

Reactive oxygen species production and callose deposition were quantified as previously described (Rodríguez-Herva et al., 2012) with slight modifications. Up to four snapshots of each specimen from equivalent areas surrounding the wound (inoculation zone) were captured with a Nikon DXM1200 camera using the Nikon ACT-1 2.70 software. The same settings and a final magnification of 40X were applied to all the samples. After calibrating all the images using the scale bar included in each picture, DAB staining and aniline blue fluorescence were quantified using the program Visilog 6.3 (Noesis, Les Ulis, France). For this purpose, the characteristic brown color of the DAB precipitate and the specific blue fluorescence of callose deposition were separated by color deconvolution using the i_classification command. Then, the stained areas were quantified, and the results were expressed in mm². Five-six images per assay were analyzed, and statistical analyses were performed using one-way ANOVA followed by post hoc comparisons using Tukey test.

Prediction of T3E genes in the draft genome sequence of P. savastanoi pv. savastanoi NCPPB 3335 previously allowed the identification of a HopAO1 homolog located in plasmid pPsv48B (Rodriguez-Palenzuela et al., 2010; Bardaji et al., 2011). A chromosomally encoded candidate HopAO2 T3E is also found in NCPPB 3335 (accession number GG774693.1, locus tag PSA3335_5047), which shows 93% amino acid identity (100% cover) with its corresponding P. syringae pv. actinidiae MAFF302091 homolog and only 33% identity with HopAO1 from NCPPB 3335 (Supplementary Figure S1). Thus, NCPBB 3335 encodes two dissimilar T3E candidates of the HopAO family (HopAO1 and HopAO2). Proteins belonging to this family share the signature sequence [LIVMF]HCxAGxxR[STC][STAG], characteristic of the active site of the protein tyrosine phosphatase (PTP) family (Fauman and Saper, 1996). Further analyses of the amino acid sequences of NCPPB 3335 HopAO1 and HopAO2 showed that, in fact, both of them harbor this conserved motif.

With the aim of extending the analysis of the prevalence of HopAO1 and HopAO2 homologs to all established phylogroups of the P. syringae complex (Berge et al., 2014; Baltrus et al., 2017), a total of 100 genomes corresponding to all Pseudomonas species and pathovars of the P. syringae complex which genomes have been sequenced were downloaded from GenBank and the presence of the HopAO1 and HopAO2 was analyzed using Geneious software v7.1.13². This study was facilitated by the recent publication of the draft genome sequences of numerous strains of the P. syringae species complex (Moretti et al., 2014; Bartoli et al., 2015; Nowell et al., 2016; Thakur et al., 2016). The HopAO1 phylogeny was reconstructed from the 462-468 amino acids alignment of 24 orthologous using maximum likelihood (Figure 1A). The HopAO1 sequences of P. savastanoi pv. savastanoi PseNe 107, P. savastanoi pv. nerii ICMP 16943 and P. savastanoi pv. retacarpa CECT 4861, which are possibly truncated, were not included in this analysis. The results revealed the partitioning of these strains into two main clusters, one of them predominantly including PG3 strains isolated from woody hosts. This cluster also includes PG1 strain P. syringae pv. philadelphi ICMP 8903 and PG2 strain P. syringae pv. papulans CFBP 1754, isolated from leaf spots of mock orange and apple, respectively. While ICMP 8903 is included in the “aesculi clade,” comprising all 10 P. syringae pv. aesculi strains analyzed, CFBP 1754 clusters together with P. savastanoi pv. fraxini ICMP 7711 (isolated from ash) and all three P. savastanoi pv. savastanoi strains analyzed, i.e., NCPPB 3335, DAPP-PG722 and ICMP 4352. The second cluster is composed by six strains, five of which were isolated from herbaceous hosts, i.e., P. syringae pv. tomato DC3000 (PG1), PG4 strains P. syringae pv. oryzae 1-6 and P. syringae pv. porri ICMP 8961 and LMG 28495 and, P. syringae pv. alisalensis ICMP 15200 (PG5). The remaining strain, P. syringae pv. berberidis ICMP 4116, belongs to PG1 and was isolated from leaf spots of barberry (Figure 1A).

The HopAO2 phylogeny was reconstructed from the 338-339 amino acids alignment of 30 orthologous using maximum likelihood (Figure 1B). In a similar way than HopAO1, one of these clusters is predominantly composed by PG3 strains isolated from woody organs of woody hosts, with the exception of all six P. syringae pv. actinidiae (PG1) strains analyzed, isolated from kiwifruit, and P. syringae pv. glycinea strains ICMP2189 and B076, also PG3 strains isolated from soybean (Figure 1B). The second cluster, which is divided in two main branches, is composed exclusively by 10 strains isolated from either herbaceous hosts or leaf spots of woody hosts. Although all five PG1 strains cluster together in the same sub-branch, i.e., P. syringae pv. berberidis ICMP 4116, P. syringae pv. maculicola ES4326, P. syringae pv. philadelphi ICMP 8903 and P. syringae pv. tomato Max13 and K40, the distribution of the remaining strains is not related to their PG.

The hopAO1 and hopAO2 P. savastanoi pv. savastanoi NCPPB 3335 T3E genes were used as probes in dot-blot hybridizations to ascertain their distribution among a collection of 31 P. savastanoi pv. savastanoi strains isolated in different countries and among a selection of P. syringae strains isolated from either woody or herbaceous hosts (Supplementary Table S2). The probes, which were specific for each of the hopAO alleles, were obtained by PCR using primer pairs hopAO1-F/hopAO1-R (hopAO1) and hopAO2-F/hopAO2-R (hopAO2) (Supplementary Table S4). The results showed that 30 out of the 31 P. savastanoi pv. savastanoi strains, as well as P. savastanoi pv. nerii 2, P. syringae pv. morsprunorum CFBP 2116, P. syringae pv. eriobotryae CFBP 2343 and P. syringae pv. glycinea PG4180, hybridized with both probes (data not shown). Conversely, strains belonging to P. syringae pathovars phaseolicola, syringae, lachrymans, dendropanacis, sesami, and alisalensis did not hybridize with any of these probes (data not shown) and were discarded for further analysis. Thus, with the exception of P. syringae pv. glycinea PG4180 (PG3, isolated from an herbaceous host) and P. syringae pv. morsprunorum CFBP 2116 (PG1, isolated from a woody host), all other strains encoding both HopAO1 and HopAO2 sequences were isolated from woody hosts and belong to PG3 (Figure 1).

Plasmid preparations and total DNA isolated from a selection of six P. savastanoi pv. savastanoi strains hybridizing with both hopAO1 and hopAO2 probes in the dot-blot analysis, as well as from all other strains hybridizing with at least one of the two probes, were analyzed by southern-blot hybridization to determine the localization of these genes on their replicons. While hopAO1 was found to be plasmid-encoded in five out of the six P. savastanoi pv. savastanoi strains used, hopAO2 was detected in the chromosome in the six strains (Figures 1C,D).

To determine whether the selected candidate effectors were T3SS substrates that could be translocated into plant cells, pCPP3234 derivatives expressing fusions of Bordetella pertussis Cya to the C terminus of full-length HopAO1 and HopAO2 were constructed. This system, which is based on cAMP production exclusively in the presence of eukaryotic calmodulin, has been widely used for analyzing the translocation of P. syringae T3E (Sory and Cornelis, 1994; Casper-Lindley et al., 2002; Schechter et al., 2004). N. tabacum leaves were infiltrated with either P. savastanoi pv. savastanoi NCPPB 3335 or the strain NCPPB 3335-T3, a T3SS mutant derived from wild-type NCPPB 3335 (Pérez-Martínez et al., 2010), expressing each of the two constructed Cya fusions. As illustrated in Figure 2A, significant differences in cAMP production between the wild-type strain and the T3SS mutant strain were observed for both T3E candidates tested—HopAO1, HopAO2—indicative of their translocation through P. savastanoi pv. savastanoi T3SS.

An identifiable Hrp box (Fouts et al., 2002) was found in the 100 nucleotides upstream of the start codon of both hopAO1 and hopAO2 candidates from P. savastanoi pv. savastanoi NCPPB 3335 (Supplementary Table S5). To unveil the HrpL-dependent expression of hopAO1 and hopAO2 in P. savastanoi pv. savastanoi NCPPB 3335, the expression of hopAO1 and hopAO2 was analyzed using qRT-PCR in both the wild-type and the NCPPB 3335 ΔhrpL mutant. Under inducing conditions (cells grown 6 h in HIM), the expression of the hopAO1 and hopAO2 genes decreased (0.4- and 0.02-fold, respectively) in the ΔhrpL mutant compared to the wild-type, indicating an expression dependency on HrpL (Figure 2B).

The HopAO T3E family is characterized on the basis of a conserved PTP domain (Bretz et al., 2003; Espinosa et al., 2003), with activity dependent on a cysteine located at position 378 of HopAO1 from P. syringae pv. tomato DC3000 (Bretz et al., 2003). To determine whether P. savastanoi pv. savastanoi NCPPB 3335 HopAO1 and HopAO2 are also active protein phosphatases, affinity-purified HopAO1-His₆ and HopAO2-His₆ were used in a phosphatase activity in vitro assay performed under non-denaturing conditions. We also tested the activity of site-directed HopAO1 and HopAO2 mutants involving the substitution of the PTP-related Cys by a Ser, HopAO1_C376S and HopAO2_C243S, respectively. FDP hydrolysis was clearly detected for wild-type HopAO1-His₆ and HopAO2-His₆; however, the percentage of hydrolysed FDP significantly decreased in both mutant proteins (Figure 3). The results demonstrate that the phosphatase activity shown in vitro by P. savastanoi pv. savastanoi NCPPB 3335 HopAO1 and HopAO2 is dependent on its Cys₃₇₆ or Cys₂₄₃ amino acid residue, respectively.

The suppression of early plant defense responses such as ROS production and callose deposition by HopAO1 and HopAO2 was tested using the heterologous expression system P. fluorescens 55 (Pf55) [pLN18] (Jamir et al., 2004; López-Solanilla et al., 2004; Oh et al., 2010; Matas et al., 2014). This system has been previously used to demonstrate the suppression activity of PTI by other P. savastanoi NCPPB 3335 T3SS effectors (Matas et al., 2014). P. savastanoi pv. savastanoi NCPPB 3335 HopAO1 and HopAO2 were expressed in Pf55 [pLN18], which induces a PTI response in inoculated plants while enabling the delivery of effector proteins into plant cells. After challenging N. tabacum leaves with the derivative strains, we analyzed ROS production and callose deposition. Figure 4 shows that the ROS levels determined by DAB staining were significantly reduced by the expression of HopAO1 or HopAO2 compared to the control strain Pf55 [pLN18] harboring an empty vector (Tukey’s test; P ≤ 0.01) (Figures 4A,C). Moreover, both T3Es significantly reduced the levels of callose deposition compared to the control strain (Tukey’s test; P ≤ 0.01) (Figures 4B,D). Therefore, HopAO1 and HopAO2 are able to interfere with the early innate immune responses of the plant.

Effector-triggered immunity responses constitute the second layer of immune defenses which overlap with PTI responses (Lindeberg et al., 2012). To determine if P. savastanoi pv. savastanoi NCPPB 3335 HopAO1 and HopAO2 functions interfere with this layer of defense, we expressed these two proteins in the P. syringae pv. tomato strain DC3000D28E (Cunnac et al., 2011). This polymutant strain harbors deletions in all 28 well-expressed effector genes and is considered functionally effectorless but otherwise wild-type in planta. The HR elicitation in N. benthamiana and N. tabacum can be explained by the fact that DC3000D28E has the wild-type complement of T3SS helper proteins (except HrpW1), unknown effectors or weakly expressed T3Es, which can elicit plant defenses and induce an HR response (Kvitko et al., 2007; Cunnac et al., 2011). It has been recently demonstrated that the induction of cell death by this strain in N. benthamiana is solely dependent on HopAD1 (Wei et al., 2015). We analyzed the ability of the DC3000D28E derivatives expressing HopAO1 or HopAO2 to elicit cell death in N. tabacum compared to the DC3000D28E strain. After 48 h of treatment, the polymutant strain stimulated an ETI-like response that was partially and completely inhibited by the expression of the P. savastanoi pv. savastanoi NCPPB 3335 proteins HopAO1 and HopAO2, respectively (Figure 5). These results suggest that these two effectors participate in the inhibition of the plant defense response associated with the onset of PCD.

DC3000D28E has been shown to grow better in planta when coinoculated with a strain that is able to suppress plant immunity, such as DC3000ΔhopQ1-1 (Cunnac et al., 2011). The ability of HopAO1 and HopAO2 to restore the growth of DC3000D28E in planta was tested in N. benthamiana leaves. For this purpose, competition assays between the polymutant strain (DC3000D28E) and each derivative expressing either HopAO1 or HopAO2 were conducted. N. benthamiana leaves were infiltrated with a mixed inoculum (1:1) of DC3000D28E and each of the derivatives; after 6 days, the bacteria were recovered, and viable cells were determined. Interestingly, the expression of P. savastanoi pv. savastanoi HopAO1 or HopAO2 in DC3000D28E significantly increased the competitiveness of the strain, which was reflected in a CI value (HopAO1/DC3000D28E or HopAO2/DC3000D28E) as a mean ± standard error of three replicates of 2.14 ± 0.078 and 1.82 ± 0.062, respectively. These values are significantly different from unity (statistical analyses were performed using Student’s t-test with a threshold of P = 0.05).

To analyze the impact of the immune suppression described above for HopAO1 and HopAO2 in the virulence of P. savastanoi pv. savastanoi NCPBB3335, we decided to construct mutant strains affected in each of these two genes. While construction of the ΔhopAO1 mutant strain by marker exchange was straightforward, the isolation of a ΔhopAO2 mutant using this technique was unsuccessful. Thus, we decided to approach the inactivation of hopAO2 by plasmid insertion using an internal fragment of this gene. However, all attempts to insert the plasmid into the NCPPB 3335 hopAO2 gene also failed, indicating that this gene is located in a genomic region of low recombination.

It has been previously described that the inoculation of olive plants with a P. savastanoi pv. savastanoi NCPPB 3335 mutant cured of plasmids pPsv48A and pPsv48B (strain Psv48ΔAB), which encode T3E HopAF1 and HopAO1, respectively, induced attenuated hyperplastic knots, and necrosis (Bardaji et al., 2011). Consequently, complemented Psv48ΔAB and ΔhopAO1 strains expressing HopAO1 were also constructed and inoculated in both in vitro micropropagated and lignified olive plants. In agreement with Bardaji et al. (2011), Psv48ΔAB induced less severe knots than the wild-type strain in both plant systems (Figures 6A–C). Interestingly, although the sizes of the knots induced by the ΔhopAO1 mutant at 28 dpi on the stem of in vitro-grown olive plants were not significantly different from those induced by the wild-type strain, they showed necrotic lesions similar to those from plants inoculated with the Psv48ΔAB mutant (Figure 6A). In addition, the competitiveness of the ΔhopAO1 mutant in this plant system was compromised compared with the wild-type strain (Figure 6D). Additionally, the weights of the knots induced by the ΔhopAO1 mutant in lignified olive plants, which also showed an increased necrosis, were significantly lower than those developed by the wild-type strain (Figure 6C). The ectopic expression of the hopAO1 gene in both the ΔhopAO1 mutant and the plasmid-cured strain Psv48ΔAB fully restored the wild-type appearance of the knots in both plant systems (Figures 6A,B). Complementation of the ΔhopAO1 mutant also resulted in the restoration of the competitiveness of the strain in olive plants (Figure 6D). Altogether, these results reveal a significant role of HopAO1 during the infection process in olive plants.