A Na+/Ca2+ exchanger of the olive pathogen Pseudomonas savastanoi pv. savastanoi is critical for its virulence

Summary In a number of compatible plant‐bacterium interactions, a rise in apoplastic Ca2+ levels is observed, suggesting that Ca2+ represents an important environmental clue, as reported for bacteria infecting mammalians. We demonstrate that Ca2+ entry in Pseudomonas savastanoi pv. savastanoi (Psav) strain DAPP‐PG 722 is mediated by a Na+/Ca2+ exchanger critical for virulence. Using the fluorescent Ca2+ probe Fura 2‐AM, we demonstrate that Ca2+ enters Psav cells foremost when they experience low levels of energy, a situation mimicking the apoplastic fluid. In fact, Ca2+ entry was suppressed in the presence of high concentrations of glucose, fructose, sucrose or adenosine triphosphate (ATP). Since Ca2+ entry was inhibited by nifedipine and LiCl, we conclude that the channel for Ca2+ entry is a Na+/Ca2+ exchanger. In silico analysis of the Psav DAPP‐PG 722 genome revealed the presence of a single gene coding for a Na+/Ca2+ exchanger (cneA), which is a widely conserved and ancestral gene within the P. syringae complex based on gene phylogeny. Mutation of cneA compromised not only Ca2+ entry, but also compromised the Hypersensitive response (HR) in tobacco leaves and blocked the ability to induce knots in olive stems. The expression of both pathogenicity (hrpL, hrpA and iaaM) and virulence (ptz) genes was reduced in this Psav‐cneA mutant. Complementation of the Psav‐cneA mutation restored both Ca2+ entry and pathogenicity in olive plants, but failed to restore the HR in tobacco leaves. In conclusion, Ca2+ entry acts as a ‘host signal’ that allows and promotes Psav pathogenicity on olive plants.


I N T RO D U C T I O N
Cytosolic calcium (Ca 2+ ) has essential functions in eukaryotic signalling as secondary messenger. The cytosolic Ca 2+ levels are influenced by the difference in its intracellular-to-extracellular concentration (Berridge et al., 2000;Bhosale et al., 2015;Islam, 2012;). In particular in mammals, Ca 2+ signalling is well understood with a central role in nearly all the known cellular processes ranging from egg-cell fertilization to programmed cell death (Brini et al., 2013;Rajagopal and Ponnusamy, 2017), impacting gene expression levels, heart and muscle contraction, neurotransmission and synaptic plasticity, secretion of hormones and their action, blood coagulation and other motility processes, to diverse metabolic pathways involved in the generation of cell fuels (Sharma et al., 2017). Furthermore, Ca 2+ acts both as a messenger and cofactor to coordinate many intracellular signalling pathways (Rajagopal and Ponnusamy, 2017). Noteworthy, it can already activate different cellular responses only by differences in the amplitude, frequency and duration of the intracellular Ca 2+ concentration (Rajagopal and Ponnusamy, 2017). Located predominantly in the extracellular environment, Ca 2+ entry relies in animals on membrane depolarization resulting from action potentials, where it then can perform its regulatory functions. In these eukaryotes, most ion channels as well as transporters, pumps, binding proteins and L-type voltage-dependent calcium channels have the capacity to transport Ca 2+ across the depolarized membrane (Cai and Lytton, 2004;Carafoli, 1987;Norris et al., 1996;Tsien and Tsien, 1990). In plants, Ca 2+ is present in high concentrations in the apoplast (i.e. intercellular spaces and xylem) (Fishman et al., 2018) and Ca 2+ influx can for example, activate plant defences (Aslam et al., 2008). Furthermore, Ca 2+ signalling plays an essential role in pollen tube elongation, seed germination, hyperosmotic and oxidative stresses (Sanders et al., 1999;White and Broadley, 2003).
Olive knot disease is characterized by knots or gall outgrowths on mainly twigs and young plant branches, while leaf and fruit infections are rare and only develop during wet summers. Psav survives as an epiphyte in the phyllosphere penetrating its host through wounds (Lavermicocca and Surico, 1987). Once inside host plants, the bacterium colonizes the apoplast and due to its ability to secrete the plant hormones indole-3-acetic acid (IAA) and cytokinins, it stimulates olive cells to produce new tissue giving rise to knot development and tissue overgrowth (Glass and Kosuge, 1988;Powell and Morris, 1986;Quesada et al., 2012;Ramos et al., 2012;Rodríguez-Moreno et al., 2008;Surico et al., 1985;Temsah et al., 2008). The switch from an epiphytic to endophytic (apoplastic) life style is an abrupt transition for the bacterium that requires: (i) a remarkable adaptation to an environment that is extremely different in pH, osmotic pressure, carbon sources and oxygen availability, (ii) the ability to suppress basal and induced plant defences (Rico et al., 2009). Although the bacterial signals (e.g. flagellin, elongation factors) that the plant perceives through specific receptors and via which it activates plant immunity have been extensively studied in the last decades (Buonaurio, 2008;Chisholm et al., 2006;Dangl et al., 2013;Jones and Dangl, 2006;Silva et al., 2018), little is known on the molecular signals that the phytopathogenic bacteria perceive during this transition to the apoplast. We here reveal that Ca 2+ influx in Psav is stimulated by low energy situations and that it requires a Na + / Ca 2+ exchanger that is essential for Psav virulence on olive plants.

R E S U LT S Ca 2+ entry in Psav cells is promoted under starvation conditions and is not influenced by exogenous indole-3-acetic acid
Our understanding of molecular signalling in the early phases of plant bacterial infection is limited, while this early signalling largely defines the onset of bacterial disease. Since Ca 2+ is a well-known signalling molecule in plants and animals, we here investigated the role of Ca 2+ signalling for a bacterial pathogen. We chose the olive -Psav pathosystem and used a biochemical approach to study if Ca 2+ signalling is important for pathogenicity and virulence. First, we assessed if the cytosolic Ca 2+ concentration of Psav is influenced by external Ca 2+ . To this end, we measured in Psav DAPP-PG 722 cells under basal conditions (i.e. Hanks' Balanced Salt Solution, HBSS buffer) whether an increase in external Ca 2+ resulted in an increase in the cytosolic Ca 2+ concentrations in Psav. We find that the cytosolic Ca 2+ concentrations rapidly increase in response to external Ca 2+ concentration in the medium (Fig. 1). This trend was suppressed when different carbon sources (glucose, fructose or sucrose) or ATP were added in a combination with Ca 2+ (Fig. 1). Since IAA is produced by Psav to stimulate plant cell proliferation and knot formation, we also investigated whether IAA or its precursor (L-tryptophan) influences Ca 2+ entry. However, addition of IAA or L-tryptophan to the incubation buffer did not significantly alter Ca 2+ entry in Psav DAPP-PG 722 cells (Fig. 1). Combined, these data suggest that Psav actively controls Ca 2+ entry rather than that this Ca 2+ influx represents a passive process.
Ca 2+ entry in Psav cells is mediated by the Na + /Ca 2+ exchanger CneA To determine if Ca 2+ entry depends on an ion channel, Psav DAPP-PG 722 cells were pre-treated with nifedipine, an inhibitor of the L-voltage channels responsible for the entry of the extracellular Ca 2+ in mammals (Sorkin et al., 1985) or LiCl that, in substitution of Na + in the buffer, inhibits Na + /Ca 2+ exchangers (Yanagita et al., 2007). Since Ca 2+ entry was inhibited by both nifedipine and LiCl (Fig. 2), we conclude that a Na + /Ca 2+ exchanger is potentially involved in the entry of extracellular Ca 2+ in Psav. In silico analysis of the genome of Psav DAPP-PG 722 (Moretti et al., 2014) revealed the presence of a single gene coding for a Na + /Ca 2+ exchanger (here designated as cneA; MK408668), which belongs to the ChaA antiporter superfamily (Shijuku et al., 2002). This cneA gene encodes for a protein (CneA) that encompasses the PRK10599, caca2 and Na + /Ca 2+ exchanger protein domains (Marchler-Bauer et al., 2017). In the genome of Psav DAPP-PG 722, two genes are located directly upstream of cneA gene, which encode for a guanine deaminase and hydroxydechloroatrazine ethylaminohydrolase, while downstream we find a gene encoding an iron(III) dicitrate transport system. A comparative phylogenetic analysis of the nucleotide sequences of the cneA gene was performed using the Geneious resource (Kearse et al., 2012). Sequences of this gene were retrieved from a series of strains that belong to the seven primary (monophyletic) phylogroups (PGs) described for the P. syringae complex. Homologs of the cneA gene were found to be widely distributed across the P. syringae complex. However, the branching of the cneA gene tree was not fully consistent with the previously reported phylogeny of the P. syringae species complex that is based on a multilocus sequence analysis (MLSA) of housekeeping genes (Baltrus et al., 2017). This suggests that the cneA gene has undergone horizontal gene transfer between species in this bacterial complex. For example, although PG2, PG3 and PG6 are equally distributed in a common branch in both the cneA and the MLSA phylogeny, some PG3 pathovars (e.g. P. syringae pathovars cunninghamiae, castaneae, photiniae and myricae, amongst others) have a different position in the cneA gene tree than in the MLSA tree (Fig. S1).
A Psav-cneA mutant is inhibited in Ca 2+ entry and is unable to induce both the hypersensitive response (HR) in Nicotiana tabacum and formation of knots on olive plants In order to investigate the role of the Psav-cneA gene in Ca 2+ entry, a Psav DAPP-PG 722 cneA mutant was constructed and its ability to transport Ca 2+ into the cytosol was tested in comparison to the wild-type strain. For this purpose, Psav cells were incubated in basal conditions or in the presence of glucose. The uptake of Ca 2+ was strongly impaired in the mutant cells incubated under basal condition (Fig. 3). It is worth mentioning that the in vitro growth rate of Psav-cneA mutant cells was identical to that of the Psav wild-type strain in KB medium (Likelihood ratio test, P-value = 0.85; Fig. S2).  To examine whether Ca 2+ entry was involved in Psav pathogenicity and virulence, both the Psav DAPP-PG 722 wild type and Psav-cneA mutant were: (i) infiltrated in the non-host tobacco, (ii) inoculated on 1-year-old wounded olive plants. When infiltrated in tobacco leaves, the Psav-cneA mutant was unable to induce an HR (Fig. 4A). Likewise, Psav-cneA mutant was significantly affected in the ability to induce knots on olive (Fig.  4B). In fact, olive plants inoculated with the Psav-cneA mutant showed a drastic reduction in knot overgrowth (Fig. 4C). It must be pointed out that the residual stem overgrowth seen on the Psav-cneA mutant inoculated plants was due to the formation of cicatrisation callus as a consequence of the wounding (incisions). Moreover, we found that the Psav-cneA mutant strain was unable to proliferate in olive plants in comparison with the Psav wild type (Fig. 4D).
In the Psav-cneA mutant expression of genes involved in type III secretion and phytohormone production are suppressed In order to investigate the expression of genes involved in pathogenicity and virulence of Psav, the promoter activity of the hrpL, hrpA, iaaM and ptz genes was determined via transcriptional fusions of their gene promoters with the promoterless lacZ gene. Although β-galactosidase levels associated to the hrpL and hrpA promoters were very low under the conditions tested ( Fig. 5A and B), transcription from the hrpL promoter was significantly reduced in the cneA mutant grown either in KB or HBSS media (Fig. 5A). When the Psav-cneA mutant strain was grown in Hrp medium (Huynh et al., 1989), the activity of the hrpA promoter was reduced in comparison to that obtained for wild-type Psav (Fig. 5B). In addition, hrpA promoter activity was significantly lower in Psav DAPP-PG 722 than in Psav NCPPB 3335 (Fig. 5B). In Psav-cneA mutant cells, the activity of the iaaM promoter was also low in all media tested (Fig. 5C). Nevertheless, the activity of this promoter was significantly lower in the cneA mutant than in the wild-type strains in Hrp medium, HBSS and HBSS amended with CaCl 2 . Furthermore, a significant reduction in the ptz promoter activity was seen in Psav-cneA mutant compared to Psav wild type in all media analysed (Fig. 5D). Together, these results suggest that Na + /Ca 2+ exchanger is needed for the proper expression of the tested pathogenicity and virulence genes under inducing conditions.

Psav-cneA mutant was restored by gene complementation
Complementation of the Psav-cneA mutant was performed using both a plasmid encoding the cneA gene expressed from the E. coli lac promoter (Psav-cneA mutant [pBBR::cneA]) or a mini-Tn7 transposon encoding cneA from its own promoter and inserted in the chromosome of the mutant strain (Psav-cneA mutant [miniTn7::cneA]). Ca 2+ entry into the complemented strains was restored to more than 60%, in the absence (Fig. 6) of glucose. Next, we assessed the virulence of the complemented strains on olive plants. The knot overgrowth generated by Psav-cneA mutant (pBBR::cneA) was not significantly different to that of Psav wild type (Fig. 7A), but it was significantly higher compared to that of Psav-cneA mutant (Fig. 7A). Also bacterial proliferation of the Psav-cneA mutant (pBBR::cneA) in olive plants was comparable to the Psav wild type (Fig. 7B). Similar results were obtained for Psav-cneA mutant (miniTn7::cneA) ( Fig. 7C and D). This means that gene complementation restored bacterial pathogenicity and virulence on olive plants to wild-type levels.

Other phenotypic characteristics of the Psav-cnaA mutant
In order to determine if the Psav-cneA mutant is impaired in other phenotypic traits important for its epiphytic and endophytic lifestyles (Ramos et al., 2012;Rodríguez-Moreno et al., 2009), several phenotypic characters were tested (Table 1). The mutant was impaired in the production of exopolysaccharides (EPSs), both in KB and LBS media, and N-acyl homoserine , red columns) and the calcium exchanger Psav-cneA mutant (blue columns). Bacterial β-galactosidase (LacZ) activity was measured 6 h after incubation in King's medium B (KB), Hrp, HBSS and HBSS+CaCl 2 media. As a negative control, Psav wt and Psav-cneA mutant strains transformed with a promoterless β-galactosidase were used. For comparison, hrpA promoter activity in Psav NCPPB 3335 strain (yellow column) was included. Each column is the mean of one experiment with three replicates ± SE. * For each medium, values recorded in the Psav-cneA mutant are statistically different (P < 0.05) respect to that of Psav wt, according to the Student's t-test. Columns capped with different letters, in Figure 5B, are significantly different (P < 0.05) according to the Duncan's multiple range test. lactones (AHLs), and it showed a higher swimming motility than the wild-type strain. No difference in proteolytic activity, siderophore production and swarming motility was observed between wild type and Psav-cneA mutant. At 24 h neither wild type nor the mutant formed biofilms under shaking or static conditions. The same results were obtained after 48 h incubation in shaking conditions. However, biofilm formation (similar to those of Pseudomonas putida KT2440; positive control) was detected in the Psav-cneA mutant strain 48 h after incubation at static conditions, while no formation was detected in the wild-type strain under these conditions (Fig. 8). Amongst the phenotypic characteristics examined, the complemented strains (Psav-cneA mutant [pBBR::cneA], Psav-cneA mutant [miniTn7:: cneA]) were not able to swim as the Psav wild type under the conditions tested, and they only recovered partially the capacity to produce EPSs (Table 1). Other phenotypes that were not restored in the complemented strains include the induction of the HR on tobacco plants, the production of AHLs (Table 1) and the inability to form biofilms under the conditions tested (Fig. 8).

D I SC U SS I O N
Based on our data, we propose that in the early phases of the Psav infection and in particular when the bacterium reaches the apoplast (intercellular spaces and xylem), the abundant presence of Ca 2+ (Stael et al., 2011) and the low concentration of sugars (Rico et al., 2009) therein permit Ca 2+ entry into the bacterial cells via the Na + /Ca 2+ exchanger cneA, which in turn induces the expression of Psav pathogenicity and virulence genes. Although the level of Ca 2+ in olive apoplast has not been reported, its concentration is likely sufficient to guarantee Ca 2+ influx in the Psav cells. In fact, the Ca 2+ concentrations used in this study are consistent with those reported in plant apoplast (Hepler, 2005;Plieth and Vollbehr, 2012), which range from 10 µM to 10 mM. In addition, during the early phase of bean infection with avirulent and virulent Pseudomonas savastanoi pv. phaseolicola strains, an increase in apoplastic Ca 2+ was documented (O'Leary et al., 2016). Our biochemical experiments demonstrated that Ca 2+ entry in Psav cells is inhibited by glucose, fructose or sucrose. Although the concentration of these sugars in olive apoplast has not been documented, their concentrations in the apoplast of other plants is low (Preston, 2017) and decrease during the early phase of a bacterial infection (O'Leary et al., 2016). Even though the level of these sugars in the olive apoplast should attenuate Ca 2+ entry, we have to consider that minimal changes in cytosolic Ca 2+ concentration can modulate gene expression (Borowiec et al., 2014;Domínguez, 2004). We therefore, hypothesize that a sugar starvation status can facilitate the entry of Ca 2+ inside Psav cells. A high degree of starvation already occurs during the epiphytic phase of Psav, which is able to live on olive leaf surfaces exploiting the poor nutrients there present (Ramos et al., 2012). We cannot exclude that in this ecological niche, Ca 2+ present in water and stored in EPSs enters into the Psav cells to regulate important processes that control in the epiphytic life style. The starvation experience during the epiphytic phase of the life cycle, is mitigated as soon as the bacteria enter the apoplast; however, a limited amount of starvation is always present in the apoplast that is considered a nutrient-limited environment (Rico et al., 2009), supporting the existence of certain starvation conditions also in this niche.
The importance of Ca 2+ for the virulence of a phytopathogenic bacterium was recently reported by Fishman et al. (2018), who characterized a two-component system of P. syringae pv. tomato DC3000 that is responsive to Ca 2+ and necessary for virulence of this bacterium. Through the use of a Na + /Ca 2+ exchanger mutant, we now identify for a related bacterium, Psav, an exchanger that is essential for Ca 2+ influx. In corroboration, we demonstrate at the biochemical and pharmacological level that Ca 2+ enters Psav bacterial cells via this Na + /Ca 2+ exchanger that belongs to the ChaA antiporter superfamily   Table 1 Phenotypic characterization of Pseudomonas savastanoi pv. savastanoi (wild type), Psav-cneA mutant, and two complemenation lines Psav-cneA (pBBR::cneA) and Psav-cneA mutant (miniTn7::cneA).

Wild type
Psav-cnea Psav-cnea (pBBR::cnea) ( Shijuku et al., 2002). Using a genomic knockout mutant and genetic complementation, we have shown that this exchanger is essential for Psav virulence on olive plants, as the mutant failed to induce knots. We also demonstrate that Ca 2+ entry stimulates the expression of both pathogenicity (hrpL, hrpA and iaaM) and virulence (ptz) genes again confirming that Ca 2+ is an important host signal that is perceived by the bacterium. We find that the Ca 2+ influx reaches its maximum levels when the energy supply is limiting. In fact, the presence of glucose, fructose, sucrose or ATP inhibited calcium entry entirely. Thus far, Psav virulence was largely linked to the bacterial secretion of the phytohormones IAA and cytokinins at the site of infection, which stimulates olive cell activity to produce new tissue and gives rise to knot development (Glass and Kosuge, 1988;Powell and Morris, 1986;Rodríguez-Moreno et al., 2008;Quesada et al., 2012;Surico et al., 1985;Temsah et al., 2008). Our results demonstrate that the presence of L-tryptophan or IAA does not alter Ca 2+ entry into the Psav cells, suggesting that there is no feedback regulation by the auxin pathway during infection. Our data imply that Ca 2+ entry regulates other virulence factors in Psav DAPP-PG 722 such as EPS, AHLs and biofilm production as well as swimming motility. In the Psav-cneA mutant lack of Ca 2+ entry affects specifically EPS and AHLs production. In the marine bacteria Pseudoalteromonas sp., in Pseudomonas putida and in Pseudomonas aeruginosa, Ca 2+ influences the production of the extracellular matrix and adhesion to seeds (Espinosa-Urgel et al., 2000;Patrauchan et al., 2005;Sarkisova et al., 2005). In Xylella fastidiosa Ca 2+ did not directly affect EPS production while being involved in the regulation of biofilm formation, cell surface attachment and twitching motility (Cruz et al., 2012(Cruz et al., , 2014(Cruz et al., , 2012(Cruz et al., , 2014Parker Jennifer et al., 2016). The opposite effect was observed in Psav-cneA mutant, i.e. enhanced biofilm formation after 48 h in static conditions and increased swimming motility. The Psav-cneA mutant strain fails to elicit disease symptoms in its host and HR on its non-host (tobacco) probably due to the suppression of type III secretion system. Here, in fact, it was noted that β-galactosidase activity associated to hrpL and hrpA promoters was statistically reduced in the Psav-cneA mutant. Wei et al. (2000) reported that HrpA may have a positive regulatory effect on hrpRS and hrpL genes expression in P. syringae pv. tomato. Based on the results obtained in this study, it can be argued that Ca 2+ positively controls expression of the genes for the type III secretion system.

Psav-cnea
Our results obtained with the deletion mutant were confirmed by plasmid and chromosomal mutant complementation, except for restoration of the HR. AHLs production and biofilm formation were also not recovered. This, may be due to the different expression levels of cneA in the complemented strain compared to the wild type (in the complementation line, Psav-cneA mutant (pBBR::cneA), the expression was driven by the lac promoter). However, it should be emphasized that complementation of the Psav-cneA mutation restored both Ca 2+ entry and pathogenicity in olive plants.

E X PE R I M E N TA L PROC E D U R E S Bacterial strains, plasmids and growth conditions
Bacterial strains and plasmids used in this study are listed in Table 2. Bacterial strains were grown at 27 °C in Luria-Bertani (LB) medium (Miller, 1972), King's B (KB) medium (King et al., 1954) or Nutrient Agar (NA). Escherichia coli was grown at 37 °C in LB broth. Chromobacterium violaceum strain CVO26 (McClean et al., 1997), used as AHL bacterial biosensor for AHL detection, was grown at 30 °C. Antibiotics were added, when required, at the following final concentrations: ampicillin 100 μg/mL, nitrofurantoin (Nitrof) 100 μg/mL, kanamycin (Km) 100 μg/mL and gentamicin (Gm) 10 μg/mL.

Recombinant DNA techniques
DNA digestions with restriction enzymes (XhoI, SpeI and EcoRI), agarose gel electrophoresis, DNA fragment purification, ligation with T4 ligase, end filling using the Klenow enzyme and E. coli transformation were performed as described by Sambrock et al. (1989). Plasmids were purified using the GenElute™ Plasmid Miniprep Kit (Sigma-Aldrich, MO, Saint Louis, USA). The genomic DNA was extracted with the GenElute Bacterial Genomic DNA Kit (Sigma-Aldrich, MO, Saint Louis, USA). Triparental mating between E. coli and Psav DAPP-PG 722 was performed using a helper E. coli strain carrying plasmid pRK2013 (Figurski and Helinski, 1979).

Determination of the cytosolic Ca 2+ levels
Cytosolic Ca 2+ levels were determined using a fluorimetric method, which employed the fluorescent probe Fura 2-AM (Fura 2-acetoxy methyl ester; Sigma-Aldrich, MO, Saint Louis, USA). Approximately 5 × 10 6 cells of Psav DAPP-PG 722 grown at 27 ± 1°C for 16 h in LB broth to a stationary phase, were suspended in 0.12 M Tris (pH 7.8) and 2 mM EGTA. At 200 s after incubation at 25 °C, 2 mM CaCl 2 was added to stop the EGTA effect as reported by Grynkiewicz et al. (1985). Then the cells were incubated for 2 h in basal condition i.e. HBSS buffer (140 mM NaCl, 5.3 mM KCl, 25 mM HEPES, pH 7.4) supplemented with 2 mM Fura 2-AM (dissolved in DMSO) or in HBSS buffer supplemented with 2 mM Fura 2-AM and different carbon sources (glucose, fructose or sucrose, 5 mM) or ATP 50 µM. The fluorescence intensities of Fura 2-AM (Ex. = 335 nm, Em. = 505 nm) were monitored with a spectrofluorophotometer (Perkin-Elmer, Waltham, Massachusetts, USA). The cytosolic Ca 2+ concentration was calculated following the formula reported by Grynkiewicz et al. (1985).

Phylogenetic analysis of the Na + /Ca 2+ exchanger gene
A comparative phylogenetic analysis of the nucleotide sequences of the cneA gene coding for the Na + /Ca 2+ exchanger was performed using the Geneious resource (Kearse et al., 2012). Blast searches were used to retrieve the close homologs of the cneA gene from different Pseudomonas species. Phylogenetic and molecular evolutionary analysis was conducted using MEGA 7 (Kumar et al., 2016) and the maximum likelihood method. Clade stability was assessed by 1000 bootstrap replications.
Construction of a P. savastanoi pv. savastanoi knockout mutant of the Na + /Ca 2+ exchanger gene cneA A genomic null mutant of the Na + /Ca 2+ exchanger gene (referred to as cneA gene) was created as follows. An internal 305 bp fragment of the cneA gene was amplified from Psav DAPP-PG 722 genomic DNA using the primers cneA For and cneA Rev ( Table 2). The amplified PCR product was cloned in plasmid pKNOCK-Km (Alexeyev, 1999), generating pKNOCK-cneA (Table 2). A Psav-cneA knockout mutant (Table 2) was generated by homologous recombination (Alexeyev, 1999) after transformation of pKNOCK-cneA in Psav DAPP-PG 722 as a suicide delivery system. Transformants were selected on KB-Nitrof + Km plates. Interruption of cneA was verified by PCR using primers specific to the pKNOCK-Km vector and to the genomic DNA sequences upstream and downstream of the targeted gene. The amplicons were sequenced at Macrogen Europe (Amsterdam, Netherlands; http://www.macrogen.com).

Plasmid and chromosomal complementation of Psav-cneA mutant
Complementation of Psav-cneA mutant with a plasmid encoding the cneA gene was performed as follows. The complete sequence of the cneA open reading frame (ORF) with its ribosome binding site was amplified from Psav DAPP-PG 722 genomic DNA using primers cneA compl 1 and cneA compl 2 (Table 2) and Q5 ® High-Fidelity DNA Polymerase (New England Biolabs, Hitchin, UK). The amplified fragment was purified from an agarose gel using the EuroGOLD Gel Extraction Kit (EuroClone, Milan, Italy) following the instructions of the manufacturer. After A-tailing (Promega, Fitchburg, WI, USA), the fragments were cloned in pGEM-T Easy vector (Promega, Fitchburg, WI, USA) and sequenced at Macrogen Europe. Having verified the correctness of the sequence, the cneA ORF was excised from pGEM-T Easy Vector using XhoI and SpeI and cloned in the corresponding sites of the plasmid pBBR MCS-5. The resulting plasmid (pBBR MCS-5-cneA; Table 2) was purified using the GenElute Plasmid Miniprep Kit (Sigma-Aldrich, MO, Saint Louis, USA) and transformed in Psav-cneA by electroporation, generating Psav-cneA mutant (pBBR::cneA) ( Table 2). Chromosomal complementation of the Psav-cneA mutant was performed using the Tn7 transposon vector pUC18R6KT-miniT-n7BB-Gm. The complete ORF of the cneA gene, including its own promoter, was amplified from Psav DAPP-PG 722 chromosomal DNA using the Expand High Fidelity PCR System (Roche, Mannheim, Germany) and the primers PromAP_Fw and AP_Rev ( Table 2). The amplified DNA fragment was cloned in the pGEM-T Easy Vector (Promega, WI, Fitchburg, USA) and sequenced at GATC Biotech (Konstanz, Germany). Once verified the correctness of the sequence, the cneA gene was excised from pGEM using EcoRI and cloned in the corresponding site of the plasmid pUC18R6KT-miniTn7BB-Gm, yielding pUC18R6KT-miniT-n7BB-cneA-Gm ( Table 2) that was electroporated in Psav-cneA. Selection of the transconjugants in KB-Gm plates yielded the complemented strain Psav-cneA mutant (miniTn7::cneA) ( Table 2). Insertion of the Tn7 transposon into the correct site was verified using the primers GlmS_savastanoi (hybridizing at the 3′ of the glmS gene) and the Tn7Rev primer (hybridizing at the Tn7R end of the integrated plasmid) ( Table 2). Only in the case of integration, a 165 bp fragment was amplified.

Phenotypic characterizations of the Psav-cneA mutant and its complemented strains
In vitro bacterial growth dynamics of wild-type Psav and the Psav-cneA mutant strains were carried out in KB liquid medium at 27 °C. Bacterial growth was spectrophotometrically followed every hour for 24 h at OD 660 and through colony counts at 4, 8, 20, 24 and 28 h post-incubation (hpi). For each bacterial strains, the relationship between the number of cells (log 10 transformed) and the hpi was investigated by means of a second-order polynomial model. Likelihood ratio test was used to assess the differences between wild type and Psav-cneA mutant strains under R statistical environment (R Core Team, 2018).
The HR assay was carried out in Nicotiana tabacum (cv. Havana 425) plants. To prepare the inoculum, the strains were grown in NA at 27 °C for 24 h, resuspended in sterile deionized water and spectrophotometrically adjusted to 10 8 CFU/mL. About 10 µL of the bacterial suspensions or water (control) was infiltrated into the mesophyll of tobacco leaves using a needleless syringe. The appearance of the HR was scored at 24 hpi.
Proteolytic activity, swarming and swimming were determined as reported by Huber et al. (2001). Qualitative analysis of EPSs was tested on KB and LB solid medium amended with 5% of sucrose (LBS). Single colonies, previously obtained from NA plates, were streaked on KB and LBS and then grown at 28 ºC for 48 h. Colonies producing EPSs showed a fluidal, mucoid appearance. Production of AHLs was performed in T-streak analysis as described by Piper et al. (1993) using the C. violaceum CVO26 as AHL biosensor. To measure biofilm formation, overnight cultures of Psav DAPP-PG 722, Psav-cneA mutant and Psav-cneA mutant (pBBR::cneA) grown in KB broth, were diluted to OD 600nm = 0.1 and loaded in a 96-well plate (150 µL per well, eight wells per strain). Plates were incubated under static or shaking conditions and biofilm formation was quantified by measurement of the A 595 after 24 h and 48 h after crystal violet staining (O'Toole and Kolter, 1998). Pseudomonas putida KT2240 was included as a positive control for biofilm formation and cell-free KB as negative control.

Pathogenicity test on olive plants
Disease severity and bacterial growth were tested in 1-year-old olive (cv. Frantoio) plants inoculated with the strains Psav DAPP-PG 722, Psav-cneA mutant, Psav-cneA mutant (pBBR::cneA) and Psav-cneA mutant (miniTn7:: cneA) ( Table 2). To prepare the inoculum, bacteria were grown on NA at 27 °C for 48 h, resuspended in sterile deionized water and adjusted spectrophotometrically to approximately 1 × 10 8 CFU/mL −1 . Also, 20 μL of bacterial suspension or water (control plant) was placed in wounds (five per plant) made in the bark of olive plants using a sterile scalpel as previously described (Moretti et al., 2008). Wounds in the inoculated and control plants were protected with parafilm (American National Can, IL, Chicago, USA) until the developing knots break it (14 to 21 days). Plants were maintained in transparent polycarbonate boxes to reach high RH values (90%-100%) and kept in a growth chamber at 22 °C to 24 °C with illumination at 70 μE/m -2 s -1 and 12 h light period. The Psav population density was calculated at 0, 7, 14, 21 and 60 dpi by serial dilution of the bacterial suspension obtained from inoculated sites excised and homogenized by mechanical disruption and plated in NA medium. Colony counts were calculated 24 h and 48 h after incubation at 27 °C. The disease severity was recorded at 60 dpi by determining the knot volume, by measuring the length, width and depth of every knot with a Vernier caliper (Moretti et al., 2008). Four plant replicates were included in each of the two in planta experiments performed.

Transcriptional analysis of Psav pathogenicity and virulence genes
To verify whether Ca 2+ entry promotes the expression of pathogenicity (hrpL, hrpA and iaaM) and virulence (ptz) genes of Psav, transcriptional fusions of their promoters were constructed with LacZ reporter gene. For amplification of the iaaM and ptz promoters, the regions upstream of the iaaM and ptz ORFs (477 bp and 373 bp, respectively) were amplified by PCR using primers iaaM For, iaaM Rev, ptz For and ptz Rev (Table 2). Amplicons were cloned into pMP220 in order to obtain promoter fusions to lacZ. The resulting plasmids and those encoding the hrpL and hrpA promoters fused to lacZ (Aragόn et al., 2014) were transferred by conjugation into both wild-type Psav DAPP-PG 722 and its cneA mutant. Cells carrying the plasmids grown overnight in KB media were diluted in the same media and incubated at 28 ºC to OD 660 of 0.5 (time = 0). The cultures were harvested by centrifugation, washed twice with 10 mM MgCl 2 and the cells were transferred to Hrp medium (Huynh et al., 1989), HBSS and HBSS amended with CaCl 2 . The cultures were adjusted to OD 660 of 0.5 and incubated for 6 h at 28 ºC. β-galactosidase enzymatic activity was measured using the methods developed (Miller, 1972) and modified previously (Maloy, 1990). Psav DAPP-PG 722 and its cneA mutant transformed with pMP220 (encoding a promoterless lacZ) were used as negative controls. To determine the activity associated exclusively to the promoter fusions to lacZ, the background activity detected in the control strains was subtracted from those obtained for each of their corresponding transformants.

SU PP O R T I N G I N F O R M AT I O N
Additional supporting information may be found in the online version of this article at the publisher's web site: