Hibiscus syriacus plant samples were collected during a survey on the cultivable bacterial population inhabiting ornamental plants. The survey was carried out in 2012 in a commercial nursery located in Pistoia, Italy, by sampling one plant per species. Ten grams of leaves, stems and buds were suspended in 90 mL of a solution containing 1% peptone and 1% Tween 90, and shaken at 200 rpm for one hour at room temperature. Three suspensions were prepared, bulked and then spread as 100 µL aliquots onto the surface of nutrient glucose agar (NGA) medium amended with cycloheximide (200 ppm). Plates were incubated at 25 ± 2°C for one week, after that bacterial colonies of different morphology were picked up and streaked (at least twice) on NGA medium for purification. Pure cultures were suspended in 30% glycerol solution and maintained at -80°C until this work was carried out.

The S. rhizophila strain that we named Ep2.2 was identified by 16S rDNA amplification with the primers fD1 (5’–GAGTTTGATCCTGGCTCAG–3’) and rP1/rP2 (5’– GGYTACCTTGTTACGACTT–3’; Y=C/T) (Pious and Thyvalappil, 2009) according to the protocol described by Krimi et al. (2016). The amplified 16S rDNA fragment was analyzed for similarity by comparing to known nucleotide sequences present in the NCBI GenBank database by BLASTn search (http://blast.ncbi.nlm.nih.gov/). The taxonomical affiliation was also determined by aligning the S. rhizophila Ep2.2 16S rDNA sequence to 16S rDNA sequences available in the NCBI database belonging to six Stenotrophomonas species, three Xanthomonas species, and Xylella fastidiosa. The evolutionary history was inferred using the Neighbor-Joining method (Saitou and Nei, 1987). The evolutionary distances were computed using the Maximum Composite Likelihood method (Tamura et al., 2004) and expressed in the units of the number of base substitutions per site. Evolutionary analyses were conducted in MEGA 6 (Tamura et al., 2013). The S. rhizophila Ep2.2 16S rDNA sequence was deposited in the EMBL/GenBank/DDBJ nucleotide databases under the accession number MZ841807. S. rhizophila Ep2.2 is included in the microbial collection of the IPSP-CNR (Sesto Fiorentino, Italy).

S. rhizophila Ep2.2 was assessed for Gram reaction by the KOH test (Buck, 1982), for catalase and oxidase activities (Schaad et al., 2001), for siderophore production on King agar B medium (Sigma-Aldrich, MO, USA) following the procedure of Mikiciński et al. (2016), and for glucanase activity on tryptic soy agar medium (Sigma-Aldrich, MO, USA) amended with glucane 0.1%. Proteolytic, lipolytic and chitinolytic activity tests were performed on skim milk agar, LB agar (Sigma-Aldrich, MO, USA) amended with 1% Tween 40, and Chitin azure medium (Sigma-Aldrich, MO, USA), respectively. Exopolysaccharide (EPS) production was assayed by using the protocol described by Tallgren et al. (1999). The ability to grow at 4, 30, 37 and 40°C was assessed in LB medium after one week of incubation. 1-aminocyclopropane-1-carboxylic acid (ACC) deaminase activity was detected by means of M9 minimal medium with ACC as unique N source (Penrose and Glick, 2003). The ability to solubilize P was verified by using the NPRBB growth medium (Nautiyal, 1999). Indole-3-acetic acid (IAA) production was tested on LB medium amended with L-triptophan according to Bric et al. (1991). The bacterial strains used as reference for the different biochemical tests are reported in Supplementary Table S1.

Swimming, swarming and twitching motility, and ability to form biofilm, were determined according to Déziel et al. (2001). Motility was assessed in comparison to S. maltophilia LMG 958, Xanthomonas campestris pv. campestris Xcc1, Erwinia amylovora E1 and Agrobacterium tumefaciens C58. The biofilm production assay was modified to be performed in microtiter plates (Costar assay plate, Corning, NY, USA).

The metabolic profile of S. rhizophila Ep2.2 was analyzed with the BIOLOG system by using GEN III MicroPlate (Catalog No. 1030, Rigel, Italy), according to the protocol provided by the manufacturer (BIOLOG, USA). The GEN III MicroPlate includes 94 phenotypic tests: 71 carbon source utilization assays and 23 chemical sensitivity assays, and allows identifying bacteria at the species level. The analysis of the metabolic fingerprint was performed by incubating at 30°C for 75 h in the OmniLog device (BIOLOG, USA), which yields colorimetric curves indicating utilization of the carbon sources or resistance to inhibitory chemicals, as a result of cell respiration. Respiration causes reduction of a tetrazolium redox dye and formation of purple color. The test was carried out in comparison with S. maltophilia type strain LMG 958 in order to highlight metabolic differences with S. rhizophila Ep2.2.

The antagonistic activity of S. rhizophila Ep2.2 was evaluated by dual culture assay against four fungal plant pathogens: Alternaria alternata, Botrytis cinerea, Fusarium oxysporum f. sp. lycopersici and Rhizoctonia solani. Briefly, 5 mm plugs were cut from the margin of fresh fungal colonies grown on potato dextrose agar (PDA; VWR Chemicals, Belgium) medium and placed in the middle of Petri dishes containing the same medium. Two 50 µL aliquots of S. rhizophila Ep2.2 suspension (OD_{600 nm} = 0.1, corresponding to 1x10⁸ cells/mL) were streaked at the two opposite sides of the fungal plug. The plates were incubated at 26°C for six days before measuring fungal colony diameters and comparing to control cultures grown in the absence of S. rhizophila Ep2.2.

Subsequently, to demonstrate the inhibitory role of volatile compounds produced by S. rhizophila Ep2.2 on B. cinerea growth, Petri dishes, prepared as described above, were either sealed with three layers of parafilm or kept unsealed, and incubated at 26°C for 2, 3 or 7 days. Diameters of fungal colonies grown both on sealed and unsealed plates were then measured and compared.

In order to exclude inhibition by diffusible molecules, a culture filtrate of S. rhizophila Ep2.2 was prepared and tested against B. cinerea as described below. S. rhizophila Ep2.2 was grown on Nutrient Broth (Scharlab S.L., Spain) amended with 2.5 g/L glucose (NGB) on an orbital shaker at 25 ± 2°C, 120 rpm, for 48 h. The bacterial suspension was then centrifuged at 10,000 rpm for 10 min. and the supernatant was then collected and sterilized through a Millipore 0.2 µm filter. The sterile filtrate (1.5 mL) was mixed with 15 mL of PDA cooled at 55°C. After medium solidification, a 0.5 mm agar plug was cut from the margin of a colony of B. cinerea grown on PDA and placed in the middle of the Petri dish. The test was conducted in triplicate. B. cinerea was grown on unamended PDA as a control. Plates were incubated at 26°C for 4 days after that the inhibitory activity was evaluated by measuring the diameter of B. cinerea colonies.

Two-compartment 92-mm-diameter Petri dishes with ventilation cams and common headspace (Sarstedt, Nümbrecht, Germany) were used to determine the inhibiting activity of volatile compounds produced by S. rhizophila Ep2.2 against the four different phytopathogenic fungi listed in the previous paragraph. An aliquot (7 mL) of agarized NGB medium (NGA) was poured into one of the two compartments of the plate, while 7 mL of PDA medium were poured into the other compartment. Once dried, 50 μL of S. rhizophila Ep2.2 bacterial suspension (OD_{600 nm} = 0.1) were spread on NGA medium, whereas a 5-mm diameter plug was cut from the edge of the fungal colony grown on PDA and placed at the center of the other compartment containing PDA. As a control, two-compartment plates containing non-inoculated NGA and fungus-inoculated PDA were used. Inoculated plates were sealed with 3 layers of parafilm to prevent dispersion of volatile compounds, and the inhibitory activity on phytopathogenic fungi was evaluated by measuring the colony diameter after 3 days of incubation at 26°C. Four replicates per each S. rhizophila Ep2.2-fungus combination were analyzed.

Solanum lycopersicum cv. Micro-Tom and cv. Marmande plants were grown in a growth room under LED lights (photoperiod 12/12 h) as previously described (Baccelli et al., 2022). The ability of S. rhizophila Ep2.2 to protect from B. cinerea infection was first tested on Micro-Tom leaves detached from plants during their second month of growth. A number of 10-13 mature and healthy leaves (selected among the 3^rd to the 5^th leaf) were cut from different undamaged plants and placed into 90 mm-Petri dishes containing a filter paper disc soaked with 1 mL sterile water to ensure high internal relative humidity (RH) conditions during incubations. S. rhizophila Ep2.2 was grown overnight in NGB medium at 28°C, centrifuged at low speed, washed twice in 10 mM MgCl₂, and finally suspended in 10 mM MgCl₂ (OD_{600 nm} = 0.1) for leaf treatment. Treatments were performed by spraying the lower (abaxial) surface of a leaf with approximately 300 μL of bacterial suspension with a 10-mL pump atomizer. Control leaves were sprayed with 10 mM MgCl₂. All the plates were sealed with parafilm and incubated at 21°C (day)/18°C (night), photoperiod 12/12 h (100 μmol m² s⁻¹) for 48 h, after that B. cinerea strain B05.10 was inoculated. For pathogen inoculation, conidia were collected from 15-20 day-old B. cinerea cultures grown in PDA under a light/dark regime (Schumacher, 2017; Baccelli et al., 2022). Conidia were suspended in potato dextrose broth (PDB, Laboratorios Conda S.A., Spain) at the concentration of 1x10⁶ conidia/mL and inoculated on leaves by applying one, or two, 10-µL droplets on each side of the midrib on the abaxial leaf surface. All the plates were sealed once again with parafilm and incubated as described above. Necrotic lesions caused by B. cinerea infection were measured after 3 days of incubation. The protective effect shown by S. rhizophila Ep2.2 against B. cinerea was validated on whole plants by using the tomato cv. Marmande. Plants were sprayed 2 weeks after germination on the adaxial leaf surfaces with S. rhizophila Ep2.2 as described above, incubated 48 h under high RH conditions, and subsequently inoculated on the adaxial surface of two opposite leaves (2^nd and 3^rd) with a drop of a conidial suspension prepared as already described. Lesions were measured after 3 days of incubation.

The expression of plant defense genes was analyzed by RT-qPCR to investigate the ability of S. rhizophila Ep2.2 to induce localized resistance in leaves. The analyses were designed to reveal either the ability of S. rhizophila Ep2.2 to induce directly defense genes before infection, or prime them for a quicker/boosted induction during B. cinerea B05.10 infection. Leaves from different 5-week-old S. lycopersicum cv. Micro-Tom plants were detached and treated as follows: a) leaves were spray-treated on their abaxial surface either with S. rhizophila Ep2.2 (OD_{600 nm} = 0.1 in 10 mM MgCl₂), or with 10 mM MgCl₂ as control, and then incubated for 48 h (samples named as “S. rhizophila” and “control”, respectively, at 48 hours post treatment, hpt); b) leaves were spray-treated on their abaxial surface with either S. rhizophila Ep2.2, or 10mM MgCl₂ as control, incubated for 48 h, and subsequently inoculated with 10-µL droplets (3-6 per leaf) of 1x10⁶ B. cinerea conidia/mL in PDB (samples named as “S. rhizophila + B.cinerea” and “control + B. cinerea”, respectively, at 6 or 24 h post infection, hpi); c) leaves were spray-treated on their abaxial surface with either S. rhizophila Ep2.2, or 10 mM MgCl₂ as control, incubated for 48 h, and subsequently mock inoculated with 10-µL droplets (3-6 per leaf) of PDB (samples named as “S. rhizophila + mock” and “control + mock”, respectively, at 6 and 24 hpi). Each biological replicate consisted of two leaves belonging to different plants, and three biological replicate per condition were prepared. Leaves were incubated into 90 mm-Petri dishes as described in the previous paragraph and frozen in liquid nitrogen upon sampling. To check that the treatments were leading to the expected reduction in disease symptoms, some “control + B. cinerea” and “S. rhizophila + B. cinerea” leaves were inoculated with a single 10-µL drop of conidial suspension prepared as described above and kept incubating for a longer period (96 hpi). Lesion diameters were then measured.

For RNA extraction, leaves were ground in liquid nitrogen and total RNA was extracted by using RNeasy Plant Mini Kit with buffer RLT (Qiagen, Italy) (Baccelli et al., 2022). The extracted RNA was quantified in a Qubit fluorometer (Thermo Fisher Scientific, MA, USA) and its integrity verified by agarose gel electrophoresis. Amplification Grade DNase I (Sigma-Aldrich) was used to degrade traces of contaminating DNA before reverse-transcription, which was performed with Maxima First Strand cDNA Synthesis Kit (Thermo Fisher Scientific, MA, USA). qRT-PCRs were performed in a StepOne Real-Time PCR System (Applied Biosystems, Thermo Fisher Scientific Inc. Waltham, MA, USA) by using Fast SYBR™ Green Master Mix (Applied Biosystems, Vilnius, Lithuania) as described in Baccelli et al. (2022).

The following genes were analyzed: Proteinase inhibitor II (PIN2), β-1,3-glucanase A (GluA), Pathogenesis-related protein 1 (PR1), 1-aminocyclopropane- 1-carboxylate oxidase 1 (ACO1), Pathogenesis-related genes transcriptional activator (PTI5), Lipoxygenase A (Lox1.1), and Polygalacturonase inhibitor protein (PGIP). The genes were selected based on their known modulation occurring during B. cinerea infection in tomato cv. Micro-Tom (Baccelli et al., 2022). Gene locus IDs and primer sequences are reported in Baccelli et al. (2022). Relative gene expression values were calculated by using the 2^-ΔΔC_T method as described in Livak and Schmittgen (2001) after melting curve analysis and amplification plot comparisons. Actin-7 was used as the endogenous reference gene for transcript normalization (Baccelli et al., 2022). Three biological replicates and two technical replicates were analyzed per each condition.

Motility of bacterial strains was analyzed by one-way ANOVA with Tukey–Kramer multiple comparison post-test (p ≤ 0.05) by performing the analysis separately per each time point. Data concerning colony and lesion diameters were analyzed by unpaired t-test (S. rhizophila-treated vs. control) and considered significant at p ≤ 0.05. Relative gene expression values were analyzed by unpaired t-test (48 hpt) or one-way ANOVA with Tukey–Kramer multiple comparison post-test (6 and 24 hpi) after normality check, and considered significantly different at p ≤ 0.05. Analyses were performed in GraphPad Prism 9 (GraphPad Software Inc., CA, USA).

The bacterial strain that we named Ep2.2 was isolated from aboveground organs of a H. syriacus plant, as described in materials and methods. The analysis of 16S rDNA sequence (GenBank acc. no. MZ841807) allowed identifying the strain as S. rhizophila (99.78% nucleotide identity). The phylogenetic analysis clustered Ep2.2 with the reference S. rhizophila strain e-p10, clearly separating them from other Stenotrophomonas and Xanthomonas species (Figure 1).

S. rhizophila Ep2.2 developed on NGA medium as a pale-yellow glistening bacterial colony, with entire margins. The pure isolate was Gram negative, catalase and oxidase positive, and unable to grow at 4°C and 37°C. The strain showed proteolytic, lipolytic and chitinolytic activities but it did not produce β-glucanases (Table 1). S. rhizophila Ep2.2 was also unable to produce siderophores, extracellular polymeric substances (EPS), indole-3-acetic acid (IAA), ACC deaminase, and to solubilize phosphate (Table 1). S. rhizophila Ep2.2 displayed a marked ability to form biofilm, similarly to S. maltophilia LMG 958, as well as significant swimming and swarming abilities which were comparable, although statistically different, to those of other tested bacterial species (Figure 2). In particular, S. rhizophila Ep2.2 showed higher swimming and swarming abilities, and lower twitching ability than S. maltophilia LMG 958 (Figure 2).

The metabolic profile of S. rhizophila Ep2.2 was unambiguously different from that of S. maltophilia LMG 958 (Figure 3). In general, the metabolic response of S. rhizophila Ep2.2 was slower or reduced in comparison to S. maltophilia LMG 958, except for D-trehalose, β-Methyl-D-Glucoside, and D-galactose on which S. maltophilia Ep2.2 displayed a better utilization capacity (Figure 3: wells A4, B4, and C4, respectively). However, unlike S. maltophilia LMG 958, S. rhizophila Ep2.2 either did not or barely utilize N-acetyl-β-D-mannosamine, D-fructose, 3-methyl glucose, D-fucose, L-fucose, L-rhamnose, inosine, D-glucose-6-phosphate, D-fructose-6-phosphate, D-serine, L-arginine, L-aspartic acid, L-glutamic acid, D-galacturonic acid, L-galactonic acid lactone, D-gluconic acid, D-glucuronic acid, glucuronamide, D-lactic acid methyl ester, α-Hydroxy-Butyric Acid, and α-Keto-Butyric Acid (Figure 3: wells B7, C3, C5, C6, C7, C8, C9, D6, D7, D9, E4, E5, E6, F2, F3, F4, F5, F6, G3, H3, and H5, respectively). S. rhizophila Ep2.2 resulted more sensitive to fusidic acid and potassium tellurite (Figure 3: wells C11 and G12). Noteworthy, the sensitivity to NaCl, previously known to discriminate S. rhizophila from S. maltophilia (Wolf et al., 2002), was similar between the two Stenotrophomonas strains tested here (i.e. tolerance to 1 and 4% NaCl; sensitivity to 8% NaCl) (Figure 3: wells B10-12). Moreover, both S. rhizophila and S. maltophilia strains showed sensitivity to minocycline (Figure 3: well D12). Distinct metabolic curves were observed concerning growth at pH 5, and regarding utilization of dextrin, sucrose, D-turanose, L-histidine, L-serine, L-lactic acid and Tween 40 (Figure 3: wells A12, A2, A7, A8, E7, E9, G4, and H1).

The inhibitory activity of S. rhizophila Ep2.2 was assessed against four fungal plant pathogens able to infect tomato plants: B. cinerea, A. alternata, F. oxysporum f. sp. lycopersici and R. solani. Two different methods were used: dual culture assay, where both S. rhizophila Ep2.2 and the pathogenic fungus were grown on the same PDA medium (in a Petri dish not sealed with parafilm), and a two-compartment assay where S. rhizophila Ep2.2 and the pathogenic fungus were grown physically separated, each one on its appropriate medium (NGA and PDA, respectively), while the headspace containing volatile compounds was shared. The two-compartment Petri dish was sealed with parafilm.

The dual culture assay highlighted a mild inhibitory activity of S. rhizophila Ep2.2 against A. alternata (~11% growth reduction), whereas B. cinerea, R. solani and F. oxysporum f. sp. lycopersici were not significantly affected by the presence of S. rhizophila Ep2.2 after 6 days of growth on PDA medium (Figure 4).

In contrast, when the two-compartment assay was performed, the growth of both B. cinerea and A. alternata was markedly inhibited by S. rhizophila Ep2.2 (~62% and ~28% growth reduction, respectively), whereas no significant reduction was detectable for R. solani and F. oxysporum f. sp. lycopersici (Figure 5). These results highlighted a role for volatile compounds in inhibiting the growth of microbial pathogens, especially B. cinerea (Figures 4, 5 and Supplementary Figure S2).

In order to confirm this clue, the dual culture assay between S. rhizophila Ep2.2 and B. cinerea was repeated by sealing the plates with parafilm, and B. cinerea growth was measured after 48, 72 and 168 h by comparing sealed with unsealed plates. As shown in Figure 6, whereas B. cinerea was not significantly affected by S. rhizophila Ep2.2 in unsealed plates, its growth was significantly inhibited in parafilm-sealed plates (Figure 6). In addition, when a culture filtrate from S. rhizophila Ep2.2 was produced and tested again B. cinerea no significant reduction in growth was observed as compared to control (Supplementary Figure S1). This result excluded an inhibitory role of diffusible molecules released by the bacterium into the medium. Overall, these results suggested the production by S. rhizophila Ep.2.2 of volatile compounds able to inhibit the growth of fungal plant pathogens (A. alternata and B. cinerea) and B. cinerea was exclusively inhibited by these compounds.

Real-time screening by PTR-MS of VOCs present in the headspace of axenic cultures of S. rhizophila revealed a complex blend, although 10 protonated ions represented 97.9% of the total (Table 2 and Supplementary Table S2). In particular, three protonated ions resulted to be mainly present: m/z = 33 (32.3 ± 4.9%), which was unambiguously assigned to methanol, followed by m/z = 97 (30.1 ± 43.3%) and m/z = 49 (21.8 ± 13.8%). Because of the high relative humidity of the headspace, the abundant presence of methanol generated a small percentage of a water-clustered methanol (i.e. methanol-H₂O) detectable at m/z = 51 (2.0 ± 15.7%). Simultaneous GC-MS analysis of the same samples (Supplementary Table S3) allowed identifying the protonated ion m/z = 97 as 2, 4-dimethyl furan, whereas m/z = 49 was assigned to methanethiol (Table 2). Among the VOCs emitted in higher percentage, other sulphur containing compounds such as dimethyl sulfide (DMS) and dimethyl disulfide (DMDS) were detected by PTR-MS at m/z = 63 and at m/z = 95, respectively, as confirmed by GC-MS identification (Table 2 and Supplementary Table S3). Moreover, the PTR-MS analysis detected a protonated ion at m/z 99 (4.0 ± 52.2%) that we assigned to a fragment of hexanoic acid rather than to hexenals, based on previous analyses (Bergamaschi et al., 2015) and because we did not detect the fragment at m/z 81, which is typically produced from the fragmentation of hexenals following proton transfer reaction (Brilli et al., 2011). We assigned to hexanals the protonated ion at m/z 101 due to the concomitant presence of m/z 83 (Supplementary Table S3), which is its main fragment (Brilli et al., 2011).

Complementary analysis by GC-MS, which has different sensitivity and selectivity than PTR-MS, highlighted the capacity of S. rhizophila to emit a wide variety of VOCs belonging to different chemical classes (Figure 7). Among those, the most abundant VOCs resulted to be the haloalkane trichloromethane (26.0 ± 15.7%), the alkanes 2,4-dimethyl heptane (9.5 ± 8.8%) and 4-methyl octane (8.0 ± 9.1%), followed by furans (furan = 2.4 ± 8.4% and 2,4-dimethyl furan = 3.0 ± 31.2%) and the organosulfur compound dimethyl sulfide (3.2 ± 8.9%). Terpenes, such as α- and β-pinene, camphene, and Δ-3-carene were present in very small percentage (< 0.1%) (Supplementary Table S3).

To assess the ability of S. rhizophila Ep2.2 to protect plant tissues against B. cinerea infection, leaves from tomato cv. Micro-Tom were detached, sprayed with S. rhizophila Ep2.2, incubated for 48 h under high RH conditions, and subsequently inoculated with B. cinerea conidia. As shown in Figure 8, the treatment with S. rhizophila Ep2.2 strongly reduced B. cinerea colonization of leaf tissues. The disease severity, as determined by the lesion size diameters, was significantly reduced after three days of incubation (~50%). After four days, the lesion size reduction was even greater (Supplementary Figure S3). The protective effect induced by S. rhizophila Ep2.2 against B. cinerea was clearly reproducible on whole tomato plants of a different cultivar (Marmande) (Supplementary Figure S4).

A gene expression analysis was performed to investigate the contribution of induced resistance to the protective effect shown by S. rhizophila Ep2.2 against B. cinerea infection in tomato leaves. This analysis was designed to highlight either the local induction of plant defense genes before infection, or their quicker or stronger expression upon infection (i.e. defense priming). Genes were selected based on their involvement in defense responses to B. cinerea as reported in Baccelli et al. (2022). Before RNA extraction, the outcome of B. cinerea infection was verified to make sure that leaf samples were displaying increased protection against B. cinerea infection (Supplementary Figure S3).

After 48 h following treatment with S. rhizophila, and prior to B. cinerea inoculation, no significant changes in the expression of defense genes were detectable in tomato leaves compared to control (Figure 9A). In contrast, 6 h after B. cinerea inoculation (Figure 9B), a significant up-regulation of the Pti5 gene was detectable only in tomato leaves previously treated with S. rhizophila (“S. rhizophila + B. cinerea” samples) (Figure 9B), thus indicating a quick response to pathogen infection in S. rhizophila-treated leaves. After 24 h, all the genes investigated were significantly modulated by the infection (24 hpi) with the exclusion of Pin2, whose levels remained unaltered (Figure 9C). The GluA gene was significantly up-regulated in B. cinerea-infected tomato leaves regardless of S. rhizophila treatment, although in S. rhizophila-treated leaves the expression level was significantly higher (Figure 9C), indicating a stronger response to pathogen infection in leaves pre-treated with S. rhizophila. In contrast, the Pti5 gene was up-regulated to a lower extent by B. cinerea infection in the leaves pre-treated with S. rhizophila (Figure 9C). The Lox1.1 gene was strongly downregulated by the infection with B. cinerea, whereas PGIP and Aco1 genes were significantly upregulated, irrespective of the treatment with S. rhizophila. After 24 h, the PR1 gene was significantly up-regulated only in the leaves pre-treated with S. rhizophila (i.e. “S. rhizophila + Mock” and “S. rhizophila + B. cinerea”) (Figure 9C), whereas no significant up-regulation was detectable in B. cinerea infected leaves which were not treated with S. rhizophila (“control + B. cinerea”), indicating that the beneficial bacterium was able to enhance PR1 gene expression.

Overall, the gene expression analyses suggested that inoculation with S. rhizophila did not cause major changes in gene expression in healthy tomato leaves, with the exception of PR1 gene up-regulation at late time points (24 hpi, i.e. 72 h after S. rhizophila inoculation on leaves, Figure 9C). However, the treatment with S. rhizophila led to the early up-regulation of the Pti5 gene (6 hpi) and to the enhanced up-regulation of GluA and PR1 genes (24 hpi) during infection with B. cinerea.