Kiwifruits cv “Hayward” [Actinidia deliciosa (A. Chev.) C.F. Liang & A. R. Ferguson] were harvested at commercial maturity (7.2° Brix) from an orchard under integrated pest management (IPM) production system located in the Friuli Venezia Giulia (FVG) region (46°02′19.08″N, 12°57′33.66″E), Italy. Fruits were selected apparently healthy, homogeneous in size, and free of lesions and stored at 0°C for 5 days with 92% RH until use.

Botrytis cinerea (Bc13), C. luteo-olivacea (Cad21), and P. synxantha (117-2b) belonged to the mycological collection of University of Udine—Di4A. Cadophora luteo-olivacea Cad21 was isolated from the tissue of symptomatic kiwifruit and molecularly characterized (Di Francesco et al., 2023). Pseudomonas synxantha 117-2b was originally isolated from the surface of kiwifruits and subsequently molecularly characterized (Di Lenarda et al., 2010). Botrytis cinerea Bc13 was isolated from infected kiwifruit, purified, and morphologically characterized. The fungal pathogens were cultivated on potato dextrose agar (PDA, 39 g L⁻¹) (Oxoid, UK) at 25°C, 1 and 2 weeks before the experiments, for B. cinerea and C. luteo-olivacea, respectively. Bacillus pumilus QST2808, the active component of the commercial product Sonata^®, was purchased from the NRRL (Northern Regional Research Laboratory, IL, USA) since its effectiveness against the target pathogens was tested in our previous study (Di Francesco et al., 2023) and it was used as a positive control in the present work. For the assays, both bacterial strains were grown on nutrient agar (NA, 13 g L⁻¹) (Oxoid, UK) at 25°C and the cells from 2-day-old NA plates were suspended in potassium phosphate buffer (PPB, 70 mL of 0.2 M KH₂PO₄, 30 mL of 0.2 M K₂HPO₄, and 300 mL of deionized water v/v/v, pH 6.5) and adjusted to the final concentration of 1×10⁸ cells mL⁻¹.

To verify P. synxantha 117-2b (Di Francesco et al., 2023) VOCs’ efficacy against C. luteo-olivacea (Cad21) and B. cinerea (Bc13) mycelial growth, a double Petri dish assay was performed as reported by Di Francesco et al. (2019). Plates were incubated at 25°C under dark conditions for 14 and 5 days for C. luteo-olivacea and B. cinerea, respectively. Bacillus pumilus QST2808 (1×10⁸ cells mL⁻¹) was inoculated as reported above and used as a positive control. NA plates inoculated with 100 μL of sterile distilled water (SDW) represented the negative control. The sample unit of each treatment was represented by five replicates and the experiment was performed twice. The inhibition rate of mycelial growth was calculated using the equation (Dennis and Webster, 1971):

where d1 and d2 are the control colony and the treated colony diameters, respectively.

To identify VOCs produced by P. synxantha (117-2b) (Di Francesco et al., 2023), sterile glass vials containing 10 mL of NA were inoculated with 50 μL of P. synxantha suspension (1×10⁸ cells mL⁻¹). Vials were incubated at 25°C for 2 days and then the VOCs were extracted and analyzed by a gas chromatograph (7890B Agilent Technologies, Santa Clara, USA) coupled with a mass spectrometer detector (5977A, Agilent Technologies, Santa Clara, USA) equipped with an autosampler for SPME automatic injections (PAL RSI 85, CTC). An SPME fiber with 50/30 μm DVB/CAR/PDMS coating was selected and conditioned before use. The method used for the extraction and absorption of volatiles provided an incubation of the vial at 40°C for 5 min, followed by fiber exposure of 30 min at the same temperature. For the desorption of the volatiles, the fiber was introduced into the injector at 250°C for 5 min. The separation of analytes was performed on a 60-m-long fused silica capillary column, with an internal diameter of 0.50 mm and a polar stationary phase (DB-WAX, Agilent Technologies, Santa Clara, USA) with a film thickness of 0.5 µm. The GC oven temperature program was as follows: 40°C for 3 min, raised from 40°C to 220°C by a constant ramp of 4°C min⁻¹; 220°C for 1 min, raised from 220°C to 250°C at 10°C min⁻¹; and 250°C for 1 min. Helium was used as carrier gas with a constant column flow rate of 1.5 mL min⁻¹. The transfer line temperature was maintained constant at 250°C. Ion source and quadrupole were set at 230°C and 150°C, respectively. Upon exiting the column, compounds were ionized via electron impact at 70 eV and detected with a quadrupole mass spectrometer in the range of a mass/charge ratio (m/z) from 35 to 500. Chromatograms were processed and analyzed by Mass Hunter (Agilent, USA). Compound identification was achieved by comparing the spectra with the NIST Standard Reference Database (NIST2014) and by linear retention indices (LRI) calculated under the same chromatographic conditions, injecting C7–C30 n‐alkane series (Supelco, Milan, Italy). VOC content was reported as absolute peak area (Aprea et al., 2012). The sample unit consisted of four replicates.

Two synthetic compounds (1-butanol, 3-methyl and 1-nonene) (Merck, Germany) identified as the main VOCs produced by P. synxantha 117-2b were selected and singularly used. For each pure compound, different concentrations (15, 25, and 50 µL, corresponding to 0.34, 0.56, and 1.12 µL mL⁻¹ headspace, respectively) were settled on sterile filter paper (Whatman^®, 90 mm Ø), previously placed inside the Petri dishes (90 mm Ø) containing PDA inoculated with fungal plugs (6 mm Ø) of each pathogen. The dishes were immediately double-sealed with Parafilm^® and incubated at 25°C for 14 and 5 days, respectively, for C. luteo-olivacea and B. cinerea. The effect of each compound was detected by measuring the diameters of each fungal colony (Youssef and Hashim, 2019). The control was represented by plates containing un-wet sterile filter papers. Each VOC concentration was represented by three plates per pathogen. The experiment was performed twice.

Biofumigation assay was conducted to verify the effect of volatile compounds produced by the BCAs in increasing the fruit resistance to gray mold and skin-pitting during the cold storage. In vivo assay was conducted by using sterile plastic boxes (29 × 18 × 10 cm, L × W × H) containing 150 mL of NA at the bottom. Six hundred microliters of P. synxantha 117-2b suspension (1×10⁸ cells mL⁻¹) was spread on the medium. The boxes were immediately double-sealed with Parafilm^® and incubated at 25°C for 2 days. Kiwifruits were surface-sterilized for 1 min with sodium hypochlorite (0.1% v/v), rinsed with distilled water for 1 min, and air-dried at room temperature. Fruits were artificially wounded (2 × 2 × 2 mm) at the equatorial part with a sterile nail and inoculated with 20 μL of conidial suspension of C. luteo-olivacea (Cad21) (1×10⁵ conidia mL⁻¹) or B. cinerea (Bc13) (1×10⁵ conidia mL⁻¹). Once dried, fruits were placed on sterile grills inside the boxes to be separated from the medium. Boxes, incubated at 15°C and 85% RH, were exposed to bacterial VOCs for 96 h. After the biofumigation, treated fruits were removed from the boxes and stored at 0°C for 3 months and 1 month for C. luteo-olivacea and B. cinerea, respectively. As a positive control, B. pumilus QST2808 was used as a biofumigant treatment. Boxes containing just NA were considered the negative control. Each sample unit was represented by three boxes (containing eight kiwifruit) for each pathogen and treatment.

Data were statistically handled by one-way analysis of variance (ANOVA) using MiniTab 17. Mean statistical comparison of the effect of VOCs on the fungal colony diameter and on the incidence and disease severity was carried out by using Tukey’s HSD test (α = 0.05), as reported by Abo-Elyousr et al. (2020). For gene expression analysis, comparison of the means within the control and treated samples was performed using the Student’s t-test (α ≤ 0.05), and that between different time points for control and treated samples was carried out using Tukey’s HSD test (*α ≤ 0.05, ** α ≤ 0.01).

The double Petri dish assay displayed the potential in vitro efficacy of VOCs produced by P. synxantha 117-2b against C. luteo-olivacea and B. cinerea. Figure 1 reports the results of the experiment that showed a significant inhibitory effect of P. synxantha 117-2b and B. pumilus QST2808 VOCs against pathogens’ mycelial diameter. Pseudomonas synxantha 117-2b showed an inhibition of 56% and 42.8%, with respect to the relative control, against C. luteo-olivacea and B. cinerea, respectively. Likewise, B. pumilus QST2808 showed a significant inhibition, with respect to the relative control, against both fungal pathogens, of 52% and 44.9%, respectively.

The volatile profile of P. synxantha 117-2b was investigated on NA medium through SPME–GC/MS ( Supplementary Table S2 ). VOCs emitted by the bacterial antagonist exhibited a miscellaneous range of chemical compound classes, particularly alcohols (1-butanol, 3-methyl and phenylethyl alcohol), alkenes (1-nonene), and ketones (2-pentanone) ( Table 1 ). VOCs that showed a higher absolute area (AA) were 1-nonene and 1-butanol, 3-methyl.

Among the detected VOCs, those with a higher AA such as 1-butanol, 3-methyl and 1-nonene were selected for assessing their effect on C. luteo-olivacea ( Figure 2A ) and B. cinerea ( Figure 2B ) colony growth. The compound 1-butanol, 3-methyl became more effective at 1.12 µL mL⁻¹ headspace, showing an inhibition by 55.3% and 100% of C. luteo-olivacea and B. cinerea colony growth compared to the control, respectively ( Figure 2C ). The alkene 1-nonene did not exhibit any significant inhibitory effect against both fungal pathogens.

VOCs produced by P. synxantha 117-2b showed a significant inhibitory effect against the targeted pathogens ( Supplementary Figures 1A, B ). Regarding C. luteo-olivacea, results were reported as severity (mm) of the disease ( Figure 3A ), and for B. cinerea, results were reported as percentage (%) of the disease incidence ( Figure 3B ), after 3 months and 1 month, respectively, of cold storage after 96 h of bacterial VOC biofumigation. The symptoms caused by the two pathogens were differently detected. The diameter of gray mold lesions was not easily detectable on kiwifruit, and conversely for the skin-pitting. Moreover, as regards C. luteo-olivacea, the incidence inhibition of the symptoms on fruits was not reported as not statistically different from the control. Regarding the skin-pitting symptoms, P. synxantha VOCs reduced the disease severity by 28.5% with respect to the control. For B. cinerea, the displayed inhibition of the disease incidence was 66.6% with respect to the control. VOCs produced by the reference strain B. pumilus QST2808, showed, in the case of skin-pitting, the same efficacy of the tested BCA, and conversely significantly higher in the case of gray mold incidence (90.3%).

With the aim of determining the expression pattern of defense-related genes activated in kiwifruit as a response to P. synxantha 117-2b VOC exposition, the expression levels of several genes ( Supplementary Table S1 ) at six different time points (3, 6, 12, 24, 48, and 96 h) after biofumigation were analyzed. Results revealed that the bacterial VOCs, in general, promoted the upregulation of the target genes ( Figure 4 ). Specifically, the results for CAT expression levels ( Figure 4A ) revealed a significant increase in its expression at 3, 24, 48, and 96 h from the treatment with bacterial VOCs, compared to the control. Notably, the highest expression level was observed at 48 h post-treatment, which was 1.2-fold higher than the expression in the treated sample compared to the control. However, no significant differences in CAT gene expression compared to the control were observed after 6 and 12 h from the treatment. Moreover, a significant increase in the expression level of the CAT gene was also observed in the control samples from 24 to 96 h. Similarly, the expression of CHI gene ( Figure 4B ) started to increase significantly at 3 h from the treatment and continued to rise for 12 h. However, after the 12-h time point, the expression levels of the CHI gene were sustained, although a significantly higher expression occurred at 24 and 96 h, compared to the control. In the control fruit, the CHI gene showed a higher expression after 48 and 96 h compared to the earlier time points. In the case of GLU ( Figure 4C ), a significant increase in the expression was observed at 6 h from the treatment compared to the untreated fruit, increasing its expression for 96 h. No significant increase was detected in the control samples over time. The results of the NPR gene expression level ( Figure 4D ) revealed a significant increase along with the treatment time points and the maximum induction took place after 96 h from the treatment compared to the control. Conversely, no significant differences in NPR gene expression were observed at 6 and 12 h with respect to the control. The exposition of fruits to bacterial VOCs caused a significant increase in POD gene expression ( Figure 4E ) at 6 h, 24 h, 48 h, and 96 h compared to the control. Specifically, a notable upregulation occurred at 96 h, being 2.3-fold higher compared to the control. However, if compared to the control, no significant differences in POD expression at 3 and 12 h were detected. Regarding the SOD gene ( Figure 4F ), the treatment induced its expression levels after 12 h, and this pattern was sustained for 96 h. SOD gene expression increased in the control samples at 96 h compared to 0-h time point. The fruit PAL gene ( Figure 4G ) was notably upregulated after bacterial VOC exposition at 24, 48, and 96 h compared to the earlier time points. In contrast, no significant differences in PAL expression were observed in control samples, except for 96 h, in which gene expression levels were upregulated compared to the other time points.