Xylella fastidiosa subsp. pauca strain A0PT1, isolated from olive trees affected by the olive quick decline syndrome (OQDS) in southern Italy, Apulia region, was used in all experiments. The strain was stored in Pierce disease broth (PD2) (Davis et al., 1980) supplemented with glycerol (50%) and maintained at −80°C. When needed, aliquots were cultured in buffered charcoal yeast extract (BCYE) (Wells et al., 1981) agar plates and grown at 28°C for 10 to15 days. Cell suspensions were prepared using phosphate buffered saline (PBS, pH 7.4, 0.01 M) and adjusted to 10⁸ CFU/mL (OD600 ≅ 0.32) (El Handi et al., 2022). L. lactis subsp. lactis ATCC 11454 was grown on YPGA plates (5.0 g/L yeast extract, 5.0 g/L peptone, 10.0 g/L glucose, and 15.0 g/L agar) for 2 days at 28°C.

Nisin (ITSISLCTPGCKTGALMGCNMKTATCHCSIHVSK), extracted and purified from L. lactis, was purchased from Sigma-Aldrich (Merck KGaA, Darmstadt, Germany). According to the manufacturer, the formulation contains 2.5% (w/w) pure nisin with potency ≥900 IU/mg. Lyophilized nisin was solubilized in sterile Milli-Q water to a stock concentration of 12.5 mg/mL and filter sterilized through a 0.22 μm nylon Acrodisc^® syringe filter (Merck, Rome, Italy).

The antagonistic effect of L. lactis subsp. lactis ATCC 11454 against Xf subsp. pauca was investigated as follows: four drops of Xf suspension (10⁸ CFU/mL), each containing 30 μL, were positioned at the upper portion of the BCYE agar plates, with approximately 1.5 cm between them. These drops were allowed to slowly descend to the opposite side, forming four parallel rows of Xf cultures. Subsequently, two drops (10 μL each) of Lactococcus lactis suspension (10⁸ CFU/mL), prepared in PBS, were administered to the center of the rows of Xf cultures after drying under the laminar flow hood. PBS served as the negative control. Following an incubation period of 10 days at 28°C, the antagonistic activity of Lactococcus lactis was perceived as an inhibition zone of Xf growth, which was measured using a digital caliper. This experiment was conducted in triplicate for reliability.

Reversed-phase liquid chromatography (RPLC) coupled with electrospray (ESI) and tandem mass spectrometry (MS/MS) was used to explore the proteins involved in the inhibitory effect induced by L. lactis against Xf. For inhibition effect examination, two pieces of culture medium from the inhibition zone were cut and subjected to extraction with 200 μL of 10 mM Tri-HCl for an easy protein/peptide recovery. In addition, 100 μL of L. lactis suspension (10⁸ CFU/mL) was challenged by 100 μL of Xf suspension (10⁸ CFU/mL) in 2 mL of YPG liquid medium for 24 h at 28°C. After incubation, the mixture was passed through a 0.22 μm pore filter and subjected to RPLC-ESI-MS/MS.

For analysis, water, acetonitrile, and formic acid LC-MS grade were purchased from Merck (Milan, Italy). Five microliters of the extract or standard solutions were injected in the UHPLC Ultimate 3000 system (Dionex Thermo Fisher Scientific, Italy) coupled to HESI (heated electrospray ionization) interface and VelosPro mass spectrometer (Thermo Scientific, Waltham, MA, United States) equipped with a double linear trap mass analyzer. The reversed-phase (RP) chromatographic separations were accomplished at 40°C using a Phenomenex Aeris WIDEPORE 200 Å C18 column (250 × 2.1 mm, 3.6 μm) equipped with Phenomenex AJO 8783 WIDEPORE C18 (2 × 2.1 mm ID) security guard cartridge and a mobile phase based on H₂O (solvent A) and ACN (solvent B) both containing 0.1% of formic acid. Specifically, the gradient used during each chromatographic run, at a flow rate of 0.200 mL/min, was the following: 0–3 min linear from 5 to 20% solvent B; 3–5 min 20% solvent B; 5–12 min linear from 20 to 85% (v/v) solvent B; 12–16 min isocratic at 85% of solvent B; 16–17 min back to the initial composition, followed by 5 min equilibration time. Mass spectrometry analyses were carried out in full scan and in selected reaction monitored (SRM) transitions mode in positive polarity or in targeted MS/MS mode.

Given the unique characteristics of Xf (slow growing propriety), an initial in vitro screening was performed to determine an approximate initial concentration of nisin suitable for subsequent challenges to Xf cells. This preliminary screening was assessed by a contact test coupled with viable-quantitative PCR (v-qPCR) using the PMAxx^™ (Biotium, Rome, Italy) (Baró et al., 2020). PMAxx^™ is a photo-reactive dye that forms a covalent bond with disrupted DNA, rendering it un-amplifiable by PCR. This unique feature makes PMAxx^™ extremely useful in selective detection of bacterial cells with intact plasma membranes by qPCR. Briefly, 50 μL of Xf cells with OD at 600 nm of 0.32 measured by NanoDrop^™ One/OneC Microvolume UV–Vis Spectrophotometer (Thermo Fisher Scientific), were treated with 50 μL of nisin at 6, 3, 1.5, 1, 0.8, 0.6, 0.4, 0.2, 0.1, and 0.05 mg/mL. Consequently, series of dilutions of Xf ranging from OD = 0.32 to OD = 0.02 were employed as controls, and samples were incubated for 3 h at 28°C. After incubation, samples were treated with PMAxx^™ at a final concentration of 7.5 μM, incubated in the dark at room temperature for 8 min, and followed by a 15 min photoactivation step using the PMA-Lite^™ LED Photolysis Device. Genomic DNA of all samples was extracted following the CTAB protocol (Burbank and Ortega, 2018). For analysis, a TaqMan-based qPCR assay was carried out in a thermocycler apparatus (Bio-Rad CFX96, BioRad, Milan, Italy), utilizing specific primers Xf-F: 5′-CACGGCTGGTAACGGAAGA-3′ and Xf-R: 5′-GGGTTGCGTGGTGAAATCAAG-3′, along with the probe Xf-Prb: 5′-6FAM-TCGCATCCCGTGGCTCAGTCC-BHQ-1-3′ (Burbank and Ortega, 2018).

To corroborate the findings derived from the v-qPCR assay pertaining to the lethal concentrations of nisin against Xf, a spot assay was conducted as follows: Xf subsp. pauca was cultured at 28°C on BCYE agar plates for up to 10 days. Then, the cultures were suspended in PBS and 30 μL of bacterial suspension (OD₆₀₀ = 0.32) were mixed with 30 μL of nisin at 3, 1.5, 1, 0.8, and 0.6 mg/mL. Subsequently, the mixtures were spotted onto the surface of the BCYE agar plates and allowed to air-dry within the laminar flow hood. Negative controls consisted of PBS, while positive controls involved 30 μL of Xf mixed with 30 μL of PBS (Mourou et al., 2022). Growth inhibition was assessed 10 days post-incubation at 28°C, checked by the level of Xf growth. The experiment was conducted with three independent replicates.

The effect of the minimal lethal concentration (MLC) of nisin on Xf cells in liquid medium was assessed by quantifying the degree of turbidity at OD₆₀₀ in both nisin-treated and untreated Xf suspensions over 3 h. Briefly, a 50 μL aliquot of Xf suspension in PD2 (Pierce disease 2 media) was mixed with 50 μL of nisin at 0.6 mg/mL, resulting in a final turbidity of Xf with an OD₆₀₀ of 0.41. Untreated control involved 50 μL of Xf mixed with 50 μL of PBS with a final turbidity of 0.41 at OD₆₀₀. Bacterial turbidity was assessed by tracking the OD₆₀₀ using the NanoDrop^™ One/OneC Microvolume UV–Vis Spectrophotometer at 0 min, 1 h, 2 h, and 3 h of incubation at 28°C. The experiment was conducted twice, with three replicates for each treatment.

Fluorescence microscopy (FM) provides an accurate assessment methodology for the bacterial lysis process by enumerating the populations of intact and permeable cells at various time intervals. In this context, 100 μL Xf suspension (OD₆₀₀ = 0.32) was mixed with 100 μL of nisin at its MLC (0.6 mg/mL) and incubated at room temperature for 3 h. Untreated control consisted of bacterial suspension treated with sterile distilled water. To assess the viability of Xf cells exposed to nisin, the LIVE/DEAD^® BacLight^™ viability kit (provided by Molecular Probes) was employed. This kit contains two nucleic acid dyes, SYTO 9 and propidium iodide (PI), which enable the differentiation of live cells with intact plasma membranes (visualized in the green channel) from bacteria that have succumbed to membrane compromise due to nisin activity (visible in the red channel). Photomicrographs were captured at 15 min, 30 min, 1 h, 2 h and 3 h post-incubation, using a Nikon E800 microscope equipped with fluorescein isothiocyanate (480/30 excitation filter, DM505 dichroic mirror, 535/40 emission filter) and tetramethyl rhodamine isothiocyanate (546/10 excitation filter, DM575 dichroic mirror, 590 emission filter) fluorescence filter sets.

To scrutinize the antibacterial properties of nisin against Xf cells, a suspension of Xf (OD₆₀₀ = 0.32) was exposed to nisin at 0.6 mg/mL for a duration of 3 h at room temperature. Representative images were taken after 3 h via transmission electron microscopy (TEM) (FEI MORGAGNI 282D, United States) using the dip method. Briefly, carbon-coated copper/rhodium grids were immersed in both the nisin-treated and untreated bacterial suspensions for 5 min, followed by a rinse with 200 μL of distilled water. To achieve negative staining, the grids were floated on 200 μL of a 0.5% w/v UA-zero EM stain solution (Agar-Scientific Ltd., Stansted, United Kingdom), and observed under an accelerating voltage of 80 kV.

The nisin’s spectrum of action at MLC (0.6 mg/mL) was assessed by targeting various pathogenic and beneficial plant bacteria using the spot assay. Therefore, 16 phytopathogenic bacterial isolates, including three different isolates of Xf subsp. pauca, together with one epiphytic (Pantoea agglomerans) and three endophytic bacteria (Bacillus subtilis, Paenibacillus rigui, and Bacillus pumilus) isolated from olive trees, which are reported in the literature to exhibit antibacterial activity against Xf subsp. pauca (Mourou et al., 2022), were examined (Table 1). For Xf isolates, spot tests were carried out on BCYE agar plates as described earlier. For other bacteria, spot tests were performed on YPGA plates as follows: bacterial isolates were cultured at 28°C on YPGA plates for up to 2 days. Then, the cultures were suspended in sterile distilled water and 200 μL of bacterial suspension (OD₆₀₀ = 0.2) were mixed with 6 mL of YPG soft agar (i.e., YPG supplemented with 0.7% agar), poured into YPGA plates, and allowed to dry. Subsequently, drops of 10 μL of nisin were spotted onto the surface of the plates. Spots were dried at room temperature and the plates cultured for up to 2 days at 28°C. The presence of a clear zone in the spot area was indicative of bacterial susceptibility to nisin. The experiment was replicated three times for validation.

To explore the antagonistic effect of nisin on Xf within plant tissues, an in-planta assay was conducted. This involved the Xf infection and the application of nisin to the stems of Nicotiana benthamiana (N. benthamiana), which was validated as a model plant host for Xf (Baró et al., 2022). Briefly, one-month-old N. benthamiana plants were inoculated using a 0.1 mL insulin syringe, with the needle penetrating approximately half the plant stem diameter to access the vascular system. Two inoculations (25 μL each) of Xf suspension (OD₆₀₀ = 0.32), prepared in PBS, were administered on the same side of the stem in a section of 4 cm at around 10 cm above the soil level. Two applications of 50 μL each of the nisin at 6 mg/mL (100 μL of nisin/plant, corresponding to 0.6 mg of nisin per plant) were performed at the same side of the stem in section of 8 cm at around 8 cm above the soil level (Figure 1). Preventive treatment with nisin was performed 24 h prior to Xf inoculation, whereas curative treatment was carried out 24 h after Xf inoculation. Each treatment consisted of 10 plants, including both positive and negative control groups, and inoculated plants were maintained in a controlled quarantine laboratory environment at 25°C. Two separate experiments were conducted to ensure the reliability of our findings.

Over a period of 74 days post inoculation (dpi), inoculated plants were visually inspected for the development of characteristic symptoms of Xf (i.e., leaf scorch), and movement of Xf within the plant was ascertained through TaqMan qPCR assay applied on genomic DNA extracted from leaves above the inoculation points (Figure 1). At the end of the experiment, leaves exhibiting suspected symptoms were analyzed for the presence of Xf using TaqMan qPCR assay as described earlier. Briefly, 1 g of leaf tissue was grinded with 2 mL of CTAB buffer, followed by heating at 65°C for 30 min. The plant extract was then centrifuged at 13,000 g for 15 min, and the resulting supernatant was washed with chloroform and precipitated in cold isopropyl alcohol. The resulting pellet, which contained the total DNA, was diluted in 100 μL of sterile water. Subsequently, 50 ng of this DNA was used for the TaqMan qPCR.

To explore the stability and translocation of nisin throughout the plant parts, one-month-old N. benthamiana plants (3 replicates) were inoculated as described in in planta section with 100 μL of nisin at 6 mg/mL. The plants were divided into five sections, with nisin inoculation carried out in the stem of section 3 (Figure 2). After 24 h, 48 h, 6 days, and 9 days, stems and leaves of each section were collected separately, ground in a mortar with 1 mL of sterile Milli-Q water, and then transferred to a centrifuge tube. After centrifugation at 3,000 g for 5 min at 5°C, the supernatant was filtered through a 0.22 μm nylon filter for RPLC-ESI-MS/MS analysis, using the same instrumentation as described above, with the following modifications. The gradient used for nisin quantification was: 0–2 min at 20% solvent B; 2–6 min linear from 20 to 85% (v/v) of solvent B; 6–9 min isocratic at 85% of solvent B; 9–10 min back to the initial composition, followed by 5 min equilibration time. For MS acquisitions, parameters were set as follows: sheath gas flow rate 35 arbitrary units (a.u.); auxiliary gas flow rate 5 a.u.; spray voltage 3.5 kV; capillary temperature 320°C; S-Lens radio frequency level 60 a.u. Acquisitions were performed in positive ion mode within the m/z interval 200–2000 and collisionally induced ionization (CID) energy was set as 35% a.u. for MS/MS ions fragmentation. Nisin was detected as tetra-charged [M + 4H]⁴⁺ or tri-charged [M + 3H]³⁺ ion, respectively, at m/z 839.3 and m/z 1118.7. Data were acquired in positive ionization with two transitions: 839.3 → 1080.7 and 1118.7 → 1451.7. The nisin standard A from Lactococcus lactis was used for calibration directly in plant sap to compensate the matrix effect exploring concentration from 6 μg/mL down to 0.06 ng/mL. The linearity of the method was determined by analyzing the standard plots associated with an 8-point standard calibration curve and the method was found to be linear in 3 orders of magnitude in the range 60–0.06 ng/mL. The LOD was calculated as 3 × the SD of the intercept or slope of the calibration curve, and LOQ was calculated as 10 × the SD of the intercept or slope of the calibration curve. The software Xcalibur 2.2 SP1.48 (Thermo Scientific) was used for raw RPLC-MS data.

The inhibitory activity of L. lactis against Xf growth was strong with an inhibition zone of 33.2 mm, highlighting the efficacy of L. lactis in restraining the Xf growth (Figure 3). In addition, L. lactis exhibited a lateral inhibition expansion, hindering the growth of Xf in the adjacent PBS-treated row C (Figure 3), demonstrating a high antagonism between L. lactis and Xf.

The RPLC-ESI-MS/MS analyses conducted on the inhibition zone, to investigate about the nature of compounds involved in the inhibitory activity of L. lactis against Xf, revealed the existence of nine peptides reported in Table 2. These peptides derive from different protein sources associated to plasma membrane or trans-membrane proteins of L. lactis. Some of these proteins are related to ATP-binding or hydrolysis activity, other to sugar, amino-sugar, and hexosamine metabolism (Table 2). The analyses could not reveal the presence of nisin in the inhibition zones, but the amino acid sequence of the transmembrane nisin-resistance protein was detected, which is a normally overexpressed by L. lactis to protect itself from the inhibitory effect of nisin. However, the RPLC-ESI-MS/MS has detected the presence of nisin, in a concentration of 345 ng/mL in the liquid medium after 24 h of incubation. This finding was taken as empirical evidence on the involvement of nisin in the inhibition of Xf growth. Based on this, we proceeded forward in the experimental design using pure nisin A in our research endeavors.

During this investigation, the antimicrobial efficacy of nisin against Xf was systematically assessed across different concentrations (6, 3, 1.5, 1, 0.8, 0.6, 0.4, 0.2, 0.1, 0.05 mg/mL). Employing v-qPCR as an initial in vitro screening method (Figure 4), the study aimed to determine the MLC of nisin for Xf. The obtained results were subsequently validated through a spot assay (Figure 5). According to the v-qPCR, nisin exhibited exceptional potency, effectively lysing Xf at different concentrations, with 0.6 mg/mL as the MLC. The spot assay results also supported this finding, showing the absence of Xf growth at concentration of nisin above 0.6 mg/mL.

The response of Xf to nisin at 0.6 mg/mL was assessed during a 3 h incubation period. In summary, nisin showed almost complete reductive activity against Xf cells within the first hour of exposure, with a reduction of up to 90% (Figure 6). Whereas, this reductive activity endured for the entire 3 h post-contact period, resulting in the complete elimination of Xf cells (Figure 6). The efficacy of nisin against Xf cells was also investigated and confirmed by fluorescence microscopy (FM) using two dyes capable of distinguishing between live cells with intact plasma membranes (green channel) and dead bacteria with membranes compromised by nisin (red channel). The micrographs obtained corroborate the results of the turbidity assay, confirming once again that nisin exhibits significant and rapid destructive activity on Xf, as evidenced by the intense red channels, indicative of substantial lysing of Xf cells (Figure 7).

Analyzing the impact of nisin on Xf cells using TEM, micrographs disclosed remarkable alterations at the structural level, leading to fragmentation of the cell wall membrane (Figure 8G). Specifically, the nisin-treated Xf cells presented cytoplasmic condensation, fragmented cell walls, disrupted internal morphology, and alterations in the outer membrane structure; all resulting in destruction of both inner and outer bacterial membranes (Figure 8). Moreover, membranes of Xf-nisin treated cells showed serial pores and protrusions, accompanied by condensed cytoplasmic outflow from the outer membrane in the form of vesicles (Figure 8H).

For a sustainable use of nisin in an integrated pest management strategy, the antibacterial activity of nisin was tested against four Xf antagonistic bacteria (Table 1). The results showed that nisin had no antibacterial effect on them, suggesting its potential use, in combination with these antagonistic bacteria as part of integrated biocontrol approach against Xf. Similarly, nisin did not showed any effect on the 16 phytopathogenic bacteria tested, except for strains of Xf subsp. pauca (Table 1). This emphasizes the specific antibacterial activity of nisin at its MLC against Xf subsp. pauca among the tested bacteria.

The antibacterial effect of nisin against Xf in plant was evaluated in N. benthamiana plants treated with nisin before Xf infection (preventive) and after Xf infection (curative). Results of TaqMan-based qPCR performed 20 dpi on leaves collected above the inoculation sites, showed that Xf translocated only within the nisin-untreated infected plants (positive controls) (data not shown). By 40 dpi, the positive controls displayed typical Xf symptoms, i.e., leaf scorch (Figure 9B). In contrast, infected plants treated preventively and curatively with nisin demonstrated no typical symptoms of Xf (Figures 9C,D), indicating that nisin effectively countered the pathogen and prevented its impact on tobacco plants. At the end of the experiment (74 dpi), leaves with suspicious symptoms were analyzed for the presence of Xf using TaqMan-based qPCR assays, which showed the absence of the bacterium in all treatments; except leaves of positive controls (Figure 10).

To assess the stability and the translocation of nisin within N. benthamiana, RPLC-ESI-MS/MS analyses were carried out on sap recovered from different plant sections of three plants, as described above. Typically, low collisional energy and selected reaction monitoring (SRM) acquisition mode are employed in proteomic studies to enhance sensitivity and were thus also applied in the present case. When multi-charged molecules are analyzed, the transitions to a higher m/z ratio are considered highly. Therefore, the two most intense fragment ions with a higher m/z ratio than the precursor ions were chosen to develop the SRM method for nisin A. The highly specific transitions from the precursor ions [M + 4H]⁴⁺ and [M + 3H]³⁺ to higher m/z ratios were identified; the most intense one 839.3 → 1080.7 was used as quantifier while the other 1118.7 → 1451.7 was used as a confirmatory transition. As shown in Figure 11, a peak in the chromatogram appeared at a retention time of 7.8 min both in the standard spiked solution (A), and in a sample derived from stem (section 4) after 24 h (B). We also found that the tandem mass spectrum obtained in the stem of section 4 after 24 h (Figure 11C) had product ions identical to literature (Lavanant et al., 1998) and to those retrieved for standard solution (data not shown).

We were able to qualitatively detect nisin A in stems, roots, and leaves; but to quantify the residual intact of nisin, we constructed an appropriate calibration curve spiking nisin A in blank plant sap. The linearity of the method was assessed at concentrations ranging from 0.06 to 60 ng/mL (5 levels) with a correlation coefficient of 0.9983 (Figure 12). The estimated LOD and LOQ values for nisin A were 2.1 and 7.1 ng/mL, respectively. Most of the values were below the LOD, some residual values above the LOD (2–4 ng/mL) were recovered in leaves (sections 3, 4 and 5) up to 9 days while the highest reliable concentration was found after 24 h in stem just above the injection point (section 4, 23.1 ng/mL). This experiment provides preliminary results on the stability of nisin in the plant, which can remain intact for up to 9 days after application, the period studied in our experiment.