In October 2024, cauliflower (Brassica oleracea var. botrytis) and broccoli (Brassica oleracea var. italica) samples were collected from a weekly open-air market located at 41.04378° N, 16.88420° E in Valenzano (BA, Italy) for the purpose of isolating antagonistic bacteria against Xfp and Xanthomonas campestris pv. campestris (Xcc). The choice of cauliflower and broccoli, both susceptible to Xcc, which is phylogenetically related to Xf, was intended to enhance the likelihood of isolating antagonistic bacteria with potential activity against Xf. Stalk tissues of three plants from each broccoli and cauliflower species were used for bacterial isolation. Briefly, 1 g of plant tissue was macerated in 4 mL of sterile distilled water, followed by six 10-fold serial dilutions. Aliquots of 50 μL from each dilution were plated onto yeast extract peptone glucose agar (YPGA) (5.0 g/L yeast extract, 5.0 g/L peptone, 10.0 g/L glucose, and 15.0 g/L agar) for bacterial isolation. Plates were incubated at 28°C for 24 h, after which morphologically distinct colonies were purified and stored at −80°C in 25% glycerol prepared in YPG broth for further study.

Bacterial isolates obtained from cauliflower and broccoli plant samples were initially subjected to antagonistic screening against Xcc strain CFBP 1710 and Xfp strain A0PT1 (Sequence Type, ST53) using dual-plate confrontation assays, following the methodology described by Sabri et al. (2025). One isolate, subsequently identified as D. fangzhongdai strain C7, was prioritized for whole-genome sequencing (WGS) due to its strong antagonistic activity against Xf on BCYE agar plates (Wells et al., 1981), to determine its taxonomic identity. The antagonistic effect of strain C7 against Xfp was assessed as follows: four 30 μL drops of Xfp suspension (OD₆₀₀ = 0.32), prepared in sterile water, were placed along the upper section of BCYE agar plates, with approximately 1.5 cm spacing between drops. These drops were allowed to slowly descend to the opposite side, forming four parallel rows of Xfp cultures. After the drops had dried under a laminar flow hood, 10 μL of C7 suspension (OD₆₀₀ = 0.2), prepared in sterile water from an overnight culture, was deposited at the center of one Xfp row. Sterile water served as negative control and the plates were incubated for 10 days at 28°C. Genomic DNA of strain C7 was extracted using the cetyltrimethylammonium bromide (CTAB) method, following the protocol described by Wilson (2001). The WGS was performed by Eurofins Genomics (Ebersberg, Germany), using Illumina HiSeq 2,500 paired-end sequencing technology, with an average read length of 150 bp. The reads were quality checked and trimmed, and de novo assembled using the Geneious Prime 2025.1.1 software (Biomatters, Ltd., San Diego, CA, USA). The complete genome of strain C7 was annotated using Prokka v1.14.6¹. Clustered Regularly Interspaced Palindromic Short Repeats (CRISPRs) were identified using CRISPRCas++², and prophage genes were predicted using the online tool PHASTER³. Additionally, genomic islands (GIs) were predicted using IslandPath-DIMOB, as implemented in IslandViewer 4.⁴

To determine the phylogenetic position of strain C7, MLSA was performed using four housekeeping genes (atpD, gyrB, infB, and rpoB) as described by Chen et al. (2024). Briefly, DNA sequences of the four housekeeping genes were aligned using the Geneious Alignment algorithm within the Geneious software platform. Concatenated nucleotide sequences of the four genes were used to construct a phylogenetic tree using the maximum likelihood method derived from MEGA 11 software, with 1,000 bootstrap replicates to assess branch support. Sequences from representative Dickeya species were included for comparison.

The strong antagonistic activity of C7 against Xfp, observed under both direct contact and physically separated conditions using a cut agar barrier, suggested the involvement of volatile antimicrobial compounds. Therefore, the production of VOCs by C7 was investigated using headspace solid-phase microextraction and gas chromatography coupled with mass spectrometry (SPME/GC–MS). Briefly, 5 mL of YPG broth were placed in SPME glass vials of 20 mL and inoculated with 200 μL of C7 suspension (10⁸ CFU/mL). The vials were incubated at 28°C for 24, 48, and 72 h. At each time point, three replicates per incubation period, along with a control containing only YPG broth, were placed on a robotic autosampler Pal Combe XT. The SPME extraction was performed at 50°C constant temperature, 2 min of equilibration and agitation, 15 min of extraction by exposing a SPME fiber DVB/CAR/PDMS, 50 μm (DVB layer), 30 μm (CAR/PDMS layer) (Supelco) in the headspace. The VOCs were then desorbed in a split injector of a GC/MS system made of a Clarus 680 GC equipped with an VF-WAXms fused silica capillary column (20 m x 0.1 mm and 0.1 μm film thickness) and interfaced with a single quadrupole mass spectrometer Clarus SQ8C (Perkin Elmer). Mass spectra of target compounds were obtained by electron impact ionization with standardized ionization energy of 70 eV. Helium 5.5 was used as a carrier gas at a constant flow rate of 0.37 mL/min. The injection was performed in splitless (closed split valve for 0.1 min) at 230°C. The oven temperature was programmed from 60°C to 240°C at 20°C/min, then hold to 240°C for 2 min. Transfer line and source temperatures were set at 250°C. Data were collected in full scan mode in the range 33–300 m/z. Qualitative results include compound identification and area percentage of related peak in the total ions chromatogram. Compound identification was performed by mass spectra (MS) search in NIST and Wiley databases. The retention index for the identified compounds was calculated according to van den Dool and Kratz (1963) and, to the best of our knowledge, for the first time using the innovative cross linked, polar stationary phase of the chromatographic column, with fast chromatography. Thus, the reported information can be useful in the future research for identification purposes.

The pathogenicity of D. fangzhongdai strain C7 was tested on cauliflower, broccoli, and Havana tobacco (Nicotiana tabacum) plants, as well as on potato tubers (Solanum tuberosum cv. Monalisa), pear fruits (Pyrus communis L. cv. Comice), and onion bulbs (Allium cepa cv. Belendina). For cauliflower, broccoli, and Havana tobacco plants, 50 μL of C7 suspension (10⁸ CFU/mL) were injected into the stems using a sterile syringe. Ten one-month-old plants from each species were inoculated, and an additional 10 plants injected with sterile water served as negative controls. Plants were maintained in a controlled quarantine laboratory environment at 26°C for 22 days. In addition, 10 potato tubers, 10 pear fruits, and 10 onion bulbs were surface disinfected by immersion in 70% ethanol for 3 min, followed by rinsing with sterile distilled water. Wounds approximately 2 mm in diameter were created using a sterile scalpel, and 50 μL of C7 suspension (10⁸ CFU/mL) was inoculated into each wound. The inoculated and control samples were subsequently placed in sterile, high-humidity chambers and incubated at 28°C for 5 days. All experiments were conducted in duplicate.

To evaluate the persistence of C7 within cauliflower and broccoli plants 22 days post-inoculation (dpi), stems of three plants from each species were separately ground in 2 mL of sterile water and centrifuged at 1,000 g for 5 min to remove plant debris. The resultant supernatant was diluted six times using a 10-fold serial dilution, and 50 μL of each dilution was plated onto YPGA plates. The plates were incubated at 28°C for 24 h and C7-like colonies were purified on YPGA plates and subjected to PCR testing using D. fangzhongdai-specific primers (DF-F:5′-GTTGGCCTTCGAACTCGGTA-3′; DF-R:5′-TCAAACTGCGCCGGATAACT-3′), which were designed in this study using Geneious software. The PCR cycles consisted of an initial denaturation step at 94°C for 5 min; 35 cycles of 94°C for 30 s, 60°C for 30 s, and 72°C for 30 s; and a final elongation step at 72°C for 5 min.

The data related to the VOCs determination obtained by SPME-GC/MS were analyzed using SPSS v22 (IBM). The mean values of the area percentage related to each identified compound together with the total peaks area amount at different incubation times were compared using one-way ANOVA followed by Tukey’s post hoc test, with a significance threshold of p < 0.05.

The complete genome of D. fangzhongdai strain C7 was sequenced using Illumina HiSeq technology, yielding 259,846 high-quality reads, corresponding to approximately 29,844 kb of raw sequence data. De novo assembly was performed using Geneious assembler, and the generated contigs were mapped against eight related reference genomes of D. fangzhongdai (accession no: CP020872.1, CP080400.1, CP087226.1, CP092458.1, CP092460.1, CP094338.1, CP107067.1, and CP109768.1). The assembled genome revealed a single circular chromosome with a total length of 5,045,655 bp and a GC content of 56.9%. The average nucleotide identity (ANI) value between C7 and other D. fangzhongdai strains was over 98.0%, exceeding the 95% species demarcation threshold, and thereby confirming its taxonomic placement within the species. The annotation of the genome predicted 4,311 coding sequences (CDS), 25 rRNA genes, 81 tRNA genes, and one tmRNA gene (Figure 1). In addition, the genome of C7 harbors 9 CRISPR arrays, two distinct CRISPR-associated (Cas) gene clusters, 10 GIs, and 12 prophage gene clusters. The complete genome sequence of C7 has been deposited in the GenBank under the acc. no: CP193488. Furthermore, the phylogenetic analysis placed strain C7 close to strain B16 from Slovenia (acc. no: NZ_CP087226), indicating a high level of genetic similarity and suggesting a close evolutionary relationship between the two strains (Figure 2).

Colony morphology of strain C7, grown on YPGA plates for 24 h, was characterized by a circular shape with a smooth glistening surface, cream to off-white in color, and a convex elevation. Colonies reached 2–4 mm in diameter with entire to slightly undulate margins (Figure 3A). After 72 h of incubation, pigmentation was observed, with colonies showing a pale yellow to cream coloration and a distinct bright orange central zone developed. The average colony diameter ranged from approximately 1 to 2.5 cm (Figure 3B). Remarkably, strain C7 showed a strong antagonistic effect against Xfp under in vitro conditions, in contrast to the uninhibited growth observed in non-challenged Xfp cells (Figure 4A). This complete inhibition suggested the potential synthesis of antibacterial metabolites by strain C7.

The dual-culture assay was performed on solid medium, wherein the agar separating the two bacterial strains was physically removed, thereby disrupting potential diffusion pathways for non-volatile metabolites through the agar matrix. Notwithstanding this spatial segregation, the growth of Xfp remained entirely inhibited (Figure 5), suggesting that the anti-Xylella activity of strain C7 is largely dependent on the synthesis of VOCs. The mechanism of VOC-mediated antagonism is a well-documented phenomenon among plant-associated bacteria (Weisskopf et al., 2021), and this finding accentuates the potential significance of airborne bioactive molecules in the suppression of Xfp. The production of VOCs by C7 strain was further investigated by SPME and GC/MS. The analyses showed that 15 VOCs were produced in the headspace of the cultured strain C7 and are reported in Table 1. The identified VOCs belong to the chemical classes of esters, aldehydes, ketones, alcohols and short chain fatty acids as the results of microbial catabolism. Considering the area percentage, the most abundant compounds were ethyl acetate, butyl acetate, isoamyl acetate among esters and two short chain fatty acids, acetic and valeric acids. Even if SPME is generally considered not trustworthy for quantitative analysis, the area of each peak in the chromatogram is certainly correlated with the concentration of VOCs in the headspace. Therefore, by comparing the total area of the identified peaks in the headspace of the cultured strains during the incubation period, it is noteworthy that the production rate of VOCs in the investigated strain was very fast, reaching a constant level at 48 h of incubation (Figure 6).

The pathogenicity assessment of strain C7 showed that all inoculated tissues exhibited rapid and progressive symptom development, except for broccoli plants, which remained asymptomatic (Figure 7). Tobacco stems showed initial blackleg and necrotic lesions within 24 h post-inoculation, evolving into extensive tissue damage by 48 h (Figure 7). Potato tubers, pear fruits, and onion bulbs demonstrated the onset of symptoms at approximately 48 h, characterized by localized softening and water-soaked lesions (Figure 7). By day 5 post-inoculation, all tissues exhibited severe maceration, accompanied by a strong, putrid odor, characteristic of advanced bacterial decomposition (Figure 7). Furthermore, C7-inoculated cauliflower plants developed symptoms characterized by vascular tissue browning by day 22 (Figure 7). Furthermore, strain C7 was successfully reisolated from inoculated cauliflower and broccoli plants on YPGA plates at 22 dpi (data not shown), and PCR assays of purified colonies confirmed the presence of the strain (Figure 8). These findings highlight the high virulence of strain C7 and its broad host tissue aggressiveness, underscoring its phytopathogenic potential across multiple economically relevant plant species.