Biochemical tests were performed with the BIOLOG GENIII plates (Biolog Inc., Hayward, CA, USA) according to the manufacturer’s instructions, using inoculation fluid A. Briefly, overnight cultures of bacteria on MHII plates were resuspended in 0.85% saline solution. A volume of 100 μl of bacterial suspension was added into each of the microplate wells. The change in color of the wells was evaluated with the naked eye. The reference or type strains of Pectobacterium species listed in Supplementary Table S1 were used for comparison. Additionally, the production of acid from α-methyl glucoside or reducing substances from sucrose, utilization of citrate, dulcitol and dextrose as the only carbon source, fermentation of 10 sugars, inulin, mannitol and sorbitol and production of 17 enzymes using the DIATABS™, Diagnostic Tablets (RoscoDiagnostica A/S, Taastrup, Denmark) were assessed.

The growth of P. betavasculorum strains in a microtiter plate was determined by absorbance (OD) measurement at 600 nm each hour for 72 h of incubation at 28°C using a microplate reader InfiniteM200Pro (Tecan, Männedorf, Switzerland).

The assays were performed in 96-well titration plates. The 200-μL of tryptic soy broth (TSB) medium with various pH, salinity, and PEG concentrations was inoculated with 5 μL of bacterial suspension with an optical density of 0.5 McF. The plates were incubated with shaking at 28°C. The absorbance readings at 600 nm were made after 0, 6, 24, and 48 h of incubation, using the Infinite M200 Pro (Tecan, Männedorf, Switzerland). The experiments with two replicates were performed twice. The effect of temperature on the growth of P. betavasculorum strains was investigated in the TSB medium. Plates were incubated with shaking at 15°C, 20°C, 28°C, and 37°C. The effect of pH on bacterial growth was studied in a TSB medium under pH values of 4, 5, 6, 7, 8, 9, 10, and 11, respectively. The ability to grow in various salinity conditions was conducted in TSB medium supplemented with 0 g L⁻¹, 10 g L⁻¹, 20 g L⁻¹, 30 g L⁻¹, 40 g L⁻¹, 50 g L⁻¹, 60 g L⁻¹, 70 g L⁻¹ NaCl, and 80 g L⁻¹. The tolerance for limited water availability was estimated in TSB medium supplemented with 0.0 g L⁻¹, 50.0 g L⁻¹, 75.0 g L⁻¹, 100.0 g L⁻¹, 200.0 g L⁻¹, 300.0 g L⁻¹, 400.0 g L⁻¹, and 500.0 g L⁻¹ of polyethylene glycol (PEG).

The metabolic activity of P. betavasculorum strains cultivated in media with different sugars and plant extracts was assessed by the polypyridyl complex of Ru(II) method as described in (Jońca et al., 2021). The following media were used: M63 medium supplemented with 0.5% solutions of four polyols, namely, arabitol, erythritol, sorbitol, and xylitol,; eight monosaccharides, namely, ribose, arabinose, xylose, glucose, fructose, fucose, mannose, rhamnose, nine disaccharides, sucrose, lactose, maltose, isomaltose, melibiose, cellobiose, turanose, trehalose, pallatinose; one trisaccharide–raphinose; three polymers and starch such as (consisting of numerous glucose units), inulin (composed mainly of fructose units—fructans), pectin (pectic polysaccharide, rich in galacturonic acid), and taurine (non-protein amino acid). Moreover, M63 medium supplemented with 10% of extracts obtained from plants such as sugar beet, fodder beet, calla lily, blackberry, pitaya, ground cherry, black nightshade, orchid, viviparous, and Arabidopsis thaliana were used for the experiments. Plant extracts were prepared according to the procedure previously described by Jońca et al. (2021). Absorbance was read at 600 nm and fluorescence at 480 nm, each hour for 72 h of incubation at 28°C using a microplate reader InfiniteM200Pro (Tecan, Männedorf, Switzerland). The experiment was performed with three replicate samples.

The antibiotic susceptibility of P. betavasculorum strains was tested by a standard disk diffusion method. Antibiotic disks containing ampicillin 10 µg, erythromycin 15 µg, gentamicin 250 µg, kanamycin 30 µg, streptomycin 10 µg, and tetracycline 15 µg (Biomaxima, Gdansk, Poland) were used. Briefly, 100 µL of bacterial suspension was spread on a Mueller–Hinton (MH) medium with an optical density of 0.5 McF. After applying the antibiotic disks, the plates were incubated for 24 h at 28°C. Then, the zone of bacterial growth inhibition around the antibiotic disk was assessed. Moreover, the production of beta-lactamases and carbapenemases was determined with the application of chromogenic medium, ESBL (extended-spectrum beta-lactamases), and KPC (Klebsiella pneumoniae carbapenemase) plates (Graso Biotech, Owidz, Poland) as described previously (Smoktunowicz et al., 2022).

Pathogenicity of the P. betavasculorum strains was assessed for chicory leaves and potato tubers. Tests were performed as described previously (Lee et al., 2013; Smoktunowicz et al., 2022). Briefly, plant leaves and tuber slices were surface sterilized by soaking in 5% (v/v) sodium hypochlorite (NaOCl) and then rinsed thrice with distilled water. Bacterial cultures were grown on CVP medium for 48 h and then harvested and resuspended in Ringer solution to an approximate cell density of 10⁸ CFU mL⁻¹. A small incision was made on each chicory leaf using a pipette tip, and 25 μL of each bacterial suspension was inoculated into the damaged section. Pipette tips containing 50 μL of each bacterial suspension were driven into the flesh of the surface disinfected potato slices. Plants were placed in sterilized plastic boxes with sufficient moisture and stored at 28°C for several days. Plants inoculated with Ringer solutions were used as a negative control. The experiments were repeated three times, each with three replicate samples.

The activity of PCWDEs was assessed as described previously (Smoktunowicz et al., 2022; Andro et al., 1984; Miller, 1972; Tindall et al., 2014; Himpsl and Mobley, 2019). Briefly, for polygalacturonase assay, isolates were incubated on a solid M63 medium (Miller, 1972) containing polygalacturonic acid. After 48 h incubation at 28°C, plates were stained by flooding with 10% (w/v) copper acetate, which forms a blue complex with the polymer, leaving clear haloes around colonies that produce pectolytic enzymes. Cellulase activity was assessed on a medium containing 0.1% (w/v) carboxymethylcellulose (CMC). After 48 h of incubation at 30°C, the plates were stained with an aqueous 0.1% (w/v) Congo red solution for 1 h at room temperature and washed with 1 M NaCl. Cellulase-producing colonies formed “halo” zones. Protease, lipase, and oligo-1,6-glucosidase activities were assessed on skimmed milk, egg yolk agar, and starch agar, respectively (Tindall et al., 2014). Siderophore production was assessed with chrome azurol S (CAS) agar plate assay (Himpsl and Mobley, 2019). The diameter of “halo” zones around bacterial colonies was measured in each case. Experiments were performed in two repetitions, and the means were calculated for each strain. Gelatinase production was assayed with the standard gelatin stab method (Tindall et al., 2014). Liquefaction of the medium was assessed after 1 week of incubation. Malonate utilization as a sole carbon source was evaluated using the malonate broth (Tindall et al., 2014). A shift in the pH indicator color from green to dark violet indicated malonate utilization.

The capability of the strains to produce N-acyl homoserine lactone (AHL) was assessed using Chromobacterium violaceum CV026 as a biosensor, as described by Ravn et al. (2001). Briefly, tested strains were streaked on the LA plates and then incubated overnight at 28°C. After that time, C. violaceum CV2026 was streaked on the plates in parallel. Plates were then once again incubated at 30°C overnight. After that time, the plates were evaluated for the purple violacein pigment that is produced by C. violaceum in the presence of AHLs.

The auxin production by P. betavasculorum strains was assessed by confirming with a colorimetric assay using the Salkowski reagent method (Gang et al., 2019). Strains were cultivated overnight at 30°C in TSB medium supplemented with 0.5 g of tryptophan. After centrifugation, 100 µL of supernatant and 100 µL of Salkowski reagent were mixed on a 96-well plate incubated for 30 min in the dark. The color intensity was measured spectrophotometrically at a wavelength of 536 nm. A non-inoculated medium was used as a control. An Escherichia coli strain ATCC8339 was used as a positive control, while ΔtnaA deletion mutant, which lost l-tryptophan degradation activity (Shimazaki et al., 2012), was used as a non-producing auxins control.

Strains of P. betavasculorum were stored in 20% glycerol at −80°C. For the preparation of genomic DNA, bacteria were first grown for 48 h on CVP medium (Hélias et al. 2012) at 28°C. A volume of 10 ml of the liquid TSB medium (bioMerieux, Marcy-l’Étoile, France) was inoculated with a single colony from the CVP plate and grown overnight at 28°C with shaking. Obtained cultures were centrifuged for cell harvest for 20 min at 3,200g at 4°C and resuspended into TE buffer (10 mM Tris–HCl, pH 7.5, 1 mM EDTA). The DNA was extracted according to the cetyltrimethylammonium bromide (CTAB) protocol (www.jgi.doe.gov, accessed on 5 August 2023). The quantity and quality of the DNA were assessed by the spectrophotometric analysis with a microplate reader InfiniteM200Pro (Tecan, Männedorf, Switzerland) and the 1% agarose gel electrophoresis.

A total genomic DNA fingerprinting was performed by the repetitive element PCR fingerprinting (rep-PCR) using enterobacterial repetitive intergenic consensus (ERIC) primers (Versalovic et al., 1991). Analysis was performed with BioNumerics V6.6 (http://www.applied-maths.com). The neighbor-joining cladogram was created with a band-based Jaccard coefficient.

The sequences of 13 housekeeping genes (acnA, 2,692 bp; gapA, 1,005 bp; gyrA, 2,649 bp; gyrB, 2,418 bp; icdA, 1,254 bp; mdh, 940 bp; mtlD, 1,149 bp; pgi, 1,644 bp; proA, 1,259 bp; recA, 1,074 bp; recN, 1,662 bp; rpoA, 990 bp; and rpoS, 1,514 bp) commonly used for multi-locus sequence analysis of Pectobacterium (Ma et al., 2007; Waleron et al., 2018), were extracted from genomes of six P. betavasculorum strains (listed in Table 1). Their sequences were merged and aligned with sequences retrieved from genomes of the type strains of each of Pectobacterium species that are available at the GenBank database with using the MUSCLE algorithm with the default settings in Geneious Prime (www.geneious.com).

Phylogenetic analyses were performed on the concatenated data set (20,770 bp) with the Maximum-Likelihood method and the General Time Reversible as the best nucleotide substitution model selected using jModeltest 2.1.9 software. Bootstrapping was executed with 1,000 replications. The gene sequences of D. solani IFB0099 (JXRS00000000) were used as an outgroup.

The genomes of two P. betavasculorum strains, CFBP1520 and SF142.2, were sequenced using the Illumina MiSeq technology (Table 2). Quality and adapter trimming of the raw reads was performed using Trimmomatic v0.32 (http://www.usadellab.org/cms/?page=trimmomatic, accessed on 5 August 2023); kmer length and distribution were analyzed using KmerGenie v1.6949 (http://kmergenie.bx.psu.edu/, accessed on 5 August 2023). De novo assembling was performed using SPAdes v3.10.1 (http://cab.spbu.ru/software/spades/, accessed on 5 August 2023), Velvet v1.2.10 (https://www.mybiosoftware.com/velvet-1-1-07-sequence-assembler-short-reads.html, accessed on 5 August 2023), and Ray v2.3.1 (http://denovoassembler.sourceforge.net/, accessed on 5 August 2023). The contigs were integrated using CISA v1.3 (http://sb.nhri.org.tw/CISA/en/CISA, accessed on 5 August 2023) software and scaffolded with SSPACE v3.0 (https://github.com/nsoranzo/sspace_basic, accessed on 5 August 2023). The final assembly was evaluated using Quast v4.5 (http://quast.sourceforge.net/quast, accessed on 5 August 2023) software. Another two strains, CFBP3291 and Ecb168, were sequenced in a hybrid mode using the Illumina HiSeq 2000 and Oxford Nanopore platforms (Table 2). An Illumina-compatible paired-end sequencing library was constructed using the NEB Ultra II FS Preparation Kit (New England Biolabs, Beverly, USA) according to the manufacturer’s instructions. The library was sequenced using an Illumina MiSeq platform (Illumina, San Diego, CA, USA) with 2 × 300 paired-end reads using v3 600-cycle sequencing kit. Sequence quality metrics were assessed using FASTQC (http://www.bioinformatics.babraham.ac.uk/projects/fastqc/), and quality was trimmed using fastp (Chen et al., 2018). Prior to long-read library preparation, genomic DNA was sheared into ~30 kb fragments using 26G needle, followed by size selection using a Short Read Eliminator kit (Circulomics, Baltimore, MD, USA). Five micrograms of the recovered DNA was taken for 1D library construction using SQK-LSK109 kit, and 0.8 µg of the final library was loaded into R9.4.1 flow cell and sequenced on GridION sequencer (Oxford Nanopore Technologies, Oxford, UK). Raw nanopore data was basecalled using Guppy v5.0.7 in super accuracy mode (Oxford Nanopore Technologies, Oxford, UK). After quality filtering using NanoFilt (De Coster et al., 2018) and residual adapter removal using Porechop (https://github.com/rrwick/Porechop), the obtained dataset was quality checked using NanoPlot (De Coster et al., 2018). Long nanopore reads were then assembled in hybrid mode using Unicycler (Wick et al., 2017). The remaining ambiguities in the genome assemblies were verified by the PCR amplification of DNA fragments, followed by Sanger sequencing with an ABI3730xl Genetic Analyzer (Life Technologies, Carlsbad, CA, USA) using BigDye Terminator Mix v. 3.1 chemistry (Life Technologies, Carlsbad, CA, USA). All of the possible sequence errors and miss-assemblies were further manually corrected using Seqman software (DNAStar, Madison, WI, USA) to obtain the complete nucleotide sequence of bacterial genomes). novo assembling was performed using Flye v 2.9.1 (https://github.com/fenderglass/Flye, accessed on 5 August 2023) and Unicycler v 0.5.0 (https://github.com/rrwick/Unicycler, accessed on 5 August 2023). The final assembly was evaluated using Quast v4.5 (http://quast.sourceforge.net/quast, accessed on 5 August 2023) software.

Annotation was performed by the NCBI PGAPpipeline (https://www.ncbi.nlm.nih.gov/genome/annotation_prok/, accessed on 5 August 2023) and RAST (Rapid Annotation using Subsystem Technology) (http://rast.nmpdr.org/, accessed on 5 August 2023). The genomic sequence was additionally annotated with BlastKOALA (KEGG Orthology And LinksAnnotation) (https://www.kegg.jp/blastkoala/, accessed on 5 August 2023).

Comparative genomics was performed with the use of Pathway Tools Software v27.0 (http://bioinformatics.ai.sri.com/ptools/, accessed on 14 August 2023), Bacterial Pangenome Analysis (BPGA) pipeline (Chaudhari et al., 2016) (https://iicb.res.in/bpga/, accessed on 13 September 2023), and a GET_HOMOLOGUES software package (Contreras-Moreira and Vinuesa, 2013). The obtained results were subjected to the BLAST analysis against the UniProt protein database (https://www.uniprot.org/, accessed on 13 September 2023). The entire set of genes identified in all P. betavasculorum genomes was recognized as a pangenome. The core of the pangenome consists of the gene families that were present on all six analyzed genomes. While the accessory genome contains gene families that were present in at least two up to five of six analyzed P. betavasculorum genomes. The unique genes were only found in a single genome.

The visual comparison of genome homology was done with BRIG (BLAST Ring Image Generator) (http://sourceforge.net/projects/brig, accessed on 5 August 2023) with the default settings (Alikhan et al., 2011). The genomes of P. betavasculorum NCPPB2795^T (JQHM00000000) and P. betavasculorum NCPPB2793 (JQHL00000000) have been included for genomic comparison.

The phylogenomic analysis based on the core proteins sequence comparison was done using PhyloPhlAn computational pipeline (https://huttenhower.sph.harvard.edu/phylophlan, accessed on 4 August 2023). The core protein analysis was performed using the sequences of other Pectobacterium members available on GenBank and the phylophlan library of 400 conserved proteins of Prokaryotes and Archea. The sequence of Duffyella gerundensis AR (CP073262) was used as an outgroup. The bootstrap consensus was inferred from 1,000 replicates. To establish the relationship between P. betavasculorum strains isolated from different plant species, the ML tree based on 3,573 core genes per genome (3,588,885 bp per genome) have been created with using the EDGAR platform for comparative genomics (https://edgar3.computational.bio.uni-giessen.de/cgi-bin/edgar_login.cgi). The calculation of the average nucleotide identity (ANI), the average amino-acid identity (AAI), and the Pairwise Percentage of Conserved Proteins (POCP) was performed using the EDGAR pipeline (Dieckmann et al. 2021). To evaluate the assignment of the studied strains to species, in silico and the average nucleotide identity (ANI) were calculated between the genomes of P. betavasculorum and the genomes of the other type or reference strains of the members of Pectobacteriaceae family with using g JSpeciesWS: a web server (Richter et al., 2016).

The presence of ICE and IME elements in the P. betavasculorum genomes was investigated using ICEfinder, a web-based tool (https://bioinfo-mml.sjtu.edu.cn/ICEfinder/ICEfinder.html, accessed on 20 September 2023).

The biosynthetic gene clusters encoding secondary metabolites related to the synthesis of phytotoxins, antibiotics, and antimicrobial compounds were predicted using antiSMASH 4.0 (Blin et al., 2023) (https://antismash.secondarymetabolites.org/#!/about, accessed on 20 September 2023).

The genetic diversity of 14 P. betavasculorum strains originating from four different plant species and three different continents was assessed by the ERIC-PCR method and multi-locus sequence analysis.

Three ERIC profiles were discriminated against (Table 1, Figure 4). For strains CFBP1520 and SF142.2 isolated from sunflower and artichoke, the same fingerprint pattern was observed; however, it was distinct from those strains derived from sugar beet and potato. Furthermore, the sugar beet strains showed genetic diversity as two patterns were distinguished. Strain Ecb168 differed from the other strains isolated from this host plant.

Multi-locus sequence analysis (MLSA) analysis was performed for seven P. betavasculorum strains, namely, Ecb168, NCPPB2795^T, NCPPB2794, CFBP3291, CFBP1520, SF142.2, and NCPPB3075, that were selected based on ERIC fingerprinting profiles. A total of 13 housekeeping genes, namely, acnA, gapA, gyrA, gyrB, icdA, mdh, mtlD, pgi, proA, recA, recN, rpoA, and rpoS, were merged. Next, the Maximum Likelihood phylogenetic analysis on concatenated gene sequences (20,770 bp) was performed to evaluate the evolutionary relationships among P. betavasculorum strains and other species within the Pectobacterium genus. The generated tree has shown the presence of two major clusters. The species P. betavasculorum belongs to the first of these and is most closely related to P. zantedeschiae, P. atrosepticum, and P. peruviense, while P. polonicum, P. punjabense, P. parmentieri, and P. wasabiae constitute their sister clade. All P. betavasculorum strains form two subgroups that form one clearly separated branch.

The second clade included all the other species of the genus Pectobacterium described so far except for P. fontis and P. cacticidum. These two species, in turn, are grouped separately (Supplementary Figure S11). Moreover, the ML tree showed that P. betavasculorum strains are separated into two groups. Strain CFBP3291 isolated from potato groups together with two sugar beet strains, NCPPB2793 and NCPPB2795^T. At the same time, sugar beet strain Ecb168 is more related to strains CFBP150 and SF142.2 that originated from sunflower and artichoke, respectively.

The phylogenomic analysis based on core proteins showed similar results as MLSA and fingerprinting methods. All P. betavasculorum strains form one group (marked in coral pink) that is separated from other Pectobacterium species. P. betavasculorum is more closely related with species P. peruviense, P. atrosepticum, and P. zantedeschiae and more distantly with P. polonicum, P. punjabense, P. parmentieri, and P. wasabiae (Figure 5).

The ML tree for six genomes, built out of a core of 3,573 genes per genome (3,588,885 bp per genome) to establish a relationship between P. betavasculorum strains isolated from different plant species, was created. Strains CFBP150 and SF142.2 isolated from sunflower and artichoke are most similar to each other. Strain Ecb168 is most closely related to type strain NCPPB2795^T, while strain CFBP3291 from potato is more related to strain NCBBP2793 from sugar beet (Supplementary Figure S12A). Additional genomic analyses, Percent of Conserved Proteins (POCP), Average Nucleotide Identity (ANI), and Average Amino acid Identity (AAI) were performed to resolve this dilemma also. According to POCP analysis, strains CFBP1520 and SF142.2 derived from sunflower and artichoke are most similar to each other and are a bit distant from strains from sugar beet NCPPB2793, NCPPB2795^T, and Ecb168, and strain CFBP3291 from potato (Supplementary Figure S12B). In contrary to ANI, AAI analyses showed that strains CFBP150 and SF142.2 isolated from sunflower and artichoke are more similar to two strains isolated from sugar beet, Ecb168 and NCPPB2793, and all together are a bit distant from strains NCPPB2795^T from sugar beet and CFBP3291 derived from potato, which are most similar to each other (Supplementary Figures S12C, D).

The genomes of four P. betavasculorum strains, CFBP1520 originated from sunflower in Mexico, SF142.2 isolated from artichoke on the French island La Reunion, CFBP3291 from potato grown in Romania, and Ecb168 from sugar beet in the USA, were sequenced, assembled, annotated, and deposited in GenBank under the following accession numbers CP118484, CP129215, JANIMZ000000000, and JARDVT000000000, respectively. The final genome sequences of strains Ecb168 and CFBP3291 are circulated and closed, while sequences of strains SF142.2 and CFBP1520 contain 48 or 63 scaffolds representing one chromosome each. The genome size ranged from 4.76 kBP to 4.93 kBP, and GC % was between 51.0% and 51.2% (Table 2).

Among the six P. betavasculorum isolates, only two strains, CFBP3291 and CBPPB2795^T, harbor 23,763 bp long plasmid that is similar in 94.27% (100% coverage) with 21,742 bp long plasmid IncFII(pCRY) (NC005814) harboring a cryptic type of IVA secretory system (T4ASS) from Yersinia pestis biovar Microtus str. 91001 (Song et al., 2004). Plasmids pPB3291 (CP129216) and pPB2795 (JQHM00000026) from P. betavasculorum strains CFBP3291 and CBPPB2795^T, respectively, are identical in 100%. Likewise, plasmid pCRY and both plasmids from P. betavasculorum strains contain 10 VirB proteins from T4SS (VirB1, TrbC/VirB2, ATPase VirB3, minor pilin of type IV secretion complex VirB5, VirB6, VirB8, VirB9, VirB10, and VirB11) and VirD, sequences for plasmid mobilization relaxosome protein MobB and MobC, RepA protein, chromosome partitioning protein ParA, sequences for dopa decarboxylase, conjugal transfer protein, Cag pathogenicity island protein Cag12, transcriptional antiterminator NusG, transcriptional regulator, EexN lipoprotein, conjugal transfer protein TrbL, and membrane protein. Moreover, the contig 4 corresponding to plasmid pPB3291 contains a thermonuclease, type II toxin-antitoxin system RelE/ParE, while scaffold 26, corresponding to plasmid pPB2795, contains additionally TriB protein ATP-binding protein, toxin-antitoxin system, toxin component HigB, micrococcal nuclease-like protein, and seven hypothetical proteins. The plasmids of P. betavasculorum strains are similar in 97.84% with plasmid pRHBSTW-00515_4 from Klebsiella pneumoniae strain RHBSTW-00515 (CP056428) isolated from livestock host (AbuOun et al., 2021, Microb Genom.) and in 94.8% (92% coverage) with numerous plasmids from clinical isolates of Citrobacter freundii (CP110909, CP110906, and CP110893).

The newly obtained genomic sequences were compared with genomes of two strains, NCPPB2793 and NCPPB2795^T, isolated from sugar beet in the USA, which sequences are available in the GenBank under accession numbers JQHL00000000 and JQHM00000000, respectively. Genomes of strains isolated from sugar beet were slightly smaller (~4.7 kBP) than those of strains originating from other host plant species (~4.9 kBP). Consequently, this has resulted in a lower number of genes encoding proteins in the case of three strains isolated from sugar beet, NCPPB2793 (4,126), NCPPB2795^T (4,321), and Ecb168 (4,329), than for strains isolated from sunflower (4,581), artichoke (4,554), and potato (4,587). However, it should be noted that four out of six of the analyzed genomes are not closed. Summary information is presented in Table 2.

The pangenomic analysis revealed that the pangenome of P. betavasculorum is open and comprises, on average, 4,361 gene families. Of these, 87% of genes are the core genome, and the unique genes are only 2% of the entire pangenome of this species (Figure 8).

The largest number of genes, 4,525, was recorded for the strain CFBP1520 from sunflower and the smallest, 4,265, for the strain NCPPB2793 isolated from sugar beet. Interestingly, the number of genes of strains CFBP1520 and SF142.2 originating from sunflower and artichoke, respectively, were bigger than those observed for strains NCPPB2793, NCPPB2795^T Ecb168, and CFBP3291 isolated from sugar beet and potato (Table 5). The complete genomes of these bacteria share 3,807 gene families, which represent over 80% of CDS in each genome. The highest unique CDS number was observed in the case of two strains isolated from sugar beet Ecb168 and NCPPB2793 (239 and 177, respectively), which was at least 10 times more than the unique CDS detected in genomes of four strains. In the case of strains SF142.2 NCPPB2795^T and CFBP3291, all unique CDS are annotated as hypothetical proteins. For the strain Ecb168, there were 239 unique CDS. The majority of them, 181 (70%), are coding hypothetical proteins. The CDS that are unique for this strain are genes encoding phage-related proteins, and three different type II toxin-antitoxin systems, RelB/DinJ, YafO/YafM, and HicA/B, VirB family type IV secretion system proteins. Among 177 CDS unique for strain NCPPB2793, 65 encoded hypothetical proteins, six CDS of membrane proteins, three genes for integrases, three encoding flavodoxin, one gene encoding flavodoxin/nitric oxide synthase, and two genes for aldehyde oxidase, transposases, DNA primases, transcriptional regulators, cold-shock proteins, non-ribosomal peptide synthetase module, LysR family transcriptional regulators, and type IV secretion protein Rhs, and 83 single genes listed in Supplementary Table S4.

The independent pangenomic analysis was performed with the implementation of the GET_HOMOLOGUES software package (Contreras-Moreira and Vinuesa, 2013), which allows the calculation of core genome with the application of three different methods (COG, BBDH, and OrthoMCL) and the discrimination of intersections between strains of different origins. (Supplementary Figure S13).

The genomes of three P. betavasculorum strains, Ecb168, NCPPB2793, and NCPPB2795^T isolated from sugar beet, shared only one common and unique gene that was not present in the genomes of strains isolated from other vegetables. Unfortunately, the function of this protein is not known.

Among five genes that were unique for three strains CFBP3291, CFBP1520, and SF142.2 originated from plants other than sugar beet, there were two hypothetical proteins, AlpA family transcriptional regulator, phage-related DUF4222 domain-containing protein and zinc-ribbon domain-containing protein.

Interestingly, genomes of two strains, CFBP1520 and SF142.2, isolated from plants whose tissues do not contain high sugar concentration (artichoke and sunflower) share 391 common CDS that are not present in other strains from sugar beet or potato. Many of them, 138 (35%), are hypothetical proteins. Interestingly, only these two strains harbored genes nifA, nifB, nifD, knife, nifH, nifL, nifM, nifN, nifQ, nifT, nifU, nifW, nifX, and nifZ that formed nif operon and are involved in nitrogen fixation (Figure 7, Supplementary Figure S14). Moreover, they shared two different type II TA systems, YdaS/YdaT and yafN/yafO, type IV TA system AbiEJ/AbiG, three restriction endonucleases including Eco47II-like restriction-modification (R-M) system type II, genes related to prophages, seven integrases, proteins Ync and Ynd being a part of conjugal type IV macromolecular transfer systems secretion system, outer membrane ferripyoverdine receptor from Ton and Tol transport systems, DNA-cytosine methyltransferase (EC 2.1.1.37), and two enzymes responsible for salicylate catabolism, salicylate esterase (SalE), and fumarylacetoacetate hydrolase (Fhf), and transporters and transcription regulators (Supplementary Table S4).

Four strains, CFBP3291, Ecb168, NCPPB2793, and NCPPB2795^T, derived from plants with high sugar content (potato and sugar beet) share only 31 unique CDS. Among them are genetic determinants of enzymes responsible for histidine degradation [forminoglutamic iminohydrolase (EC 3.5.3.13), histidine ammonia-lyase (EC 4.3.1.3), histidine utilization repressor, imidazolonepropionase (EC 3.5.2.7), N-formylglutamate deformylase (EC 3.5.1.68), urocanate hydratase (EC 4.2.1.49)] (Supplementary Figure S15), ABC transporters (substrate-binding protein and permease-cluster 13, osmolytes), stresses-induced protein Ves (HutD), formiminoglutamic iminohydrolase (EC 3.5.3.13), and genes for phage-related proteins and 10 hypothetical proteins with unknown functions.

Moreover, comparative analysis allows us to identify 162 genes that are unique for two strains CFBP3291 and NCPPB2795^T containing plasmid. Among them, 71 are coding hypothetical proteins and genes of phage-related proteins, genes mcrBC and mrr encoding type IV RM system, and genes hsdM, hsdR, and hsdS coding complete RM system type I. They shared two different type II TA systems, yafN/yafO and relE/rarE, and the solitary gene of yeeU encoding antitoxin to YeeV, and the genes of flagellin lysine methyltransferase FliB, cag pathogenicity island Cag12 family protein, and aromatic l-amino acid decarboxylase. On the contrary, four strains without plasmid have only 32 unique genes mainly encoding hypothetical proteins, phage-related proteins, integrases, relaxase, and transcriptional regulators (Supplementary Table S4).

Functional annotation of genes within the core genome, accessory, and unique genes according to both COG and KEGG databases showed that most functions focused on metabolic processes (Figure 9). However, twice as few genes determining metabolic processes belong to the accessory part of the genome and four times fewer among unique genes. The opposite trend can be observed in the case of genes related to environmental information processing and information storage processing. In the core genome, the four COGs categories are the most numerous: [J] Translation ribosomal structure and biogenesis, [P] Inorganic ion transport and metabolism, [H] Coenzyme transport and metabolism, and [F] Nucleotide transport and metabolism. In the accessory part of the genome, the most abundant is the [Q] Secondary metabolites biosynthesis transport and catabolism category, while among the unique genes, the largest are two categories, [U] Intracellular trafficking secretion and vesicular transport and [V] Defense mechanisms (Figure 10). Using KEGG database, over 70% of genes from each category belonged to the core genome. The same KEGG sub-category related to carbohydrate metabolism is the largest in core, accessory, and unique part of P. betavasculorum genomes. The second largest subcategory in the core genome gathers genes related to membrane transport; in the accessory part of the genome, it was the amino acid metabolism subcategory, while within the unique genes, replication and repair category (Supplementary Table S3).

Comparative genomics analyses were also performed to further identify distinctive traits between the P. betavasculorum and other Pectobacterium species for which genomes are available in the GenBank. Using the function-based comparison tool under RAST [29], 59 genes were predicted to be specific for P. betavasculorum (Table 6, Supplementary Table S2). Among them, we found 25 genes encoding hypothetical proteins, three genes encoding ABC transporters, two genes of transposases of insertion sequence (IS), two genes encoding major facilitator family (MSF) transporters, four encoding transcriptional regulators from LysR and GntR families, efflux transporter periplasmic adaptor subunit (hlyD), hemagglutinin repeat family protein, thiol: disulfide intercharge protein, or two phage-related proteins. Moreover, unique CDs encode proteins involved in amino acid synthesis and transport. Interestingly, P. betavasculorum is the only species of the genus Pectobacterium with genes encoding isopentyl transferase, allantoin permease, urea carboxylase and allophanate hydrolase gene cluster, genetic determinants of acetoin and butanediol biosynthesis, or genes responsible for sulfoquinovose metabolism (Supplementary Figure S16).

The genomic comparison of P. betavasculorum strains revealed the presence of 106 genes responsible for the synthesis of various virulence factors like plant cell wall-degrading enzymes (51) and toxins (25), and 106 genes encoding proteins budling six secretion systems (T1SS–T6SS) and outer membrane vesicles (known as T0SS) (Table 7).