Seven originally identified Xep strains were selected based on the year and location of isolation (Lai et al., 2021). Two strains, T0709-01 and T0709-03, were isolated in 2016, and four strains, T0319-01, XpT2, XpT32, and XpT39, were isolated in 2018. An Xep isolate, XTN47, obtained in 1996 (Leu et al., 2010), was included in this study. Strains XTN47, T0709-01, and T0709-03 were obtained from Chiayi, T0319-01 from Shueishang, XpT2 from Taipao, and XpT32 and XpT39 from Liujiao. All strains were isolated from tomato leaves with symptoms of bacterial spot, and their pathogenicity was confirmed by fulfilling Koch’s postulates (Table 1).

Genomic DNA (gDNA) from the seven strains was prepared using the NautiaZ Bacteria/Fungi DNA Mini Kit (Nautia Gene, Taipei, Taiwan). The quality of the gDNA was estimated using a Qubit fluorometer (Qiagen), and the fragment size was measured using a capillary electrophoresis instrument TapeStation (Agilent Technologies). Sheared gDNA was selected using BluePippin with a 0.75% agarose gel cassette (Sage Science) to obtain gDNA fragment sizes ranging from 6 to 20 kb. Nanopore sequencing libraries were prepared using the PCR-free ligation-based sequencing kit (SQK-LSK109) with native barcoding expansion (EXP-NBD104) for sample multiplexing. Nanopore sequencing was performed using an Oxford Nanopore GridION device (R9.4 flow cell FLO-MIN106D). In total, we obtained 161 K–183 K Nanopore reads for each strain, with an average Nanopore read length of 8,875 bp and an L50 of 48,138 bp (Supplementary Table S1). For the Illumina sequencing, genomic DNA libraries were prepared for high-throughput sequencing using the Illumina DNA Prep Kit (Illumina, San Diego, CA, United States). Whole-genome shotgun sequencing was performed using an Illumina NovaSeq 6,000 platform (Illumina, San Diego, CA, United States) to generate 150 bp paired-end reads. Sequencing depth achieved an average of 2.8–3.1 million paired-end reads per genome (Supplementary Table S1).

Sequencing read quality was evaluated using FastQC (v.0.11.9) (Andrews et al., 2023). Low-quality reads were subsequently removed using Trimmomatic (v.0.36) (Bolger et al., 2014) for Illumina reads and NanoFilt (v.2.6.0) (De Coster et al., 2018) for Nanopore reads. The base quality of Nanopore reads greater than 1 kb was further improved by the corresponding high-quality Illumina reads using FMLRC (v.1.0.0) (Wang et al., 2018). Each strain was assembled using three different assemblers, Canu (v.1.8) (Koren et al., 2017), Flye (v.2.5) (Kolmogorov et al., 2019), and wtdbg2 (v.2.5) (Ruan and Li, 2020), for their ability to handle repetitive regions. The draft assemblies for each strain were then compared and merged using BLASTN (v.2.10.1) (Camacho et al., 2009) with manual inspection to produce a consensus assembly. The per-base accuracy of the assemblies was improved using Pilon (Wang et al., 2014) with high-quality Illumina reads. On average, the final coverage of the assembled genomes exceeded 200X. Genomic sequences of the seven strains were resolved on a single circular chromosome. All plasmid sequences were fully circularized. Genome assembly quality in terms of completeness and contamination was assessed based on the presence and duplication of marker genes using CheckM (Parks et al., 2015) and BUSCO (v5.5.0, −l xanthomonadales_odb10) (Manni et al., 2021) (Supplementary Table S1).

To facilitate whole-genome comparison, the dnaA gene was set as the start of the complete genome assembly. Three conserved DnaA boxes have been identified in the oriC region between dnaA and dnaN (da Silva et al., 2002; Yen et al., 2002; Qian et al., 2005). The terminator of replication for each strain was located at approximately 2.5 Mb of the chromosome sequence. Replichores of the three strains were identified using a GC skew plot.

Genome annotation was performed using the local installation of the NCBI Prokaryotic Genome Annotation Pipeline (PGAP) (Tatusova et al., 2016), followed by manual curation. Specifically, the Docker image of PGAP (v.2019-08-22.build3958) was used for the initial genome annotation. The annotations generated by PGAP were then manually inspected using Artemis (Carver et al., 2012) to verify the correctness of the start and stop codons, among other features. A comparative analysis with published X. euvesicatoria genomes was conducted using BLASTP (Camacho et al., 2009) to further refine and validate the annotations.

Despite the publication of the first Xanthomonas genome (X. citri 306) (da Silva et al., 2002) in 2002, followed by numerous subsequent genome sequences and studies, a significant portion of the genome remains functionally uncharacterized. To enhance functional annotation, we used a combined approach with UniProt (Coudert et al., 2023), Blast2GO (Gotz et al., 2008), and Intrproscan (Jones et al., 2014). However, this combined strategy resulted in approximately 20% of the protein-coding genes remaining classified as hypothetical, indicating an unknown function.

Two complementary approaches based on the diversity of nucleotide sequences and protein-coding genes have been used for genome-wide analyses to understand overall genome variations. The assembled chromosome and plasmid sequences were compared with published 132 Xanthomonas genomes (Supplementary Table S2) to identify conserved and novel sequence elements (Barak et al., 2016; Constantin et al., 2016; Bansal et al., 2018; Fan et al., 2022). The whole-genome similarity metrics of all strains were estimated using the alignment-free approximate sequence mapping algorithm FastANI (v.1.20) (Jain et al., 2018), with default settings. To confirm the taxonomic nomenclature and classification of these strains, genome-based taxonomy was performed using the GGDC database (Meier-Kolthoff et al., 2022).

To further understand the evolution of protein-coding genes in the Xanthomonas genomes, we analyzed genes from 139 Xanthomonas strains (Supplementary Table S2) to infer orthologous relationships using the Markov Cluster Algorithm (MCL) (Enright et al., 2002). Briefly, genes were compared using all-against-all BLASTP (e-value 10⁻⁵) (Camacho et al., 2009), and the results were analyzed using the scalable unsupervised cluster algorithm MCL (inflation value 5). Orthologous and paralogous relationships are presented as networks based on e-values. MCL clustering resulted in 8,580 gene families, which were presented as a gene number matrix of 139 genomes. Gene families present in more than one genome were considered the ′pan-genome′ whereas those shared in at least two genomes were considered the ′core-genome′ (Méric et al., 2014).

A total of 1,346 orthologous genes with a strict one-to-one relationship as the ′single copy core gene′ in 139 Xanthomonas strains were used to construct the phylogenetic tree. Multiple sequence alignments of the protein sequences in these single-copy core genes were individually performed using MUSCLE (Edgar, 2004), and gaps in the alignments were removed using trimAl (Capella-Gutiérrez et al., 2009). Individual sequence alignments were concatenated to create a single protein sequence for each strain. Maximum likelihood phylogenetic trees were inferred using RAxML-NG (Kozlov et al., 2019) with 1,000 bootstrap replicates (v.1.0.1, −model LG + G8 + F –seed 2 –bs-trees 1,000) and visualized using FigTree. ² Genome collinearity analysis was performed using i-ADHoRe 3.0 (Proost et al., 2012) to infer the syntenic regions.

To investigate the evolutionary divergence and better understand the genomic basis of Taiwanese X. euvesicatoria pv. perforans (Xep) strains, we performed whole-genome sequencing of seven Xep strains collected from different locations during 1996–2018 (Table 1). Three typical Xep strains (T0709-01, T0709-03, and XpT2) and four atypical Xep strains (XpT32, XpT39, T0309-01, and XTN47 (Leu et al., 2010)) were selected. Among these, one typical strain and three atypical strains were characterized as Cu^R strains, capable of growing on NA plates supplemented with 0.8 mM CuSO (Lai et al., 2021). Conversely, two other typical Xep strains, T0709-03 and XpT2, and an atypical strain isolated in 1996, XTN47, were sensitive to copper.

We used a hybrid assembly approach that combined short-read (Illumina) and long-read (Oxford Nanopore) sequencing technologies in an integrated bioinformatics workflow (Huang et al., 2021). This method allowed us to complete the genomes of all seven strains into gap-free and circularized chromosomes and plasmid sequences (Figure 1). The assembled genome sizes of the seven strains showed little variation, ranging from 4.915 to 5.073 Mb, with a consistent GC content between 64.9 and 65.0%. Gene prediction analysis identified between 4,100 and 4,289 genes per strain, including rRNA genes and pseudogenes. These values are comparable to those of previously published X. euvesicatoria genomes (Timilsina et al., 2020; Tables 1; Supplementary Table S1). To assess the quality of the genome assemblies, we evaluated metrics such as coverage depth, contiguity, base accuracy, CheckM (Parks et al., 2015), and BUSCO (Manni et al., 2021). The final assembled genomes achieved depths exceeding 200×, ensuring high confidence in assembly accuracy (Supplementary Table S1). CheckM and BUSCO were used to determine genome completeness and contamination. The hybrid approach enhanced assembly contiguity and accuracy, particularly in resolving repetitive regions and plasmid sequences. We identified 1–3 plasmids in each strain, both in the typical and atypical Xep strains. The plasmid size varied substantially, ranging from 10 kb to over 255 kb. Notably, all plasmid sequences were completely circularized (Supplementary Table S1), providing a comprehensive view of the plasmid architecture in these strains.

To clarify the phylogenetic positions and genetic diversity of the sequenced Taiwanese Xep strains, we used multiple complementary approaches to analyze their genome sequences. We compared the average nucleotide identity using FastANI (v.1.2) (Jain et al., 2018) and digital DNA–DNA hybridization (dDDH) (Meier-Kolthoff et al., 2022) of 139 strains, including Xep reference strain 91–118 and Xanthomonas euvesicatoria pv. euvesicatoria (Xee) reference strain 85–10 (Supplementary Table S2). The commonly accepted thresholds for species delineation are ANI ≥ 95% and dDDH ≥70% (Chun et al., 2018). The three typical Xep strains from Taiwan (XpT2, T0709-01, and T0709-03) and strain 91–118 had ANI scores of 100.0% and dDDH values of 99.8–100.0%, indicating great similarity in their genomes (Figure 2). These findings are consistent with those of previous studies (Barak et al., 2016; Constantin et al., 2016; Bansal et al., 2018; Jibrin et al., 2018; Fan et al., 2022), indicating that Xep differs from Xee based on an ANI score > 98.5% and a dDDH value >87.3% (Figure 2).

Surprisingly, the four atypical strains sequenced in this study (T0319-01, XpT39, XpT32, and XTN47) and the Nigerian strain NI1 had ANI scores of 99.8–100.0% and dDDH values of 98.6–100.0%, revealing high similarity among their genomes (Figure 2). These four strains were the most closely related to X. euvesicatoria pv. allii CFBP6369 when compared with other X. euvesicatoria strains. The ANI scores between the four strains were 99.0–99.1% and the dDDH values were 91.3–92.3%, supporting the results of the phylogenomic analysis (see Phylogenomics analysis). Additionally, the four strains had ANI scores of 98.5–98.6% and dDDH values of 88.0–88.4% compared to Xee strains, and ANI scores of 98.8–99.0% and dDDH values of 90.7–91.8% compared to Xep strains (Figure 2).

To further understand the phylogenetic positions of these Taiwanese Xep strains and their relationship with other pathovars of X. euvesicatoria, we compared the protein-coding genes across 139 strains. Using the ortholog inference method Markov Cluster Algorithm (MCL) (Enright et al., 2002), 8, 580 gene families were identified. Based on the presence or absence of gene families in each genome, we identified 1,642 core gene families (present in all strains, regardless of copy number) and a pangenome of 6,938 gene families (present in at least one strain) (Figure 3; Supplementary Figure S2). For phylogenetic analysis, 1,346 orthologous genes with a strict one-to-one single-copy relationship were used to construct the phylogenetic tree. In agreement with the nucleotide sequence comparison, the three typical Xep strains of this study (XpT2, T0709-01 and T0709-03) formed a well-supported clade with published Xep strains, including the reference strain 91–118, and were separated from Xee. On the other hand, the four atypical Xep strains of this study (T0319-01, XpT39, XpT32, and XTN47) and seven published strains (XVP-272, XVP-270, XVP-273, XVP-271, XVP-267, XVP-265 and Xant-477) (Chen et al., 2024) that were also collected in Taiwan and the atypical Xep strain NI1 from Nigeria formed a separate branch on the phylogenetic tree (Figure 3; Supplementary Figure S2). This suggests that, together with the published NI1 genome, these strains form a new lineage of X. euvesicatoria. This new lineage is phylogenetically closer to X. euvesicatoria pv. allii than to Xep (Figure 3; Supplementary Figure S2).

We further studied the presence and absence of variations in gene families among Xee, Xep, and the new lineages (Figure 4; Supplementary Figures S3, S4). In total, 3,553 gene families were shared among the three lineages and comprised the core genome (Figure 4A). The new lineage had more unique gene families (274 gene families) than did the other two lineages (219 and 246 in Xee and Xep, respectively). The number of gene families shared between the new lineage and Xep (281) is much higher than those shared between the new lineage and Xee (52) or between Xee and Xep (73). Gene families unique to the new lineage were further analyzed (Figure 4B; Supplementary Figure S4). Of the 274 lineage-specific genes, a limited number were shared between two or more strains, and 41 genes (15%) were shared by five strains. Strains XTN47 and NI1 had 73 and 83 unique genes, respectively, but only 5–8 unique genes were found in each of the other three strains (Figure 4B). The lineage-specific genes in the new Xep strains were functionally diverse, with many being involved in key biological processes, such as recombinational repair (GO:0000725), carbohydrate metabolism (GO: 0005975), catalytic activity (GO:0003824), zinc ion homeostasis (GO:0006829), and electron transport chain (GO:0022900). Additionally, these genes contribute to critical cellular functions, such as rRNA processing (GO:0006364), cilium- or flagellum-dependent cell motility (GO:0001539), and response to oxidative stress (GO:0006979), highlighting their potential roles in the adaptation and survival of the new lineage under varying environmental conditions and host interactions (Supplementary Figure S3).

To infer recombination events, we reconstructed a reference phylogeny using single-copy gene families across Xep strains. Recombination was assessed by analyzing concatenated universal genomic regions common to the compared genomes in the whole-genome alignment. ClonalFrameML was employed to estimate the rates of homologous recombination among the three lineages of X. euvesicatoria. The analysis revealed that within atypical Xep strains, including NI1, the rate of homologous recombination involving imported DNA was less than half the rate of mutation (R/θ < 0.5; Table 2). However, when we included other Xep strains in the comparison, the recombination rate increased, suggesting more frequent genetic exchange across these groups. Conversely, when atypical Xep strains were excluded, the remaining Xep strains exhibited a recombination rate 1.5 times higher than their mutation rate (R/θ > 1.5). Notably, the average size of the recombination fragments (δ) decreased significantly upon the inclusion of atypical and other typical Xep strains. This reduction indicated substantial recombination events between the lineages of atypical Xep strains and other typical Xep strains, possibly due to horizontal gene transfer. In all scenarios, the effect of recombination on nucleotide variation (measured as R/m) was greater than that of mutations (R/m > 1), underscoring the dominant role of recombination in the genetic diversity of X. euvesicatoria populations.

Based on the gapA phylogeny, Xep strains isolated in Taiwan in 1996 formed three distinct groups. In addition to the two typical groups of Xep strains, several strains isolated in Taiwan since 1996 are phylogenetically close to the representative atypical strain NI1 from Nigeria (NZ_NISG01000001.1; Timilsina et al., 2015), suggesting a potential origin for atypical Xep strains in Taiwan. The atypical Xep strains of the new lineage were amylolytic, indicating that they can break down starch into simpler sugars through the action of amylase enzymes. This characteristic is shared by Xep, but not in the Xee (Supplementary Table S3). Additionally, these strains differed from Xee and Xep in their carbon utilization patterns (Supplementary Table S3). These phenotypic differences suggest that atypical Xep strains may represent a divergent lineage, potentially of Taiwanese origin. Together with the previous analysis, these findings further support the idea that these strains form a new lineage and do not belong to the known pathovars of X. euvesicatoria. Considering the recent progress in defining intra-species units (Rodriguez-R et al., 2024; Viver et al., 2024), we propose to consider these atypical strains as X. euvesicatoria genomovar taiwanensis (Xet, See Discussion).

We identified eight plasmids in three typical Xep strains and 10 plasmids in four Xet strains (Supplementary Table S1). All strains carried at least one additional plasmid in addition to the Cu^R plasmid. In typical Xep strains, a consistent 82-kb plasmid was identified, appearing to be a fusion of a 63-kb plasmid from Xet strains and a 64-kb plasmid from X. campestris pv. campestris CFBP6943 (GenBank accession number: CP066920.1) (Dubrow et al., 2022; Table 1). Smaller plasmids of 28 kb and 10.2 kb were also identified, with the 10.2 kb plasmid being a derivative of the 28 kb plasmid (Webster et al., 2020). Notably, the plasmid pXpT2.1, derived from the Cu^R plasmid, lacked a 160-kb region carrying heavy metal resistance genes (Figure 5A).

In Xet strains, 63-kb plasmids, similar to pXAC64 of X. citri pv. citri (GenBank: NC_003922.1) were identified (56–61% coverage and > 95% identity) (da Silva et al., 2002) (Figure 5B). Strain XTN47, isolated in 1996, carried a 97-kb plasmid, pXTN47, but lacked a 63-kb plasmid. Sequence analysis revealed that pXTN47 is a fusion of the 63-kb plasmid and a 34-kb region from the 67-kb plasmid pLMG696-1 of X. citri pv. durantae LMG696 (GenBank accession number: CP066344.1) (Rana et al., 2022), suggesting a homologous recombination-mediated plasmid fusion. Collectively, either the 63-kb plasmid or its fused derivative was present in all the sequenced strains. Additionally, the 115-kb plasmid pT0319-01.2 was highly similar to the plasmid pF of X. citri pv. phaseoli var. fuscans (GenBank: CP021016.2; 94% coverage, 99.9% identity) and pF of X. citri pv. citri DAR73910 (GenBank accession number: CP060477.1) (Webster et al., 2020) (Figure 5B). A 30-kb plasmid pXpT39.3 was similar to pXAG27 of X. citri pv. glycines (GenBank accession number: CP026335.1; 94% coverage and > 96.6% identity) (Carpenter et al., 2019).

The four Cu^R strains possessed a large ~250 kb plasmid containing the Cu resistance gene cluster and were absent in the Cu^S strains. This megaplasmid showed high nucleotide sequence identity (>99%) with Cu^R strains in pT4p2 of X. citri subsp. citri (Gochez et al., 2018; Huang et al., 2021) and pLH3.1 of Xep (Richard et al., 2017) (Figure 5A). Each plasmid contained a complete set of copper resistance gene clusters, including copLABGF, but lacked copCD (Huang et al., 2021). Additionally, a mercury-resistance operon (merRTPA) was flanked by an ISL3-like transposon, suggesting transposition. Unlike pT4p2, these plasmids did not contain an ~40 kb inverted repeat or arsenate resistance gene cluster.

The copper resistance plasmids also contained genes encoding heavy metal resistance ATP-binding cassette (ABC) transporters (cusABC and czcABC clusters) and a CadR/PbrR-like transcription regulator near a czcABC cluster. A complete Tra gene cluster encoding 16 Tra proteins essential for plasmid mobility was found in the Cu^R plasmids and pXpT2.1. The presence of Tra genes with high sequence similarity to other plasmids suggested that heavy metal resistance genes were likely horizontally transferred between xanthomonads.

To further test the mobility of copper resistance genes, they were transferred from Cu^R strain T0319-01 to Cu^S strain XTN47Rif through bacterial conjugation. The resulting transconjugants showed the same level of copper resistance as donor strain T0319-01. The frequency of conjugative gene transfers ranged from 10⁻⁷ to 10⁻⁵ transconjugants per recipient.

We also screened two populations of X. euvesicatoria strains collected from Taiwan at different times (Chen et al., 2024; Parajuli et al., 2024). The analysis revealed that the frequency of the copper resistance gene cluster increased in the strains collected in recent years, whereas earlier collections showed a lower frequency (Supplementary Figure S5). This indicates the recent acquisition and spread of copper resistance among X. euvesicatoria populations in Taiwan, likely driven by selective pressure from the use of copper-based bactericides.

Lipopolysaccharides (LPSs) produced by Xee 85–10 are known to trigger basal defense responses in pepper (Keshavarzi et al., 2004). Comparative genomic analysis showed that LPS biosynthetic gene clusters, located between the conserved genes cystathionine gamma-lyase (metB) and electron transport flavoprotein subunit A (etfA), exhibited high variability among the four lineages of bacterial spot Xanthomonas. Specifically, the atypical Xep strain NI1 differed significantly from the reference strain 91–118 in both gene content and cluster number (Potnis et al., 2011; Jibrin et al., 2018; Jibrin et al., 2022). Three typical Xep strains from Taiwan had LPS biosynthetic gene clusters identical to those of Xep 91–118 (Supplementary Figure S6). Additionally, a unique LPS biosynthetic gene cluster in NI1 was consistently present in the Xet lineage (Supplementary Figure S6). These findings highlight the distinct composition of LPS biosynthetic gene clusters in the new lineage compared to typical Xep strains.

In Xanthomonas, the type III secretion system (T3SS) is essential for pathogenicity because it modulates host plant physiology and enables evasion of host immune responses (Timilsina et al., 2020). All seven sequenced strains processed conserved T3SS gene clusters encoding hrp (hypersensitive response and pathogenicity), hrc (hrp conserved), and hpa (hrp associated), indicating that T3SS is highly conserved among these strains.

We examined the allelic diversity of three effector genes, xopAF, xopJ2a, and xopJ4, which determine the tomato races and host ranges in Xep strains (Astua-Monge et al., 2000a; Astua-Monge et al., 2000b; Timilsina et al., 2016). The three typical Xep strains contained an intact xopJ4 gene (syn. AvrXv4), which specifically determines the tomato race T4 in Xep (Astua-Monge et al., 2000b). An early stop codon was found in xopAF (syn. AvrXv3), the effector responsible for eliciting Xv3 resistance (Astua-Monge et al., 2000a), classifying these Xep strains as the tomato race T4 (Jibrin et al., 2022) (Supplementary Figure S7).

In contrast, five strains of the new lineage Xet, including the Nigerian strain NI1 (Jibrin et al., 2022), possess xopAF, but lack xopJ4. Within this lineage, two distinct xopAF alleles were observed. Strains XTN47 and NI1 have an intact xopAF gene, classifying strain XTN47 as tomato race T5, similar to NI1 (Jibrin et al., 2022). However, in the other three strains, the C-terminus of the xopAF coding sequence was truncated due to the insertion sequence ISXac4. Additionally, all seven sequenced strains and NI1 contained xopJ2b, a homolog of xopJ2a with 71% amino acid identity (Iruegas-Bocardo et al., 2018; Potnis, 2021; Jibrin et al., 2022). The xopJ2b gene was consistently present in these strains, located on the 63-kb plasmid or its derived plasmids, whereas xopJ2a was absent (Figure 6). These findings highlight conserved T3SS gene clusters but significant allelic diversity and host specificities among different X. euvesicatoria lineages.

Further analysis showed that the xopJ4 gene in the three typical Xep strains was located within a P2-like temperate prophage region between the gpJ and gpS genes (Figure 6). This structure, including xopJ4, is conserved in typical Xep strains and the reference strain 91–118 (Timilsina et al., 2019) but is absent in the Xet strains. Interestingly, another P2-like prophage sharing 92% identity with the prophage carrying xopJ4 was identified in the same chromosomal region of the new lineage strains T0319-01 and XTN47, but no effector genes were found. The plasmid-borne xopJ2b gene is consistently flanked by two insertion sequence elements. These findings suggest significant recombination events and horizontal gene transfer involving type III effectors among the different X. euvesicatoria lineages.

We identified two copies of the VirB/VirD4 type IV secretion system (T4SS) cluster in typical Xep strains from Taiwan. The first one, a complete cluster with 12 genes ending with virD4, was located on the chromosome and showed high protein sequence similarity (90–100%) with the T4SS of the Xep reference strain 91–118 (Supplementary Table S4). The second cluster, consisting of nine genes lacking virB1 and virD4, was found on the 82-kb plasmids (Supplementary Table S4). These plasmid-borne T4SS genes have low protein sequence similarity with those of reference strains Xep 91–118, Xee 85–10, and X. citri subsp. citri 306. Notably, the three typical Xep strains lacked the Dot/Icm T4SS cluster (Supplementary Table S5).

In contrast, each strain of the new lineage harbored at least two distinct VirB/VirD4 T4SS clusters (Supplementary Table S4). The first T4SS cluster, a chromosomal cluster of 12 genes ending with virD4, shares 90–100% protein sequence similarity with the chromosomal T4SS in Xep, indicating strong conservation within the lineage (Supplementary Table S5). This cluster was also present on the NI1 chromosome, further distinguishing this new lineage from typical Xep strains. Additionally, a second VirB/VirD4 T4SS cluster, lacking virD4 and consistsing of 10 genes, was consistently found on the 63-kb plasmids and the 97-kb plasmid pXTN47, which includes the 63-kb plasmid backbone. This plasmid-borne cluster shares high protein sequence similarity with the T4SS genes of Xee 85–10 but differs from those in Xep 91–118, particularly in the VirB4, VirB5, and VirB6 proteins (Supplementary Table S4).

Interestingly, the plasmid-borne T4SS genes in the new lineage were distinct from those in the Xep strains in both gene composition and sequence. Proteins such as VirB1, VirB4, VirB6, VirB9, VirB10, and VirB11 showed greater similarity to those in X. citri subsp. citri 306 than to those in Xee 85–10, suggesting a distinct evolutionary pathway (Supplementary Table S4). The third T4SS cluster, found in the plasmids of strains XTN47 and T0319-01, originated from the pLMG696-1 plasmid of X. citri pv. durantae LMG696, further emphasizing the diversity of T4SS clusters in these strains. These findings highlight the diversity and complexity of T4SS clusters among different Xep strains and the new lineage, suggesting their potential role in pathogenicity and host interactions.

Plasmids in Xanthomonas species that cause bacterial leaf spot in tomato and pepper vary widely in size and often carry virulence or heavy metal resistance genes (Minsavage et al., 1990; Thieme et al., 2005; Roach et al., 2019; Huang et al., 2021). Unfortunately, detailed information regarding plasmids from bacterial spot Xanthomonas, particularly Xep, remains limited (Potnis et al., 2011; Roach et al., 2019; Chen et al., 2024; Parajuli et al., 2024). In this study, we found that the 63-kb plasmid in the Xet lineage resembles the pathogenicity plasmid pXAC64 of X. citri subsp. citri (da Silva et al., 2002). The 82-kb plasmid in three typical Xep strains likely originated from a recombination event, leading to plasmid cointegration, a phenomenon also observed in X. citri subsp. citri strains in Taiwan (Huang et al., 2021).

The copLABGF cluster associated with heavy metal resistance and the Tra cluster for plasmid mobility are located on an approximately 250-kb megaplasmid. Previous conjugative transfer experiments confirmed the mobility of this plasmid, enabling the horizontal transfer of copper resistance genes between Xet strains (Basim et al., 2005; Behlau et al., 2012). Additionally, the presences of the mercury resistance gene cluster (merRTPA) instead of the arsenate resistance gene cluster (Huang et al., 2021), suggests transposition events mediated by insertion elements. These findings indicate significant variation in plasmid distribution among X. euvesicatoria strains depending on their geographical origin and host, highlighting the diversity and evolutionary significance of plasmid profiles in Xanthomonas (Burlakoti et al., 2018; Lai et al., 2021; Xu et al., 2023).

Recombination and other mechanisms likely contribute to the gain and loss of type III secretion system (T3SS) effector genes in Xanthomonas. The xopJ4 gene is carried by a P2-like temperate prophage in the Xep genomes, reflecting lysogenic conversion (Harrison and Brockhurst, 2017), which facilitates the horizontal transfer of virulence factors and enhances the fitness of pathogenic bacteria. This process has been observed in both animal pathogens (Harrison and Brockhurst, 2017) and the phytopathogen Ralstonia solanacearum (Askora et al., 2017), suggesting its role in the evolution of Xep races.

In contrast, the xopJ2b gene is consistently flanked by two insertion sequence elements and is located on a virulence-related plasmid. Insertion sequence-mediated transposition may be associated with the gain and loss of xopJ2a or xopJ2b in the Xep strains (Klein-Gordon et al., 2021). The XopJ family of type III effectors, which includes XopJ1, XopJ2a (syn. AvrBsT), XopJ3 (syn. AvrRxv), and XopJ2b (syn. XopJ6 homolog) has been identified in Xee and Xep (Minsavage et al., 1990; Ma and Ma, 2016; Iruegas-Bocardo et al., 2018; Osdaghi et al., 2021; Potnis, 2021) and plays a role in enhancing bacterial fitness and determining the host range (Timilsina et al., 2016; Osdaghi et al., 2021; Potnis, 2021). XopJ2a can trigger a hypersensitive response (HR) in pepper, limiting the host range of Xep to tomato and enhancing Xep fitness in tomato plants (Minsavage et al., 1990; Timilsina et al., 2016; Abrahamian et al., 2018). XopJ2b also induces HR in peppers (Iruegas-Bocardo et al., 2018), although its role is not well understood (Sharma et al., 2024).

Since all strains sequenced in this study carried the 63-kb plasmid or its derivatives housing the xopJ2b gene, we propose that these plasmids may contribute to the virulence and fitness of the Xep and Xet strains. Further research is needed to clarify the distribution and roles of xopJ2b and these plasmids in X. euvesicatoria genomovars (Sharma et al., 2024).

The VirB/VirD4 type IV secretion system (T4SS) clusters are widely distributed among xanthomonads responsible for diseases such as bacterial spot in tomato and pepper, citrus canker, and rice bacterial blight (da Silva et al., 2002; Thieme et al., 2005; Potnis et al., 2011; Kaur et al., 2019; Jibrin et al., 2022). Our characterization of these clusters revealed that the new genomovar Xet differed significantly from Xee 85–10 and Xep 91–118. Protein sequence comparisons suggested that the chromosomal T4SS cluster of Xet is a hybrid cluster that combines elements from the T4SS clusters of Xep and X. citri subsp. citri.

Interestingly, the Taiwanese strains studied here exhibit substantial variation in their plasmid-borne VirB/VirD4 T4SS proteins compared to each other and Xep 91–118, suggesting that they carry a novel VirB/VirD4 T4SS in their plasmids. Furthermore, the plasmid-borne T4SS cluster in Xet may have evolved through recombination between the clusters found in the plasmids of Xee and X. citri subsp. citri. Notably, the hybrid plasmid pXTN47 of strain XTN47 harbors two VirB/VirD4 T4SS clusters: one identical to the T4SS cluster of Xet strains and the other identical to that in pLMG696-1 of X. citri pv. durantae LMG696. These findings suggest that plasmids carrying distinct T4SSs may play a critical role in the evolution and adaptation of plant-pathogenic xanthomonads (Rana et al., 2022).

Additionally, strain T0319-01 was the only strain found to carry the Dot/Icm T4SS cluster within the 112-kb plasmid p0319-01.2, which shows high similarity to the plasmid pF of X. citri but differs markedly from pXCV182 of Xee 85–10. This suggested that X. euvesicatoria strains may acquire plasmids carrying the Dot/Icm T4SS cluster from various donors via horizontal gene transfer or other unknown mechanisms (Drehkopf et al., 2023).