Phylogenomic Analysis Supports the Transfer of 20 Pathovars from Xanthomonas campestris into Xanthomonas euvesicatoria

: The Gram-negative bacterial genus Xanthomonas includes numerous infra-speciﬁc taxa known as pathovars, which are deﬁned primarily on host range and disease symptoms. With the advent of molecular sequence data, many pathovars have been transferred from X. campestris into other Xanthomonas species to better harmonise taxonomy and phylogeny. We performed whole-genome shotgun sequencing on pathotype strains of the following X. campestris pathovars: blepharidis , carissae , clerodendri , convolvuli , coriandri , daturae , euphorbiae , ﬁci , heliotropii , ionidii , lawsoniae , mirabilis , obscurae , paulliniae , pennamericanum , spermacoces , uppalii , vernoniae , viegasii and zingibericola . These genomes showed more than 98% average nucleotide identity with the type-strain of X. euvesicatoria and less than 88% with the type-strain of X. campestris . We propose the transfer of these pathovars into X. euvesicatoria and present an emended species description for X. euvesicatoria .


Introduction
The genus Xanthomonas contains Gram-negative plant-associated bacteria, many of which are pathogens of crops, ornamentals, or other plants [1,2]. An important feature of Xanthomonas pathogen classification and nomenclature is the inclusion of infra-specific taxa known as pathovars, defined primarily on their host range. The taxonomy of pathovars is governed by the International Standards for Naming Pathovars of Phytopathogenic Bacteria [3]. These standards are devised by the Committee on the Taxonomy of Plant Pathogenic Bacteria under the auspices of the International Society for Plant Pathology.
Recently, there have been significant efforts to reconcile the taxonomy of Xanthomonas with phylogenetic relationships inferred from molecular sequence data as well as biochemical traits. Nevertheless, the taxonomic positions of some taxa are incongruent with their apparent evolutionary relationships. This is the case for more than 20 infra-specific taxa (i.e., pathovars) that are classified within the species X. campestris, yet appear to be only distantly related to the typestrain of this species, according to preliminary phylogenetic analysis based on the gyrB genetic locus [4]. In the present study, we perform a more robust phylogenetic analysis based on whole-genome sequencing rather than just that single locus. Thereby, we clarify the phylogeny and propose taxonomic revisions to reflect this.
The current inconsistencies between phylogeny and taxonomy can be explained by the historical context. Collectively, Xanthomonas bacteria cause disease in hundreds of plant species. However, most of these bacterial strains each causes disease on a narrow range of hosts, often a single plant genus or species. Before the advent of molecular methods, Xanthomonas pathogens were difficult to distinguish phenotypically other than by their host ranges and disease symptoms. Historically, the commonly accepted approach to taxonomy was to propose a new Xanthomonas species for each new host. This led to a proliferation of

Materials and Methods
The bacterial strains used in this study are listed in Table 1 and were purchased from the National Collection of Plant Pathogens (NCPPB), York, United Kingdom. For the preparation of genomic DNA, bacteria were grown in plates with King's B agar medium [13] at 28 • C for 48 h. Single colonies were picked and transferred to Universals with 10 mL King's B broth medium and incubated overnight at 28 • C at 220 rpm. Next, 1.8 mL of culture was transferred to a 2 mL Eppendorf tube and centrifuged at 5000× g for 10 min. The supernatant was discarded, and another 1.8 mL of culture was added to this pellet, and tubes were centrifuged for 2 min at 5000× g. The pellets were used for DNA extraction immediately or flash-frozen in liquid nitrogen and stored at −20 • C until extraction. This procedure for bacterial growth is documented in protocols.io at https: //dx.doi.org/10.17504/protocols.io.ewov1nr92gr2/v1 (accessed on 19 November 2022). Table 1. Bacterial pathotype strains used in this study. Fatty acid methyl ester (FAME) clusters are shown according to their designation in the study by Yang and colleagues [9]. Those that belong to Yang's FAME clusters 1 and 14 are indicated with those respective numbers. (1) Strains are related (>0.4 similarity index) to FAME cluster 1. Strains are remotely related (0.2-0.4 similarity index) to FAME cluster 1 [1]. Sequence reads were assembled de novo using SPAdes version 3.15.1 [46] as described in protocols.io at https://dx.doi.org/10.17504/protocols.io.kxygxzrqzv8j/v1 (accessed on 19 November 2022). The resulting genome assemblies were annotated by the NCBI Prokaryotic Genome Annotation Pipeline (PGAP) [47]. Assembly quality was assessed using CheckM version 1.2.2 [48], selecting its markers for genus Xanthomonas. Pairwise average nucleotide identities (ANI) between genome assemblies were calculated using FastANI [49]. Phylogenomic analysis used PhaME [50] and FastTree [51] as described in the protocol at https://dx.doi.org/10.17504/protocols.io.261geny57g47/v1 (accessed on19 November 2022). The resulting phylogenetic tree was visualised using the Interactive Tree of Life [52].

Genome Sequencing and Assembly
We sequenced the genomes of the pathotype strains for each of 32 pathovars of X. campestris, which are listed in Table 1. These sequenced strains included 20 pathovars whose gyrB sequences suggested a closer relationship to X. euvesicatoria rather than to X. campestris [4]. It also includes, as controls, strains that have recently been transferred from X. campestris into X. euvesicatoria [10] or X. citri [11]. A further set of controls comprised seven pathovars that are closely related to the type strain of X. campestris [4]. Finally, we include the pathotype strain of X. axonopodis pv. passiflorae [53], synonymous with X. campestris pv. passiflorae, which has been mentioned in the literature as "X. phaseoli pv. passiflorae" [54].
The raw sequencing reads and draft-quality genome assemblies are available in the Sequence Read Archive [55,56] via BioProject accessions PRJNA742925 and PRJNA774128. Summary statistics for the genome assemblies are provided in Table 2. Assemblies consisted of between 18 and 136 contigs with N 50 lengths ranging between 93 and 996 kb. CheckM reported each assembly as more than 99.6% complete and no more than 1.71% contamination (see Table A2 in Appendix B).

Phylogenomic Reconstruction
The phylogenetic positions of many of these pathovars had previously been inferred on the basis of partial sequences of the gyrB locus. Therefore, we used our newly generated genome assemblies to explore the levels of genomic similarity, going beyond that single locus. We used PhaME [50] to infer the phylogeny based on whole-genome sequence data ( Figure 1). The resulting tree included a clade that includes the type-strain of X. euvesicatoria together with the pathotype strains of X. euvesicatoria pathovars perforans, alangii, allii, raphani and physalidis ( Figure 1). Into this clade also fell pathovars blepharidis, carissae, clerodendri, convolvuli, coriandri, daturae, euphorbiae, fici, heliotropii, ionidii, lawsoniae, mirabilis, obscurae, paulliniae, pennamericanum, spermacoces, uppalii, vernoniae, viegasii and zingibericola, suggesting that these 20 pathovars phylogenetically fall within the species X. euvesicatoria. Consistent with previous studies [4,11], pathovars merremiae and trichodesmae showed phylogenetic proximity to X. citri. Pathovars aberrans, armoraciae, barbaraeae, incanae, papavericola, plantaginis and raphani all fall close to the type-strain of X. campestris, as expected. Pathovar passiflorae falls closer to the type-strain of X. phaseoli than to X. axonopodis. Figure 1. Phylogenetic tree, based on core genome sequences, for the newly sequenced strains, species type strains and relevant pathovar pathotype strains of Xanthomonas, generated using PhaME [50] and FastTree [51]. The tree was graphically rendered using the Interactive Tree of Life [52].
Configuration and tree files are available from https://github.com/davidjstudholme/phylogenomics-Xanthomonas-2 (accessed on 19 November 2022). Grey circles indicate genomes sequenced in the present study. Clades corresponding to X. euvesicatoria, X. citri, X. phaseoli and X. campestris are indicated by strips of orange, cyan, brown and green, respectively. A list of accession numbers and references for the genome sequences is provided in Appendix A as Table A1.

Species Delineation Based on Average Nucleotide Identity
The results of the genome-based phylogenetic analysis supported the proposition Figure 1. Phylogenetic tree, based on core genome sequences, for the newly sequenced strains, species type strains and relevant pathovar pathotype strains of Xanthomonas, generated using PhaME [50] and FastTree [51]. The tree was graphically rendered using the Interactive Tree of Life [52]. Configuration and tree files are available from https://github.com/davidjstudholme/phylogenomics-Xanthomonas-2 (accessed on 19 November 2022). Grey circles indicate genomes sequenced in the present study. Clades corresponding to X. euvesicatoria, X. citri, X. phaseoli and X. campestris are indicated by strips of orange, cyan, brown and green, respectively. A list of accession numbers and references for the genome sequences is provided in Appendix A as Table A1.

Species Delineation Based on Average Nucleotide Identity
The results of the genome-based phylogenetic analysis supported the proposition that 20 pathovars are evolutionarily much closer to X. euvesicatoria than to X. campestris ( Figure 1). However, to delineate the bounds of bacterial species, the standard approach is to use genome-wide average nucleotide identity (ANI). Therefore, we calculated pairwise ANI between each of the pathotype strains and type strains of relevant Xanthomonas species. These ANI values are summarised in Table 3.

Discussion
The species Xanthomonas campestris has encompassed numerous pathovars, which are assemblages of strains sharing similar host ranges and pathology. With the advent of cheap and easy molecular sequencing, it has become apparent that there is great genetic heterogeneity among strains classified as X. campestris and at least 38 species have been described for this genus [64]. Previous studies [4,8,10,12,65] have highlighted that many pathovars classified within X. campestris are more closely related to other Xanthomonas species than to the type strain of X. campestris. Consequently, recent taxonomic revisions have transferred many X. campestris pathovars into different species [10][11][12]66]. Nevertheless, there remain X. campestris pathovars whose taxonomy remains to be resolved in the light of genetic and genomic evidence.
Here, the results of our genome sequencing and phylogenomic analysis are consistent with recently published taxonomic revisions that place pathovars merremiae and trichodesmae within X. citri and pathovars alangii and physalidis in X. euvesicatoria [10]. These results further support the transfer of pathovar passiflorae within X. phaseoli and a further 20 pathovars into X. euvesicatoria.
We also sequenced type strains of pathovars aberrans, armoraciae, barbaraeae, campestris, incanae, papavericola, plantaginis, and raphani and investigated their evolutionary relationships. Consistent with previous studies that were limited to sequencing a single genetic locus [4], we find that these are closely related to the type strain of X. campestris and fall within the boundaries of this species as delineated by ANI [67]. Therefore, we do not here propose any changes to the taxonomy of these pathovars. However, we previously noted [68] confusion and redundancy in the nomenclature for X. campestris isolates that cause nonvascular leaf spot disease on Brassica spp. We reiterate the previous proposal that such isolates should be classified as X. campestris pv. raphani rather than armoraceae [68].
In summary, 20 X. campestris pathovars examined in the current study have not been previously transferred from X. campestris, though analysis of limited available DNA sequence suggested that phylogenetically there would be a case for doing so [4]. Here, we present draft genome assemblies for the pathotype strains of these 20 pathovars for the purpose of assigning them to species. All of these could thereby be unambiguously assigned to either X. euvesicatoria on the basis of a 96% threshold for genome-wide ANI. Below, we present emended taxonomic descriptions to implement the proposed taxonomic transfers.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.