Xanthomonas hortorum – beyond gardens: Current taxonomy, genomics, and virulence repertoires

Abstract Taxonomy Bacteria; Phylum Proteobacteria; Class Gammaproteobacteria; Order Lysobacterales (earlier synonym of Xanthomonadales); Family Lysobacteraceae (earlier synonym of Xanthomonadaceae); Genus Xanthomonas; Species X. hortorum; Pathovars: pv. carotae, pv. vitians, pv. hederae, pv. pelargonii, pv. taraxaci, pv. cynarae, and pv. gardneri. Host range Xanthomonas hortorum affects agricultural crops, and horticultural and wild plants. Tomato, carrot, artichoke, lettuce, pelargonium, ivy, and dandelion were originally described as the main natural hosts of the seven separate pathovars. Artificial inoculation experiments also revealed other hosts. The natural and experimental host ranges are expected to be broader than initially assumed. Additionally, several strains, yet to be assigned to a pathovar within X. hortorum, cause diseases on several other plant species such as peony, sweet wormwood, lavender, and oak‐leaf hydrangea. Epidemiology and control X. hortorum pathovars are mainly disseminated by infected seeds (e.g., X. hortorum pvs carotae and vitians) or cuttings (e.g., X. hortorum pv. pelargonii) and can be further dispersed by wind and rain, or mechanically transferred during planting and cultivation. Global trade of plants, seeds, and other propagating material constitutes a major pathway for their introduction and spread into new geographical areas. The propagules of some pathovars (e.g., X. horturum pv. pelargonii) are spread by insect vectors, while those of others can survive in crop residues and soils, and overwinter until the following growing season (e.g., X. hortorum pvs vitians and carotae). Control measures against X. hortorum pathovars are varied and include exclusion strategies (i.e., by using certification programmes and quarantine regulations) to multiple agricultural practices such as the application of phytosanitary products. Copper‐based compounds against X. hortorum are used, but the emergence of copper‐tolerant strains represents a major threat for their effective management. With the current lack of efficient chemical or biological disease management strategies, host resistance appears promising, but is not without challenges. The intrastrain genetic variability within the same pathovar poses a challenge for breeding cultivars with durable resistance. Useful websites https://gd.eppo.int/taxon/XANTGA, https://gd.eppo.int/taxon/XANTCR, https://gd.eppo.int/taxon/XANTPE, https://www.euroxanth.eu, http://www.xanthomonas.org, http://www.xanthomonas.org/dokuwiki


DIA et Al.
A group of strains causing bacterial spot of tomato and pepper (Solanum lycopersicum and Capsicum annuum) was originally named "Pseudomonas gardneri" (Šutic, 1957). Some years later, it was suggested to be part of genus Xanthomonas (Dye, 1966), but it was not formally described as X. gardneri until the beginning of the 21st century (Jones et al., 2004). The taxonomical history of the Xanthomonas strains causing bacterial spot of tomato and pepper has been thoroughly reviewed (Osdaghi et al., 2021;Potnis et al., 2015).
Strains associated with bacterial bract spot of artichoke (Cynara scolymus) were first reported in the 1950s as members of the Xanthomonas genus (Ridé, 1956), yet the official species description as X. cynarae was only provided in 2000 (Trébaol et al., 2000). Although a few phylogenetic studies demonstrated the high F I G U R E 1 The taxonomical history of Xanthomonas hortorum, outlining official taxonomical descriptions and changes, as well as first reports or suggested reclassifications of the various pathovars genetic relatedness between X. hortorum, X. cynarae, and X. gardneri Young et al., 2008), they were only recently formally accepted as the same taxonomic entity (Morinière et al., 2020;Timilsina et al., 2019).
Genomic, phenotypic, and pathogenicity analyses were first used to prove the synonymy of X. cynarae and X. gardneri and reclassify them as pathovars of X. cynarae (Timilsina et al., 2019). In that same study, X. hortorum and X. cynarae were acknowledged to be paraphyletic species but were kept separate, based on previous wet-lab DDH results. However, only the type strain of X. hortorum was included in the 2019 study. A comprehensive analysis revisited the taxonomy of those strains, and included all type, pathotype, or representative strains of X. hortorum and X. cynarae (Morinière et al., 2020). Standard genome-to-genome comparison parameters, such as average nucleotide identity (ANI), in silico DDH (isDDH), and tetranucleotide frequencies (Tetra), between X. hortorum and X. cynarae fell into the transition zone of the species boundary (Morinière et al., 2020), a concept described previously (Richter & Rosselló-Móra, 2009;Rosselló-Móra & Amann, 2015). Phylogenetic reconstructions suggested a continuous evolution and diversification of pathovars and phenotypic data did not reveal stable diagnostic traits allowing distinction between X. cynarae and X. hortorum strains.

| HOS T R ANG E
Making a distinction between natural and experimental hosts of plant-pathogenic bacteria is important to better understand the extent of their host range (Bull & Koike, 2015). The natural host range of a pathogen consists of naturally infected plants (i.e., in nonexperimental settings), and is the criterion for pathovar identification and classification (Dye et al., 1980). The experimental host range includes plants that show symptoms after artificial inoculation. Its scope depends on the choice of plant species and of inoculation procedures.
The experimental host range provides invaluable information on the pathogen's potential to adapt to new host plants (Jacques et al., 2016).
Each X. hortorum pathovar has its own natural host range and the experimental host ranges of multiple pathovars have been studied.
Additionally, many unassigned strains within X. hortorum have also been isolated from multiple different plants (e.g., wheat, peony, and hydrangea). Most of the reported natural hosts of X. hortorum belong to the Geraniaceae, Araliaceae, and Asteraceae families, while most of the reported experimental hosts of the pathogen belong to Asteraceae (Table 1). X. hortorum affects more than 65 plant species in 15 botanical families, as summarized in Table 1.
To our knowledge the only known hosts of X. hortorum pvs carotae and taraxaci are their respective initial hosts of isolation. X. hortorum pv. carotae is pathogenic on wild carrot (Daucus carota) and its cultivated subspecies (D. carota subsp. sativus) (Kendrick, 1934;Myung et al., 2014;Temple et al., 2013), while Russian dandelion (Taraxacum kok-saghyz) is the only reported host of X. hortorum pv.
Several unclassified X. hortorum strains cause disease on other plant species such as peony (Paeonia spp.) (Klass et al., 2019;Oliver et al., 2012) and sweet wormwood (Artemisia annua) (Ssekiwoko et al., 2009) (Table 1). Strains causing an unknown disease of lavender (Lavandula dentata, L. angustifolia, and L. × intermedia) were first identified as X. campestris (Koike et al., 1995), but reclassified as X. hortorum based on sequence data (Roberts & Parkinson, 2014;Rotondo et al., 2020). Strains reported as closely related to X. hortorum are sometimes unavailable in public or private strain collections, as is the case for angular leaf spot disease of oak-leaf hydrangea (Hydrangea quercifolia) observed in Georgia, USA (Uddin et al., 1996).
Recently, similar strains were reported from leaf spot symptoms on hydrangea in Flemish (Belgium) nurseries (Cottyn et al., 2021).

| DIS E A S E SYMP TOMS
The pathovars of X. hortorum can cause bacterial spot and/or bac- symptoms caused by all these pathogens share common characteristics but also have some subtle differences.
The presence of a chlorotic halo around spots or lesions is pathovar-or plant/cultivar-dependent. For example, a chlorotic halo is present around lesions caused by X. hortorum pvs pelargonii, hederae, and taraxaci, but its presence varies in angular leaf spot caused by X. hortorum pvs carotae, vitians, and gardneri (Daughtrey & Wick, 1995;Gilbertson, 2002;Myung et al., 2014;Nameth et al., 1999;Pruvost et al., 2010). In advanced infection stages, lesions and spots usually turn dark in colour (brown to black) on plant parts affected by X. hortorum pvs pelargonii, hederae, carotae, gardneri, and vitians. They can also coalesce (e.g., in the presence of X. hortorum pvs hederae, gardneri, and vitians), giving a papery appearance to leaves affected by X. hortorum pv. vitians (Bull & Koike, 2005). In final infection stages, leaves usually harden and dry and, in the case of leaves affected by X. hortorum pv. hederae, a red-purple margin might appear on their upper surface (Suzuki et al., 2002).

F I G U R E 4 Distribution of the seven Xanthomonas hortorum pathovars. Map from the ggmap R package (Kahle et al., 2019) and data adapted from the European and Mediterranean Plant Protection Organization (EPPO). Location is an approximation based on literature available
The two primary sources of inoculum of X. hortorum pathovars are seeds and cuttings, although they can be disseminated through other means as well (e.g., insects, rain, and irrigation water) and can survive on weeds, crop debris, or in soils. Seed is a main source of inoculum for bacterial spot and blight caused by X. hortorum pvs carotae, gardneri, and vitians (Barak et al., 2001(Barak et al., , 2002Kendrick, 1934;Kuan, 1985;Mtui et al., 2010;Sahin & Miller, 1997;du Toit et al., 2005du Toit et al., , 2014. Contaminated seed, stecklings, or seedlings may initiate an epidemic in grower fields (McDonald & Linde, 2002;du Toit et al., 2005), which could result in a nonnormal pathogen distribution, as observed for X. hortorum pv. carotae populations (Scott & Dung, 2020). This can pose a challenge to the development of detection methods and durable resistant cultivars.
Crop residues can allow X. hortorum pvs carotae and vitians to overwinter for several months or until the following growing season (Christianson et al., 2015;Sahin et al., 2003). X. hortorum pv.
carotae can persist in infected carrot foliage on soil for up to a year (Gilbertson, 2002). X. hortorum pv. vitians can survive in crop debris for up to 1 month, in both summer and winter months (Barak et al., 2001;Fayette et al., 2018). X. hortorum pvs vitians and gardneri can survive epiphytically or infect weeds, respectively (Araújo et al., 2015;Barak et al., 2001;Fayette et al., 2018). Soil or crop debris also act as an important inoculum source for X. hortorum pv. carotae, where it can survive for up to 3 months (Kendrick, 1934), and for pv. pelargonii, which can survive in soils for up to a year (Gilbertson, 2002). Survival in weeds, plant residues, and soils can serve as a secondary inoculum source in the presence of favourable hosts and environmental conditions (Gitaitis & Walcott, 2007). If bacterial populations are high, they can re-emerge from inside the plant tissue and serve as a secondary inoculum on the plant itself or on nearby hosts.
X. hortorum pvs gardneri, carotae, and vitians are also disseminated by wind or rain, or mechanically transferred during planting and cultivation du Toit et al., 2005). X. hortorum pv. carotae has been observed in aerosolized debris generated by carrot seed threshers during field operations (du Toit et al., 2005).

| IDENTIFI C ATI ON AND DE TEC TI ON
Visual symptom assessment is the first step to detect a suspected X. hortorum infection and subsequent identification is based on pathogen isolation. X. hortorum strains are readily isolated from infected plant tissue using serial dilution plating. Growth media used can be nonselective (e.g., nutrient agar, sucrose peptone, or yeastdextrose-calcium carbonate [YDC] agar) or semiselective (Saddler & Bradbury, 2015). Irrespective of medium type, X. hortorum colonies are yellow, mucoid, and convex (Saddler & Bradbury, 2015).
Antibodies were used to detect X. hortorum pv. pelargonii.
Several DNA-based molecular assays have been developed over recent decades to identify and detect X. hortorum strains ( Table 2).
The available methods are limited to four of the seven pathovars, as diagnostics methods are unavailable for X. hortorum pvs hederae, taraxaci, and cynarae at the time of writing. Several PCR detection protocols are available to amplify DNA from many Xanthomonas species, including one or more X. hortorum pathovars, by targeting the 16S rRNA gene (Maes, 1993), the hrp gene cluster (Leite et al., 1994), or gumD, fyuA, and the internal transcribed spacer (ITS) (Adriko et al., 2014). However, these general protocols are usually not implemented for the identification and detection of X. hortorum pathovars. Instead, targeted assays allowing specific detection and identification at the pathovar level are often preferred.
The first targeted detection DNA-based assays were mostly derived from DNA fingerprint methods, and were used to study the genetic diversity of X. hortorum pathovars (Barak & Gilbertson, 2003  TA B L E 2 (Continued) Hamza et al., 2012;Sahin et al., 2003;Sulzinski, 2001). For example, a PCR for X. hortorum pv. carotae was developed from random amplified polymorphic DNA (RAPD) analysis (Meng et al., 2004). Similarly, diagnostic PCR tests for X. hortorum pv. pelargonii were developed from specific DNA fragments identified by RAPD analysis, enterobacterial repetitive intergenic consensus (ERIC) PCR or repetitive extragenic palindromic (REP) PCR (Chittaranjan & De Boer, 1997;Manulis et al., 1994;Sulzinski, 2001;Sulzinski et al., 1995Sulzinski et al., , 1996Sulzinski et al., , 1997Sulzinski et al., , 1998. A multiplex PCR scheme for the simultaneous detection of X. hortorum pv. pelargonii and members of the Ralstonia solanacearum species complex, the second major bacterial pathogen of geranium, was developed from one of the two previously identified ERIC-PCR fragments (Glick et al., 2002). The same molecular region was used to develop a real-time quantitative PCR (qPCR) assay, allowing quantification of the pathogen (Farahani & Taghavi, 2016).
For X. hortorum pv. gardneri, a marker identified using amplified fragment polymorphism (AFLP) was initially used to design a diagnostic PCR assay (Koenraadt et al., 2009 Partial gene sequence of gyrB offers a sufficient resolution for the identification of xanthomonad isolates at the species level (Parkinson et al., 2007. MLSA is preferred to single-gene (e.g., gyrB) to outline the precise phylogeny of X. hortorum (Morinière et al., 2020). However, MLSA schemes are based on different partial gene sequences, (sub)sets of partial genes, and trimming settings, which complicates the analysis by not allowing proper comparison between studies (Catara et al., 2021). The sequencing of the first draft genome of X. hortorum pv. carotae in 2011 allowed the first use of comparative genomics to develop two new diagnostics assays to detect this pathovar (Kimbrel et al., 2011;Temple et al., 2013). The first assay used a TaqMan qPCR, whereas the second relied on loopmediated isothermal amplification (LAMP) (Temple et al., 2013). The latter method showed superior performance compared to qPCR because of its robustness in the presence of inhibitors, and its rapidity, versatility, and usefulness in facilities with limited resources (Kimbrel et al., 2011). Both assays were also the first ones to be used as viability assays (i.e., detection of viable bacterial cells) with a xanthomonad, by including a propidium monoazide treatment prior to DNA extraction.
Two other isothermal amplification methods were recently published for the in-field detection of X. hortorum pv. gardneri (Table 2), with an emphasis on differentiation from the other xanthomonad species responsible for tomato bacterial spot. The first method is based on recombinase polymerase amplification (RPA) and targets hrcN (hrpB) (Strayer-Scherer et al., 2019). The second, based on LAMP, targets partial hrpB gene sequence (Stehlíková et al., 2020).  (Table 3). Their completeness varies, and 74% (n = 26) are incomplete (i.e., assembled at the scaffold or contig levels; and included 370 protein-coding genes. These genes were mostly associated with critical cellular processes (e.g., translation, energy production, lipid transport), with 355 and 334 of them conserved within X. hortorum and the Xanthomonadaceae family, respectively.

| Lipo-and exo-polysaccharides
Lipopolysaccharides (LPSs) are involved in biofilm formation and protecting pathogens from their environment (Corsaro et al., 2001;Newman et al., 2002Newman et al., , 2007 The number of CDS is a direct output from NCBI and three numbers appear to overestimate the actual number. stress (Kakkar et al., 2015;Kamoun & Kado, 1990;Sutherland, 1993). The EPS gene cluster of X. hortorum pvs carotae and vitians is arranged similarly to that of X. campestris pv. campestris and contains all 12 genes from the gumB-gumM cluster (Kimbrel et al., 2011;Morinière et al., 2021). Unlike in other Xanthomonas species (Katzen et al., 1998;Kim et al., 2008), only the mutations of gumE, gumI, and gumJ were lethal in X. hortorum pv. vitians LM 16734 (Morinière et al., 2021). The presence of a tRNA gene flanking the cluster in some Xanthomonas genomes suggests a horizontal transfer acquisition (Lu et al., 2008). However, no evidence of insertion elements was found in the EPS gene cluster of X. hortorum pvs carotae and vitians (Kimbrel et al., 2011;Morinière et al., 2021).

| Secretion systems
Secretion systems and their effector proteins are crucial determinants of virulence in the Xanthomonas genus (Büttner & Bonas, 2010). There are two types of type II secretion system (T2SS) clusters within Xanthomonas: the T2SS-xps, directly involved in virulence, and the T2SS-xcs, which has seemingly no direct virulence function (Szczesny et al., 2010). The pathotype strains of X. hortorum pvs hederae, gardneri, and cynarae, in addition to strain B07-
The type III secretion system (T3SS) delivers effector proteins that, in turn, can suppress or trigger plant defence mechanisms (Büttner, 2016;White et al., 2009). The T3SS is found in most Xanthomonas strains, including X. hortorum (Timilsina et al., 2020). The T3SS of X. hortorum pv. gardneri ATCC 19865 PT is a mosaic hrp cluster, with elements like that of X. campestris pv. campestris ATCC 33913 T , but also including novel effectors (see Molecular host-pathogen interactions) . X. hortorum pv. carotae M081 has a complete hrp cluster and is predicted to be functional (Kimbrel et al., 2011). Furthermore, a recent study reported that the T3SS hrp cluster in X. hortorum pv. gardneri ATCC 19865 PT , cynarae CFBP 4188 PT , hederae CFBP 4925 T , and carotae M081 are similar, with some differences in the two 20 kb regions flanking the cluster (Merda et al., 2017).
The type IV secretion system (T4SS) is involved in protein transfer as well as bacterial conjugation (Guglielmini et al., 2014;Lawley et al., 2003;Llosa et al., 2002). X. hortorum pv. gardneri ATCC 19865 PT has two plasmidborne and one chromosomal T4SS clusters   ., 2011). The presence of a T4SS cluster in X. hortorum pv. carotae M081 was suggested by the detection of virB genes scattered over three different contigs but its functionality was inconclusive (Kimbrel et al., 2011).
The type V secretion system (T5SS) is responsible for the secretion of various proteins, including adhesins, which are important for host colonization as they are among the first contact points between pathogen and host (Meuskens et al., 2019). The members of T5SS are autotransporters, with the exception of type 5b, which is formed of two proteins (Guérin et al., 2017). In Xanthomonas spp., T5SS clusters belong to categories 5a, 5b, and 5c (

| Copper resistance and homeostasis
Copper resistance is attributed to the acquisition of a copper resistance gene cluster through horizontal gene transfer (Behlau et al., 2008;Bender et al., 1990;Cooksey, 1994). Copper resistance is usually plasmid encoded (Stall et al., 1986) and can thus be acquired via conjugation by other bacteria (Basim et al., 1999). Because copperbased solutions have been extensively used for controlling bacterial spot diseases, with recommendations going back to the 1920s (Abrahamian et al., 2020;Higgins, 1922;Obradovic et al., 2008), copper-resistant strains pose a challenge for disease management (see Disease control and management).
X. hortorum pv. gardneri strains differed in their response to copper. For example, strain ATCC 19865 PT has copLAB homologs on the chromosome (cohLAB) and is homeostatic to copper, growing in copper concentrations up to 75 mg/L . In contrast, strains JS749-3 and ICMP 7383 have plasmidborne copLAB and copMGCDF genes Richard, Ravigné, et al., 2017), as well as cusAB/smmD systems, involved in heavy metal efflux resistance and originally described in Stenotrophomonas maltophilia (Crossman et al., 2008).
Knocking out rpfF and rpfC in X. hortorum pv. pelargonii Xhp305 altered in planta motility, decreased disease severity on pelargonium plants, and disrupted the plant colonization pattern (Barel et al., 2015). The resulting inability of X. hortorum pv. pelargonii to switch back and forth between biofilm and planktonic lifestyles is thus DSF-dependent (Barel et al., 2015), and this shift is essential for pathogenicity (He & Zhang, 2008). Furthermore, in the rpfF and rpfC mutants, genes gumM, pilC, and pilT were down-regulated compared to the wild type, suggesting that gumM expression and biofilm production, and the type 4 pilus apparatus are DSF-dependent in X. hortorum pv. pelargonii (Barel et al., 2015).
Effectors are used by Xanthomonas species to trigger or suppress host defence mechanisms. Repertoires of effectors (effectomes) have been suggested to play a role determining host specificity (Hajri et al., 2009). Within X. hortorum, effector-related work is mainly focused on pv. gardneri, but there are also reports on pv. carotae strains (more information below). The T3SS of X. hortorum pv. gardneri ATCC 19865 PT was associated with hrpW , a gene predicted to encode a pectate lyase (White et al., 2009), involved in plant tissue maceration and rotting (Collmer & Keen, 1986). The function of effector gene xopZ2, located downstream of hrpW, was suggested by avrBs2 reporter gene fusion . Other T3SS effectors (T3Es) in X. hortorum pv. gardneri strains were also reported: XopAM, XopAO (homolog of AvrRpm1 from P. syringae), XopAQ (homolog of Rip6/ Rip11 from R. solanacearum), and XopAS (homolog of HopAS1 from P. syringae). Effectors XopAM and XopAO were demonstrated to be dependent on the T3SS using the AvrBs2 reporter system . In addition, four novel T3Es were reported in multiple field strains of X. hortorum pv. gardneri: a second XopE2 paralog, in addition to XopJ and two predicted effectors, named T3EP and PTP, with homologs in R. solanacearum and X. campestris pv. campestris, respectively (Schwartz et al., 2015).
Effector AvrHah1, a transcription activator-like (TAL) effector of the AvrBs3/PthA family (Schornack et al., 2008), was the first characterized effector of X. hortorum pv. gardneri. AvrHah1 was able to trigger a Bs3-dependent hypersensitive response (HR) on pepper plants (Schornack et al., 2008). Gain-of-function experiments with a X. euvesicatoria pv. euvesicatoria strain revealed that avrHah1 is responsible for enhanced water-soaking in pepper leaves, a phenotype typical for the compatible interaction of X. hortorum pv. gardneri, the donor pathogen (Schornack et al., 2008). The virulence function of AvrHah1, triggering enhanced water-soaking in its known hosts tomato, pepper, and Nicotiana benthamiana, was attributed to the movement of water into the infected apoplast (Schwartz et al., 2017). Gene avrBs7 was also identified in X. hortorum pv. gardneri as another avirulence gene as its product triggered an HR in Capsicum baccatum var. pendulum (Potnis et al., 2012). When the corresponding single dominant resistance gene Bs7 was introgressed into C. annuum 'Early Calwonder' (ECW), the resulting near-isogenic line ECW-70R was resistant to strains harbouring avrBs7.
Twenty-one candidate T3E genes were identified in X. hortorum pv. carotae M081, and the products of two of them, AvrBs2 and XopQ, were found to elicit effector-triggered immunity (Kimbrel et al., 2011

| D IS E A S E CONTROL AND MANAG EMENT
An integrated control programme that focuses on excluding, reducing, or eradicating the pathogen, in combination with various methods like biological control and host resistance breeding, is the most suitable to manage bacterial spot pathogens like X. hortorum (Agrios, 2005;Marin et al., 2019 Physical or chemical treatment of the planting material can decrease pathogen inoculum (Janse & Wenneker, 2002). Hot-water seed treatment reduced X. hortorum pvs carotae and gardneri infections. However, hot-water seed treatment can sometime be unsuitable. For example, a treatment at 50°C for 2 h of lettuce seeds against X. hortorum pv. vitians significantly reduced seed germination (Carisse et al., 2000). Some chemical seed treatments against this pathovar, such as soaking in 1% sodium hypochlorite for 5 min (Carisse et al., 2000), in 3% hydrogen peroxide for 5 min, or in suspensions of copper hydroxide plus mancozeb (Pernezny et al., 2002), were more effective in reducing seed contamination than others (copper hydroxide alone, benzoyl peroxide, or calcium peroxide). The Crop rotations or fallow periods could be used to eliminate contamination in plant debris by overwintering pathovars (Barak et al., 2001). In addition, good weed control and removing diseased plants can reduce inoculum amount (Barak et al., 2001;Toussaint et al., 2012), keeping in mind that some X. hortorum pathovars can survive epiphytically on weeds, and even infect them (see Epidemiology). Another good practice for decreasing the risk of disseminating X. hortorum pv. pelargonii involves not growing perennial Geranium spp. near greenhouse facilities producing Pelargonium spp. (Nameth et al., 1999).
However, a limitation of ASM is its adverse effect on tomato growth and yield, which may be attributed to the energy cost associated with resistance induction (Romero et al., 2001).

| HOS T RE S IS TAN CE
cerasiforme line PI 114490, possessed broad-spectrum resistance to multiple xanthomonad pathogens, including X. hortorum pv. gardneri.
Resistance in PI 114490 is quantitatively inherited, and QTL-3 locus and allele at QTL-11 are major contributors of resistance against X. hortorum pv. gardneri (Bernal et al., 2020).
A mutation in DMR6 (downy mildew resistance 6) in Arabidopsis, conferring broad-spectrum resistance to various Xanthomonas and Pseudomonas phythopathogens, was tested in tomato (Thomazella et al., 2016). The stable transgenic tomato plants were resistant against X. hortorum pv. gardneri and were not compromised in their growth and development.
In pepper, two dominant resistance genes, Bs3 and Bs7, are known to confer resistance against X. hortorum pv. gardneri strains carrying avirulence genes avrHah1 and avrBs7, respectively (Potnis et al., 2012;Schornack et al., 2008). However, the plasmidborne nature of both avirulence genes suggests vulnerability to resistance breakdown, so they have not been further considered in breeding programmes. Screening of core pepper germplasm collection against X. hortorum pv. gardneri revealed that more than 40 PI lines of C. baccatum in greenhouse conditions and multiple PI lines of C. annuum showed promising resistance levels (Potnis et al., 2012). A total of 20 significant single nucleotide polymorphisms (SNPs), co-located within 150 kb of 92 unique genes, were recently identified against the pathovar .
Regarding X. hortorum pv. vitians, different lettuce genotypes (L. sativa) show differential responses to the pathogen. For example, romaine and butterhead lettuce cultivars are among the highly susceptible ones (Pernezny et al., 1995). Moderately resistant cultivars include both green-leaf (e.g., Waldmann's Green and Grand Rapids) and red-leaf (e.g., Red Line) cultivars (Carisse et al., 2000), although other studies have noted their susceptibility (Bull et al., 2007). Such discrepancy could be due to differences in the experimental setups of the studies, such as using different strains for the pathogenicity tests.

Other moderately resistant cultivars include Little Gem and Reine des
Glaces (Batavia crisphead) (Bull et al., 2007). These two latter cultivars were deemed to be promising in breeding resistant cultivars against X. hortorum pv. vitians (Hayes, Trent, Mou, et al., 2014;Hayes, Trent, Truco, et al., 2014). However, undesirable traits (e.g., small size and low yield) associated with cv. Little Gem are of concern. Furthermore, this cultivar has also shown variable resistance in separate studies, making it an unattractive candidate (Bull et al., 2007;Lu & Raid, 2013). This difference could be due either to virulence dissimilarities at the strain level or to host susceptibility variation as a result of different environmental conditions at the cultivar evaluation locations (Lu & Raid, 2013).
In addition, resistance of cv. Reine des Glaces was also highly dependent on environmental conditions (Bull et al., 2007).
Genetic maps of various wild lettuce species like L. serriola, L. saligna, and L. virosa, have revealed multiple genes conferring broad resistance (McHale et al., 2009;Truco et al., 2013). However, L. saligna and L. virosa have compatibility issues, making hybridization difficult. The broad resistance in wild lettuce species have yet to be tested against X. hortorum pv. vitians.
The high genetic variability of the pathogen population is a challenge for breeding cultivars with durable resistance. Resistance against MLSA-based groups B, D, and E of X. hortorum pv. vitians was identified to be controlled by a single dominant locus, Xanthomonas resistance 1 (Xar1), in the Batavia heirloom cv. La Brillante Hayes, Trent, Mou, et al., 2014;Hayes, Trent, Truco, et al., 2014). Two other cultivars, Little Gem and Pavane, carry Xar1 alleles and are resistant to Californian isolates of X. hortorum pv. vitians.
Another locus identified as X. campestris vitians resistance (Xcvr) was found in the same linkage group (LG2) during the mapping of a PI 358001-1 × Tall Guzmaine population. The durability of Xar1 and Xcvr resistances in cv. La Brillante and PI 358001-1 raised concerns because of the high variability in the pathogen population (Hayes, Trent, Mou, et al., 2014;Hayes, Trent, Truco, et al., 2014). Major and minor quantitative trait loci (QTLs) controlling this resistance were identified and co-located in the same region of LG2 as previously identified with Xar1 and Xcvr (Sandoya et al., 2019).
A germplasm screening of carrot species (e.g., PI lines, public inbred lines, commercial cultivars, and wild varieties) indicated four PI lines (PI 263601, PI 418967, PI 432905, and PI 432906) and two of the wild relatives, Ames 7674 and SS10 OR, were the most resistant against X. hortorum pv. carotae (Christianson et al., 2015). The resistant PI lines are promising for use in commercial breeding programmes (Christianson et al., 2015).

| RE S E ARCH PER S PEC TIVE S
Several advances improving our understanding of the X. hortorum species have recently been published. However, some knowledge gaps remain, mainly related to extent of host range, detection, and control methods, including host resistance. Given the broad and diverse host range described for this species, it is likely that unreported hosts remain to be identified in various ecosystems. Further investigation of the natural and experimental host ranges of X. hortorum could provide insight into its evolutionary history and determine if plant domestication influenced host specialization of the pathovars of X. hortorum.
The recent increased availability of genomic data for X. hortorum will help in the identification of novel isolates from new natural hosts through establishing quick, field-deployable detection methods. Such tools will also be very beneficial for phytosanitary control, especially as prevention strategies are preferred to formulation applications and are less costly than containment and eradication measures.
There is a significant need to conduct a comprehensive comparative genomics analysis of this species, especially in view of the recent taxonomical changes. Because plasmids offer a potentially large source of variation in the species, determining the plasmid content of strains and their contribution to pathogenicity is highly relevant. The recent application of a TnSeq analysis in X. hortorum pv. vitians paves the way to functional genomics analysis of other X. hortorum members. Aside from providing insights into essential bacterial genes in different in vitro and in planta conditions, TnSeq would also considerably improve our understanding of X. hortorum biology.
Most important commercial varieties are still highly susceptible to diseases caused by X. hortorum. The inefficiency of applicationbased control strategies further consolidates host resistance as a promising area for devising practical and durable disease control solutions against X. hortorum. Because nonhost resistance is more durable than host resistance, screening more nonhost species for their disease response to X. hortorum could uncover broad, nonhost resistance genes against the pathovars. Furthermore, exploring recent advancements in the field of host resistance against bacteria, such as CRISPR/Cas9-mediated gene mutations, also sound promising for breeding X. hortorum-resistant plant cultivars. However, as highlighted throughout this pathogen profile, the high genetic variability of these phytopathogens affecting several plant families represents a real challenge for long-term resistance.

ACK N OWLED G EM ENTS
This article is based on work from COST Action CA16107 EuroXanth, supported by COST (European Cooperation in Science and Technology).