Xanthomonas hydrangeae sp. nov., a novel plant pathogen isolated from Hydrangea arborescens

20 This paper describes a novel species isolated in 2011 and 2012 from nursery-grown 21 Hydrangea arborescens cultivars in Flanders, Belgium. After four days at 28°C, the strains 22 yielded yellow, round, convex and mucoid colonies. Pathogenicity of the strains was 23 confirmed on its isolation host, as well as on Hydrangea quercifolia. Analysis using MALDI24 TOF MS identified the Hydrangea strains as belonging to the genus Xanthomonas but 25 excluded them from the species Xanthomonas hortorum. A phylogenetic tree based on gyrB 26 confirmed the close relation to X. hortorum. Three fatty acids were dominant in the 27 Hydrangea isolates: anteiso-C15:0, iso-C15:0 and summed feature 3 (C16:1ω7c/C16:1ω6c). 28 Unlike X. hortorum pathovars, the Hydrangea strains were unable to grow in presence of 29 lithium chloride and could only weakly utilize D-fructose-6-PO4 and glucuronamide. 30 Phylogenetic characterization based on multilocus sequence analysis and phylogenomic 31 characterization revealed that the strains are close to, yet distinct from, X. hortorum. The 32 © Dia et al., 2021. The definitive peer reviewed, edited version of this article is published in the International Journal of Systematic and Evolutionary Microbiology, 71, 12, 2021, doi:10.1099/ijsem.0.005163.


INTRODUCTION
The genus Xanthomonas, first described by Dowson in 1939, consists of Gram-negative plant pathogenic bacteria that are rod-shaped with a single flagellum and which produce yellow-pigmented colonies [1].The genus currently comprises 29 validly published species (https://lpsn.dsmz.de/genus/xanthomonas;accessed in September 2021), causing diseases in more than 400 plant hosts, including stone fruits, citrus, ornamental plants, and grasses such as rice and wheat [2].The original classification of the genus was based on a phenetic approach focusing on one phenotypic trait, namely host specificity.However, many Xanthomonas species were demonstrated to be heterogeneous based on a number of techniques, e.g., protein electrophoresis [3], restriction fragment length polymorphism (RFLP) on genomic DNA [4] and DNA-DNA hybridization [3,5].This led to an extensive revision of the Xanthomonas taxonomy, resulting in 20 nomenspecies based on DNA-DNA hybridization by Vauterin et al. [6].Parkinson et al. [7,8] later demonstrated the genetic relatedness between the different Xanthomonas spp.using partial gyrB gene sequences.In these studies, some strains were clustered into different species-level clades (slc), the members of which share a common ancestor before final speciation [8].One example of a species-level clade is the Xanthomonas hortorum slc.The relatedness of members of the X. hortorum slc was confirmed by two multilocus sequence analysis (MLSA) studies using partial sequences of four (dnaK, fyuA, gyrB and rpoD) [9] or even six (atpD, dnaK, efp, glnA, gyrB, and rpoD) [10] housekeeping genes.
In this study, we report on four strains isolated in Belgium between 2011-2012 from container-grown Hydrangea arborescens plants [21].The objective of this study was to characterize these strains through a polyphasic approach combining phylogenetic characterizations based on 16S rRNA gene sequencing, a MLSA of seven housekeeping genes (atpD, dnaK, efp, glnA, gyrB, lepA and rpoD) and a phylogenomic analysis.

ORIGIN AND ISOLATION
Hydrangea arborescens plants with leaf spot symptoms were received at ILVO's Plant Diagnostic Center (PDC) from three Belgian ornamental nurseries in 2011 and 2012 [21].
The symptomatic plants were from four different cultivars (i.e., cvs.Bella Anna, Annabelle, Invincibelle and Incrediball).Bacteria were isolated from leaf spots at ILVO-PDC.Four representative isolates were selected for further analysis: LMG 31884 T (= CCOS 1956 T , GBBC 2123 T ), LMG 31885 (= CCOS 1954, GBBC 2128), LMG 31886 (= CCOS 1957, GBBC 2199) and LMG 31887 (= CCOS 1955, GBBC 2202).LMG 31884 T was isolated in 2011 and originated from a Belgian nursery which had obtained rooted cuttings from its subsidiary company in Ethiopia.The mother plants to produce cuttings at the Ethiopian subsidiary were purchased in the USA.LMG 31887 was isolated in 2012 and was obtained from a second nursery that had received planting stock from the first nursery.LMG 31885 and LMG 31886 were isolated in 2011 and 2012, respectively, and originated from a third nursery which had imported the planting stock from the Netherlands.
Tracking and tracing of the origin of the plants suggests that there was more than one introduction of the pathogen.

GROWTH AND INITIAL CHARACTERIZATION
The bacterial isolates were first grown on Pseudomonas agar F medium (Life Technologies Europe BV, Merelbeke, Belgium) for four days at 28°C.They were subsequently grown on nutrient-yeast glycerol agar (NYGA, 5 g l −1 peptone, 3 g l −1 yeast extract, 20 g l −1 glycerol and 15 g l −1 agar) for two days at 28°C.On both media, yellow, round, convex and mucoid colonies were observed.
Isolates were stored in a 1:1 (v:v) solution of nutrient-yeast glycerol broth (NYGB) and glycerol at -80°C.Strains were routinely plated on NYGA and incubated at 28°C for 48 h.
The strains were deposited at the BCCM/LMG (Belgian Co-ordinated Collections of Microorganisms / Laboratory of Microbiology, Ghent University, Belgium) and the CCOS (Culture Collection of Switzerland, Wädenswil, Switzerland).Information on their host, country and year of isolation is given in Table 1.
The partial gyrB gene of the isolates was amplified using gyrB primers XgyrPCR2F and X.gyrrsp1, as described by Parkinson et al. [7], and purified PCR products were sequenced (Genewiz, Leipzig, Germany).Based on the gyrB phylogenetic tree, the isolates likely represented a new species, with X. hortorum as the closest relative [21].

PATHOGENICITY TESTS
The pathogenicity of two isolates (LMG 31884 T , and LMG 31886) was tested by pressurized spray inoculation on the leaves of H. arborescens cv.Annabelle, H. quercifolia, Hydrangea macrophylla cv.Zorro and Hydrangea paniculata cv.Pinky Winky, in addition to Hedera cv.Hibernica (Atlantic ivy), Pelargonium zonale, Taraxacum kok-saghyz (Russian dandelion) and Daucus carota cv.Nerac (carrot).Fresh, pure bacterial cultures of the bacterial strains were grown for 48 h at 28°C on NYGA-Sucrose media, and were used to prepare the inoculum into a mixture of 10 mM Phosphate-Buffer Saline (PBS) -0.02% Tween.The inoculum was diluted to 10 6 cfu/ml based on OD 600 measurement.The abaxial side of the leaves were pressurized-sprayed with either the inoculum (three plants/inoculum) or the PBS-Tween solution, serving as a control (one plant/inoculum).The plants were placed in plastic boxes in a greenhouse; the boxes were covered with a plastic sheet for two days to maintain high humidity.Plants were checked daily for symptoms and irrigated when needed.
Regular leaf spots developed only on H. arborescens at 28 days post-inoculation (Fig. 1a).
Few spots developed on H. quercifolia (Fig. 1b).No spots developed on H. macrophylla and on H. paniculata, while symptoms were not detected on the tested host plants for X. hortorum pathovars (Atlantic ivy, Pelargonium zonale, Russian dandelion, or carrot; data not shown).Furthermore, X. hortorum pv.carotae strain CFBP 7900 did not develop spots on H. arborescens (data not shown).The results show that the homologous plant host, H. arborescens, is the most susceptible of the Hydrangea species tested, but only for the Hydrangea strains.To fulfil Koch's postulates, selected spots were macerated in PBS and then plated in serial dilutions.Partial sequences of gyrB or rpoD from yellow, mucoid characteristic Xanthomonas colonies were determined and the sprayed strains were confirmed to be the causal agents of the observed symptoms (data not shown).

MALDI-TOF MS ANALYSIS
The four Hydrangea strains were analyzed through a Bruker MALDI-TOF MS Biotyper (Bruker Daltonics, Billerica, MA) identification system.Strains were grown in tryptic soy broth at 28°C with shaking at 220 rpm for 24 h before centrifuging 1 ml at 11,000 g for 2 min and inactivating the pellet by resuspending in 250 μl 70% ethanol for a minimum of 1 min.The inactivated cells were then pelleted as described above.The pellet was dried for a minimum of 5 min at room temperature, resuspended in 25 μl of 70% formic acid and diluted v/v with acetonitrile.After another centrifugation step as above, 1 μl supernatant was spotted on a steel standard target and air-dried completely.The spot was then overlayed with 2 μl of a 40 mg ml -1 alpha-cyano-4-hydroxy-cinnamic acid matrix and air-dried before analysis.Log(score) values were calculated against the Bruker database v..1 (6,903 strains, including 36 Xanthomonas species).
High confidence identification scores (2.00 to 3.00; secure species) were only obtained for X. hortorum pv.hederae CFBP 4925 T , reaching between 2.01 and 2.27 with various strains of X. hortorum pv.pelargonii and X. hortorum pv.cynarae (Table S1).The four Hydrangea strains yielded lower scores (scores ≤ 2.07; Table S1), with the majority of scores between 1.70 and 1.99 (low confidence identification; secure genus).The identification based on Bruker Biotyper thus shows that the strains undeniably belong to the genus Xanthomonas, with X. hortorum as the closest relative, but there is substantial doubt that the Hydrangea strains belong to X. hortorum as the log(score) values were too low to support a solid inclusion in the species.
vitians LMG 838 neoPT , were grown on NYGA plates at 28°C for 48 h before FAME analysis.
The fatty acid methyl esters were then extracted and identified following the protocol of Sherlock Microbial Identification System (MIDI, Newark, Delaware, USA).Peaks were identified using the TSBA6 method and the Sherlock library database v.6.1.0.

PHENOTYPIC CHARACTERIZATION
Utilization of carbon sources by strains LMG 31884 T and LMG 31885 and their resistance to certain chemical compounds were characterized using GEN III MicroPlates (Biolog, Hayward, United States).Experiments were carried out in triplicates.Strains were subcultured from glycerol stock, grown on NYGA medium for 48 h at 28°C and then grown overnight on Biolog Dehydrated Growth agar at 28°C.Fresh colonies were transferred into Biolog Inoculating Fluid A using cotton swabs.The inoculum density was adjusted to a transmittance of 95-98%.A volume of 100 µl of inoculum was pipetted into wells of the MicroPlate, which was incubated at 28°C.The MicroPlates were measured using a MicroStation 2 Reader (Biolog) after 72 hours of incubation.The utilization data was first compared to the water no-substrate control (NSC); results were considered positive if 160% higher than the NSC and negative if 130% or less than the NSC values, to subtract the nosubstrate baseline.Results falling between those two values were qualified as weak [22].
Differential phenotypes are listed in Table 2, while the full phenotypic profiles are reported in a supplementary table (Table S3).Across all triplicates, both LMG 31884 T and LMG 31885 unequivocally utilized 33 substrates, weakly utilized six substrates, and could not utilize five substrates (Table S3).For the rest of the substrates, the results were inconclusive as they were inconsistent across the replicates.Strains LMG 31884 T and LMG 31885 can be distinguished from all X. hortorum pathovars by their inability to grow in the presence of lithium chloride and their ability to weakly utilize D-fructose-6-PO 4 and glucuronamide (Table 2).

GENOME FEATURES
The DNA extraction, genome sequencing and subsequent assembly and annotation were carried out as previously described [23].The genome sequences of strains LMG 31884 T , LMG 31887, LMG 31885 and LMG 31886 were deposited in EMBL, and their accession numbers are ERZ1690675, ERZ1690678, ERZ1690676 and ERZ1690677, respectively (Table 3).The genome of LMG 31884 T is complete and consists of a circular chromosome and four circular plasmids pLMG31884_p76, pLMG31884_p73, pLMG31884_p46 and pLMG31884_p23 (Table 3).A total of 4,703 CDS were annotated for the LMG 31884 T genome.Its G+C content of 63.6 mol% is within the typical range of Xanthomonas genomes.
The genomes of the three other Hydrangea strains are complete, except one gap in strain LMG 31885 that could not be resolved, yielding the genome of this strain to be not circular.
Strains LMG 31887 and LMG 31885 harbor two plasmids each, while strain LMG 31886 has only one plasmid.
The partial 16S rRNA gene sequences of the type strains of Xanthomonas species, in addition to one Xyllela fastidiosa strain, ATCC 35879 T , serving as an outgroup, were retrieved from the List of Prokaryotic names with standing in nomenclature (LPSN; access date January 2021).The partial 16S rRNA gene sequences of the four Hydrangea strains and 13 X.hortorum strains were retrieved from their genomes.The 16S rRNA gene sequences were aligned using MEGA X v.10.1.7 [24] and a Maximum-Likelihood phylogenetic tree was constructed (1,227 positions, 1,000 bootstraps).The 16S rRNA genebased phylogeny (Fig. 2) displays two main clades within the genus Xanthomonas, with one of them including all X. hortorum slc strains.
The 16S rRNA gene sequences of strains LMG 31884 T , LMG 31887, LMG 31885 and LMG 31886 differed by one nucleotide at position 364: G for strains LMG 31884 T and LMG 31887, and A for strains LMG 31886 and LMG 31885.The 16S rRNA gene sequences of the Hydrangea strains, the X. hortorum strains and Xanthomonas type species X. campestris pv.campestris strain ATCC 33913 T differed by three nucleotides at positions 342, 364 and 894, corresponding to a variation of less than 0.3%.Furthermore, there were only 52 variable nucleotides (4.24% variation) over the full 16S rRNA gene sequences of the Xanthomonas strains used in this phylogenetic study.The 16S rRNA gene is not variable enough within the genus Xanthomonas to be sufficiently discriminative between species [25].
MLSA with multiple housekeeping genes provided higher discriminative phylogeny resolution within the genus [26].Twenty strains, namely the four strains from Hydrangea, thirteen strains of X. hortorum and three outgroup strains (X.arboricola pv.juglandis CFBP 2528 T , X. populi CFBP 1817 T and X. campestris pv.campestris ATCC 33913 T ), were used to conduct MLSA based on the almost-complete concatenated sequences of seven housekeeping genes atpD (1,407 bp), dnaK (1,926 bp), efp (567 bp), glnA (1,404 bp), gyrB (2,442 bp), lepA (1,791 bp) and rpoD (1,878 bp).Sequences of the seven housekeeping genes of X. hortorum pv.carotae M081 (CFBP 7900; Table S4) were used as a query to retrieve closest orthologs (Table S5) from the genomes using tBLASTn (50% coverage, 60% hsp identity).A maximum likelihood phylogenetic tree (Fig. 3) was constructed after alignment in MEGA X [24] using the General Time Reversible model and a Gamma distribution with invariant sites [27].Phylogeny was tested using 1,000 bootstrap replications, and bootstrap values below 50% were removed.The alignment of almost-complete concatenated sequences was used to trim the loci to the partial gene sequences previously described [9].A similar topology was obtained (Fig. S1).Trimming settings and housekeeping gene lengths are reported in Table S6.
The four Hydrangea strains form a separate cluster within the X. hortorum slc; the closest X. hortorum strain is X. hortorum pv.pelargonii CFBP 2533 PT , while its closest non-slc relative is X. populi CFBP 1817 T .This is again indicative that the four Hydrangea strains may represent a separate taxon.

WHOLE-GENOME PHYLOGENY
Genome sequences of the aforementioned 20 strains were uploaded to the Type (Strain) Genome Server (available at https://tygs.dsmz.de)for a whole genome-based taxonomic analysis [28].Intergenomic distances were used to infer a balanced minimum evolution tree with branch support via FASTME 2.1.4including SPR postprocessing [29].Branch support was inferred from 100 pseudo-bootstrap replicates each.The tree (Fig. 4) was rooted at the midpoint [30] and visualized with PhyD3 v.1.3[31].
Whole-genome phylogeny reveals that strains of the X. hortorum pathovars form two major subclusters: subcluster A, encompassing X. hortorum pv.gardneri, pv.cynarae, pv.vitians and pv.taraxaci; and subcluster B, including X. hortorum pv.pelargonii, pv.hederae and pv.carotae.X. hortorum subcluster B is the closest phylogenetic relative of the Hydrangea strains (Fig. 4).Thus, both the MLSA-based and whole-genome phylogenies revealed three clusters within the X. hortorum slc: X. hortorum subclusters A and B and the cluster formed by the four Hydrangea strains.

OVERALL GENOME-RELATEDNESS INDICES
To investigate the degree of relatedness of these three clusters, overall genome-relatedness indices ANI and isDDH values were calculated for the same 20 genomes used in the previous analyses (Fig. 5).ANI values were calculated using fastANI, an alignment-free whole-genome ANI method [32], and isDDH was calculated using the genome BLAST distance phylogeny approach (GBDP) [33].

X. hortorum subcluster B is the closest relative to the genomes of the four strains from
Hydrangea.Between those two entities, fastANI and isDDH values ranged between 94.84 and 95.19% ANI, respectively, and between 58.20 and 59.40% isDDH, respectively (Fig. 5).
The isDDH values were well below the thresholds for species delineation, and even below the isDDH transition zone of 60-70%.Some of the fastANI values were above 95%, with the highest being 95.19%.The fastANI and isDDH values for X. hortorum subcluster A, in comparison to the Hydrangea strains, ranged between 94.35% and 94.99% ANI, and between 55.70% and 58.60% isDDH (Fig. 5).Both ANI and isDDH ranges fall below the species threshold values of 95% and 70%, respectively, and are also below the transition thresholds.
Since our genome dataset includes genomes with varying degrees of completeness, fastANI is the preferred method as it is more robust than alignment-based methods like ANIm (ANI calculation based on MUMmer) or ANIb (ANI calculation based on BLAST+ alignments) in the presence of draft genomes [32].
Nevertheless, to further investigate the relatedness between the Hydrangea strains and its closest phylogenetic relative X. hortorum subcluster B, alignment based ANIm and ANIb values between those two clusters (Table S7) were calculated using JSpeciesWS v.3.8.2 [35], as these two methods are still very commonly used for Xanthomonas.
ANIm values between the Hydrangea strains and the three X.hortorum pathovars in subcluster B ranged between 94.87 and 95.15% ANIm, with 78.75 to 84.95% of the sequences aligned.ANIb values between those same groups ranged from 94.16 to 94.5% ANIb, with 82.6 to 75.9% of sequences aligned.ANIm values are in accordance with the alignment-free fastANI values; indeed, ANIm values between X. hydrangeae and X. hortorum pvs.carotae and hederae, similarly to those calculated using fastANI, fall in the transition-zone, while ANIb values are below 95%.
Between the two methods that rely on sequence alignment, ANIm is the most robust as evidenced by the increased sequence alignment percentage when compared to ANIb.This is consistent with published results, reporting that ANIm yields more accurate results than ANIb especially when genomes compared have ANI values higher than 90% [34].
When ANI and/or isDDH values between strains fall in the transition zone, the decision to consider the strains as either the same or a distinct species is based on other genomic, phylogenetic and/or phenotypic analyses.Thus, even though some ANI values between the Hydrangea strains and X. hortorum were at the lower range of the "transition zone", the remaining analyses, such as isDDH values below the "transition zone" and phenotypic traits discriminating between X. hydrangeae and all the X. hortorum pathovars, suggest the Hydrangea strains represent members of a novel species.By contrast, within X. hortorum pathovars, ANI and isDDH values do not warrant their separation into distinct species since other genomic, phenotypic, and phylogenetic analyses strongly point to the pathovars belonging to X. hortorum [12].

LOOP-MEDIATED ISOTHERMAL AMPLIFICATION-BASED DIAGNOSTICS ASSAY
For taxonomy, two aspects are important: definition of a species as given above, but also diagnostics, e.g., for the identification of this organism in laboratory settings or more preferably in the field.We developed a highly species-specific loop-mediated isothermal amplification (LAMP) assay that can distinguish the Hydrangea strains from all the pathovars of its closest relative X. hortorum and around 70 other Xanthomonas and non-Xanthomonas strains.This genomics-informed LAMP assay serves as an alternative to the phenotypic characterization.

In silico assay development
The assay was developed using a singleton coding sequence (CDS) approach on the EDGAR 3.0 platform [36].Four CDS were retained for further analysis (Table S8).LAMP primer sets (F3/B3, LoopF/LoopB and FIP/BIP) were designed for each of the four specific CDS using LAMP Designer v.1.16[37].After a primer-BLAST analysis, one primer set, Xhyd-B (Table 4), was selected for the in vitro testing of the assay.Primer set Xhyd-B targets a 150 bp region of a 1,116 bp chromosomal CDS (annotated as hypothetical protein).

Performance of LAMP assay in a laboratory setting
Amplification was carried out on a LightCycler480 (Roche, Basel, Switzerland).Primer set Xhyd-B was first tested on boiled cells of the Hydrangea strains and had a mean amplification time of 9.62 min.This primer set was further tested on the normalized genomic DNA (1 ng/µl) of 88 strains (Table S9), namely ten strains isolated from Hydrangea plants including the four strains characterized in this study, 39 X. hortorum slc strains and 39 strains representing 30 non-X.hortorum slc species.Only the target strains were amplified within 20 min (Table S9), thereby confirming the presence of six further strains closely related to the Hydrangea ones.Assay performance metrics were calculated as reported by Blaser et al. [38].The specificity, sensitivity and efficiency of the Xhyd-B LAMP assay were 100%, 97.3% and 97.7%, respectively.

Performance of LAMP assay on blind samples
The performance of the LAMP assay was tested on thirteen blind spot samples.Briefly, Hydrangea plants were inoculated at ILVO (Merelbeke, Belgium) as previously described [21].Upon being received at ZHAW, each spot specimen (ca. 1 cm 2 ) was diagonally cut, and one part of the spot was transferred into a 15 ml tube with 5 ml of NYGB and incubated for 48 h with shaking at 220 rpm at 28°C.Around 1-2 ml of suspension was then transferred into a 1.5 ml tube in duplicates and centrifuged at 20,000 g for 5 min.The pellets were resuspended in 200 µl of ddH 2 O.The suspensions were boiled at 95°C and directly used as DNA template for the LAMP reaction as previously described.For each sample, two suspensions were prepared, and measurements were taken in triplicates.A sample was only considered positive if all six instances were positive (e.g., amplification within 15-20 min).
Samples with inconclusive results (e.g., inconsistent results between the biological duplicates or the experimental triplicates) were measured a second time.
Results of the LAMP assay were compared with isolation assays performed at ILVO from identical spot symptoms on leaves from the same plant sent to ZHAW.The re-isolated strains were confirmed to be identical to the spray-inoculated strains based on their gyrB sequences.The LAMP assay correctly assessed 11 out of the 13 samples, an accuracy based on blind samples of around 85% (data not shown).One sample was a false negative and the other one a false positive.

Evidence for assigning the Hydrangea strains to a new species
We have provided evidence which shows that the four Hydrangea strains represent members of a novel species, phylogenetically close, yet distinct to X. hortorum.Of the tested plants, H. arborescens is the only host for the strains.Outcomes of the Bruker MALDI-TOF MS analysis showed that the Hydrangea strains should be included in the genus Xanthomonas, but not within the species X. hortorum.This was confirmed by ANI and isDDH values for the Hydrangea strains cluster, which were well below the thresholds for species delineation.Although only few phenotypic traits allow a discrimination from X. hortorum, the phylogenetic and phylogenomic analyses showed that the Hydrangea strains consistently form a distinct, distant cluster to the species X. hortorum.For this reason, we propose the name Xanthomonas hydrangeae sp.nov.as a new member within the genus Xanthomonas to accommodate the Hydrangea strains, with strain LMG 31884 T (CCOS 1956 T =GBBC 2123 T ) as its type strain.Along with the four strains characterized in this study, six further strains were identified by the LAMP assay that is highly specific for this novel species.
Cells are Gram-negative, motile, non-sporulating straight rods, and form colonies that are round, convex and mucoid when grown on Pseudomonas agar F for 4 days or nutrient-yeast glycerol agar for 2 days at 28°C.Produce yellow pigment.Catalase-and oxidase-negative.
The optimal temperature for growth is between 25°C and 28°C.Grow at pH 7.0 and 6.0 but not at pH 5.0, and in the presence of NaCl concentrations up to 1%.Weak growth was observed at NaCl concentrations of 4%, while no growth was obtained at 8%.Grow in the presence of 1% sodium lactate, tetrazolium violet or tetrazolium blue, and in presence of antibiotics rifamycin or lincomycin, but not in presence of lithium chloride, sodium bromate, and potassium tellurite.is marked with a triangle, and X. hortorum subclusters A and B are marked with a diamond and a square, respectively.The phylogeny was inferred from the genome BLAST distance phylogeny approach (GBDP).The approach consists of two steps: a BLAST genome comparison followed by distance matrix computation to construct a phylogenetic tree.The substrates that can distinguish the Hydrangea strains from X. hortorum are indicated in bold.The complete results for the biochemical panels tested below are listed in Table S3.Strains: 1: LMG 31884 T ; 2: LMG 31885; 3: X. hortorum pv.pelargonii CFBP 2533 PT ; 4: X. hortorum pv.carotae CFBP 7900; 5: X. hortorum pv.hederae CFBP 4925 T ; 6: X. hortorum pv.taraxaci NCPPB 940 PT ; 7: X. hortorum pv.vitians LMG 938 neoPT ; 8: X. hortorum pv.cynarae CFBP 4188 PT ; 9: X. hortorum pv.gardneri ATCC 19865 PT ; 10: X. campestris pv.campestris ATCC 33913 T .

Furthermore, pathogenicity
tests, Matrix Assisted Laser Desorption Ionization -Time of Flight Mass Spectrometer (MALDI-TOF MS) peptide mass fingerprinting, phenotypic profiling and a fatty acid methyl esters (FAME) analysis were carried out.Overall-genome relatedness indices (i.e.Average Nucleotide Identity ANI and in silico DNA-DNA-Hybridization isDDH) were calculated.Moreover, we propose a comparative genomicsbased, loop-mediated isothermal amplification (LAMP) diagnostics method for the rapid and selective identification and detection of this new species.

Fig. 2 .
Fig. 2. Maximum-likelihood tree based on partial 16S rRNA gene sequences, constructed using the General Time Reversible model.There were a total of 1,227 positions in the final dataset.Evolutionary analyses were conducted in MEGA X. Xanthomonas hydrangeae strains are in bold font and strains of the Xanthomonas hortorum species level clade are marked with rectangle boxes.The 16S rRNA gene sequence and contig accessions are noted in grey.When no 16S rRNA gene accession is available, contig accession in addition to the gene's position on the contig is reported between brackets.Percent bootstrap support values calculated for 1,000 iterations are indicated near nodes only when over 51.

Fig. 3 .
Fig. 3. Maximum-likelihood tree based on MLSA using concatenated sequences of atpD, dnaK, efp, glnA, gyrB, lepA and rpoD (total 11,415 bp).The tree highlights the phylogenetic relationship of Xanthomonas hydrangeae LMG 31884 T , LMG 31887, LMG 31885 and LMG 31886 (bold, circle), in relation to Xanthomonas hortorum subclusters A and B, marked with a diamond and a square, respectively.The X. hortorum species level clade is marked with an arrow.Percent bootstrap support values calculated for 1,000 iterations are indicated near nodes only when over 51.

Table 1 .
Isolation information of Xanthomonas hydrangea strains used in this study.

Table 2 .
Selected differential phenotypic characteristics of Hydrangea strains and its closest phylogenetic relative Xanthomonas hortorum characterized in triplicates using Biolog GEN III MicroPlates.

Table 3 .
Genome metrics of Xanthomonas hydrangea strains.

Table 4 .
LAMP primer set Xhyd-B used for the identification and detection of Xanthomonas hydrangea.

Table 1 .
5. Bi-directional table showing average nucleotide identity (fastANI, upper half of table) and in silico DNA-DNA hybridization (isDDH, lower half of table) values between Xanthomonas hydrangeae, its closest phylogenetic relative Xanthomonas hortorum and three outgroup strains, Xanthomonas populi, Xanthomonas arboricola and Xanthomonas campestris.List of Xanthomonas hydrangeae strains used in this study.BCCM/LMG: the Belgian Coordinated Collections of Microorganisms/ bacteria collection Laboratory of Microbiology at Ghent University, Belgium; CCOS: Culture Collection of Switzerland in Wädenswil, Switzerland; GBBC: Gewasbescherming Bacteriële Collectie (English: Plant Health Unit Bacterial Collection) at ILVO, Merelbeke, Belgium.

Table 2 .
Selected differential phenotypic characteristics of Hydrangea strains and its closest phylogenetic relative Xanthomonas hortorum characterized in triplicates using Biolog GEN III MicroPlates.