Genome-Based Taxonomy of Species in the Pseudomonas syringae and Pseudomonas lutea Phylogenetic Groups and Proposal of Pseudomonas maioricensis sp. nov., Isolated from Agricultural Soil

Species in the phylogenetic group Pseudomonas syringae are considered the most relevant plant pathogenic bacteria, but their taxonomy is still controversial. Twenty named species are validated in the current taxonomy of this group and in recent years many strains have been genome-sequenced, putative new species have been proposed and an update in the taxonomy is needed. A taxonomic study based on the core-genome phylogeny, genomic indices (ANI and GGDC) and gene content (phyletic pattern and Jaccard index) have been applied to clarify the taxonomy of the group. A phylogenomic analysis demonstrates that at least 50 phylogenomic species can be delineated within the group and that many strains whose genomes have been deposited in the databases are not correctly classified at the species level. Other species names, like “Pseudomonas coronafaciens”, have been proposed but are not validated yet. One of the putative new species is taxonomically described, and the name Pseudomonas maioricensis sp. nov. is proposed. The taxonomies of Pseudomonas avellanae and Pseudomonas viridiflava are discussed in detail as case studies. Correct strain identification is a prerequisite for many studies, and therefore, criteria are given to facilitate identification.


Introduction
The genus Pseudomonas encompasses a large number of species, with more than 270 named species in the current taxonomy, as stated in the List of Prokaryotic Names with Standing in Nomenclature [1] (https://www.bacterio.net/,last accessed in 22 October 2023).Pseudomonas species are ubiquitous, but some are preferentially present in the environment, including soil (Pseudomonas putida group) or fresh water (Pseudomonas fluorescens group); others are known to be pathogenic to humans (Pseudomonas aeruginosa group) or insects (Pseudomonas entomophila), and finally, some are related to plants, including those that are beneficial (Pseudomonas protegens group) and those that are pathogens (Pseudomonas syringae group) [2].Pseudomonads can cause diseases in an array of plant species [3][4][5], and P. syringae is considered the top species within the top 10 plant pathogenic bacteria [6]; however, its taxonomy is controversial, and many strains are included in the so-called P. syringae species complex, which is phenotypically divided into more than 60 pathovars that can cause diseases in monocots, dicots and woody plants [7].Pathovars are not always supported by phylogenetic traits, but at least a total of thirteen phylogenetic groups have been devised based on the similarity of housekeeping genes [8].
Several phylogenetic groups of species can be distinguished in the genus Pseudomonas [2], and most of the plant pathogenic species are included in the P. syringae phylogenetic group, comprising the P. syringae species complex and the named Pseudomonas species, including P. amygdali (which includes P. meliae, P. ficuserectae and P. savastanoi), P. asturiensis, P. avellanae, P. cannabina, P. caspiana, P. caricapapayae, P. cerasi, P. cichorii, P. congelans, P. syringae, P. tremae and P. viridiflava, as described by Gomila et al. in 2017 [9].Since that publication, nine new species have been described in the group (P.alliivorans, P. capsici, P. floridensis, P. folii, P. lijiangensis, P. quasicaspiana, P. triticifolii and P. ovata)."P.coronafaciens" also belongs to the group, but the name has not been validated.Numerous strains have been genomesequenced, and putative new species have been proposed, highlighting the necessity for an updated taxonomy because genomic analysis has profoundly changed the taxonomy of the genus [2,10].With this objective in mind, a genome-based taxonomic study was carried out, including species in the closely related phylogenetic group of Pseudomonas lutea that includes the following Pseudomonas species: P. abietaniphila, P. bohemica, P. graminis, P. lutea and P. petrae.One of the new putative species, represented by strain S25T, is phenotypically, chemotaxonomically and genomically described to propose Pseudomonas maioricensis sp.nov., with strain S25T (= CCUG 69272T, CECT 30911T) as the type strain.

Pseudomonas Strains and Growth Conditions
The strains studied are detailed in Table 1, which includes all type strains of the P. syringae and P. lutea groups, the related reference strains considered in the GTDB taxonomy, and other Pseudomonas strains not classified at the species level or incorrectly classified and whose genomes are available in the NCBI database (https://www.ncbi.nlm.nih.gov/, last visited in 22 October 2023).Two strains considered to be putative new species in previous publications were included in the analysis: strains S25T [9] and JDS28PS106 [11].Strain S25T was isolated in September 2010 from an agricultural soil sample collected at the campus of the University of the Balearic Islands in Mallorca, Spain (39 • 38 11.8 N, 2 • 38 50.1 E), during a screening for environmental bacteria.From a suspension of 2 g of soil in 8 mL of Ringer (Merck), 100 µL was plated on Middlebrook agar supplemented with Middlebrook Oleic Albumin Dextrose Catalase Growth Supplement (Middlebrook OADC enrichment, Difco, Madrid, Spain) under aerobic conditions.The plates were incubated at 30 • C for four days.Colonies were checked for purity on agar-solidified LB medium (lysis broth) [12].Strain JDS28PS106 was isolated from a water sample taken from the Danube River and proposed as the new phylospecies PS8 in a multilocus sequence analysis [11].Bacteria were cultured routinely at 30 • C on LB media.

Genome Sequencing
Genomic DNA was isolated from the S25T and JDS28PS106 strains using a Wizard Genomic DNA Purification Kit (Promega) according to the manufacturer's instructions.The paired-end library reads obtained from the Illumina HiSeq 2000 platform were assembled de novo using the Newbler Assembler v2.7 program (Roche).The draft genomes were annotated using the NCBI Prokaryotic Genome Annotation Pipeline (PGAP).The wholegenome shotgun sequencing data of strains S25T and JDS28PS106 have been deposited into DDBJ/EMBL/GenBank under the accession numbers LOHG00000000.1 (BioProject PRJNA305113) and JBAJGT000000000 (BioProject PRJNA1048145), respectively.

Phylogenomic Analysis
The genomic relatedness of all species type strains, GTDB reference strains and closely related strains in the P. syringae and P. lutea phylogenetic groups of publicly available species (comprising a total of 74 genomes) was determined based on the genome-aggregated average nucleotide identity (ANI) determined with the BLAST algorithm.It was calculated using the JSpecies software tool accessible at http://jspecies.ribohost.com/jspeciesws/(last accessed in 22 October 2023) [13].Additionally, digital DNA-DNA hybridisation between the selected strains was performed by the genome-to-genome distance (GGDC) method.The GGDC was calculated using the web service http://ggdc.dsmz.de,last accessed in 22 October 2023 [14], and the recommended BLAST method.The GGDC results presented herein were based on the recommended Formula 2. The similarity of the genomes of closely related species-type strains was also calculated using the Type Strain Genome Server (TYGS), a free bioinformatics platform (https://tygs.dsmz.de,accessed on 22 October 2023) [15].The recommended thresholds considered for species differentiation are 95-96% for ANI [13] and 70% for GGDC [14].
The M1CR0B1AL1Z3R web server (https://microbializer.tau.ac.il/, last accessed in 22 October 2023) [16] was used to identify orthologous genes in the genomes.Briefly, the server extracts the orthologous sets of genes in each genome and analyses the gene presence-absence patterns.The settings applied were as follows: maximal e-value cut-off: 0.01; percent identity cut-off: 70.0%; and minimal percentage for the core: 100.0%.The concatenated sequences of the resulting core proteomes were used to infer the phylogeny of the strains, and the results were visualised by the server in a RAxML tree.Two strains were considered members of the same phylogenomic species (pgs) when both were affiliated with the same phylogenetic branch, and the average ANI and GGDC values were above the established species thresholds.
The Jaccard similarity index implemented in the PAST package of programs [17] was used as a measure of similarity in pairwise comparisons.The Jaccard index of similarity was calculated as SJ = a a+b+c+d , in which a is the number of genes that were present in both genomes of each pair, b and c are the number of genes present in one strain but absent in the other, and d is the number of orthologues in the group of strains studied that are absent in both of the compared strains.The final matrix was represented in a UPGMA dendrogram with PAST [17].The orthologous genes shared by the pairs of strains were also represented in a split-tree decomposition.SplitsTree software (version 5) was used for computing unrooted phylogenetic networks from molecular sequence data, as discussed by Huson and Bryant [18].The phyletic pattern obtained with M1CR0B1AL1Z3R, that is, the presence or absence of the orthologous genes in each strain, was also represented in a heat plot generated using the PAST program.A Venn diagram was constructed on the web page https://bioinfogp.cnb.csic.es/tools/venny/last accessed on 22 October 2023 [19] to differentiate the closely related strains in the P. caspiana phylogenetic branch and to identify exclusive and shared genes.Individual genes were identified by BLAST in the NCBI database (https://www.ncbi.nlm.nih.gov/, last accessed on 22 October 2023) and in the Pseudomonas genome database (https://www.pseudomonas.com/,last accessed on 22 October 2023) [20] or were annotated with Rapid Annotation using Subsystems Technology (RAST; https://rast.nmpdr.org/,last accessed on 22 October 2023) [21].Phylogenies were also constructed with the autoMLST website [22] as previously reported [2] and were also retrieved from the GTDB website (https://gtdb.ecogenomic.org/,last accessed in 22 October 2023) [23].

Genomic Insights
The following databases were used to detect diagnostic or relevant genes related to the taxonomy of the new species according to the S25T genome: the NCBI; the Kyoto Encyclopedia of Genes and Genomes (KEGG; https://www.genome.jp/kegg/,last accessed in 22 October 2023) [24]; the Virulence Factors of Pathogenic Bacteria database (http://www.mgc.ac.cn/VFs/, last accessed in 22 October 2023) [25]; the P. syringae Genome Database (http://www.pseudomonas-syringae.org/,last accessed in 22 October 2023); the PHASTER (PHAge Search Tool Enhanced Release) web server for the identification and annotation of prophage sequences within bacterial genomes (http://phaster.ca/,last accessed in 22 October 2023) [26]; and the toxin-antitoxin Finder (TA Finder; http: //202.120.12.133/TAfinder/index.php, last accessed in 22 October 2023).If not otherwise stated, a gene was considered present following the 50/50 criterion (similarity higher than 50% in at least 50% of the sequence coverage) with a gene in the database.The predicted functional categories and COG classifications were established via protein BLAST against all coding gene sequences in the COG database available at NCBI.Antibiotic resistance genes were identified in the genome with the Comprehensive Antibiotic Resistance Database (CARD) [27].

Characterisation of Strain S25T
Cell size, morphology and flagellar insertion were determined via the transmission electron microscopy of cells from the exponential growth phase in LB.A Talos F200i electron microscope (Thermo Fisher, Barcelona, Spain) was used at 80 kV.The samples were negatively stained with uranyl acetate in aqueous solution (2%) as described by Lalucat [28].The presence of fluorescent pigments was tested on King B media (Pseudomonas agar F, Difco, Madrid, Spain), and pyocyanin production was tested on King A media (Pseudomonas agar P, Difco).The S25T strain and other type strains in the study were characterised phenotypically using API 20 NE strips (bioMérieux) and Biolog GN2 MicroPlates (Biolog, Hayward, CA, USA).Plant pathogenicity tests of the strain S25T were performed on citrus leaves as described by Beiki and collaborators [29].Briefly, the strains were cultivated on nutrient agar at 25 • C for 24 h, after which the cells were suspended in sterile distilled water at approximately 1 × 10 8 CFU/mL.One hundred microlitres of solution was injected into the intercellular space of the orange leaves with a 0.5 mm needle, after which the plants were incubated at 20 • C. Symptoms were observed for one week.
To determine the whole-cell protein profile, a matrix-assisted laser desorption/ionisation time-of-flight mass spectrometry (MALDI-TOF MS) analysis of the S25T strain was performed, together with its closely related strains, at the Scientific-Technical Services of the University of Balearic Islands (Spain) as described by Sanchez et al. [30].The profile obtained for each species was analysed and compared, and the corresponding dendrogram was generated using MALDI BioTyper software (version 1.0; Bruker Daltonics).
Whole-cell fatty acid methyl ester (FAME) analysis was performed at the Spanish Type Culture Collection (CECT), Valencia, Spain.The strains were cultured for 24 h in Trypticase Soy Agar at 28 • C. Fatty acids were extracted and prepared according to standard protocols as described for the MIDI Microbial Identification System [31].The cellular fatty acid content was analysed via gas chromatography with an Agilent 6850 unit, the MIDI Microbial Identification System using the RTSBA6 method [32] and the Microbial Identification Sherlock software package version 6.1.
Experimental DNA-DNA hybridisation was performed as previously described [30].

Phylogenomic Analysis
The core proteome was calculated with the M1CR0B1AL1Z3R web server for all the genomes using the P. aeruginosa type strain as an outgroup.The concatenated gene sequences of the 1505 core genes (a total core length of 552,494 nt) allowed us to infer the phylogenetic tree depicted in Figure 1 with RAxML.All branches were supported by high bootstrap values (95 of the values were 100, and only 4 values were 97-99 based on 100 replicates).The strains were clustered into five main branches represented by P. syringae, P. cichorii, P. ovata, P. caspiana and P. lutea.Two strains, v388 and JDS28PS106, were located on a different branch.The OWCr strain classified as P. caspiana was clearly outside these groups belonged to the P. mandelii subgroup of the fluorescens group and was not further studied.To obtain a greater number of orthologous genes shared by each pair of strains, the analysis was also repeated without an outgroup.The core proteome consisted of 2015 genes (a total core length of 727,877 nt), and the main branches were maintained as depicted in Figure 2. Clustering was in accordance with the GTDB taxonomy, the autoMLST phylogeny and previous studies, but the branching order was slightly different.The P. lutea and P. ovata branches were located between the P. syringae and the P. caspiana branches when the tree was constructed without an outgroup.The ANI in the pairwise comparisons was calculated for all the genomes.The matrix is represented in the corresponding UPGMA dendrogram in Figure 3.The main branches were maintained.In total, 50 phylogenomic species could be differentiated.P. meliae, P. savastanoi and P. ficuserectae belonged to the P. amygdalii phylogenomic species that was previously reported.The five putative novel species proposed in a previous publication [9], species A, B, C, D and E, were confirmed to be different pgs.In the present study, species A was included in Pseudomonas_E syringae_M (strain B728a), B was included in Pseu-domonas_E syringae_Q (strain CC1583), C was included in P. triticifolii (strains CC1417 and CC1524), species D was included in Pseudomonas_E sp002699985 (strain UB246) and E was included in the newly proposed species P. maioricensis (strain S25).Twenty-two of the 50 pgs defined are currently named species; the rest are representatives of potential novel species and were mainly singletons.A high percentage of the strains whose genomes have been deposited in the NCBI database are not correctly classified at the species level in the current taxonomy, as indicated in Table 1, and will be discussed later.For example, 13 P. syringae strains were phylogenetically related to 10 different pgs, some of which were in the P. caspiana phylogenetic branch.Additionally, 17 deposited genomes from strains not identified at the species level corresponded to 13 potential new species.The ANI in the pairwise comparisons was calculated for all the genomes.The matrix is represented in the corresponding UPGMA dendrogram in Figure 3.The main branches were maintained.In total, 50 phylogenomic species could be differentiated.P. meliae, P. savastanoi and P. ficuserectae belonged to the P. amygdalii phylogenomic species that was previously reported.The five putative novel species proposed in a previous publication [9], species A, B, C, D and E, were confirmed to be different pgs.In the present study, species A was included in Pseudomonas_E syringae_M (strain B728a), B was included in Pseudomonas_E syringae_Q (strain CC1583), C was included in P. triticifolii (strains CC1417 and CC1524), species D was included in Pseudomonas_E sp002699985 (strain UB246) and E was included in the newly proposed species P. maioricensis (strain S25).Twenty-two of the 50 pgs defined are currently named species; the rest are representatives of potential novel species and were mainly singletons.A high percentage of the strains whose genomes have been deposited in the NCBI database are not correctly classified at the species level in the current taxonomy, as indicated in Table 1, and will be discussed later.For example, 13 P. syringae strains were phylogenetically related to 10 different pgs, some of which were in the P. caspiana phylogenetic branch.Additionally, 17 deposited genomes from strains not identified at the species level corresponded to 13 potential new species.Notably, the species P. avellanae was one of the most abundant.The GTDB taxonomy includes 353 strains in the species cluster, but only 17 are classified in the NCBI taxonomy as P. avellanae.The other strains, including pathovar reference strains, were assigned to different species: 310 in P. syringae, 20 in P. viridiflava, and 6 in P. amygdali, reflecting difficulty in species identification.A detailed study of 44 P. avellanae strains by M1CR0B1AL1Z3R, ANI and GGDC (Supplemental Figure S1) revealed that two branches are phylogenetically differentiated.One group was represented by P. avellanae BPIC631T, and the other was represented by P. syringae pv.tomato DC3000.The ANI and GGDC values among members of both groups were less than 96% ANI and less than 70% GGDC; the intragroup values were higher than 96,5% ANI and 78% GGDC in the P. avellanae group and higher than 98% ANI and 90% GGDC in the DC3000 group.
The strains analysed that were assigned to "P. coronafaciens" in the NCBI taxonomy were classified as P. tremae in the present study and in the GTDB taxonomy.Therefore, additional strains classified as "P.coronafaciens" in the NCBI taxonomy were analysed on the autoMLST web server, together with the P. tremae and "P.coronafaciens" type strains.The results are given in Supplementary Figure S2.Both species are monophyletic, and the ANI indices among all the strains are greater than 98%.Additionally, these species must be considered members of the same phylogenomic species.

P. viridiflava Case Study
One thousand five hundred and twenty-two genome-sequenced strains of the Bioproject PRJEB24450 assigned taxonomically to the species P. viridiflava [33] were analysed in the present study to test the superiority of the phylogenomic approach in the current taxonomy.The phylogeny was inferred using the GTDB and the AutoMLST website.Most sequences from the RefSeq collection in the NCBI database were suppressed because of an unverified organism source, although the isolate identities were reported on the web site and in the original publication [33]; therefore, they were included in our study.As indicated in Table 2, our results demonstrated that most of the strains (90%) were correctly assigned to P. viridiflava, but 164 (10%) belonged to 37 different pgs (15 named species and 22 putative new species).Eleven pgs were phylogenetically placed in the P. lutea/P.syringae groups (43 strains), but the majority of the strains belonged to the P. fluorescens group or were located close to other phylogenetic groups (Table 2).

Gene Content Comparisons
The core proteomes of the 74 genomes in the P. syringae and P. lutea groups calculated by M1CR0B1AL1Z3R consisted of 2015 genes of the 23,319 orthologues.The genes shared by each pair of strains were studied in three different ways: the Jaccard index, split tree decomposition and heatmap analysis.The grouping of strains by the three methods confirmed the main phylogenetic branches defined by the core proteome and are depicted in Figure 4a-c (Jaccard, Splits and heatmap, respectively).The strains in each cluster were at least 0.55 similar according to the Jaccard index.The presence/absence of the 23,319 orthologous genes in each strain is clearly presented in the matrix plot (Figure 4c heatmap).A total of 2015 core genes were present in all the strains, and the others were preferentially distributed within the phylogenetic branch; however, it was not possible to detect a significant group of genes that were present in all the strains in a branch and that were absent in the other groups.1.
Strain S25T was clearly representative of a phylogenomic species that is different from the others; therefore, it was taxonomically described.Strain S25T was proposed in the present study as the type strain of a novel species, Pseudomonas maioricensis.The closest type strain species identified in the ANI and GGDC analyses were P. caspiana (86% ANI and 32.7% GGDC) and P. quasicaspiana (88 ANI and 33.4 GGDC).Experimental DNA-DNA hybridisation (DDH) was performed to confirm the genomic differentiation of strain S25T from P. caspiana strains.The labelled DNA of strain S25T and P. caspiana FBF102T strain was hybridised separately with DNA from selected type strains of the P. syringae phylogenetic group and from other P. caspiana strain.The labelled DNA of the S25T and FBF102T strains presented DDH values less than 60%, with all the type strains tested, and did not meet the 70% threshold established for species DNA-DNA similarity.The DDH values between S25T and FBF102T were less than 50%.These experimental results confirmed the results of the digital DNA-DNA hybridisation: strain S25T is not a member of any other Pseudomonas species tested (Table S1).

Genome Insights
The gene content of strain S25T was studied in detail by comparing this strain with four other selected strains in the P. caspiana branch, including the three species type strains and the strain proposed as "P.phytophila".The analysis performed with the M1CR0B1AL1Z3R web server identified 4018 genes shared by all five strains and a significant number of exclusive genes (called "orphan genes" or singletons) that were not shared with any other strain in the study (Table S2).The 281 exclusive genes of strain S25T were analysed manually.Those containing a high number of unresolved nucleotides (N) were discarded, and the rest of the sequences were concatenated and annotated with RAST.The results are summarised in Table S3.Two hundred and twenty-five proteins were annotated as hypothetical proteins; 43 were assigned a function, and 5 were predicted to be phage proteins.The orthologous genes shared among the five selected strains are represented in a Venn diagram (Figure 5).The exclusive genes of each strain in the diagram are orthologues also present in some other strains in the set of 74 genomes studied in the P. syringae-P.lutea phylogenetic groups.Given the importance of the strains of this group as potential plant pathogens, the genes involved in pathogenesis were studied in detail.Supplementary Table S5 summarises the relevant genes found in the genome of strain S25T: (i) a complete set of genes for flagellation and chemotaxis; (ii) essential genes for pyoverdin synthesis; and (iii) genes involved in adherence, such as a complete set of alginate genes (regulation, polymerisa- The analysis performed with the KEGG website allowed us to infer the metabolic characteristics of strain S25T.No significant differences could be found between strain S25T and the other four closely related strains included in the study.All the strains contained enzymes for a functional Entner-Doudoroff pathway for the catabolism of sugars, for starch hydrolysis, for the ortho cleavage pathway, for the degradation of aromatics, and for the reduction of nitrate to ammonia but not for denitrification.The five strains had secretion systems of types 2 and 6 (T2SS and T6SS).
To identify strain-or species-specific genes, a prophage survey was performed on the PHASTER server.As shown in Table S4, no complete prophage was found in strain S25T, although two incomplete phages were detected (9.8 Kb-7 proteins; and 8.1 Kb-9 proteins).No prophages were found in the genomes of P. caspiana FBF102T or P. quasicaspiana CDFA 553T.A region with a complete phage with 49 proteins was found in P. folii DOAB 1069T, and two intact and two incomplete phages were found in "P.phytophila" ICMP 23753.
Given the importance of the strains of this group as potential plant pathogens, the genes involved in pathogenesis were studied in detail.Supplementary Table S5 summarises the relevant genes found in the genome of strain S25T: (i) a complete set of genes for flagellation and chemotaxis; (ii) essential genes for pyoverdin synthesis; and (iii) genes involved in adherence, such as a complete set of alginate genes (regulation, polymerisation, acetylation, epimerisation and the export of alginate), and type IV pili synthesis with a role in biofilm formation; and (iv) genes for antimicrobial activities that can confer a competitive advantage.Several efflux pumps related to fluoroquinolone transport were found and are listed in Table S6.The AcrAB genes encode the acriflavine resistance protein and, together with TolC, constitute a tripartite complex of an efflux pump that may mediate resistance against host-derived antimicrobial peptides and is associated with antibiotic resistance (Table S6).At least three types of possible toxin and antitoxin systems were found using the KEGG website: HigA/HigB, HigA-1/HigB-1 and mqsA/mqsR.

Cell Morphology, Physiology and Biochemical Characterisation
The S25T strain is a Gram-negative, rod-shaped bacterium with polar flagellation, as shown by the electron microscopy images in Figure S3.After incubating for 48 h at 30 • C, the colonies were round (1-1.5 mm in diameter), flat, beige coloured, had regular margins and were translucent.The S25T strain was positive for catalase and oxidase activities, was strictly oxidative (not fermentative) and was not a denitrifier.The S25T strain was able to grow in LB media at 4-30 • C for 24-48 h after one week of incubation.No growth was detected at 42 • C or at 37 • C. Growth was observed on nutrient broth in the presence of NaCl until a concentration of 6% (w/v) was reached and the tolerated pH ranged from 5 to 9. Strain S25T showed fluorescent pigmentation on Pseudomonas agar F and did not produce pyocyanin on Pseudomonas agar P when cultured for 24-48 h at 30 • C. A complete list of the results is given in Table S8.Strain S25T was not pathogenic to the citrus plants according to the tests performed, as shown in Supplementary Figure S4.The relevant differential characteristics of strain S25T are given in Table 3 and in the protologue Table 4.
Microorganisms 2024, 12, x FOR PEER REVIEW 16 of 23 identity with P. viridiflava).No T6SS-II type has been found in strain S25T, although it has been reported in the phylogenetically closest species P. caspiana FBF102T [34].

Cell Morphology, Physiology and Biochemical Characterisation
The S25T strain is a Gram-negative, rod-shaped bacterium with polar flagellation, as shown by the electron microscopy images in Figure S3.After incubating for 48 h at 30 °C, the colonies were round (1-1.5 mm in diameter), flat, beige coloured, had regular margins and were translucent.The S25T strain was positive for catalase and oxidase activities, was strictly oxidative (not fermentative) and was not a denitrifier.The S25T strain was able to grow in LB media at 4-30 °C for 24-48 h after one week of incubation.No growth was detected at 42 °C or at 37 °C.Growth was observed on nutrient broth in the presence of NaCl until a concentration of 6% (w/v) was reached and the tolerated pH ranged from 5 to 9. Strain S25T showed fluorescent pigmentation on Pseudomonas agar F and did not produce pyocyanin on Pseudomonas agar P when cultured for 24-48 h at 30 °C.A complete list of the results is given in Table S8.Strain S25T was not pathogenic to the citrus plants according to the tests performed, as shown in Supplementary Figure S4.The relevant differential characteristics of strain S25T are given in Table 3 and in the protologue Table 4.The dendrogram obtained by WC-MALDI-TOF MS analysis showed that the S25T strain, P. quasicaspiana strains and P. foli strains clustered together at 70%.The similarity of the strains in the P. caspiana and P. syringae groups was lower than 45% (Figure S5).The S25T strain was located on an independent branch of the P. caspiana strains at a 10% distance, so that it could be easily differentiated in the dendrogram.

Discussion
The value of the genomic approach in the current taxonomy of species within the P. syringae and P. lutea phylogenetic groups is highlighted in the present study.All the data presented demonstrate phylogenetic consistency and clear genomic indices for species differentiation.At least ten not-yet-named novel phylogenomic species could be delineated in the P. lutea group, six in the P. syringae branch, five in the P. caspiana branch, two in the P. cichorii branch and one in the P. ovata branch.The other two pgs (v388 and JDS28PS106) were not affiliated with any of the branches.This substantial increase in the number of species within the group will undoubtedly hinder species identification via routine analysis.These 26 potential novel species have been misidentified thus far or have been classified only at the genus level.It is also important to emphasise that the named species P. meliae, P. savastanoi and P. ficuserectae are synonyms of P. amygdali.This result was already published by Gardan et al. in 1999 after experimental DNA-DNA hybridisations that grouped the four species type strains into the same genomospecies [36].The genome sequence of strain ICMP 23,753 was deposited in the NCBI as "P.phytophila", but the name has never been formally proposed, and the present study demonstrated that it belongs to the species P. quasicaspiana.
However, the classification of the important pathovar reference strain P. syringae pv.tomato DC3000 is controversial [37].It is a model plant pathogen that causes disease in tomato and in the model plants Arabidopsis thaliana and Nicotiana benthamiana.Strain DC3000 does not taxonomically belong to P. syringae.The GTDB taxonomy includes DC3000 in P. avellanae; however, the P. avellanae strains are divided into two clear phylogenetic groups separated by the ANI and GGDC indices: the 96% ANI and 70% GGDC thresholds.As proposed in a previous publication [9], the group represented by strain DC3000 could be classified into a different species, subspecies or new genomovar; however, a definitive taxonomic status requires additional investigations and a formal taxonomical proposal.As stated by Scortichini et al. [37], "Name changes may be confounding and should be attempted with caution".
This challenge in identification is exemplified in the important study on the population dynamics of P. viridiflava developed by Karasov et al. [33], in which 1522 strains were initially grouped by their 16S RNA sequence and identified as P. viridiflava; however, the GTDB phylogenomic identification in the present study assigned 10% of them to 15 named species and 22 pgs that require further characterisation.
Notably, "P.coronafaciens" is not included in the List of Bacterial Names with Standing in the Nomenclature.The species was initially proposed by Schaad and Cunfer [38] but was not included in the Approved Lists of Bacterial Names [39].Dutta et al. [40] proposed the revival of the name "P. coronafaciens", but the genome sequence of the proposed type strain (NCPPB 600T = CFBP 2216T = LMG 5060T = ICMP 3316T) was not included in their analysis.The genome sequence of "P.coronafaciens" LMG 5060T (GCA_000773135.1)together with 76 other genomes of "P.coronafaciens" strains are available on the NCBI website, but these strains are classified as P. tremae in the GTDB taxonomy because the type strain P. tremae ICMP 9151T is included in the cluster."P.coronafaciens" is not considered in the GTDB taxonomy.The "P. coronafaciens" LMG 5060T genome remains "undefined" in the GTDB taxonomy because the sequence failed the quality check (1981 contigs).Additionally, in a previous publication [9], the genome sequence of the P. tremae ICMP 9151T strain was revised due to inconsistencies in the multilocus sequence analysis, and the genome failed the quality analysis with the GTDB taxonomy.In our phylogenomic study, the strains "P.coronafaciens" LMG 5060T and P. tremae ICMP 9151T were identified in the same species, as well as in other P. coronafaciens and P. tremae strains.However, further genomic studies are needed to determine whether the name "P. coronafaciens" has to be revived or considered synonymous with P. tremae.
Routine biochemical tests often fail to differentiate many species in the so-called P. syringae complex [41].Consequently, molecular identification techniques have become essential, and phylogenomic classification is currently considered the most effective method.Simultaneously, genomic techniques allow us to at least partially justify the enormous phenotypic diversity observed within the group.In this context, the study presented highlights the considerable number of genes present in the strains of the group (pangenome).Additionally, this study revealed how these genes can be shared and exchanged among strains.Despite this genomic fluidity, the combination of specific genes defines the distinct phylogenetic branches and, by extension, the individual species within the complex.
As an example of the use of the proposed phylogenomic approach in the description of new Pseudomonas species, we have presented the characterisation of the strain S25T.Pseudomonas sp.S25T was initially described as a putative new species in the P. syringae phylogenetic group of species in 2017 [9].The differentiation of this species from its closest relatives was confirmed in the present study by new molecular and phenotypic data from a polyphasic perspective.Its phenotypic characteristics are corroborated in part by genomic insights, as well as by the potential for its presence in the genome or phage genes for strain differentiation.An interesting point is the existence of a T6SS that is used by many Gram-negative bacteria to deliver toxins that seem to confer competitive advantages in both environmental and pathogenic bacteria and are present in the majority of plant pathogens [42].It cannot be ruled out that the T6SS has other effects in addition to pathogenicity since many plant-growth-promoting bacteria also possess one or more T6SSs [42].A single strain may harbour multiple T6SSs [34,42], such as P. caspiana FBF102T, the closest plant pathogen relative to S25T.P. maioricensis S25T possesses a single T6SS with a small difference (a gene duplication) with respect to one of the two T6SSs of P. caspiana FBF102T, but S25T is not pathogenic according to laboratory tests.It is possible that both subtypes of the T6SS are necessary or that gene duplication may have a negative effect.However, the role of the T6SS in S25T needs further experimental investigation.
The rapid increase in the number of Pseudomonas species described in recent years presents a significant challenge in the routine identification of new isolates, both pathogenic and environmental saprophytic.Therefore, it is necessary to maintain an updated taxonomy to facilitate accurate identification.The present study serves to clarify the classification and provides identification criteria for species delineation within one of the most important groups of plant pathogenic bacteria.The sequencing of genomes and selected genes, such as the rpoD gene, in combination with phenotypic identification based on whole-cell protein profiles obtained by MALDI-TOF MS is currently recognised as the most valuable taxonomic tools.

Description of Pseudomonas maioricensis sp. nov.
The morphological, physiological, biochemical, chemotaxonomic, phylogenetic and genomic characteristics support the proposal of a new bacterial species based on the type strain S25T (= CCUG 69272T, CECT 30911T).The complete description is given in the protologue presented in Table 4.

23 Figure 1 .
Figure 1.RAxML phylogenetic tree (1505 core genes) of representative strains in the P. syringae and P. lutea phylogenetic groups with P. aeruginosa as the outgroup.The numbers in the nodes are bootstrap values of 100 replicates.Main phylogenetic branches are labelled in different colours.

Figure 1 .
Figure 1.RAxML phylogenetic tree (1505 core genes) of representative strains in the P. syringae and P. lutea phylogenetic groups with P. aeruginosa as the outgroup.The numbers in the nodes are bootstrap values of 100 replicates.Main phylogenetic branches are labelled in different colours.

Figure 2 .
Figure 2. RAxML phylogenetic tree (2015 genes of the core proteome) of representative strains in the P. syringae and P. lutea phylogenetic groups without an outgroup.Main phylogenetic branches are labelled in different colours.

Figure 2 .
Figure 2. RAxML phylogenetic tree (2015 genes of the core proteome) of representative strains in the P. syringae and P. lutea phylogenetic groups without an outgroup.Main phylogenetic branches are labelled in different colours.

Figure 3 .
Figure 3. Dendrogram of the ANI values among the strains studied.Type strains are highlig bold.Main phylogenetic branches are labelled in different colours.Red line: species threshold ANI distance.

Figure 3 .
Figure 3. Dendrogram of the ANI values among the strains studied.Type strains are highlighted in bold.Main phylogenetic branches are labelled in different colours.Red line: species threshold.Scale: ANI distance.

Figure 4 .
Figure 4. Gene content distribution of the 23,319 orthologues calculated by M1CR0B1AL1Z3R, represented by the Jaccard index (a) (discontinuous red line indicates the 6 main phylogenetic branches), a split-tree decomposition (b) and a heatmap (c).Each phylogenetic branch is highlighted with a different colour.Numbers indicate the strain in Table1.

Figure 4 .
Figure 4. Gene content distribution of the 23,319 orthologues calculated by M1CR0B1AL1Z3R, represented by the Jaccard index (a) (discontinuous red line indicates the 6 main phylogenetic branches), a split-tree decomposition (b) and a heatmap (c).Each phylogenetic branch is highlighted with a different colour.Numbers indicate the strain in Table1.

Microorganisms 2024 , 23 Figure 5 .
Figure 5. Venn diagram of the orthologous genes shared by the strains studied in the P. caspiana branch.

Figure 5 .
Figure 5. Venn diagram of the orthologous genes shared by the strains studied in the P. caspiana branch.

Figure 6 .
Figure 6.Organisation of the P. maioricensis S25T T2SS (a) and T6SS-I (b) clusters and flanking regions.Arrows indicate the detected ORFs and their direction of transcription.The red square indicates a possible gene duplication in P. maioricensis S25T that was absent in P. caspiana FB102T.

Figure 6 .
Figure 6.Organisation of the P. maioricensis S25T T2SS (a) and T6SS-I (b) clusters and flanking regions.Arrows indicate the detected ORFs and their direction of transcription.The red square indicates a possible gene duplication in P. maioricensis S25T that was absent in P. caspiana FB102T.

Table 1 .
List of bacterial strains and genome accession numbers, NCBI identification, GTDB classification and proposed identification in the present study.Nr: strain number in the present study.Species type strains are labelled in bold.Ps_E: abbreviation of the genus Pseudomonas_E in the GTDB taxonomy.

Table 2 .
[33]ogenomic identification of the P. viridiflava strains studied by Karasov et al.[33]that were assigned to a different species in the present study.

Table 3 .
Differential physiological and biochemical characteristics of Pseudomonas maioricensis S25T and closely related Pseudomonas named species type strains in the P. caspiana branch.P. syringae is included as the reference species in the group.+, positive; w, weak; −, negative.

Table 3 .
Differential physiological and biochemical characteristics of Pseudomonas maioricensis S25T and closely related Pseudomonas named species type strains in the P. caspiana branch.P. syringae is included as the reference species in the group.+, positive; w, weak; −, negative.