Pseudomonas coronafaciens sp. nov., a new phytobacterial species diverse from Pseudomonas syringae

We propose Pseudomonas coronafaciens sp. nov. as a new species in genus Pseudomonas, which is diverse from P. syringae. We also classified strains from onions which are responsible for yellow bud (YB) disease as P. coronafaciens. Sequencing of 16S rRNA gene and multi-locus sequence analysis (MLSA) of housekeeping genes (gyrB, rpoD, gltA and gap1 genes) for the P. syringae pv. coronafaciens strains along with other strains of P. syringae pathovars resulted in a distinct cluster separate from other P. syringae pathovars. Based on DNA-DNA relatedness, pathotype strain of P. syringae pv. coronafaciens (CFBP 2216PT) exhibited ≤35.5% similarity with the pathotype strains of P. syringae pv. syringae (CFBP 1392PT, 4702T) but exhibited ≥90.6% with the YB strains (YB 12–1, YB 12–4, YB 09–1). Also, the YB strains (YB 12–1, YB 12–4, YB 09–1) were able to infect only onion but not oat, rye and Italian ryegrass (common hosts for P. syrinage pv. coronafaciens). Contrastingly, P. syringae pv. coronafaciens strains (NCPPB 600PT, ATCC 19608, Pcf 83–300) produced typical halo blight symptoms on oat, rye and Italian rye grass but did not produce any symptoms on onion. These results provide evidence that P. syringae pv. coronafaciens should be elevated to a species level and the new YB strains may potentially be a novel pathovar of hereto proposed P. coronafaciens species.


Introduction
The taxonomy of Pseudomonas syringae sensu lato and its pathovars has evolved and been a matter of debate for last 34 years [1]. In the 8 th edition of Bergey's Manual of Determinative Bacteriology, P. syringae was widely accepted as a species and was comprised of fluorescent phytopathogenic Pseudomonas nomenspecies [2,3]. Furthermore, a revised taxonomic classification placing 41 nomenspecies of P. syringae as pathovars, was proposed in the 1 st edition of Bergey's Manual of Systematic Bacteriology [4]. This proposal was supported by the International Society for Plant Pathology, subcommittee on taxonomy of plant pathogenic bacteria [5]. The descriptions of most P. syringae pathovars were based on limited cross-pathogenicity a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 tests on different hosts. As a result, overlap in host-range among different P. syringae pathovars often occur. Moreover, routine biochemical tests do not differentiate many of the P. syringae pathovars creating problems in correct identification of pathogens [6,7]. Gardan et al. (1999) [8] identified nine 'genomospecies' of the P. syringae complex in a comprehensive DNA-DNA re-association study. Among different genomospecies, genomospecies 4 (also called phylogroup 4) included pathovars of graminaceous species [P. syringae pv. coronafaciens (Elliott) Young et al., P. syringae pv. atropurpurea, P. syringae pv. striafaciens, P. syringae pv. oryzae, and P. syringae pv. zizaniae,], P. syringae pv. garcea [coffee (Coffea arabica; Rubiaceae)], and P. syringae pv. porri [leek (Allium ampeloprasum; Liliaceae)]. However, genomospecies classifications of P. syrinage pathovars were not supported by their ribotyping or substrate utilization studies. Detailed polyphasic study using genetic approaches were not conducted. Hence, Gardan et al. (1999) [8] refrained from making a formal proposal to elevate P. coronafaciens to species level and it remained as a pathovar of P. syringae (P. syringae pv. coronafaciens). Prior to study by Gardan et al. [8], Schaad and Cunfer (1979) [9] tried to differentiate P. syringae pv. coronafaciens, P. syringae pv. zea, P. syringae pv. atropurpurea and P. syringae pv. striafaciens; however, they later concluded that these bacterial species/pathovars are synonymous. These strains did not differ in their physiological, immunological and substrate utilization tests. In addition, little to no differences in their host range was reported as these strains were able to infect oat (Avena sativa), rye (Secale cerale), wheat (Triticum aestivum), barley (Hordeum vulgare), smooth bromegrass (Bromus inermis), Japanese brome (B. japonicas), chess brome (B. secalinus), cheatgrass (B. tectorum), quackgrass (Agropyron repens), maize (Zea mays).
In this paper, we propose the elevation of P. syringae pv. coronafaciens to a species level as P. coronafaciens, which was confirmed by various molecular and biochemical methods including sequencing of the 16S rRNA gene, and multi-locus sequence analysis (MLSA) based on sequences of housekeeping genes gyrB, rpoD, gltA, and gap1, substrate utilization tests (BIOLOG), polymerase chain reaction (PCR) analysis using plasmid (pCOR1), coronafactate ligase (cfl) and HrpZ effectors genes-specific primers, and DNA-DNA-hybridization. We also characterized strains from onion which are responsible for yellow bud (YB) disease [10] and concluded that they may potentially be a novel pathovar of hereto proposed P. coronafaciens species.

DNA-DNA hybridization and determination of DNA G+C content
High-quality DNA for DNA-DNA hybridization was prepared by the method of Wilson (1987), with minor modifications [20,21]. DNA-DNA hybridization was performed using the microplate method with some modifications [20,21]. The hybridization temperature was 45±1˚C. The strains were labeled with 4-methylumbelliferyl-beta-D-galactoside and the fluorescence intensity was measured. Reciprocal reactions were performed for select hybridization pairs and variation within the limits of this method [22]. The DNA G+C contents for the pathotype strain of P. syringae pv. coronafaciens (NCPPB 600 PT ) and an onion strain (YB 12-1) was measured by HPLC [23,24].

Fatty acid analysis
The whole-cell fatty acid methyl ester (FAME) composition was determined for the type strain of P. coronafaciens (NCPPB 600 PT ) and an onion strain (YB 12-1). Strains were cultured on tryptic soy broth agar for 24 h at 28˚C, and whole-cell fatty acids were saponified, methylated, and extracted as described previously by Miller and Berger (1985) [26]. FAME analysis was conducted using the Microbial Identification System, Sherlock version 3.10 (MIDI).

Phenotypic characteristics
The most useful phenotypic characteristics for the differentiation of the P. syringae pv. coronafaciens strains from P. syringae pathovars are listed in Table 6. Trigonelline is a key substrate that differentiates P. syringae pv. coronafaciens from P. syringae pathovars with the latter being able to utilize the substrate.

Pathogenicity test
Seedlings (of tested hosts) inoculated with PBS did not produce symptoms at 15 DPI. One hundred percent of the seedlings of oat, rye and Italian ryegrass produced halo blight symptoms when inoculated with the strains of P. syringae pv. coronafaciens (NCPPB 600 PT , ATCC 19608, Pcf 83-300) ( Table 7). However, symptoms were not produced when strains of P. syringae pv. coronafaciens were inoculated on onion seedlings. In contrast, 100% of the onion seedlings displayed symptoms, when inoculated with the YB strains (12-1, 09-1, 12-4) or P. syringae pv. porri (NCPPB 3364 PT ) ( Table 7). However, symptoms produced by YB strains were different (intense chlorosis in emerging leaves and severe blight in the older leaves) than those produced by P. syringae pv. porri (NCPPB 3364 PT ) (water-soaked necrotic lesions on younger leaves). Unlike P. syringae pv. coronafaciens, YB strains or P. syringae pv. porri (NCPPB 3364 PT ) strain did not produce any symptoms on the seedlings of oat, rye and Italian ryegrass. P. syringae pv. syringae strains (NCPPB 281 PT , Pss 87-300) did not produce symptoms on any of the inoculated plants (Table 7). Subsequent bacterial isolation and reidentification for all bacterial strain-plant species inoculation combination reconfirmed the association of symptoms with typical bacterial strain inoculated.

Discussion
The taxonomy of Pseudomonas syringae and its pathovars has changed and been a matter of confusion for three decades [1]. Among the identified nine 'genomospecies' of the P. syringae  [6,8]. An attempt was made by Schaad and Cunfer (1979) [9] to differentiate P. syringae pv. coronafaciens, P. syringae pv. zea, P. syringae pv. atropurpurea and P. syringae pv. striafaciens; however, the authors found these bacterial species/pathovars to be synonymous. The authors couldn't differentiate these strains based on physiological, immunological, substrate utilization and host-range tests. Apart from these two studies, detailed investigation is lacking on characterization of P. syringae pv. coronafaciens and P. syringae pv. syringae. Despite being in genomospeies 4, "Pseudomonas syringae pv. coronafaciens" was not included in the Approved List of Bacterial Names and hence is not recognized as a valid species name [27]. Polyphasic approach of taxonomic classification was not adopted when species designations were made. The current study adopted a polyphasic approach to re-characterize P. syringae pv. coronafaciens strains (hosts: oat, rye and onion; and also a pathotype strain) and observed them to be distinct from the pathotype strains of P. syringae and P. syringae pathovars. Hence, it is recommended to elevate P. syringae pv. coronafaciens to a species level as P. coronafaciens. Furthermore, polyphasic approach was also used to identify an unknown bacterial pathogen that was responsible for a new disease in onion (yellow bud), to a species (P. coronofaciens).  Pseudomonas coronafaciens sp. nov., a new phytobacterial species diverse from Pseudomonas syringae Phylogenetic analysis based on 16S rRNA sequences indicate that strains of P. syringae pv. coronafaciens [NCPPB 600 PT , ATCC 19608, 93-2, 83-300] and YB from onion (09-1, 12-1, 12-4, and 12-5), and the pathotype strains of P. syringae pv. porri (NCPPB 3364 PT ), P.syringae pv. oryzae (NCPPB 3683 PT ), and P.syringae pv. garcea (NCPPB 588 PT ) formed a clade that was distinct from other P. syringae pathovars. Sequencing and concatenation of four housekeeping gene loci gltA, gap1, gyrB, and rpoD resulted in a distinct clade that comprised of four YB strains (12-1, 09-1, 12-4, 12-5) and P. syringae pv. coronafaciens strains [NCPPB 600 PT , ATCC 19608, 93-2, 83-300]. The pathotype strains of P. syringae pv. porri (NCPPB 3364 PT ), P.syringae pv. oryzae (NCPPB 3683 PT ), P.syringae pv. striafaciens (NCPPB 1898 PT ) and P.syringae pv. garcea (NCPPB 588 PT ) were also grouped in this clade, and were separated from other P. syringae pathovars. These results suggest that strains from P. syringae pv. coronafaciens clade are closely related and are different from other P. syringae pathovars. Similar observations were made by Gomila et al. (2017) [28] where the authors compared whole genomes and pan-genomes of 139 Pseudomonas pathovars. They observed that P. syringae pv. coronafaciens along with P. syringae pv. garcea, P. syringae pv. oryzae, P. syringae pv. striafaciens, and P. syringae pv. porri formed a distinct cluster different from other P. syringae pathovars [28]. Rombouts et al., (2015) [29] also demonstrated separate grouping of P. syringae pv. garcea, P. syringae pv. oryzae, P. syringae pv. striafaciens, and P. syringae pv. porri strains from the P. syringae pathovars using rpoD based sequencing and DNA fingerprinting by BOX-PCR. Although in above studies MLSA or 16S rRNA sequencing were not used but the conclusions derived from these independent studies were similar.
Plasmids not only govern bacterial host range, and microbial evolution but in some cases can be utilized in bacterial taxonomy. The knowledge of plasmid profile (quantity and type) may help in understanding bacterial phylogeny and taxonomy. However, sole or heavy reliance of plasmid diversity in bacterial taxonomy can be misleading as it can be easily transferred or lost [30]. Nevertheless, plasmid pCOR1 is common among the coronatine (a chlorosis producing phytotoxin) producing Pseudomonas sp. including P. syringae pv. Pseudomonas coronafaciens sp. nov., a new phytobacterial species diverse from Pseudomonas syringae coronafaciens and P. syringae pathovars (pvs. atropurpurea, maculicola, glycinea and morsprunorum) [16]. Further characterization of P. syringae pv. coronafaciens strains using pCOR1based PCR assay resulted in a positive amplification from the YB strains along with pathotype strain of P. syringae pv. coronafaciens (NCPPB 600 PT ). However, P. syringae pathovars used in this study were not amplified indicating close relationship of the YB strains to P. syringae pv. coronafaciens. Based on hrpZ group specific PCR assay, it was observed that the YB strains along with P. syringae pv. coronafaciens strains belonged to group IV, which is distinct from P. syringae and P. syringae pathovars. These results indicate that the YB strains have a close relationship with the pathotype strain of P. syringae pv. coronafaciens (NCPPB 600) and also they are different from P. syringae pathovars. However, we acknowledge that PCR based assays reported above reflect mere presence/absence of gene or genes but they do not truly reflect their functionality. Profile of effector genes (type) tend to be similar to some extent in closely related phytopathogenic bacterial species. The YB and P. syringae pv. coronafaciens strains possessed similar effector genes; avrPto, avrD1, avrAE1, hopA1, hopB1, hopD1, and hopAF1 genes. In contrast, P. syringae pv. syringae possessed effector genes (avrPto, avrD1, avrAE1, hopC1, and hopAN1) whereas it lacked genes; hopA1, hopB1, hopD1, hopF2, hopG1, hrpK1, and hopAF1. These results suggest that the YB strains were similar to P. syringae pv. coronafaciens with respect to the presence of effector genes. Despite differences in effector profile between P. syringae pv. coronafaciens and P. syringae pv. syringae, we acknowledge that such differences may not truly reflect species level distinction. Further detailed investigation on determining effector profiles of multiple P. syringae pv. coronafaciens and P. syringae pv. syringae may throw some light on this perspective.
DNA-DNA relatedness is a good indicator of species delineation and in some cases is better than 16S rRNA and MLSA. Moreover, DNA-DNA relatedness has been demonstrated to carry similar weight as that of whole genome sequencing [22]. Goris et al. (2007) [22] examined the quantitative relationship between DNA-DNA relatedness values and genome sequencederived parameters, such as the average nucleotide identity (ANI) of common genes and the percentage of conserved DNA. The authors observed a close relationship between DNA-DNA relatedness values and ANI and the percentage of conserved DNA for each pair of strains. The authors recommended that cut-off point of 70% DNA-DNA relatedness values for species delineation more likely corresponds to 95% ANI and 69% conserved DNA. It would be interesting to evaluate relationships among the pathotype strains in genomospecies 1 including P. syringae pv. syringae and pathotype strains of genomospecies 4 including P. syringae pv. coronafaciens strains using genome sequence-derived parameter like ANI of common genes.
Future studies should include comparative genomics of pathotype strains of genomospecies 1 and 4.
Pathovar is a bacterial classification that plant pathologist and applied plant microbiologists often use to differentiate bacterial strains based on their ability to cause infection on different plant host/hosts [5,19]. In this study, host range for the YB strains was determined on common hosts known for P. syringae pv. coronafaciens (oat, rye and Italian ryegrass) and also on an isolated host 'onion'. As expected P. coronafaciens strains (NCPPB 600 PT , ATCC 19608, Pcf 83-300) produced typical halo blight symptoms on oat, rye and Italian rye grass but did not produce any symptoms on onion. Contrastingly, the YB strains (12-1, 09-1, 12-4) and a pathotype strain of P. syringae pv. porri (NCPPB 3364 PT ) produced symptoms on onion but did not produce symptoms on any of the other tested hosts (oat, rye and Italian ryegrass). However, symptoms produced by YB strains were different (intense chlorosis in emerging leaves and severe blight in the older leaves) than those produced by P. syringae pv. porri (NCPPB 3364 PT ) (water-soaked necrotic lesions on younger leaves). These observations suggest that the YB strains, although belong to P. syringae pv. coronafaciens (identified in this study), did not share the common host range (oat, rye, Italian ryegrass). Also, these results indicate that the YB strains can potentially be a novel pathovar of P. syringae pv. coronafaciens infecting onion. This is the first report that of any P. syringae pv. coronafaciens infecting a member of Alliacea family (onion).
The "P. syringae pv. coronafaciens" strains belong to genomospecies 4 according to Garden et al. [8]. The species "P. syringae pv. coronafaciens" has been proposed by Schaad and Cunfer (1979) [9] based on phenotypic characteristics. Recently, whole genome and pan-genome comparison of 139 Pseudomonas pathovars revealed that P. syringae pv. coronafaciens belonged to a distinct cluster different from other P. syringae pathovars and hence the authors proposed to revive "P. syringae pv. coronafaciens" as a nomenspecies [28]. However, the study by Gomila et al. (2017) [28] lacked relevant information on phenotypic and genotypic characterization of P. syringae pv. coronafaciens strains that we provide in the current study and thereby proposing to designate and revive P. coronafaciens as a separate species.