Housekeeping Gene Sequencing and Multilocus Variable-Number Tandem-Repeat Analysis To Identify Subpopulations within Pseudomonas syringae pv. maculicola and Pseudomonas syringae pv. tomato That Correlate with Host Specificity

Abstract

Pseudomonas syringae pv. maculicola causes bacterial spot on Brassicaceae worldwide, and for the last 10 years severe outbreaks have been reported in the Loire Valley, France. P. syringae pv. maculicola resembles P. syringae pv. tomato in that it is also pathogenic for tomato and causes the same types of symptoms. We used a collection of 106 strains of P. syringae to characterize the relationships between P. syringae pv. maculicola and related pathovars, paying special attention to P. syringae pv. tomato. Phylogenetic analysis of gyrB and rpoD gene sequences showed that P. syringae pv. maculicola, which causes diseases in Brassicaceae, forms six genetic lineages within genomospecies 3 of P. syringae strains as defined by L. Gardan et al. (Int. J. Syst. Bacteriol. 49[Pt 2]:469–478, 1999), whereas P. syringae pv. tomato forms two distinct genetic lineages. A multilocus variable-number tandem-repeat (VNTR) analysis (MLVA) conducted with eight minisatellite loci confirmed the genetic structure obtained with rpoD and gyrB sequence analyses. These results provide promising tools for fine-scale epidemiological studies on diseases caused by P. syringae pv. maculicola and P. syringae pv. tomato. The two pathovars had distinct host ranges; only P. syringae pv. maculicola strains were pathogenic for Brassicaceae. A subpopulation of P. syringae pv. maculicola strains that are pathogenic for Pto-expressing tomato plants were shown to lack avrPto1 and avrPtoB or to contain a disrupted avrPtoB homolog. Taking phylogenetic and pathological features into account, our data suggest that the DC3000 strain belongs to P. syringae pv. maculicola. This study shows that P. syringae pv. maculicola and P. syringae pv. tomato appear multiclonal, as they did not diverge from a single common ancestral group within the ancestral P. syringae genomospecies 3, and suggests that pathovar specificity within P. syringae may be due to independent genetic events.

INTRODUCTION

Pseudomonas syringae is a complex bacterial species that groups plant-pathogenic bacteria with a huge host range. Strains of P. syringae cause diseases in both monocotyledonous and dicotyledonous plants. However, high specificity is observed at the strain level. This specificity led plant pathologists to create the term “pathovar” to refer to a set of strains with the same or similar pathological characteristics, such as the disease facies they cause and their host ranges (12). The host range of plant-pathogenic bacteria is a major issue for plant health, because the main strategies to control bacterial diseases are based on the understanding of host specificity in the absence of chemical control. Seed and plant certification, and introgression of resistance genes to reduce the host range, are the most efficient ways to maintain crop health. Epidemiological features are highly dependent on host range, and the emergence of new bacterial diseases is sometimes correlated with broadening of the host range (5). Type III secretion system effectors (T3SE) are known to play a key role in race/cultivar specificity in P. syringae (31). Repertoires of T3SE were recently reported to be involved in adaptation to the host by P. syringae (47) and to be correlated with pathovar division in Xanthomonas axonopodis (17). AvrPto and AvrPtoB are well-known examples of type III effectors. AvrPto and AvrPtoB trigger resistance in tomato plants carrying cognate protein Pto, a serine/threonine kinase, and Prf, a nucleotide-binding leucine-rich repeat protein (33, 34).

Pseudomonas syringae pv. maculicola (McCulloch, 1911) is the causal agent of bacterial leaf spot in Brassicaceae worldwide. During the last decade, P. syringae pv. maculicola has caused recurrent outbreaks in radish crops in the Loire Valley, resulting in financial losses for growers. Understanding the genetic diversity of bacterial populations and the bacterial epidemiological cycle are preconditions for developing specific detection systems, identifying inoculum reservoirs, and eradicating the pathogen from the crop. The roles of seeds and weeds in the contamination of plants are still unclear, and genetic markers are needed to develop monitoring techniques to characterize the epidemiological traits of the disease.

Bacteria identified as P. syringae pv. maculicola and those identified as P. syringae pv. tomato are very similar both phenotypically and genetically, and it has even been suggested that they are synonymous (10, 20, 50, 57). P. syringae pv. maculicola and P. syringae pv. tomato are both pathogenic for tomato, and all strains of P. syringae pv. maculicola are pathogenic for Brassicaceae. However, the host range of some strains identified as P. syringae pv. tomato (such as the strain DC3000) overlaps with that of P. syringae pv. maculicola.

The P. syringae sensu lato group includes nine confirmed plant-pathogenic species (Pseudomonas amygdali, Pseudomonas avellanae, Pseudomonas cannabina, Pseudomonas caricapapayae, Pseudomonas ficuserectae, Pseudomonas meliae, Pseudomonas savastanoi, Pseudomonas tremae, and P. syringae) (ISPP Taxonomy of Plant-Pathogenic Bacteria Committee at http://www.Isppweb.org/names [58]). Nine genomospecies were distinguished in an extensive DNA-DNA hybridization study (15). In addition, phylogenetic analysis comparing sequence alignments of loci of protein-coding genes can identify relatedness at a higher resolution (1, 41, 47, 48, 56) than that obtained with 16S rRNA gene loci and previous characterization methods (8, 40). Since the whole genome sequence is available for P. syringae pv. tomato DC3000 (NC_004578.1), it is possible to use multilocus variable-number tandem-repeat (VNTR) analysis (MLVA) to study the genetic structure at both infraspecies and infrapathovar levels. This technique is very promising for fine-scale epidemiology studies. Its advantages include its discriminatory power, maximum reproducibility of results, and the portability of the equipment (40a).

In the present study, we showed that P. syringae pv. maculicola, which causes diseases on Brassicaceae, comprises six genetic lineages within genomospecies 3 of P. syringae, whereas P. syringae pv. tomato comprises two distinct genetic lineages. MLVA confirmed the genetic structure obtained with rpoD and gyrB sequence analyses. Our results provide promising tools for fine-scale epidemiological studies on diseases caused by P. syringae pv. maculicola and P. syringae pv. tomato. The two pathovars have distinct host ranges, as only P. syringae pv. maculicola strains are pathogenic for Brassicaceae. Taking phylogenetic and pathological features into consideration, our data suggest that the strain DC3000 belongs to P. syringae pv. maculicola.

MATERIALS AND METHODS

Bacterial strains.

The bacterial strains used in this study are listed in Table 1. The strains were obtained from the Collection Française de Bactéries Phytopathogènes (CFBP), Institut National de la Recherche Agronomique (INRA), Angers, France, or isolated from affected radishes in the vicinity of Nantes, France. Pseudomonas syringae pv. maculicola and Pseudomonas syringae pv. tomato strains were routinely cultured at 27°C on King's medium B (30). Bacteria were freeze-dried for long-term conservation or kept at −80°C in YPG broth (yeast extract, 7 g/liter; Bacto peptone, 7 g/liter; glucose, 7 g/liter) supplemented with glycerol (50%, vol/vol) for short-term storage.

Table 1.

Pseudomonas syringae strains used in this study^a

Strain name	Host plant	Name(s) in other collections	Geographical origin	Yr of isolation
Pseudomonas syringae pv. maculicola
CFBP1657Pt	Brassica oleracea	LMG 5071, ICMP 3935, NCPPB 2039	New Zealand	1965
CFBP1738	Brassica oleracea	NCPPB 1776	United Kingdom	1965
CFBP1740	Brassica oleracea	NCPPB 2704	Zimbabwe	1970
CFBP3716	Brassica nigra		United Kingdom	1967
CFBP3717	Brassica oleracea		Denmark	1987
CFBP3718, CFBP3719, CFBP3720	Brassica oleracea
CFBP3721, CFBP3722, CFBP3723, CFBP3724, CFBP3725, CFBP3726	Brassica oleracea		United Kingdom	1987
CFBP4313	Brassica oleracea	NCPPB 952	USA	1937
CFBP4314, CFBP4315	Brassica oleracea	NCPPB 1766, NCPPB 1777	United Kingdom	1965
CFBP4316	Brassica oleracea	NCPPB 1886	New Zealand	1958
CFBP4317	Brassica oleracea	NCPPB 2038	United Kingdom	1966
CFBP4318	Brassica oleracea	NCPPB 2040	New Zealand	1965
CFBP4464	Brassica nigra	ICMP 2744	United Kingdom	1968
CFBP6860, 13269	Raphanus sativus		France	2003
CFBP6861, CFBP6862, CFBP6888, CFBP6889	Raphanus sativus		France	2004
CFBP6923	Brassica oleracea		France	2004
CFBP6928	Matthiola incana	ICMP 14577	Australia	2000
CFBP6931, CFBP6932	Brassica rapa	ICMP 11281, ICMP 11282	China
CFBP6948	Brassica napus	IBSBF 1115	Brazil	1994
CFBP7018	Brassica napus	IBSBF 1612	Brazil	1993
12797, 12798	Brassica oleracea		France	2002
12803	Brassica oleracea		France	2005
12804, 12805, 12818, 12819, 12820, 12821, 13270	Raphanus sativus		France	2005
13260, 13261, 13262, 13265, 13266, 13267, 13268	Raphanus sativus		France	2007
13263, 13264, 13274, 13275, 13276, 13277, 13278, 13279, 13280, 13281, 13282	Raphanus sativus		France	2008
13271	Raphanus sativus		France	2006
13272, 13273	Raphanus sativus		France	1993
MAFF301174	Brassica rapa		Japan	1973
MAFF301175	Brassica rapa		Japan	1976
MAFF302539	Brassica rapa		Japan	1985
MAFF302724	Raphanus sativus		Japan	1982
MAFF302783	Brassica rapa		Japan	1983
Pseudomonas syringae pv. tomato
CFBP1320	Lycopersicon esculentum	NCPPB 2424	Switzerland	1969
CFBP1324, CFBP1325	Lycopersicon esculentum		France	1971
CFBP1426, CFBP1427	Lycopersicon esculentum		France	1972
CFBP1636	Lycopersicon esculentum	NCPPB 1367	Canada	1944
CFBP1696	Lycopersicon esculentum	NCPPB 269	Denmark	1949
CFBP1698	Lycopersicon esculentum	ICPB PT 9, NCPPB 1008	USA	1942
CFBP1785	Lycopersicon esculentum	NCPPB 2683	New Zealand	1972
CFBP1918	Lycopersicon esculentum		Canada	1978
CFBP2212Pt	Lycopersicon esculentum	NCPPB 1106, ICMP 2844, LMG 5093	United Kingdom	1961
CFBP2545, CFBP6876	Lycopersicon esculentum		France	1979
CFBP3728	Lycopersicon esculentum		Yemen	1988
CFBP4408	Lycopersicon esculentum		France	1984
CFBP4409	Lycopersicon esculentum		France	1987
CFBP5421	Lycopersicon esculentum		Macedonia	1996
DC3000b	Lycopersicon esculentum		United Kingdom	1960
Pseudomonas amygdali
CFBP3205T	Prunus amygdalus	NCPPB 2607, ICMP 3918, ATCC 33614	Greece	1967
Pseudomonas cannabina pv. alisalensis
CFBP6866Pt	Brassica rapa	ATCC BAA-566, ICMP 15200	USA	1995
Pseudomonas cannabina pv. cannabina
CFBP1631	Cannabis sativa	ICPB PC 114, ATCC 13436, NCPPB 2069	Yugoslavia	1968
Pseudomonas cichorii
CFBP2101T	Cichorium endivia	NCPPB 943, ATCC 10857, ICMP 5707, LMG 2162		1929
Pseudomonas syringae pv. actinidiae
CFBP4909Pt	Actinidia deliciosa	ICMP 9617, NCPPB 3739	Japan	1984
Pseudomonas syringae pv. antirrhini
CFBP1620Pt	Antirrhinum majus	LMG 5057, ICMP 4303, NCPPB 1817	United Kingdom	1965
Pseudomonas syringae pv. apii
CFBP2103Pt	Apium graveolens	LMG 2132, ICMP 2814, ATCC 9654, NCPPB 1626	USA	1942
Pseudomonas syringae pv. avii
CFBP3846Pt	Prunus avium	NCPPB 4290, ICMP 14479	France	1991
Pseudomonas syringae pv. berberidis
CFBP1727Pt	Berberis sp.	ICMP 4116, LMG 2147, NCPPB 2724,	New Zealand	1972
Pseudomonas syringae pv. coriandricola
CFBP5010Pt	Coriandrum sativum	ICMP 12471	Germany	1990
Pseudomonas syringae pv. delphinii
CFBP2215Pt	Delphinium sp.	NCPPB 1879, ICMP 529, LMG 5381	New Zealand	1957
Pseudomonas syringae pv. papulans
CFBP1754Pt	Malus sylvestris	LMG 5076, ATCC 19875, ICMP 4048, NCPPB 2848	Canada	1973
Pseudomonas syringae pv. persicae
CFBP1573Pt	Prunus persica	ICMP 5846, LMG 5184, NCPPB 2761	France	1974
Pseudomonas syringae pv. philadelphi
CFBP2898Pt	Philadelphus coronarius	ICMP 8903, NCPPB 3257	United Kingdom
Pseudomonas syringae pv. porri
CFBP1908Pt	Allium porrum	NCPPB 3364, ICMP 8961	France	1978
Pseudomonas syringae pv. spinaceae
CFBP5524Pt		MAFF 211266	Japan
Pseudomonas syringae pv. tagetis
CFBP4093	Tagetes erecta	LMG 5684	Australia	1976
Pseudomonas syringae pv. viburni
CFBP1702Pt	Viburnum sp.	ICPB PV 7, ATCC 13458, ICMP 3963, LMG 2351, NCPPB 1921	USA
Pseudomonas viridiflava
CFBP1590	Prunus cerasus		France	1974

^aPt, pathotype strain; ^T, type strain; CFBP, Collection Française de Bactéries Phytopathogènes, INRA, Angers, France; ICMP, International Collection of Microorganisms from Plants, Manaaki Whenua Landcare Research, Auckland, New Zealand; NCPPB, National Collection of Plant Pathogenic Bacteria, York, United Kingdom; ATCC, American Type Culture Collection, Manassas, VA; LMG, Laboratorium Microbiology Gent Culture Collection, Rijksuniversiteit, Belgium; ICPB, International Collection of Phytopathogenic Bacteria, Davis, CA; IBSBF, Phytobacteria Culture Collection of Instituto Biológico, Campinas, SP, Brazil; MAFF, Ministry of Agriculture, Forestry and Fisheries Culture Collection, Tsukuba, Ibaraki, Japan.

^bDC3000 is a rifampin-resistant mutant of CFBP2212, which is the pathotype strain of P. syringae pv. tomato.

Strains of P. syringae pv. persicae, P. syringae pv. avii, P. syringae pv. spinaceae, P. syringae pv. apii, P. syringae pv. antirrhini, P. syringae pv. delphinii, P. syringae pv. viburni, P. syringae pv. actinidiae, P. syringae pv. coriandricola, P. cannabina pv. cannabina, P. cannabina pv. alisalensis, P. syringae pv. tagetis, P. syringae pv. papulans, P. amygdali, P. syringae pv. porri, Pseudomonas viridiflava, and Pseudomonas cichorii were included for purposes of comparison.

gyrB and rpoD sequence analysis.

Gene fragments of gyrB and rpoD were amplified (56) from boiling suspensions of Pseudomonas cells. All PCR products were sent to the Plateforme de Séquençage et Génotypage Biogenouest, Nantes, France, for sequencing. The DNA sequences obtained were corrected, translated, and aligned using the BioEdit Sequence Alignment Editor software (18). Two nucleotide sequences were obtained, 816 bp long for gyrB and 726 bp long for rpoD genes. A phylogenetic tree was constructed on gyrB-rpoD concatenated sequence data with MEGA 4.0.2 software (51), using the neighbor-joining method (46) based on Kimura's two-parameter model (29) and 1,000 bootstrap replicates. The nucleotide sequences from P. cichorii strain CFBP2101 were used as an outgroup to build the phylogenetic tree.

Pathogenicity tests.

Pathogenicity tests were performed on radish (Raphanus sativus variety Kocto), cabbage (Brassica oleracea variety capitata cv. Bartolo), and four tomato cultivars, Lycopersicon esculentum cv. Monalbo, Rio Grande, Rimone, and Ontario 7710. Cultivars Monalbo and Rio Grande devoid of the Pto gene are sensitive to P. syringae pv. tomato, while cultivars Ontario 7710 and Rimone, which is a quasi-isogenic line from Rio Grande, have the Pto resistance gene to P. syringae pv. tomato. Plants were grown in a greenhouse under a temperature regime of 18°C at night and 24°C during the day. Tomato plants were inoculated 3 to 4 weeks after sowing. Radish and cabbage were inoculated when plantlets reached the three-leaf stage. Inoculum consisted of approximately 10⁷-CFU/ml suspensions of 24-h-old bacterial cultures in sterile distilled water. The inoculum concentrations were determined by dilution plating. For negative controls, plants were inoculated with sterile distilled water. Inoculum was rubbed onto the lower surface of the leaves with a swab, and carborundum was added in radish. Two plant replicates were performed for each strain. After inoculation, plants were placed in a high-humidity growth chamber for 48 h and incubated in the growth chamber for five additional days, with 14 h at 24°C during the day and 7 h at 20°C at night. The first leaf spot symptoms appeared after 4 days, and the disease occurrence was monitored 7 days after inoculation.

avrPto1 and avrPtoB detection.

The avrPto1 gene was detected by PCR using two different primer sets. One primer pair designed by Yan et al. (57) is located up- and downstream from the avrPto1 gene, while 8Pto1F and 1Pto1R were designed inside the gene with the Primer3 program (http://frodo.wi.mit.edu/primer3/).

Primers avrPtoBF and avrPtoBR were designed in conserved regions from the aligned sequences of P. syringae pv. tomato DC3000 (AY074795.1), P. syringae pv. tomato T1 (DQ133533.1), and P. syringae pv. avellanae (AM410895.1). One 1,632-bp-long DNA fragment and one 1,710-bp-long DNA fragment were amplified using these primers within the avrPtoB coding sequences of P. syringae pv. tomato DC3000 and P. syringae pv. tomato T1, respectively. Four primer pairs were specially designed to target each avrPtoB homolog hopAB2 and hopAB3 in order to check for the absence of the gene. The primer sequences used in this study are listed in Table S1in the supplemental material.

Gene fragments of avrPto1 and avrPtoB were amplified from boiled suspensions of P. syringae cells. All PCRs were performed on a Veriti 96-well thermal cycler (Applied Biosystems) in a 20-μl reaction volume with 1× colorless GoTaq Flexi buffer (Promega, Madison, WI); 200 μM each deoxynucleoside triphosphate, 3 mM MgCl₂ for 8Pto1F/1Pto1R primer pair, and 2.5 mM MgCl₂ for the other PCR; 0.5 μM each 8Pto1F and 1Pto1R; 0.3 μM each avrPtoBF, avrPtoBR, avrPtoB2F, avrPtoB2R, avrPtoB3F, avrPtoB3R, PtoBT1F1, PtoBT1R1, PtoBT1F2, and PtoBT1R2; and 0.75 U of Taq DNA polymerase (Promega). Most primer pairs were used with a 60°C annealing temperature except for avrPtoBF/avrPtoBR, which were used at 62°C.

Amplicons avrPtoB were sequenced (GenoScreen, Lille, France) for three P. syringae pv. tomato strains, CFBP1698, CFBP1324, and CFBP4408, and three P. syringae pv. maculicola strains, CFBP6862, CFBP1657, and CFBP3719, using avrPtoBF and avrPtoBR as sequencing primers.

MLVA scheme.

A search for potential VNTR candidates in the genome of P. syringae pv. tomato DC3000 was carried out (http://minisatellites.u-psud.fr) (32). Eight loci (see Table S2 in the supplemental material) of 6- to 15-bp tandem-repeat (TR) units were selected. The conservation of the TR flanking regions of each locus in other bacteria was determined with BLAST (2) on NCBI data and then aligned with ClustalW (53). Primer pairs targeting single-locus alleles were manually designed in the conserved regions to obtain amplicons of not more than 450 bp in length. Primer quality was checked using Amplify 3× software. Boiled cell suspensions were used as templates for multiplex PCR in two different 20-μl reaction mixtures.

Reaction 1 contained 3 μl of boiling suspensions, 1× colorless GoTaq Flexi buffer (Promega), 2.5 mM MgCl₂, 0.75 U of Taq DNA polymerase (Promega), 200 μM each deoxynucleoside triphosphate, 0.2 μM each PtM8F, PtM8Rb–6-carboxyfluorescein (FAM), PtM13F-VIC, PtM13R, PtM15F-PET, PtM15R, PtM16F-PET, and PtM16R, and sterile distilled water. One of each primer pairs was marked with one of the fluorescent dyes FAM (Eurogentec, Angers, France), NED, PET, and VIC (Applied Biosystems, Courtaboeuf, France).

Reaction 2 contained 3 μl of boiling suspensions, 1× colorless GoTaq Flexi buffer; 3 mM MgCl₂; 0.75 U of Taq DNA polymerase (Promega); 200 μM each deoxynucleoside triphosphate; 0.2 μM each PtM2F-FAM, PtM2Ra, PtM5F-FAM, PtM5R, PtM6F-FAM, PtM6R, PtM14F-NED, and PtM14R; and sterile distilled water.

PCR amplifications were performed under the following conditions: a 5-min denaturation step at 95°C; 30 cycles of 95°C for 30 s, annealing at 60°C for 30 s, and 72°C for 30 s; and a final extension step at 72°C for 7 min.

Amplified products of each reaction were pooled and diluted 1/60 with sterile distilled water. Then 2.5-μl aliquots were mixed with 9.45 μl of Hi-Di formamide and 0.3 μl of GeneScan 500 LIZ internal lane size standard (Applied Biosystems). Capillary electrophoresis was performed in an ABI PRISM 3130 genetic analyzer (Applied Biosystems). P. syringae pv. tomato strains DC3000 and CFBP2212 were used as controls in each experiment.

MLVA data analysis.

Output data from capillary electrophoreses were managed with the BioNumerics software package (version 6.5; Applied-Maths, St-Martens-Latem, Belgium), and chromatograms were also checked with GeneMapper software (version 4.0; Applied Biosystems). The allele scores based on the fragment sizes were converted into repeat numbers and used as character data for cluster analysis. A minimum spanning tree (MST) was generated using the categorical coefficient and the maximum number of single-locus variants (SLVs) as a priority rule (44). An equal weight was assigned to each of the eight VNTR loci. Nei's unbiased estimates of genetic diversity (H_e) and allelic richness (R_s) were calculated for each locus and sample using FSTAT 2.9.3 (http://www2.unil.ch/popgen/softwares/fstat.htm) (16). R_s was calculated to overcome the problem of the sample size and to adapt the rarefaction index (21).

The Bayesian clustering approach implemented in STRUCTURE, version 2.3, was used to infer population structure and assign individuals to groups characterized by distinct allele frequencies (43). The method estimates a probability of ancestry for each individual in each of the groups. Individuals are assigned to one of the groups or populations or jointly to two or more populations if their genotypes indicate that they are admixed. Thirty independent runs of STRUCTURE were performed by setting the number of subpopulations or groups (K) from 1 to 15, with 500,000 burn-in replicates and a run length of 1,500,000 replicates to decide which value of K fitted the data best. K was selected by examining the estimates of the posterior probability of the data for a given value of K, Pr(XK) (where X represents the number of genotypes in the sample), as a guide and by estimating the modal value of the distribution of ΔK [calculated from the STRUCTURE output Pr(XK)], which is a good indicator of the real K (13). We examined the clustering of P. syringae pv. maculicola and P. syringae pv. tomato strains for the inferred number of groups. The admixture model was used for MLVA data, and clustering was performed without prior population information under the F model, which assumes that the allele frequencies in the populations are not correlated.

Nucleotide sequence accession numbers.

All gyrB and rpoD gene sequences were deposited in GenBank under the accession numbers JN185719 to JN185907 and JN190410 to JN190430. Allelic sequences of avrPtoB genes (hopAB2 and hopAB3) were deposited in GenBank under the accession numbers JQ480304 to JQ480309.

RESULTS

Strains isolated from radish in France belong to P. syringae pv. maculicola and form several genetic lineages within genomospecies 3.

The analysis of the concatenated gyrB and rpoD gene sequences of strains isolated from radish disease enabled us to identify the pathogen as P. syringae pv. maculicola. These analyses were performed on several related P. syringae strains in order to study the genetic diversity of the P. syringae pv. maculicola strains in comparison with the other pathovars grouped in genomospecies 3. A total of 18 P. syringae pv. tomato strains, known to be very closely related to P. syringae pv. maculicola (7, 9, 19, 48), were included in the study for the purpose of comparison.

Phylogenetic analyses distinguished several groups with high bootstrap values (Fig. 1) in a cluster grouping genomospecies 3 and 8 as defined by Gardan et al. (14, 15, 41). Within this cluster, 16 genetic lineages (GL) were distinguished with highly significant bootstrap values (>98%). Strains of P. syringae pv. spinaceae, P. syringae pv. apii, P. syringae pv. antirrhini, P. syringae pv. delphinii, and P. syringae pv. actinidiae were shown to be singletons. Strains of P. syringae pv. berberidis and P. syringae pv. philadelphi grouped in one cluster, and P. syringae pv. persicae and P. syringae pv. avii grouped together as well. In contrast, P. syringae pv. tomato and P. syringae pv. maculicola strains were shown to be distributed in eight different lineages.

Fig 1.

Phylogenetic tree, disease responses, and detection of avrPto1 and avrPtoB avirulence genes in the collection of P. syringae pv. maculicola and P. syringae pv. tomato strains. Neighbor-joining tree of concatenated nucleotide sequences of gyrB and rpoD genes representing the genetic lineages of P. syringae pv. maculicola and P. syringae pv. tomato in genomospecies 3. Numbers above the nodes are bootstrap values based on 1,000 pseudoreplicates, and only bootstrap values up to 70 are given. For the disease responses in radish, cabbage, susceptible tomato (Pto−), and resistant tomato (Pto+), black, gray, and white grids represent disease symptoms, no disease symptoms, and not tested, respectively. For the detection of the avrPto avirulence gene, black and gray grids represent presence and absence, respectively; for the detection of the avrPtoB avirulence gene, the gray grid represents absence of the gene, and presence is encoded as follows: = hopAB2, = hopAB2_Δ306–383, = hopAB3, and = hopAB3 + ISPsy9.

P. syringae pv. maculicola strains formed six different genetic lineages. Five of them (GL3.1, GL3.2, GL3.3, GL3.5, and GL3.7) contained only P. syringae pv. maculicola strains, whereas GL3.8 contained strains identified as P. syringae pv. maculicola (CFBP7018, 12803, and CFBP6948) and strains identified as P. syringae pv. tomato (DC3000, CFBP2212, and CFBP1636) as well. The other P. syringae pv. tomato strains were clustered in two different GLs: GL3.4 and GL3.6. Most P. syringae pv. maculicola strains isolated from radish were grouped in GL3.1, whereas GL3.2, GL3.3 and GL3.5 grouped P. syringae pv. maculicola strains isolated from various Brassicaceae plants, and GL3.7 contained only one P. syringae pv. maculicola strain (CFBP6928) isolated from wallflower.

P. syringae pv. maculicola and only those strains of P. syringae pv. tomato that belong to GL3.8 were shown to be pathogenic for all plants tested, i.e., radish, cabbage, and tomato.

To assess the pathogenic properties of the P. syringae pv. maculicola and P. syringae pv. tomato strains in the collection, tomato, radish, and cabbage plants were inoculated. All the plants developed dark necrotic lesions, most of which had bright-yellow borders when inoculated with P. syringae pv. maculicola and P. syringae pv. tomato strains (Fig. 2). Disease symptoms were even more typical on the lower surface of the leaves.

Fig 2.

Symptoms observed after inoculation of P. syringae pv. maculicola strain 13278 on radish (a) and cauliflower (b) and P. syringae pv. tomato strain CFBP1918 on tomato Pto− (c).

Only P. syringae pv. maculicola strains and P. syringae pv. tomato DC3000, CFPB2212, and CFBP1636 were shown to be pathogenic for radish, cabbage, and tomato, while the other P. syringae pv. tomato strains were virulent only on tomato plants but avirulent on radish and cabbage. All P. syringae pv. maculicola strains tested that belonged to GL3.1 (except CFBP6862) were able to infect tomato plants that contained the Pto resistance gene.

A subpopulation of P. syringae pv. maculicola, pathogenic for Pto-expressing tomato, lacked avrPto1 and avrPtoB or contained a disrupted avrPtoB homolog.

AvrPto1 is a widely studied type 3 effector protein. It contributes to virulence in susceptible tomato cultivars (7, 34) and triggers plant defenses in tomato plants that carry the Pto resistance gene (45). The avrPto1 avirulence gene was detected only in P. syringae pv. maculicola and P. syringae pv. tomato strains in genomospecies 3 (Fig. 1). However, P. syringae pv. maculicola strains belonging to GL3.1 and GL3.5 as well as two P. syringae pv. tomato strains (CFBP4408 and CFBP4409) were shown to be devoid of the avrPto1 gene.

The avrPtoB gene was first detected by PCR with the avrPtoBF/avrPtoBR primer set in all P. syringae pv. tomato strains except for CFBP1636, in all P. syringae pv. maculicola strains that belong to GL3.2, GL3.3, GL3.5, GL3.7, and GL3.8, and in CFBP6862 that belongs to GL3.1. Sequencing of avrPtoB from P. syringae pv. maculicola strains (CFBP6862, CFBP1657, and CFBP3719) and P. syringae pv. tomato strains (CFBP1324, CFBP1698, and CFBP4408) revealed the presence of two avrPtoB homologs: hopAB2 and hopAB3. The absence of each avrPtoB homolog in the P. syringae pv. tomato and P. syringae pv. maculicola strain collection was checked by PCR using the avrPtoB2F/avrPtoB2R and avrPtoB3F/avrPtoB3R primer sets to detect hopAB2 and the PtoBT1F1/PtoBT1R1 and PtoBT1F2/PtoBT1R2 primer sets to detect hopAB3.

All P. syringae pv. maculicola strains belonging to GL3.2, GL3.3, GL3.5, and GL3.8 contained hopAB2. In the CFBP3719 and CFBP4313 strains, the gene was deleted from nucleotide positions 306 to 383, which matches an HopAB2_Δ103–128 protein. The CFBP6928 and CFBP6862 strains contained hopAB3. The other P. syringae pv. maculicola strains that belong to GL3.1, except strains 13264 and 13263, which do not possess avrPtoB, had an hopAB3 gene with an insertion sequence (IS) between nucleotide positions 408 and 415, resulting in a modification of the protein sequence starting from amino acid position 136. This IS was identified as ISPsy9 (4, 24), which is present in a single copy in the DC3000 genome. ISPsy9 is a 1,452-bp-long IS belonging to the IS3 family and has two consecutive partially overlapping reading frames, orfA and orfB, that encode three peptides: OrfA, OrfB, and OrfAB (38, 49). A search for highly similar nucleotide sequences in P. syringae pv. tomato DC3000, K40, Max13, and T1 whole-genome sequences was performed on the Pseudomonas-Plant Interaction website (http://pseudomonas-syringae.org/). The IS was found only in P. syringae pv. tomato DC3000. P. syringae pv. tomato strains belonging to GL3.4 and GL3.6 also contained the hopAB3 gene, while P. syringae pv. tomato strains belonging to GL3.8 contained the hopAB2 gene (except for CFBP1636). ISPsy9 sequences of 3 P. syringae pv. maculicola strains (13267, 13270, and CFBP6888) had seven extra nucleotides in nucleotide position 636 of the IS, which matches a 2-unit tandem repeat and causes a frameshift that prematurely terminates translation of the integrase coding sequence.

Only one P. syringae pv. tomato strain (CFBP4409) containing the avrPtoB gene caused disease symptoms on Pto-expressing tomato leaves. This exception in the avr/R scheme has already been reported for the P. syringae pv. tomato T1 strain (28). Lin et al. (33) reported that the P. syringae pv. tomato T1 strain that contains the avrPtoB gene and a functional type III secretion system (T3SS) accumulated AvrPtoB transcripts but failed to accumulate AvrPtoB protein due to an unknown mechanism.

MLVA distinguished several clonal complexes in maculicola and tomato pathovars.

A total of 50 MLVA types were detected among the 87 P. syringae pv. maculicola and P. syringae pv. tomato strains when data from the eight loci were pooled, with allele numbers per locus ranging from two (for PtM6) to 16 (for PtM8) (Table 2). Nei's index of diversity (H_e) ranged from 0.045 (for PtM6) to 0.878 (for PtM13), and allelic richness (R_s) ranged from 1.608 (for PtM6) to 10.898 (for PtM8). Among the strains in the collection, the PtM2 locus was not amplified in seven P. syringae pv. maculicola isolates from Zimbabwe, China, and Japan. Respectively, 36 and 14 haplotypes were observed in P. syringae pv. maculicola and P. syringae pv. tomato strains with no shared haplotype. Nei's index of diversity ranged from 0.209 (for PtM6) to 0.791 (for PtM13) for P. syringae pv. tomato strains, from 0 (for PtM6) to 0.881 (for PtM13) for P. syringae pv. maculicola strains, from 0 (for PtM6) to 0.907 (forPtM13) for P. syringae pv. maculicola strains isolated from Brassica plants, and from 0 (for PtM6, PtM2, PtM15) to 0.738 (for PtM13) for P. syringae pv. maculicola strains isolated from radish. Allelic richness was higher in P. syringae pv. maculicola strains than in P. syringae pv. tomato strains except in PtM6 and PtM15. A lower allelic richness was observed in P. syringae pv. maculicola strains isolated from radish than in P. syringae pv. maculicola strains isolated from Brassica.

Table 2.

Numbers of alleles, Nei's index of diversity (H_e), and allelic richness (R_s) for each VNTR locus used to study 87 P. syringae pv. maculicola (Psm) and P. syringae pv. tomato (Pst) strains

VNTR name	Range of repeat numbers	Pst and Psm strains (n = 87)			Pst strains (n = 18)			Psm strains (n = 69)			Psm strains isolated from Brassica (n = 33)			Psm strains isolated from radish (n = 35)
		No. of alleles	He	Rs	No. of alleles	He	Rs	No. of alleles	He	Rs	No. of alleles	He	Rs	No. of alleles	He	Rs
PtM8	9–31	16	0.839	10.898	7	0.745	7	13	0.775	9.537	11	0.856	10.200	5	0.518	4.887
PtM13	7–19	13	0.878	10.662	4	0.791	4	13	0.881	10.689	11	0.907	10.301	9	0.738	8.008
PtM5	1–10	8	0.794	7.002	4	0.595	4	7	0.749	6.246	5	0.759	4.987	6	0.697	5.665
PtM14	3–10	7	0.755	5.729	4	0.477	4	5	0.749	4.454	4	0.722	3.993	3	0.592	2.990
PtM6	1–3	2	0.045	1.608	2	0.209	2	1	0	1	1	0	1	1	0	1
PtM2	0–8a	6	0.346b	3.806	3	0.307	3	4	0.211c	3.191	4	0.452d	3.903	1	0	1
PtM15	3–7	5	0.503	4.734	4	0.595	4	3	0.239	2.874	3	0.447	2.999	1	0	1
PtM16	1–18	13	0.814	10.217	6	0.732	6	11	0.729	8.661	9	0.881	8.835	3	0.165	2.717

^aSeven P. syringae pv. maculicola strains (CFBP1740, CFBP6931, CFBP6932, MAFF301174, MAFF301175, MAFF302539, MAFF302783) did not produce any amplicon for the PtM2 locus.

^bCalculated for n = 80 strains.

^cCalculated for n = 62 strains.

^dCalculated for n = 26 strains.

The relationships identified among the haplotypes are shown by an MST (Fig. 3A). Clonal complexes grouped haplotypes that differ from one another by at most two loci. P. syringae pv. maculicola strains formed five clonal complexes and seven singletons, while P. syringae pv. tomato strains formed two clonal complexes and two singletons. Results obtained with MLVA are consistent with the phylogenetic analysis of gyrB and rpoD sequences. The previously defined genetic lineages were structured in clonal complexes and singletons. All genetic lineages defined by housekeeping gene sequence analysis were distinct from some of the others in MLVA analysis. P. syringae pv. maculicola strains belonging to GL3.1 and GL3.5 were represented by two unique clonal complexes, while P. syringae pv. maculicola strains belonging to GL3.2 and GL3.3 were divided into clonal complexes and singletons. P. syringae pv. tomato strains belonging to GL3.4 and GL3.6 formed two clonal complexes, while strains belonging to GL3.8 were divided into one clonal complex and three singletons.

Fig 3.

Minimum spanning tree of 87 P. syringae pv. maculicola and P. syringae pv. tomato strains based on 8 VNTR loci. Each circle represents an MLVA type with a size corresponding to the number of strains that share an identical MLVA type. MLVA types connected by a thick solid line differ from one another by 1 VNTR locus, while MLVA types connected by a thin solid line differ by 2 or 3 VNTR loci. MLVA types that differ from each other by 4 VNTR loci or more are connected by dashed and dotted lines. MVLA types were distinguished to define clonal complexes and to group in the gray zone MLVA types that differ from one another by at most 2 locus variants. MLVA results are compared with the 8 GL defined from gyrB and rpoD sequence analyses (A), the isolated hosts of the strains (B), and the presence of AvrPtoB homologs (C).

MLVA results were also analyzed in terms of host of origin (Fig. 3B). P. syringae pv. maculicola strains isolated from Brassicaceae plants formed separate clonal complexes from those formed by P. syringae pv. tomato strains isolated from tomato plants. P. syringae pv. maculicola strains isolated from radish were mostly grouped in one clonal complex corresponding to GL3.1, and three strains (12804, 12805, and CFBP6861) were grouped with P. syringae pv. maculicola isolated from Brassica belonging to GL3.5. The CFBP6928 P. syringae pv. maculicola strain isolated from wallflower (Matthiola incanae) was identified as a singleton. P. syringae pv. tomato strains were grouped in two clonal complexes corresponding to GL3.4 and GL3.6, and CFBP2212, CFBP1636, and DC3000 were identified as two singletons.

When MLVA results were compared to the presence or absence of the avrPtoB avirulence gene, three main groups were represented (Fig. 3C). The first group corresponded to strains containing an hopAB2 homolog and comprised four clonal complexes and seven singletons, including most P. syringae pv. maculicola strains isolated from Brassica, three P. syringae pv. maculicola strains (12804, 12805, and CFBP6861) isolated from radish, and P. syringae pv. tomato strains DC3000 and CFBP2212. The second group corresponded to strains containing the hopAB3 homolog. This group comprised P. syringae pv. tomato strains (except DC3000 and CFBP2212), the P. syringae pv. maculicola CFBP6928 strain isolated from wallflower, and one strain isolated from radish (CFBP6862). The third group was formed by a single clonal complex in which the majority of the strains contained a disrupted hopAB3 homolog. The avrPtoB gene (hopAB3) contained a 1,452- to 1,459-bp-long insertion sequence identified as ISPsy9. These strains were isolated from radish, except for CFBP3716 and CFBP4464, which were isolated from black mustard (Brassica nigra). No avrPtoB homolog was detected in three strains isolated from radish and tomato (P. syringae pv. maculicola 13263, P. syringae pv. maculicola 13264, and P. syringae pv. tomato CFBP1636).

Structural analysis of MLVA data revealed four ancestral populations in the collection of P. syringae pv. tomato and P. syringae pv. maculicola strains.

Bayesian clustering was used to analyze multilocus haplotypes to infer the genetic ancestry of the individual strains in both pathovars. STRUCTURE identified four ancestral groups in the collection. Observations of plateaus in the estimates of ln [Pr(X K)] and a clear modal value in the distribution of ΔK indicated values of K = 4 (Fig. 4). The assignment of strains to ancestral clusters was better correlated with the genetic lineages defined by the phylogenetic analysis of gyrB and rpoD nucleotide sequences (Fig. 4B) than with the isolated hosts (Fig. 4A).

Fig 4.

Population structure of the collection of P. syringae pv. tomato and maculicola strains isolated from Lycopersicon esculentum and Brassicaceae as inferred by STRUCTURE and assignment of individuals to K ancestral groups. Numbers on the x axis above panel A indicate the isolated host of the strains as follows: 1, Lycopersicon esculentum; 2, Brassica genus with Brassica napus (a), Brassica rapa (b), Brassica oleracea (c), Brassica nigra (d); 3, Matthiola incanae; and 4, Raphanus sativus. Numbers on the x axis above panel B indicate the genetic lineages defined in the legend to Fig. 1. Each strain is represented by a vertical segment and named under the x axis. Within each segment, the length of each color section (y axis) indicates the estimated probability of assignment to each of the K ancestral groups defined. The use of the same color suggests that strains share an ancestral group.

Three populations of P. syringae pv. maculicola strains and two populations of P. syringae pv. tomato strains were identified as having a common ancestor shared by both pathovars. Two populations of P. syringae pv. maculicola strains isolated from radish were identified (Fig. 4A). Both populations shared a common ancestor with strains isolated from Brassica oleracea. One of the populations identified also shared a common ancestor with strains isolated from Brassica nigra. Two populations of P. syringae pv. tomato strains isolated from tomato plants were consistently identified. One population was unique and specific to P. syringae pv. tomato strains, while the second population shared a common ancestor with strains isolated from Brassica and Matthiola incanae.

In comparison with the phylogenetic analyses of gyrB and rpoD, only P. syringae pv. maculicola strains belonging to GL3.3 and GL3.2 were assigned to more than one population (Fig. 4B). P. syringae pv. maculicola strains belonging to GL3.1 and some P. syringae pv. maculicola strains belonging to GL3.2 and GL3.3 shared a common ancestor. P. syringae pv. maculicola strains belonging to GL3.5 and two P. syringae pv. maculicola strains belonging to GL3.2 were identified as a single population. The third population of P. syringae pv. maculicola strains that was identified included P. syringae pv. tomato and P. syringae pv. maculicola strains belonging to GL3.8, the P. syringae pv. maculicola strain (CFBP6928) belonging to GL3.7, and some P. syringae pv. maculicola strains belonging to GL 3.3.

DISCUSSION

Insights into the genetic diversity of the collection of P. syringae pv. maculicola strains.

Over the last 10 years, a bacterial disease has emerged in radish crops and has led to financial losses for market gardeners in France. The pathogenic bacterium responsible for this disease was identified as P. syringae pv. maculicola. In order to identify the genetic structure of this pathovar, we sequenced two housekeeping genes from a collection of P. syringae pv. maculicola strains representative of the outbreaks and a set of strains representative of the taxonomically closely related pathovars, with special emphasis on P. syringae pv. tomato, which has sometimes been considered synonymous to P. syringae pv. maculicola (50). According to Sarkar and Guttman (48), understanding bacterial clonal population structure is an essential component of understanding bacterial pathogenesis. Knowing how clonal lineages arise in bacterial populations provides a means for identifying and characterizing virulence-associated factors. Housekeeping genes belong to the core genome and provide the best indicators of the clonal evolutionary history of a bacterial species (22). Phylogenetic analysis of the gyrB and rpoD nucleotide sequences (Fig. 1) confirmed the phylogenetic structure of genomospecies 3 defined by Gardan et al. (15). The phylogenetic neighbor-joining tree drawn with sequence data of gyrB and rpoD is congruent with MLST trees previously designed on P. syringae using 7 or more loci (23, 47, 57). These two genes belong to the core genome that is less prone to horizontal gene transfer and therefore provides a good indication of the clonal evolutionary history of a bacterial species. Thus, P. syringae pv. maculicola and P. syringae pv. tomato formed polyphyletic groups: P. syringae pv. maculicola strains were distributed in five different genetic lineages, while P. syringae pv. tomato appeared in three genetic lineages, including one shared by the two pathovars. Several distinct populations likely emerged in genomospecies 3 that cause bacterial symptoms in Brassicaceae plants.

In the case of P. syringae sensu lato, little recombination was detected and mutation was found to contribute approximately four times more than recombination to variation (47). In another study, Yan et al. (57) found that recombination contributed 5.8 times more than mutation to variation between isolates that belong to the genomospecies 3. Mutation and recombination play important roles in the evolution of P. syringae pathovars. However, the genetic structure obtained with data derived from the gene sequence analysis is congruent with that defined by MLVA; therefore, the phylogenetic trees give an accurate representation of the genetic relationships within P. syringae pv. maculicola and related pathovars.

Characteristics of P. syringae pv. maculicola isolated from radish.

Strains isolated from radish were grouped mainly in GL3.1. The strains are pathogenic for radish, cabbage, and tomato and have developed an adaptive capacity to bypass the Pto resistance in tomato plants (Fig. 1). This observation is in agreement with the absence of the avrPto1 gene and the presence of a disrupted avrPtoB gene. Actually, we found an IS element, identified as ISPsy9, belonging to the IS3 family, inside the avrPtoB gene. This IS element was first described in the DC3000 P. syringae pv. tomato strain (4) but did not disrupt any avirulence gene. ISs have been widely documented as being involved in the activation and abolition of gene expression. They play an important role in adaptability of bacterial hosts by facilitating the evolution of new combinations of genes (39). It has already been reported that IS may contribute to the evolution of new virulence genes and to inactivation of virulence genes that have become a liability for the bacterial strain concerned due to R gene recognition (17, 25, 27, 37). Our study identified a P. syringae pv. maculicola strain population that can overcome the resistance encoded by the Pto tomato gene. This results in a shift from incompatible to compatible plant-pathogen interaction. Recent studies identified IS elements in seven T3SE genes from strains belonging to six Xanthomonas axonopodis pathovars (17). Moreover, four of the six ISs belonged to the IS3 family. The IS3 family has been reported to disrupt several T3SEs in several genera of plant-pathogenic bacteria, such as Ralstonia solanacearum (42), Xanthomonas axonopodis (17), Xanthomonas campestris pv. vesicatoria (26), and P. syringae pv. tomato (19). Does this IS play a role in the adaptability and virulence of P. syringae pv. maculicola strains in radish crops? P. syringae pv. maculicola strain CFBP6862, which possesses a complete avrPtoB gene, is not pathogenic on Pto-expressing tomato plants but is still pathogenic on radish. Consequently, this T3SE gene, whether functional or inactivated, did not play a role in the virulence in radish.

Genetic population analyses.

MLVA is a genotyping technique widely used in clinical microbiology (36) that has a high level of discriminatory power (52, 54). This technique has been compared with other typing techniques and its suitability demonstrated for intrapathovar analysis (3, 40a, 55). This technique is an excellent tool to describe population structure (40a). The scheme provided by STRUCTURE analysis of MLVA data supported the insights into the evolutionary relationships of pathovars maculicola and tomato already provided by Hendson et al. (20), suggesting that they may be diverging from a common ancestral group. This group presents pathogenic properties more similar to those of P. syringae pv. maculicola strains than those of P. syringae pv. tomato strains.

The MLVA scheme used in this study distinguished several clonal complexes within the P. syringae pv. maculicola and P. syringae pv. tomato strain collection, none of which grouped the two pathovars (Fig. 3A). Strains isolated from Brassicaceae were separated from those isolated from tomato, revealing a different structure depending on the host association, e.g., Brassicaceae versus Solanaceae. However, P. syringae pv. maculicola strains were assigned to two clonal complexes that contained strains isolated from both the Raphanus and Brassica genera. This suggests that the strains grouped in these clonal complexes could have been isolated from outbreaks that occurred in various Brassicaceae species and the strains grouped in the main clonal complex described in Fig. 3 could be associated with the disease observed in radish. The population structure obtained by MLVA is in agreement with the phylogenetic structure obtained with the gyrB and rpoD nucleotide sequence analyses and revealed the structure of populations within each gyrB/rpoD genetic lineage (GL) (Fig. 3B).

Lin et al. (33) described several avrPtoB homologs widely distributed among phytopathogenic bacteria, suggesting that members of this T3SE family may play an important role in bacterial virulence. It is interesting to note that the radish P. syringae pv. maculicola strains grouped in GL3.1 possessed the hopAB3 gene, as did the P. syringae pv. tomato strains grouped in GL3.4 and GL3.6, while the other P. syringae pv. maculicola strains isolated from radish and other brassicaceous plants possessed the hopAB2 gene, as did the P. syringae pv. tomato DC3000 strain. This result reveals the particularity of these P. syringae pv. maculicola strains and suggests that this avirulence gene has evolved in a different way than the core genome. In fact, hopAB2 and hopAB3 have been reported to be syntenically conserved host association factors in different genome regions (35), which could explain the difference observed between phylogenetic structure based on the analysis of core genome molecular markers and the avrPtoB distribution patterns. Concerning P. syringae pv. maculicola, the inactivation of the avrPtoB gene may have suppressed an avirulent effector in the P. syringae pv. maculicola strains. This avrPtoB inactivation is shared by the majority of the P. syringae pv. maculicola strains currently known to be pathogenic for radish, and the ISPsy9 insertion probably occurred before the emergence of GL3.1 in radish, which subsequently evolved clonally. This supports the hypothesis that the inactivation of this gene does not play a role in virulence in Brassicaceae and did not lead to a reduction in the fitness of this population associated with radish.

In this study, the MLVA was performed on a collection of strains of P. syringae pv. maculicola and P. syringae pv. tomato and provided new information on the relationship between P. syringae pv. maculicola and P. syringae pv. tomato. The minisatellite markers we selected and the primers designed for their amplification are excellent tools to perform epidemiological analyses and will help understand how seeds, weeds, water, and soil play a role in crop contamination. All P. syringae pv. maculicola strains isolated from outbreaks observed in Loire Valley from 2005 to 2008 grouped in the same genetic lineage, GL3.1, except strains 12804 and 12805, which grouped in GL3.5. Strains isolated from the same field were in the same clonal complex defined by MLVA. This suggested that strains within most fields originated from a single inoculum source, as it was previously reported on cruciferous crops in Oklahoma (59). One may speculate that inoculum would be provided by contaminated seeds used for sowing the fields. However, some P. syringae pv. maculicola strains were isolated from weeds surrounding the infected field (unpublished data), and minisatellite markers must be used on isolates from diverse sources to conclude on the origin of the infection.

P. syringae pv. tomato and DC3000.

P. syringae pv. tomato has long been a controversial pathovar because of the phenotypic diversity of the strains, particularly at the pathogenic level (10, 20, 50). Because the host ranges of P. syringae pv. tomato and P. syringae pv. maculicola strains overlap, some authors suggested grouping the two pathovars together (21, 50). Our results support the hypothesis of two populations of P. syringae pv. tomato formed by distinct genetic lineages. One population is not pathogenic for Brassicaceae, while the second population, which contains the DC3000 P. syringae pv. tomato strain, is pathogenic for both tomato and Brassicaceae, as is the case of P. syringae pv. maculicola (11, 20). DC3000 and the other P. syringae pv. tomato strains that also infect cabbage and radish belong to a genetic lineage shared with some P. syringae pv. maculicola strains isolated from Brassica, as already described by Yan et al. (57). Our results support the hypothesis proposed by Yan et al. (57) that typical P. syringae pv. tomato strains and the DC3000-like strains are two distinct pathovars. The question is whether it is better to keep the name “P. syringae pv. tomato” for a very atypical strain whose genotypic and phenotypic characteristics appear to be more similar to the pathovar maculicola than to the pathovar tomato or change the pathovar tomato type strain (CFBP 2212 from which was selected DC3000) and rename the probably most studied plant-pathogenic bacteria and run the risk of creating confusion in the phytopathologic community. We propose to identify the DC3000 strain and the strains grouped in the lineage GL3.8 as P. syringae pv. maculicola. The DC3000 strain is a spontaneous rifampin-resistant mutant of strain CFBP2212 (9), which was isolated from tomato in Guernsey Island in 1961. The Channel Islands are located at the heart of the diversification area of Brassicaceae, and many wild Brassicaceae species are endemic along the west coast of Europe. One could hypothesize that the polymorphic P. syringae pv. maculicola population is associated with Brassicaceae in this area and occasionally contaminated tomato crops.

In conclusion, P. syringae pv. maculicola and P. syringae pv. tomato show relevant genetic variation, as pointed out using housekeeping gene sequence analysis and MLVA. These two pathovars appear multiclonal, as they did not diverge from a single proper common ancestor. However, all pathovars grouped in genomospecies 3 diverge from a common ancestor. This study provides new information on the relationship between P. syringae pv. maculicola and P. syringae pv. tomato that will be useful for understanding the epidemiology of plant-pathogenic bacteria and for the development of reliable identification and detection tools.