Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Comparative Genomics of Multiple Strains of Pseudomonas cannabina pv. alisalensis, a Potential Model Pathogen of Both Monocots and Dicots

  • Panagiotis F. Sarris ,

    sarrispanos@science.agrool.gr

    Current address: The Sainsbury Laboratory, John Innes Centre, Norwich Research Park, Norwich, United Kingdom

    Affiliations Department of Plant Sciences, School of Agricultural Technology, Technological Educational Institute of Crete, Heraklion, Greece, Department of Biology, University of Crete, Heraklion, Greece

  • Emmanouil A. Trantas,

    Affiliation Department of Plant Sciences, School of Agricultural Technology, Technological Educational Institute of Crete, Heraklion, Greece

  • David A. Baltrus,

    Affiliation School of Plant Sciences, The University of Arizona, Tucson, Arizona, United States of America

  • Carolee T. Bull,

    Affiliation United States Department of Agriculture–Agricultural Research Service, Salinas, California, United States of America

  • William Patrick Wechter,

    Affiliation United States Department of Agriculture–Agricultural Research Service, Charleston, South Carolina, United States of America

  • Shuangchun Yan,

    Affiliation Department of Plant Pathology, Physiology, and Weed Science, Virginia Tech, Blacksburg, Virginia, United States of America

  • Filippos Ververidis,

    Affiliation Department of Plant Sciences, School of Agricultural Technology, Technological Educational Institute of Crete, Heraklion, Greece

  • Nalvo F. Almeida,

    Affiliation School of Computing, Federal University of Mato Grosso do Sul, Mato Grosso do Sul, Brazil

  • Corbin D. Jones,

    Affiliation Howard Hughes Medical Institute and Department of Biology, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, United States of America

  • Jeffery L. Dangl,

    Affiliations Howard Hughes Medical Institute and Department of Biology, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, United States of America, Carolina Center for Genome Sciences, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, United States of America

  • Nickolas J. Panopoulos,

    Affiliations Department of Biology, University of Crete, Heraklion, Greece, Professor Emeritus, University of Crete, Heraklion, Greece and University of California, Berkeley, California, United States of America

  • Boris A. Vinatzer,

    Affiliation Department of Plant Pathology, Physiology, and Weed Science, Virginia Tech, Blacksburg, Virginia, United States of America

  • Dimitrios E. Goumas

    Affiliation Department of Plant Sciences, School of Agricultural Technology, Technological Educational Institute of Crete, Heraklion, Greece

Abstract

Comparative genomics of closely related pathogens that differ in host range can provide insights into mechanisms of host-pathogen interactions and host adaptation. Furthermore, sequencing of multiple strains with the same host range reveals information concerning pathogen diversity and the molecular basis of virulence. Here we present a comparative analysis of draft genome sequences for four strains of Pseudomonas cannabina pathovar alisalensis (Pcal), which is pathogenic on a range of monocotyledonous and dicotyledonous plants. These draft genome sequences provide a foundation for understanding host range evolution across the monocot-dicot divide. Like other phytopathogenic pseudomonads, Pcal strains harboured a hrp/hrc gene cluster that codes for a type III secretion system. Phylogenetic analysis based on the hrp/hrc cluster genes/proteins, suggests localized recombination and functional divergence within the hrp/hrc cluster. Despite significant conservation of overall genetic content across Pcal genomes, comparison of type III effector repertoires reinforced previous molecular data suggesting the existence of two distinct lineages within this pathovar. Furthermore, all Pcal strains analyzed harbored two distinct genomic islands predicted to code for type VI secretion systems (T6SSs). While one of these systems was orthologous to known P. syringae T6SSs, the other more closely resembled a T6SS found within P. aeruginosa. In summary, our study provides a foundation to unravel Pcal adaptation to both monocot and dicot hosts and provides genetic insights into the mechanisms underlying pathogenicity.

Introduction

Pseudomonas cannabina is a Gram-negative, fluorescent, flagellated, aerobic bacterium that causes leaf and stem spot of hemp (Cannabis sativa), from which it derives its species name [1]. Although it was classified as a pathovar of Pseudomonas syringae, following extensive polyphasic taxonomic studies (including DNA-DNA hybridization, carbon source utilization and ribotyping) this pathovar was revived as a species in 1999 (P. cannabina pv. cannabina; [2]).

Pseudomonas syringae pv. alisalensis was recently transferred to Pseudomonas cannabina [3]. P. cannabina pv. alisalensis strains are not pathogenic on C. sativa but are frequently identified as the causal agent of bacterial blight diseases of field and greenhouse grown crucifers. Reciprocally, P. cannabina pv. cannabina strains are not pathogenic on crucifers. These differences in host range demonstrated that P. cannabina consists of at least two distinct pathovars, P. cannabina pv. cannabina (Pcan), and P. cannabina pv. alisalensis (Pcal) [3]. In addition to host range differences, P. cannabina pv. alisalensis also physiologically differs from P. cannabina pv. cannabina isolates with regards to carbon source utilization (Manitol, D(+) arabitol, Myoinositol, D-mannitol, D(-) tartrate), bacteriophage sensitivity, and pigment production further supporting distinct pathovars in P. cannabina [3].

Plant pathologists have only differentiated crucifer diseases caused by P. syringae pv. maculicola (Pma) and Pcal since 2000, thus some isolates from crucifers identified as Pma prior to 2000 were actually Pcal [3], [4]. For example, the well-studied P. syringae pv. maculicola strain B70 (otherwise known as CFBP 1637, M4, ES4326, and ICMP 4326), isolated from diseased radish in Wisconsin by Williams and Keen in 1965 [5] was recently shown to belong to P. cannabina pv. alisalensis using Multi-Locus Sequence Analysis (MLSA) [3]. Pcal ES4326 has been used extensively in various studies of plant-bacteria interactions and is pathogenic on several ecotypes of the model plant Arabidopsis thaliana [6][9].

Pcal has been isolated from diseased crucifers (e.g. arugula, Eruca sativa [10][12]; cabbage, Brassica oleracea var. capitata [13]; cauliflower, Brassica oleracea subsp. botrytis [14]; Brussels sprouts, Brassica oleracea subsp. gemmifera [3], [15]; rape, Brassica napus [4]; white mustard, Brassica hirta [4]; rutabaga, Brassica napus var. napobrassica [16] and radish, Raphanus sativus [3], [17]). However, Pcal can infect a wider range of hosts under experimental conditions including dicots (Solanum lycopersicum [10], [18]) as well as monocots (e.g. California brome, Bromus carinatus; oat, Avena sativa; common timothy, Phleum pretense and Bromus diandrus [3], [10], [18]). Thus, P. cannabina has the potential to be a model for studying plant-bacterial interactions across divergent hosts and can serve as a parallel model system to P. syringae for understanding virulence and host range evolution.

The Type III Secretion System (T3SS) is an essential mediator of the interaction of bacterial phytopathogens with their plant hosts and shares a common ancestor with the bacterial flagellum [19][22]. The T3SS is an inter-kingdom transfer device, which translocates a diverse array of proteins, called Type III Effector Proteins (T3EPs), through the type III pilus, either to extracellular locations or directly into the plant cells. T3EPs modulate host immune responses and determine the outcome of host-pathogen interactions [20]. Pcal strain ES4326 was the first strain of any plant pathogenic Pseudomonas species for which the majority of the T3EP repertoire was experimentally determined [9]. However, little is known about the conservation of T3EPs and other pathogenicity factors like the type VI secretion system (T6SS) and toxins throughout Pcal. The T6SS is a recently discovered protein secretion/translocation system found in a large number of bacteria, including phytopathogenic and plant-associated bacteria [23], [24]. Although T6SSs are widespread, and are often present in multiple copies in the genomes of phytopathogens, little is known about the exact role of this system in plant-bacterial interactions. The T6SS has been studied primarily in the context of mammalian pathogenic bacteria-host interactions and may also function to promote commensal or mutualistic relationships between bacteria and eukaryotes, as well as to mediate cooperative or competitive interactions between bacterial species [24]. In a recent work, Haapalainen and colleagues [25] present evidence for the requirement of one of the P. syringae T6SSs, for the survival in competition with enterobacteria and yeasts, through growth suppression of these competitors.

Comparative genomics can provide insights into host-pathogen interactions, differences in virulence factor repertoires, and pathogen evolution [26]. Several, complete or draft genome sequences have been generated for important phytopathogenic pseudomonads, including: P. syringae pv. tomato (Pst; [27], [28]), P. syringae pv. phaseolicola (Psph; [29]), P. syringae pv. syringae (Pss B728a [30], Ps FF5 [31]), P. syringae pv. oryzae (Psor; [32]), among others [33][37]. Baltrus and colleagues reported the analysis of multiple genomes within P. syringae, and including Pcal ES4326 (although this strain was referred to as P. syringae pv. maculicola), to reveal the genetic changes underlying differences in host range and virulence across host plants ranging from rice to maple trees [37].

The wide host range of Pcal provides a unique opportunity to understand virulence evolution in a closely related parallel model system to P. syringae. For this purpose we selected to sequence the genomes of four representative strains from geographically distant areas; the USA strains Pcal BS91 (CFBP 6866 pathotype strain) [3], and Pcal T3C (CFBP 7684) [37], and the Greek strains Pcal PSa1_3 (CFBP 7682) [10], and Pcal PSa866 (CFBP 7683).

The comparison of (a) the sequenced Pcal genomes with each other, and (b) with the genomes of other related pathogens, will help us to better understand the structure and function of the genomic elements that give each strain its unique virulence characteristics. More specifically, given their demonstrated importance for inter-organismal interactions, we focused our analysis on the genes coding for the T3SS, T6SSs, and T3EPs. Where possible, we also compared our findings with a draft genome sequence of the Pcan type strain (CFBP 2341 [2]), despite the poor quality of the Pcan genome sequence; and the available data from the Pcal ES4326 genome [37], [38]. Our analysis demonstrated intra-pathovar variation in T3EPs content and phylogeny revealing two distinct lineages within Pcal, in agreement with previously reported rep-PCR and MLSA results [3]. The data presented here will provide the basis for focused molecular plant-microbe investigations into patho-adaptation of Pcal strains to plants and other inter-organismal interactions involving Pcal.

Results and Discussion

Pathogenicity Tests and Determination of Host Range

A range of hosts to which the Pcal strains BS91, T3C and ES4326 are pathogenic was previously determined [3], [18], [39]. We found that the Pcal strains PSa1_3 and PSa866 had identical experimentally determined host ranges (Figure 1 and Figure S1). Disease symptoms occurred after inoculation of the dicots Eruca sativa; Brassica oleracea var. capitata, Brassica oleracea subsp. botrytis, Brassica oleracea subsp. gemmifera, Brassica napus, Brassica hirta, Brassica napus var. napobrassica and Raphanus sativus and Solanum lycopersicum, as well as the monocots Avena sativa and Bromus diandrus with all Pcal strains. All strains induced the same disease symptoms regardless of their original host of isolation, which argues against extensive specialization within Pcal across the tested hosts (Figure 1 and Figure S1).

thumbnail
Figure 1. Compatible reactions to artificial inoculations of P. cannabina pv. alisalensis (Pcal) on monocots and dicots plant species.

Artificial inoculations were performed using the sequenced Pcal strain PSa1_3 on: tomato leaves (A), tomato stem (B), Oat leaves (C) and Bromus diandrus leaves (D). Information for additional artificial inoculations on various plant species can be found in figure S1.

https://doi.org/10.1371/journal.pone.0059366.g001

Genome-wide Sequence Data

A draft genome sequence for Pcal ES4326 was recently published [37]. To characterize diversity within Pcal and to compare virulence gene repertoires within Pcal and with P. syringae pathovars, we sequenced the genomes of PSa1_3, PSa866, BS91 and T3C. Strains PSa1_3 and PSa866 were sequenced on a 314 chip of the ION Torrent PGM (Institute of Molecular Biology and Biotechnology, Greece) while strains BS91 and T3C were sequenced on an Illumina GAII (University of North Carolina, USA).

For the Pcal PSa1_3 strain a total of 655.105 DNA fragments of an average length of 107 nts were obtained, totaling 70.166.144 nts. The reads were assembled into 489 contigs yielding a total draft genome size of 6.022.892 nts and representing a genome coverage of 11,6 x. For the Pcal PSa866 strain a total of 547.652 DNA fragments of an average length of 77 nts were read, totaling 42.249.079. The reads were assembled in 777 contigs yielding a total draft genome size of 5.849.615 nts and representing a genome coverage of 7,2x (Table 1).

thumbnail
Table 1. Statistics and features of the sequenced P. cannabina pv. alisalensis strains.

https://doi.org/10.1371/journal.pone.0059366.t001

Illumina sequencing of Pcal BS91 gave an output of 12.967.070 paired-end sequences of an average length of 2×35 nts, totaling 453.847.450 nts. Reads were assembled into 381 scaffolds yielding a draft genome with size of 6.039.137 nts (75×coverage). The sequencing of the Pcal T3C strain gave 7.420.598 reads of an average length of 35 nts, totaling 259.720.930 nts. These reads were assembled into 464 scaffolds yielding a draft genome with the size of 5.821.618 nts (33×coverage).

Genome size for the four strains was within the range of previously sequenced and published P. syringae draft genomes [32][37]. Also the G+C content (Table 1) was similar to sequenced P. syringae genomes. Sequence assemblies from all Pcal strains have been deposited to JGI/IMG-ER under the following accessions: BS91 (project ID 28.208, Taxon OID 2516653056), T3C (project ID 28211, Taxon OID 2516653057), PSa1_3 (project ID 6398, Taxon OID 2512564082), PSa866 (project ID 6397, Taxon OID 2512564081).

Phylogenetic Relationship and Protein Comparison between Pcal and the P. syringae Species Complex

To determine the phylogenetic relationship and differences in protein repertoires between the sequenced Pcal strains, genomes of representative strains of the P. syringae species complex (Table S3) were compared with the annotated Pcal genomes. Results of the protein comparison can be viewed at this URL: http://pacu.facom.ufms.br/Pcal/.

Prediction of protein repertoires is limited by the draft status of most pseudomonad genome sequences so that predicted proteins encoded by genes at the end of contigs may be missing their C-termini or N-termini and may have a higher percentage of sequencing errors due to lower sequencing depth compared to those found in the middle of contigs. However, considering these limitations, there are 710 predicted protein families with members (orthologs and in-paralogs) in each Pcal and each P. syringae genome sequence and 50 predicted protein families that have members (orthologs and in-paralogs) present in each of the five Pcal genome but absent from all sequenced P. syringae genomes. Although some are short hypothetical proteins, approximately 30 proteins have predicted functions and may be indicative of Pcal-specific metabolism and/or virulence mechanisms, for example, a predicted carbamoyl transferase of the NodU family, two predicted acyltransferases, a predicted beta-lactamase, and a multidrug efflux system transmembrane protein (possibly conferring antibiotic and/or plant toxin resistance).

All proteins present exactly one time in each genome of the P. syringae species complex (including the Pcal genomes) were aligned and used to construct a phylogenetic tree (Figure 2). It can be clearly seen that Pcal strains form a separate clade compared to the rest of the P. syringae species complex but still share a more recent common ancestor with P. syringae than with P. viridiflava. P. fluorescens is the Pseudomonas species most closely related to the genome-sequenced P. syringae and P. viridiflava strains. P. fluorescens strains were thus used as outgroup.

thumbnail
Figure 2. Phylogenetic tree of P. cannabina pv. alisalensis (Pcal) and the P. syringae species complex based on core genome proteins.

A phylogenetic tree was constructed based on all proteins present exactly one time in each of the Pcal genome sequences and representative P. syringae genome sequences. The Bootstrap values for all branches of the tree are 100. The genomes and their accession numbers are listed in Table S3.

https://doi.org/10.1371/journal.pone.0059366.g002

Within the Pcal clade, strain PSa866 appears to have diverged significantly from the rest of Pcal. However, since the genome sequence of PSa866 has the lowest relative genome coverage and the PSa866 assembly is more fragmented than the other Pcal genome sequences, this apparent divergence may be the result of sequencing errors.

Alignment between the Draft Genomes of Pcal and the Complete Genomes of Pss B728a and Pst DC3000

We investigated the overall genomic differences between Pcal and Pss B728a and Pst DC3000 by aligning the previously sequenced Pcal ES4326 genome and the Pcal, PSa1_3 and PSa866 genomes sequenced here with the closed genomes of Pss B728a and Pst DC3000 using MAUVE 2.3.1 software [40] (Figure 3 and Figure S2). The Pcal PSa1_3 and PSa866 genomes were highly syntenic with Pcal ES4326 confirming their close relationship (Figure 3B). The alignments between: a) Pcal ES4326 and Pss B728a (Figure 3A), b) Pcal PSa1_3, Pcal PSa866 and Pcal ES4326 (Figure 3B), and c) Pcal ES4326, Pcal PSa1_3 and Pcal PSa866 and the complete genome of Pst strain DC3000 (Figure S2), are also presented. Since all Pcal genomes are unfinished, only genomic rearrangement that occurred within contigs are identifiable and the actual number of rearrangements may therefore be higher.

thumbnail
Figure 3. Genome alignments.

Pairwise alignment between Pcal ES4326 (previously known as P. s. pv. maculicola ES4326) and the compete genome of P. s. pv. syringae B728a (A) and the draft genomes of Lineage-I Pcal PSa866 and Lineage-II Pcal PSa1_3 (B) using the MAUVE software. Colored blocks outline genome sequence that aligned to part of another genome, and was presumably homologous and internally free from genomic rearrangement (Locally Colinear Blocks or LCBs). White regions are sequence that were not aligned and probably contain sequence elements specific to a particular genome. Blocks below the center line indicate regions that aligned in the reverse complement (inverse) orientation. The height of the profile within each LCB demonstrates the average degree of sequence conservation within an aligned region.

https://doi.org/10.1371/journal.pone.0059366.g003

General Features of the Pcal Draft Genome Sequences Compared to Pcal ES4326

Table 1 lists genome features for the sequenced Pcal strains. Alignments of the draft genome assemblies for Pcal PSa1_3, Pcal PSa866 and Pcal ES4326 suggested a high level of conservation across the chromosome. The first two strains had highly syntenic chromosomes, but many syntenic blocks between them were presented on different relative positions in the ES4326 genome (Figure 3B). Based on the total length of Pcal assemblies, as well as the size of the Pcal ES4326 draft genome (6.221.751 bp), we estimate the depths of coverage and that the 96,8% of Pcal PSa1_3 genome and 94,1% of Pcal PSa866 was represented in their respective draft genomes.

Comparison of these Pcal draft genomes to other publically available assemblies, revealed significant heterogeneity for several genomic parameters. Several differences between the Pcal genomes were found in regard to: a) prediction of protein coding genes without function; b) enzyme-encoding genes; c) protein coding genes without assigned enzymatic function, but with similarity to groups of enzymes based on KEGG Orthology (KO); and d) the predicted fusions of protein coding genes (Figure 4; Table 1). However, because of the draft status of the assembled genomes it is unclear whether these differences represented true differences between strains or they were sequencing or assembly artifacts.

thumbnail
Figure 4. Presentation of the distribution of various gene categories (COGs) in the respective Pcal strains PSa1_3, PSa866, BS91, T3C and ES4326.

https://doi.org/10.1371/journal.pone.0059366.g004

In an attempt to predict and assemble the plasmids harbored by the strains under investigation, reference assemblies utilizing publicly available plasmid sequences from strain ES4326 were used [38]. Pcal ES4326 contained five plasmids: pPCAL4326A (46.697 bp), pPMA4326B (40.110 bp), pPMA4326C (8.244 bp), pPMA4326D (4.833 bp) and pPMA4326E (4.217 bp). We found that sequences showing homology to regions of plasmids pPMA4326A and pPMA4326B were present in all Pcal strains, but no evidence was found for the presence of pPMA4326C, pPMA4326D and pPMA4326E in any of them. Surprisingly, the Pcan BS0968 strain harbor sequences showing homology to regions of three of the Pcal ES4326 plasmids (pPMA4326A, pPMA4326B and pPMA4326C) but not pPMA4326D and pPMA4326E (Table S2).

Secretion Systems in P. cannabina pv. alisalensis

All five Pcal strains (PSa1_3, PSa866, BS91, T3C and ES4326) possessed loci with similarity to structural genes involved in the biogenesis of the type I, II, III and VI secretion systems. Genes for Autotransporter-1 (AT-1) family proteins of the type V secretion system, the general secretion (Sec) pathway, and the twin-arginine translocation (Tat) pathway were also present in all five Pcal strains. All of the above mentioned secretion systems are present in several other sequenced plant pathogenic pseudomonads, with the exception of the T1SS, which is absent in the genome of Pst DC3000 [41], [42].

Complete hrp/hrc gene clusters encoding T3SSs of the “Hrp-1 family” [43] were also found in all five Pcal strains (GenBank Accession N° of Pcal PSa1_3 T3SS: JQ517282) and closely resembled the hrc/hrp cluster of P. syringae B728a. Strains Pcal PSa1_3 and Pcan BS0968 harbored a complete T4SS (classified as a “P” group T4SS according to Souza et al. [44]) as was previously reported for strain Pcal ES4326 (located on plasmid pPMA4326A [37]). However, only a portion of this T4SS was identified in the Pcal PSa866 genome. Surprisingly, the other two Pcal strains (T3C and BS91) did not appear to carry genes related to this system, even though all three appear to harbor a plasmid similar to pPMA4326A of Pcal ES4326.

Each of the five Pcal genomes contained two clusters coding for predicted T6SSs (GenBank Accession N° for Pcal PSa1_3 T6SS-I: JQ517283 and for Pcal PSa1_3 T6SS-II: JQ517284) [24]. The functionality of these two T6SSs and their exact role during pathogenesis or other aspects of Pcal biology has yet to be investigated. Interestingly, recently a role for the Pst DC3000 T6SS and the T6SS cargo protein Hcp2 in Pst survival in competition with enterobacteria and yeasts was reported [25].

Organization of the hrp/hrc Cluster in P. cannabina pv. alisalensis

The hrp/hrc clusters in all five examined Pcal genomes (PSa1_3, PSa866, BS91, T3C, and ES4326) were identical in size and orientation (Figure 5). The cluster spans 25,658 bp, and is composed of twenty-eight genes. In most of the Pcal strains T3SS genes were arranged in five operons organized in two main blocks having convergent transcription: the hrpRS, hrpZ and hrpC transcriptional units in one orientation, and the hrpU and hrpJ transcriptional units in the opposite direction. The two Pcal T3SS gene blocks were separated by a hyper variable region that was previously shown to have very low level of conservation even between taxonomically closely related bacteria [45]. In the Pcal genomes, this hyper variable region harbors two ORFs with unknown function, which were also present in Pss B728a (Psyr_1204, Psyr_1203) and in P. syringae pv. aceris str. M302273 (PSYAR_01264, PSYAR_01274) but are absent from all other sequenced P. syringae genomes. The two genes were located upstream of the hrp/hrc cluster (near hrpL and hrpK) and were not part of any operon.

thumbnail
Figure 5. Schematic representation of the hrp/hrc gene cluster coding for the synthesis of the type III secretion system of Pseudomonas cannabina pv. alisalensis.

The genes were arranged in five operons organized in two main blocks having convergent transcription: the hrpRS, hrpZ and hrpC operons in one orientation, and the hrpU and hrpJ in the opposite direction. The two blocks were separated by a hyper variable region with a very low level of conservation between closely taxonomically related bacteria.

https://doi.org/10.1371/journal.pone.0059366.g005

This kind of organization has also been reported for other P. syringae pathovars harboring the “Hrp-1 family” hrp/hrc T3SS [20].

The G+C content of the entire T3SS cluster for all five Pcal strains (57.9%) was in the range reported for other hrp/hrc clusters of phylogenetically related P. syringae pathovars (56–58% [46]). The T3SS cluster of Pcal strains appears to be syntenic with all other phytopathogenic P. syringae strains, and appears chromosomally located and present in a single copy as assessed by BlastN searches. Our analysis also revealed thirty three Single Nucleotide Polymorphisms (SNPs) between the Pcal PSa1_3, PSa866 and T3C hrp/hrc clusters (Table S1). Identity between predicted Hrp/Hrc proteins from Pcal and those from Pss B728a, Pst DC3000 and Psph 1448a range from 46% (for HrpA of Pst DC3000) to 97% (for HrpF, hrpT, and HrpV of Pss B728a) (Table 2). Of note, most Pcal Hrp/Hrc proteins shared highest identity with those of Pss B728a, which further supported the close relationship between the T3SSs of Pcal and Pss B728a. This is surprising because the Pss B728a strain is more distantly related to the Pcal strains compared to other sequenced P. syringae strains.

thumbnail
Table 2. Aminoacid sequence identity percentages between predicted Hrp/Hrc proteins from P. cannabina pv. alisalensis and those from P. syringae pv. syringae B728a, P. syringae pv. tomato DC3000 and P. syringae pv. phaseolicola 1448a.

https://doi.org/10.1371/journal.pone.0059366.t002

Phylogenetic Analysis of T3SS Clusters in P. cannabina pv. alisalensis Strains and Other Species of P. syringae sensu lato

We completed several phylogenetic analyses comparing several previously sequenced P. syringae isolates together with these five Pcal strains. The analysis included: a) comparison of the entire hrp/hrc cluster of Pss B728a, Pst DC3000 and the hrp/hrc cluster from the five Pcal strains (Figure 6) and b) phylogenetic analysis and comparison based on the nucleotide and amino acid sequences of the hrcV (Figure 7) and hrpZ and hrcC genes (Figure S3), which each represent an individual transcriptional unit of the T3SS gene cluster. The comparison of the entire T3SS cluster between investigated Pcal strains revealed significant conservation (Figure 6).

thumbnail
Figure 6. hrp/hrc clusters comparison.

Pairwise alignment between the hrp/hrc gene clusters coding for the type III secretion systems of the complete sequenced P. syringae pv. syringae B728a and P. syringae pv. tomato DC3000 and those of P. cannabina pv. alisalensis PSa1_3, PSa866 and T3C. Areas where white is visible, were not aligned and probably contained sequence elements specific to a particular genome. The height of the profile demonstrates the average degree of sequence conservation within an aligned region.

https://doi.org/10.1371/journal.pone.0059366.g006

thumbnail
Figure 7. T3SS core component HrcV phylogenetic analysis.

For the phylogenetic analysis the amino acid (A), as well as the nucleotide sequences (B), were used. Information for additional phylogenetic analysis of various T3SS core components can be found in figures S3. The various MLSA groups [43], [45] are marked in different colors.

https://doi.org/10.1371/journal.pone.0059366.g007

Phylogenetic characterization of P. syringae strains from different pathovars by Multilocus Sequence Analysis (MLSA) previously revealed five major groups of P. syringae [47], [48]. A recent study demonstrated that MLSA groups correspond to the nine genomospecies: genomospecies 1, 2, 3, 4 and 9 corresponds to MLSA groups 2, 3, 1, 4 and 5, respectively [47]. Members of genomospecies 6 and 8 are also distinct by MLSA but have not been given an MLSA group number. Likewise, studies based on the phylogenetic analysis of selected hrp/hrc genes investigated the relationship between hrp/hrc phylogenetic clustering and the MLSA grouping of P. syringae pathovars [43], [48], [49]. Pcal strain ES4326 clusters with MLSA group 5 (genomospecies 9), based on the concatenated DNA sequence of seven and four housekeeping genes, respectively [48], [49].

Our results based on the phylogenetic analysis of nucleotide as well as amino acid sequences of hrcV revealed nearly identical topologies. The Pcal T3SSs were grouped together and linked with P. syringae pathovars from the MLSA Group 2 (Figure 7; genomospecies 1). However, phylogenies based on hrpZ and hrcC amino acid and nucleotide sequences differed compared to hrcV (Figures 7 and S3). In both, hrpZ and hrcC trees, the Pcal strains clustered more closely with pv. mori and pv. lachrymans from MLSA Group 3 (genomospecies 2) with high bootstrap support (Figure S3). This unusual relationship among the MLSA groups 3 and 5 has been previously reported for the hrpZ operon and interpreted as a result of a putative recombination event [43]. Since our results revealed that the hrcC tree had the same basic topology as the hrpZ tree and since hrcC did not directly flank the hrpZ operon (Figure 5), we concluded that the recombination event similar to that inferred in [43] affected a considerably larger region than originally expected.

These data suggest that P. syringae and Pcal clades share a common ancestor that likely acquired the hrp/hrc cluster by a single Horizontal Genetic Transfer (HGT) event prior to their speciation [50], followed by localized HGT events within the T3SS.

Phylogenetic analysis also demonstrated that hrpZ and hrpA (Figure S3 and Table 2) were some of the most variable genes among the P. syringae and Pcal strains, confirming earlier reports from several P. syringae pathovars [46]. For example, the Pst DC3000 and Pss 61 hrpA genes are only 30% similar [46]. The genes coding for these two proteins are located in the hrpZ operon of the hrp/hrc cluster (Figure 5). Based on this fact, specific primers were designed to exploit this variability for diagnostic purposes in order to differentiate P. syringae strains at the pathovar level [51]. Based on our data (Table 2), the Pcal hrpA and hrpZ gene sequences would be useful targets for designing Pcal-specific primers.

Distribution of Type III Effector Genes among P. cannabina pv. alisalensis Strains

Using the Hop Database (HDB; http://www.pseudomonas-syringae.org), nucleotide and/or amino acid sequences of type III effectors (T3EPs) were used to screen the draft genome sequences of all five Pcal strains. This allowed us to identify the core T3EPs conserved in all five strains as well as strain-specific effectors (Figure 8A). A subset of 19 genes coding for T3EPs were found to form the core T3EP repertoire of Pcal: avrE1, avrPto5, hopAA1a, hopAA1b, hopAB3-1, hopAF1, hopAL1, hopAO1, hopAQ, hopAS1, hopD1, hopI1, hopM1, hopR1, hopV1, hopW1-1, hopW1-2, and hopX1a (Figure 8A). Conservation implied an important role for these T3EPs in pathogenicity of Pcal. A second set of genes coding for T3EPs (avrRpm1, hopAB3-2, hopAC1, hopAD, hopAE1, hopAH1, hopQ1, HopX2, and hopAD1) were present in all five strains but truncated or with an in-frame stop codon in at least one of the Pcal strains. A third subset of genes coding for T3EPs was strain-specific: avrPto1, hopAM1, hopAT1, hopAU1, hopAV1, hopAW1, hopAY1, hopAZ1, hopBD1, hopBD2, hopBF1, hopBG, hopE1, hopG, hopO1, hopT1, hopX1b, and hopZ1. Finally, there was a subset of truncated strain-specific effectors: hopAG1, hopAR1, hopBB1, and hopH1 (Figure 8A).

thumbnail
Figure 8. Comparison of the sequenced Pcal strains genomes, as well as of the Pcal ES4326 and P. cannabina pv. cannabina BS0968 genomes, for genes encoding candidate T3SS secreted substrates.

In addition to known hop effector genes, two genes for T3SS helper proteins are also included (hrpZ1 and hrpW1) (A). Occurrence of effector genes in the six strains is indicated: Absence (white), Presence (gray), Truncation (bordeaux), disrupted by a premature stop codon (green). Comparison tree contracted using the information presenting in the presence-absence table (part A of the same Figure) (B).

https://doi.org/10.1371/journal.pone.0059366.g008

Pcal strains formed two distinct lineages (Lineage-I and Lineage-II) according to their T3EP repertoires (Figure 8B), which mirrored the two lineages established within Pcal by rep-PCR and MLSA [3]. Over 150 strains in Lineage I, including the pathotype strain BS91, were isolated from brassica species [3], while all strains of Lineage-II (including PSa1_3 and ES4326) were isolated from radish or arugula. The only exception is one strain isolated from radish in Germany during an outbreak in 2008 that was attributed to Lineage-I based on rep-PCR [17]. All sequenced strains in Lineage-I (PSa866, T3C and BS91) had identical T3EP repertoires. However, there were some differences in the genes coding for T3EPs between strains of Lineage-II (ES4326 and PSa1_3), for example, hopAM1, hopAU1, hopG1 were present in PSa1_3 but not in ES4326. Future studies will focus on whether these differences affect virulence and host range.

Phylogenetic Relationships of P. cannabina pv. alisalensis Type III Effectors

We used amino acid sequences of 34 Pcal T3EP families, for which there are more than three orthologues in the Hop Database (HDB), for phylogenetic analysis. Cluster comparisons were based on the presence of common and consistent clades (bootstrap values>70). Pcal T3EPs were grouped with P. syringae T3EPs. Some Pcal T3EPs were identical or similar to other known effectors deposited in the HDB. This group included the T3EPs AvrPto1, HopAF1, HopAO1, HopD1, HopO1, HopQ1, HopT1, HopW1-2, HopX1b, HopX2, AvrRpm1, HopZ1, HopG1, HopAW1, HopBF1, HopAQ1, and HopAD1 (Figure 9 and Figure S4). However, HopD1 and HopAQ1 of Pcal ES4326 and PSa1_3 clustered separately from their orthologues in the other Pcal genomes and Pcan BS0968. A second group of Pcal T3EPs grouped only weakly with known effectors. This group included orthologs of the effector proteins AvrPto5, HopAZ1, HopBD1, HopR1, HopW1-1, HopE1, HopAM1 (Figure 9 and Figure S4). However, the effector protein HopAM1 of Pcal PSa1_3 was phylogenetically divergent from the other Pcal and Pcan orthologs. A third group of Pcal T3EPs were phylogenetically distant from their P. syringae orthologues. This group of proteins included the effectors: AvrE1, HopAA1a, HopAA1b, HopAB3-1, HopAB3-2, HopAL1, HopAS1, HopBD2, HopM1, HopI1, HopX1a, HopAU, HopAV1, HopAY1, HopAC1 and HopBG1. It is noteworthy that the effector HopBG1 was found only in Pcal pathovars and specifically only in Pcal ES4326 and PSa1_3 strains; therefore, due to the lack of HopBG1 orthologues in the HDB we were unable to construct a phylogenetic tree based on this protein.

thumbnail
Figure 9. T3SS effector proteins (T3EPs) phylogeny.

For the phylogenetic analysis, the amino acid sequences of every effector group were used, as they are presented in the Hop Database website. Additional information for the phylogeny of the rest of Pcal T3EPs can be found in figure S4.

https://doi.org/10.1371/journal.pone.0059366.g009

Some of the strains harbored multiple phylogenetically distinct paralogs of several effectors, which were not likely the products of duplication events. The three Pcal strains belonging to Lineage-I (T3C, BS91 and PSa866), as well as Pcan BS0968, harbored two AvrPto copies (AvrPto1 and AvrPto5), while the strains of Lineage-II (Pcal PSa1_3 and ES4326) carried only one copy (AvrPto5). On the other hand, Lineage-I strains carried three HopX xenologs loci (HopX1a, HopX1b and HopX2), while the Lineage-II strains, Pcal PSa1_3 and ES4326, as well as the Pcan BS0968, harbored only two (HopX1a and HopX2) (Figure 9). The Pcal HopX1b orthologues resembled HopX1 from Pst DC3000 that has been reported to enhance RNA-mediated gene silencing (RNAi) in Nicotiana benthamiana independently of plant R-gene recognition [52]. Therefore, it will be interesting to examine if the newly identified Pcal HopX proteins have a similar activity. All strains examined harbored two copies of the T3EP HopAA (HopAA1a and HopAA1b), HopAB (HopAB3-1 and HopAB3-2) and HopW (HopW1-1 and HopW1-2). Lastly, HopBD1 was present only in Lineage-II strains (Pcal PSa1_3 and ES4326), each carrying two divergent copies (HopBD1 and HopBD2; Figures 8A and Figure S4).

Type VI Secretion System Gene Clusters in the P. cannabina pv. alisalensis Genomes

Our analysis of Pcal genomes revealed two genomic clusters coding for putative T6SSs (T6SS-I and T6SS-II). The T6SS-I locus consisted of a set of 16 conserved T6SS core genes (Figure 10). This locus is syntenic and phylogenetically closely related to the P. aeruginosa (Paer) T6SS-I [53] (formally HSI-I; Pseudomonas spp. T6SS group 3, according to Barret et al., [54]), with the exception of a replacement of the P. aeruginosa pppB gene with impM in the Pcal genome and an insertion of two putative ORFs apparently unrelated to type VI secretion that interrupt the synteny of the hcp and impG ORFs. The HSI-I type of T6SS is also present in several other Pseudomonas species, including the plant root-associated bacterium Pseudomonas brassicacearum subsp. brassicacearum NFM421.

thumbnail
Figure 10. Schematic representation of the gene cluster coding for the type VI secretion systems (T6SSs) of Pseudomonas cannabina pv. alisalensis.

https://doi.org/10.1371/journal.pone.0059366.g010

In addition to structural core proteins that serve as landmarks of T6SS loci, a serine-threonine protein kinase gene with homology to ppkA (Stk1), as well as its cognate phosphoprotein phosphatase pppA (Stp1), were also found within the Pcal T6SS-I locus (Figure 10). We also searched for the presence of the tse (tse1, tse2 and tse3) toxin genes. Tse2 (PA2702) is a toxic protein, first described in P. aeruginosa strain PAO1, that affects growth of both prokaryotic and eukaryotic cells when expressed intra-cellularly and is a proposed substrate of the P. aeruginosa T6SS-I [55]. P. aeruginosa also codes for an immunity protein, Tsi2 (PA2703) that prevents Tse2-dependent cell death [55]. We did not identify any orthologs of either tse1, tse3, or the tse2/tsi2 regulon in any of the Pcal genomes examined. Since these genes were missing from all five draft genomes this is probably not due to poor genome coverage but rather was likely a reflection of differences in biology between P. aeruginosa and Pcal.

The second T6SS cluster (T6SS-II) in the Pcal genomes was highly similar to the T6SS-II cluster of Pst DC3000 (Pseudomonas spp. T6SS group 1.1) [23], [24], [54], except that the Pcal cluster lacked the pppA and ppkA regulatory genes. Again, the absence of these two genes from all five draft genome sequences suggested that the absence of these genes was not an artifact of poor genome coverage.

Phylogenetic Analysis of P. cannabina pv. alisalensis T6SS Proteins

Two phylogenetic trees were constructed using the T6SS core proteins ImpL (Figure S5) and ClpV/B (Figure S6). Various Pseudomonas species exhibiting different phylogenetic relatedness to Pcal were included. Both phylogenetic trees showed three deep branches. The first deep branch in the ImpL tree (marked as Paer and Pcal group), contained only Pcal T6SS-I along with T6SS-I of P. aeruginosa, which were phylogenetically separated from all other P. syringae and P. aeruginosa T6SSs. The next two branches were as described previously [23], [24]; the first (Psph group) included the P. syringae T6SSs-I, while the second branch (Pst group) comprised the Pcal T6SS-II, the P. syringae T6SS-II and T6SS-III along with the sole T6SS of Pseudomonas fluorescens Pf0-1, P. aeruginosa T6SS-II and P. entomophila T6SS-II. The P. aeruginosa T6SS-III was present as a separate group in the ImpL phylogeny. The ClpV/B tree had three branches (Figure S6). However, while the Pcal T6SS-I grouped again with Paer T6SS-I (Pcal and Paer group), the P. syringae T6SS-I branched separately as previously reported [23], [24].

Phylogenetic analysis of the T6SS loci provides some insight into the potential function of these T6SS-I: it tightly links the Pcal T6SS-I clusters with the P. aeruginosa T6SS-I, (HSI-1) which is proposed to be essential for inter-bacterial interactions [55], [56], suggesting a similar role for the Pcal T6SS-I.

However, unlike P. syringae, P. aeruginosa and other Pseudomonas species the Pcal strains did not harbor multiple copies of vgrG and hcp genes [23], [53].

Bacterial Toxin Genes and Gene Clusters

Our analyses revealed the presence of genes and gene clusters coding for: a) coronatine biosynthesis, with high levels of similarity to the Pst DC3000 coronamic acid synthetases: 100% identity with the CmaD (PSPTO_4707), 99% identity with CmaE (PSPTO_4708), 93% identity with the CmaA (PSPTO_4709), 100% identity with the CmaB (PSPTO_4710), 100% identity with the CmaC (PSPTO_4711) and finally 99% identity with CmaT (PSPTO_4712), b) An ethylene-forming enzyme (efe). Although the IAA-lysine synthetase (iaaL) was present in all five Pcal strains, with 96% identity with the IaaL ortholog of the Pst DC3000 (PSPTO_0371), we did not identify orthologs of the iaaM/iaaH genes. Furthermore, all five Pcal strains possessed a gene coding for tabtoxin resistance protein (TTR) with identity 92% to the tabtoxin resistance protein of Pseudomonas syringae pv. tabaci strain ATCC 11528 (PsyrptA_8578), an enzyme involved in self-protection found in tabtoxin-producing pathogens. Interestingly, there were several remnants of what appear to be members of the tabtoxin biosynthesis gene cluster in the Pcal genomes.

Conclusion

Comparison of five P. cannabina pv. alisalensis genomes with each other and with several P. syringae genomes revealed that virulence genes have been exchanged between these pathogens during their evolution and that virulence mechanisms and regulation of virulence gene repertoires between P. syringae and P. cannabina largely overlap.

Within Pcal, differences between strains with regard to T3EP repertoires reflected the existence of two genetic lineages that were previously identified [3]. Interestingly, although strains belonging to these two lineages were isolated from different crops no host range differences could be identified experimentally suggesting that the two lineages may have separated by factors other than host range evolution. However, experimental manipulation of T3EP repertoires may still reveal differential roles of effectors in virulence on different hosts. Likewise, analysis of Pcal genomes showed two genomic clusters coding for T6SSs (T6SS-I and T6SS-II), the first of which is syntenic and related to the P. aeruginosa (Paer) T6SS-I. However, the possible contribution of the T6SS in pathogenicity of Pcal species remains to be studied.

Pcal is a particularly interesting pathogen because it can cause disease on both monocots and dicots. Now that the virulence gene repertoires of Pcal have been predicted, roles for these genes in pathogenicity can be investigated in a variety of plant systems. Such studies will not only provide mechanistic insights into molecular interactions underlying virulence, but also provide a framework for understanding selection pressures that shape evolutionary dynamics for bacterial phytopathogens across divergent plant hosts.

Materials and Methods

Bacterial Strains

To evaluate the core genome and strain specific characters for Pcal, we sequenced four strains selected in order to differ in the genera of host plant and geographic origin: the pathotype strain Pcal BS91 (CFBP 6866) which was isolated in USA from Brassica rapa subsp. rapa [3]; the Pcal strain T3C (CFBP 7684), which was isolated from the Turnip cv. ‘topper’ in Lexington, South Carolina [39]; the strain Pcal PSa1_3 (CFBP 7682) that was isolated in Greece from Eruca sativa [10]; the strain Pcal PSa866 (CFBP 7683) a re-isolation from artificially infected tomato plants with a Pcal strain of unknown origin. Additionally, we used the genomic sequence available for Pcal ES4326 for comparisons with the other Pcal strains [37]. All strains were isolated from symptomatic plants revealing the typical symptoms and were used for successful artificial inoculations on various plant species (Figures 1 and S1).

Bacterial Cultures and Genomic DNA Preparation

The strains examined were grown at 26–28°C in King’s medium B broth for 24 h. From these cultures, cells were washed with sterile 10 mM MgCl2, and a cell suspension was prepared, which was adjusted to an OD600 of 0.4. Cultures were stored in aliquots of 500 µL KB in 2 mL cryo-tubes at −80°C. For DNA extraction, bacterial cells from over-night cultures were spun down, and the pellets kept on ice before further use. Total bacterial DNA isolation was carried out using the DNeasy Blood & Tissue Kit from QIAGEN (UK) according to the manufacturer’s instructions.

Library Preparation and Sequencing

Library preparation for Pcal PSa866 and Pcal PSa1_3 sequencing was performed with the Ion Fragment Library Kit (Life Technologies, Darmstadt, Germany) according to the protocol (part no. 4467320 rev. A, 04/2011) with minor modifications. Size selection was done with E-GelH Size Select 2% Agarose (Invitrogen, UK) for strains Pcal PSa1_3 and Pcal PSa866. Template preparation was carried out with the Ion Xpress TM Template Kit (Life Technologies) according to the Ion Xpress TM Template Kit User Guide (part no. 4467389 Rev. B, 05/2011). Emulsified Ion Sphere TM particles were collected by centrifugation (2200 g for 8 min) in a SOLiDH emulsion collection tray (Life Technologies). After centrifugation a clear oil phase developed above a white solid pellet. The oil layer was decanted and pelleted Ion Spheres were re-suspended with 700 µl of breaking solution followed by two washes of the emulsion collection tray with breaking solution. In a departure from the User Guide, all three fractions were pooled in the same 2 ml reaction tube. Washing of the recovered Ion Sphere particles was performed as described in the protocol. The Ion Sequencing Kit (Life Technologies) was used with the Personal Genome Machine TM (PGM TM) sequencer as described in the Ion Sequencing Kit User Guide (part no. 4467391 rev. B, 04/2011). Enriched ISPs were prepared for sequencing as described in the protocol and deposited on the chip in three consecutive loading cycles. Each cycle was composed of the following steps: (i) adjust sample volume to 19 µl with annealing buffer, (ii) 10 sec sonication followed by a quick spin, (iii) re-suspension by pipetting and loading of 6 µl of the sample to the chip, and (iv) 3 min centrifugation using the custom centrifuge adapter/rotor.

Genomic DNA of strains Pcal BS91 and Pcal T3C was extracted using the Gentra Puregene Yeast/Bacteria kit (Qiagen) following manufacturer’s instructions. Illumina paired-end sequencing was performed as previously described [57].

Whole-genome Assembly and Alignment of Illumina and Ion Torrent Genomes

For Pcal PSa866 and Pcal PSa1_3 paired reads of about 100 nts were assembled into contigs using the de novo assembly, as well as using the reference assembling option of the CLC genomic workbench (CLC-bio, Aarhus, Denmark) using as reference the Pcal ES4326 genome [37]. Gene calling was performed by genomic annotation tools of JGI/IMG-ER (https://img.jgi.doe.gov/cgi-bin/er/main.cgi). The genome sequence analysis was performed with the aid of the software package MAUVE [40], specialized for construction of multiple genome alignments in the presence of large-scale evolutionary events such as rearrangements and inversions. Additional assessment of the effector genes repertoire was performed based upon the methods of Baltrus [37].

For Pcal BS91 and T3C paired-end reads were assembled de novo using Velvet 0.7.55 [58] as previously described [57] with scaffolding turned on.

Phylogenetic Analyses

Phylogenetic analysis of specific genes was carried out using the sequences obtained from all sequenced strains plus corresponding sequences retrieved from the GenBank. Sequence alignments were carried out using the program CLUSTALW [59]. The phylogenetic trees were established using the Neighbour-Joining method [60]. The percentage of replicate trees in which the associated strains clustered together in the bootstrap test for 1500 replicates [61] was estimated and is shown next to the tree nodes. The trees were drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic trees. The evolutionary distances were computed using the Maximum Composite Likelihood method [62] and are in the units of the number of base substitutions per site. All positions containing gaps and missing data were eliminated from the dataset (complete deletion option). Phylogenetic analyses of specific genes were conducted in MEGA5 [63].

The whole genome phylogenetic analysis was conducted using all-against-all Blastp [64] comparisons of protein sequences of 41 genomes (http://pacu.facom.ufms.br/Pcal/genomes). Sequences were clustered in 13.607 families using OrthoMCL [65]. From this total, 710 families were selected that contained exactly one protein per genome (except for the outgroup genomes of the P. fluorescens strains, which contained a maximum of one protein per family). Each of the 710 families was then aligned using MUSCLE [66]. Non-informative columns of each alignment were removed using Gblocks [67] and all alignments were concatenated, resulting in 210,004 positions. Finally, RAxML [68] with model PROTGAMMAWAGF was used on this alignment to build the final species tree, shown in Figure 2.

Pathogenicity Tests

In preliminary assays all isolates were screened for their ability to induce hypersensitive reaction on tobacco leaves and to cause typical symptoms on their natural hosts by previously described methods [3], [10]. All plants used for inoculations, were grown in separate 20 cm diameter pots in a 3∶1:1 compost, peat and perlite mix. Plants were watered with surface drip irrigation. Fertilizer 20-20-20 (N-P-K) was applied weekly by watering. Inoculations were carried out on known host plants. Ten plants of each host were inoculated with the examined strains (Figures 1 and S1). For foliar inoculations on the host plants, a suspension of each isolate was sprayed onto leaves until run off with a hand sprayer. The bacterial inocula were prepared from 24 hrs old King’s medium B plate cultures, suspended in sterile distilled water and adjusted to approximately 106 CFU·mL−1 based on turbidity measurement at 600 nm and by dilution plate counts. Control plants were sprayed with sterile distilled water. Inoculations of tomato plants were performed by first slightly wounding the plant leaves and stems with a soft plastic brush immediately prior to spray inoculation as described above. Controls were similarly treated. All inoculated plants were held in a greenhouse (10–30°C) under intermittent mist (10 sec each hour) and symptoms were evaluated for two weeks after inoculation. Cross-inoculation tests were made on several plant species (Figures 1 and S1). Bacterial colonies were re-isolated from infected plants and the re-isolates had the same LOPAT profile as the original isolates of Pcal. Isolates from control plants did not reveal any pathogenic bacteria, and published host ranges were confirmed [3], [10], [18], [39], [69].

Supporting Information

Figure S1.

Compatible reactions to artificial inoculations of Pseudomonas cannabina pv. alisalensis (Pcal) on various plant species. Artificial inoculations were performed using the sequenced Pcal strain PSa1_3 on: Brassica oleracea, Eruca sativa, and Brassica napus. Information for additional artificial inoculations on other plant species can be found in Figure 2.

https://doi.org/10.1371/journal.pone.0059366.s001

(PDF)

Figure S2.

Pairwise alignment between the compete genome of P. s. pv. tomato DC3000 and the draft genome of Pcal ES4326 (previously known as P. s. pv. maculicola ES4326) (A) and the draft genomes of Pcal PSa1_3 and Pcal PSa866 (B) using the MAUVE software. Colored blocks outline genome sequence that aligned to part of another genome, and is presumably homologous and internally free from genomic rearrangement (Locally Colinear Blocks or LCBs). Areas that are completely white were not aligned and probably contained sequence elements specific to a particular genome. Blocks below the center line indicate regions that aligned in the reverse complement (inverse) orientation.

https://doi.org/10.1371/journal.pone.0059366.s002

(PDF)

Figure S3.

T3SS core component HrpZ and HrcC phylogenetic analysis. For the phylogenetic analysis the amino acids, as well as the nucleotide sequences were used.

https://doi.org/10.1371/journal.pone.0059366.s003

(PDF)

Figure S4.

T3SS effector proteins (T3EPs) phylogeny. For the phylogenetic analysis, the amino acid sequences of the Pcal effector as well as of other effectors, as they are presented in the Hop Database website, were used. Additional information for the phylogeny of the rest of Pcal T3EPs can be found in Figure S8.

https://doi.org/10.1371/journal.pone.0059366.s004

(PDF)

Figure S5.

Phylogenetic analysis based on the protein sequences of the T6SS component, ImpL. For the phylogenetic analysis the amino acid sequences were used. Information for additional phylogenetic analysis of various T6SS core components can be found in figure S6.

https://doi.org/10.1371/journal.pone.0059366.s005

(PDF)

Figure S6.

Phylogenetic analysis based on the T6SS protein sequences of the T6SS ATPase, ClpV/B. For the phylogenetic analysis the amino acid sequences were used. Information for additional phylogenetic analysis of various T6SS core components can be found in figure S5.

https://doi.org/10.1371/journal.pone.0059366.s006

(PDF)

Table S1.

Single nucleotide polymorphisms between Pcal PSa1_3, PSa866 and T3C hrp/hrc clusters.

https://doi.org/10.1371/journal.pone.0059366.s007

(DOCX)

Table S2.

Plasmids present in Pcal strains.

https://doi.org/10.1371/journal.pone.0059366.s008

(DOCX)

Table S3.

Genomes used for the phylogenetic tree shown in Figure 2 and for the comparison of protein repertoires that can be queried at http://pacu.facom.ufms.br/Pcal/.

https://doi.org/10.1371/journal.pone.0059366.s009

(XLSX)

Acknowledgments

The authors would like to thank Mr C. Zoumadakis and Dr. E. Stratidakis (IMBB-FORTH, Greece) for their help in Ion Torrent sequencing.

Author Contributions

Conceived and designed the experiments: PFS CDJ JLD BAV. Performed the experiments: PFS EAT DAB WPW SY NFA. Analyzed the data: PFS EAT DAB CTB SY JLD NJP BAV NFA. Contributed reagents/materials/analysis tools: FV CDJ JLD BAV DEG. Wrote the paper: PFS EAT DAB CTB NJP BAV.

References

  1. 1. Sutic D, Dowson WJ (1959) An investigation of a serious disease of hemp (Cannabis sativa L.) in Yugoslavia. Phytopath Z 34: 307–314.
  2. 2. Gardan L, Shafik H, Belouin S, Broch R, Grimont F, et al. (1999) DNA relatedness among the pathovars of Pseudomonas syringae and description of Pseudomonas tremae sp. nov. and Pseudomonas cannabina sp. nov. (ex Sutic and Dowson 1959). Int J Syst Bacteriol 49: 469–478.
  3. 3. Bull CT, Manceau C, Lydon J, Kong H, Vinatzer BA, et al. (2010) Pseudomonas cannabina pv. cannabina pv. nov., and Pseudomonas cannabina pv. alisalensis (Cintas Koike and Bull, 2000) comb. nov., are members of the emended species Pseudomonas cannabina (ex Sutic & Dowson 1959) Gardan, Shafik, Belouin, Brosch, Grimont & Grimont 1999. Syst Appl Microbiol 33: 105–115.
  4. 4. Bull CT, Rubio I (2011) First report of bacterial blight of crucifers caused by Pseudomonas cannabina pv. alisalensis in Australia. Plant Dis 95: 1027–1027.
  5. 5. Williams PH, Keen NT (1966) Bacterial blight of radish. Plant Dis Rep 50: 192–195.
  6. 6. Hendrickson EL, Guevera P, Penaloza-Vazquez A, Shao J, Bender C, et al. (2000) Virulence of the phytopathogen Pseudomonas syringae pv. maculicola is rpoN dependent. J Bacteriol 182: 3498–3507.
  7. 7. Cui J, Jander G, Racki LR, Kim PD, Pierce NE, et al. (2002) Signals involved in Arabidopsis resistance to Trichoplusia in caterpillars induced by virulent and avirulent strains of the phytopathogen Pseudomonas syringae. Plant Physiol 129: 551–564.
  8. 8. Wang L, Mitra RM, Hasselmann KD, Sato M, Lenarz-Wyatt L, et al. (2008) The genetic network controlling the Arabidopsis transcriptional response to Pseudomonas syringae pv. maculicola: roles of major regulators and the phytotoxin coronatine. Mol Plant Microbe Interact 21: 1408–1420.
  9. 9. Guttman DS, Vinatzer BA, Sarkar SF, Ranall MV, Kettler G, et al. (2002) A functional screen for the Type III (Hrp) secretome of the plant pathogen Pseudomonas syringae. Science 295: 1722–1726.
  10. 10. Sarris PF, Karri IV, Goumas DE (2010) First report of Pseudomonas syringae pv. alisalensis causing bacterial blight of arugula (Eruca vesicaria subsp. sativa) in Greece. New Disease Reports 22.
  11. 11. Bull CT, du Toit LJ (2008) First report of bacterial blight on conventionally and organically grown Arugula in Nevada caused by Pseudomonas syringae pv. alisalensis. Plant Dis 93: 109–109.
  12. 12. Bull CT, Goldman P, Koike ST (2004) Bacterial blight on arugula, a new disease caused by Pseudomonas syringae pv. alisalensis in California. Plant Dis 88: 1384–1384.
  13. 13. Mauzey SJ, Koike ST, Bull CT (2010) First report of bacterial blight of cabbage (Brassica oleracea var. capitata) caused by Pseudomonas cannabina pv. alisalensis in California. Plant Dis 95: 71–71.
  14. 14. Koike ST, Kammeijer K, Bull CT, O’Brien D (2006) First report of bacterial blight of Romanesco cauliflower (Brassica oleracea var. botrytis) caused by Pseudomonas syringae pv. alisalensis in California. Plant Dis 90: 1551–1551.
  15. 15. Bull CT, Mauzey SJ, Koike ST (2010) First report of bacterial blight of brussels sprouts (Brassica oleracea var. gemmifera) caused by Pseudomonas cannabina pv. alisalensis in California. Plant Dis 94: 1375–1375.
  16. 16. Koike ST, Kammeijer K, Bull CT, O’Brien D (2007) First report of bacterial blight of Rutabaga (Brassica napus var. napobrassica) caused by Pseudomonas syringae pv. alisalensis in California. Plant Dis 91: 112–112.
  17. 17. Rubio I, Hiddink G, Asma M, Bull CT (2012) First report of the crucifer pathogen Pseudomonas cannabina pv. alisalensis causing bacterial blight on Radish (Raphanus sativus) in Germany. Plant Dis 96: 904–904.
  18. 18. Cintas NA, Koike ST, Bull CT (2002) A new pathovar, Pseudomonas syringae pv. alisalensis pv. nov., proposed for the causal agent of bacterial blight of broccoli and broccoli raab. Plant Dis 86: 992–998.
  19. 19. Dale C, Moran NA (2006) Molecular interactions between bacterial symbionts and their hosts. Cell 126: 453–465.
  20. 20. Tampakaki AP, Skandalis N, Gazi AD, Bastaki MN, Sarris PF, et al. (2010) Playing the Harp: Evolution of our understanding of hrp/hrc genes. Annu Rev Phytopathol.
  21. 21. Troisfontaines P, Cornelis GR (2005) Type III Secretion: More systems than you think. Physiology 20: 326–339.
  22. 22. Abby SS, Rocha EP (2012) The Non-Flagellar Type III Secretion System Evolved from the Bacterial Flagellum and Diversified into Host-Cell Adapted Systems. PLoS Genet 8: e1002983.
  23. 23. Sarris P, Skandalis N, Kokkinidis M, Panopoulos N (2010) In silico analysis reveals multiple putative type VI secretion systems and effector proteins in Pseudomonas syringae pathovars. Mol Plant Pathol 11: 795–804.
  24. 24. Sarris PF, Trantas EA, Skandalis N, Tampakaki AP, Kapanidou M, et al. (2012) Phytobacterial type VI Secretion System - Gene distribution, Phylogeny, Structure and Biological functions. In: Cumagun CJ, editor. Plant Pathol: InTech.
  25. 25. Haapalainen M, Mosorin H, Dorati F, Wu R-F, Roine E, et al. (2012) Hcp2, a secreted protein of the phytopathogen Pseudomonas syringae pv. tomato DC3000, is required for competitive fitness against bacteria and yeasts. J Bacteriol.
  26. 26. Ahmed N (2009) A Flood of microbial genomes–Do we need more? PLoS ONE 4: e5831.
  27. 27. Buell CR, Joardar V, Lindeberg M, Selengut J, Paulsen IT, et al. (2003) The complete genome sequence of the Arabidopsis and tomato pathogen Pseudomonas syringae pv. tomato DC3000. Proceedings of the National Academy of Sciences of the United States of America 100: 10181–10186.
  28. 28. Almeida NF, Yan S, Lindeberg M, Studholme DJ, Schneider DJ, et al. (2008) A Draft genome sequence of Pseudomonas syringae pv. tomato T1 reveals a Type III Effector repertoire significantly divergent from that of Pseudomonas syringae pv. tomato DC3000. Mol Plant Microbe In 22: 52–62.
  29. 29. Joardar V, Lindeberg M, Jackson RW, Selengut J, Dodson R, et al. (2005) Whole-genome sequence analysis of Pseudomonas syringae pv. phaseolicola 1448A reveals divergence among pathovars in genes involved in virulence and transposition. J Bacteriol 187: 6488–6498.
  30. 30. Feil H, Feil WS, Chain P, Larimer F, DiBartolo G, et al. (2005) Comparison of the complete genome sequences of Pseudomonas syringae pv. syringae B728a and pv. tomato DC3000. Proceedings of the National Academy of Sciences of the United States of America 102: 11064–11069.
  31. 31. Sohn KH, Jones JD, Studholme DJ (2012) Draft genome sequence of Pseudomonas syringae pathovar syringae strain FF5, causal agent of stem tip dieback disease on ornamental pear. J Bacteriol 194: 3733–3734.
  32. 32. Reinhardt JA, Baltrus DA, Nishimura MT, Jeck WR, Jones CD, et al. (2009) De novo assembly using low-coverage short read sequence data from the rice pathogen Pseudomonas syringae pv. oryzae. Genome Res 19: 294–305.
  33. 33. Studholme D, Ibanez S, MacLean D, Dangl J, Chang J, et al. (2009) A draft genome sequence and functional screen reveals the repertoire of type III secreted proteins of Pseudomonas syringae pathovar tabaci 11528. BMC Genomics 10: 395.
  34. 34. Green S, Studholme DJ, Laue BE, Dorati F, Lovell H, et al. (2010) Comparative genome analysis provides insights into the evolution and adaptation of Pseudomonas syringae pv. aesculi on Aesculus hippocastanum. PLoS ONE 5: e10224.
  35. 35. Rodriguez-Palenzuela P, Matas IM, Murillo J, Lopez-Solanilla E, Bardaji L, et al. (2010) Annotation and overview of the Pseudomonas savastanoi pv. savastanoi NCPPB 3335 draft genome reveals the virulence gene complement of a tumour-inducing pathogen of woody hosts. Environ Microbiol 12: 1604–1620.
  36. 36. Qi M, Wang D, Bradley CA, Zhao Y (2011) Genome sequence analyses of Pseudomonas savastanoi pv. glycinea and subtractive hybridization-based comparative genomics with nine pseudomonads. PLoS ONE 6: e16451.
  37. 37. Baltrus DA, Nishimura MT, Romanchuk A, Chang JH, Mukhtar MS, et al. (2011) Dynamic evolution of pathogenicity revealed by sequencing and comparative genomics of 19 Pseudomonas syringae isolates. PLoS pathog 7: e1002132.
  38. 38. Stavrinides J, Guttman DS (2004) Nucleotide sequence and evolution of the five-plasmid complement of the phytopathogen Pseudomonas syringae pv. maculicola ES4326. J Bacteriol 186: 5101–5115.
  39. 39. Keinath A, Wechter W, Smith J (2006) First report of bacterial leaf spot on leafy Brassica greens caused by Pseudomonas syringae pv. maculicola in South Carolina. Plant Dis 90: 683–683.
  40. 40. Darling ACE, Mau B, Blattner FR, Perna NT (2004) Mauve: multiple alignment of conserved genomic sequence with rearrangements. Genome Res 14: 1394–1403.
  41. 41. Cunnac S, Lindeberg M, Collmer A (2009) Pseudomonas syringae type III secretion system effectors: repertoires in search of functions. Curr Opin Microbiol 12: 53–60.
  42. 42. Bronstein PA, Marrichi M, Cartinhour S, Schneider DJ, DeLisa MP (2005) Identification of a twin-arginine translocation system in Pseudomonas syringae pv. tomato DC3000 and its contribution to pathogenicity and fitness. J Bacteriol 187: 8450–8461.
  43. 43. Guttman DS, Gropp SJ, Morgan RL, Wang PW (2006) Diversifying selection drives the evolution of the type III secretion system pilus of Pseudomonas syringae. Mol Biol Evol 23: 2342–2354.
  44. 44. Souza RC, del Rosario Quispe Saji G, Costa MO, Netto DS, Lima NC, et al. (2012) AtlasT4SS: a curated database for type IV secretion systems. BMC Microbiol 12: 172.
  45. 45. Tegli S, Gori A, Cerboneschi M, Cipriani MG, Sisto A (2011) Type three secretion system in Pseudomonas savastanoi pathovars: does timing matter? Genes 2: 957–979.
  46. 46. Preston GM, Collmer A (2004) The Type III Secretion Systems of plant-associated Pseudomonads: Genes and Proteins on the move. In: Ramos J-L, editor. Pseudomonas: Virulence and gene regulation. New York: Plenum Pub Corp. 181–222.
  47. 47. Bull CT, Clarke CR, Cai R, Vinatzer BA, Jardini TM, et al. (2011) Multilocus sequence typing of Pseudomonas syringae sensu lato confirms previously described genomospecies and permits rapid identification of P. syringae pv. coriandricola and P. syringae pv. apii causing bacterial leaf spot on parsley. Phytopathology 101: 847–858.
  48. 48. Hwang MSH, Morgan RL, Sarkar SF, Wang PW, Guttman DS (2005) Phylogenetic characterization of virulence and resistance phenotypes of Pseudomonas syringae. Appl Environ Microbiol 71: 5182–5191.
  49. 49. Sarkar SF, Guttman DS (2004) Evolution of the core genome of Pseudomonas syringae, a highly clonal, endemic plant pathogen. Appl Environ Microbiol 70: 1999–2012.
  50. 50. Naum M, Brown EW, Mason-Gamer RJ (2009) Phylogenetic evidence for extensive horizontal gene transfer of type III secretion system genes among enterobacterial plant pathogens. Microbiology 155: 3187–3199.
  51. 51. Inoue Y, Takikawa Y (2006) The hrpZ and hrpA genes are variable, and useful for grouping Pseudomonas syringae bacteria. Journal of General Plant Pathology 72: 26–33.
  52. 52. Sarris PF, Gao S, Karademiris K, Jin H, Kalantidis K, et al. (2011) Phytobacterial type III effectors HopX1, HopAB1 and HopF2 enhance sense-post-transcriptional gene silencing independently of plant R gene-effector recognition. Mol Plant Microbe Interact 24: 907–917.
  53. 53. Sarris PF, Scoulica EV (2011) Pseudomonas entomophila and Pseudomonas mendocina: Potential models for studying the bacterial type VI secretion system. Infect Genet Evol 11: 1352–1360.
  54. 54. Barret M, Egan F, Fargier E, Morrissey JP, O’Gara F (2011) Genomic analysis of the type VI secretion systems in Pseudomonas spp.: novel clusters and putative effectors uncovered. Microbiology 157: 1726–1739.
  55. 55. Hood RD, Singh P, Hsu F, Guvener T, Carl MA, et al. (2010) A type VI secretion system of Pseudomonas aeruginosa targets a toxin to bacteria. Cell Host Microbe 7: 25–37.
  56. 56. Mougous JD, Cuff ME, Raunser S, Shen A, Zhou M, et al. (2006) A virulence locus of Pseudomonas aeruginosa encodes a protein secretion apparatus. Science 312: 1526–1530.
  57. 57. Cai R, Lewis J, Yan S, Liu H, Clarke CR, et al. (2011) The plant pathogen Pseudomonas syringae pv. tomato Is genetically monomorphic and under strong selection to evade tomato immunity. PLoS Pathog 7: e1002130.
  58. 58. Zerbino DR, Birney E (2008) Velvet: algorithms for de novo short read assembly using de Bruijn graphs. Genome Res 18: 821–829.
  59. 59. Thompson JD, Higgins DG, Gibson TJ (1994) Clustal-W - Improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res 22: 4673–4680.
  60. 60. Saitou N, Nei M (1987) The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol Biol Evol 4: 406–425.
  61. 61. Felsenstein J (1985) Confidence-Limits on Phylogenies - an Approach Using the Bootstrap. Evolution 39: 783–791.
  62. 62. Tamura K, Nei M, Kumar S (2004) Prospects for inferring very large phylogenies by using the neighbor-joining method. Proceedings of the National Academy of Sciences of the United States of America 101: 11030–11035.
  63. 63. Tamura K, Peterson D, Peterson N, Stecher G, Nei M, et al. (2011) MEGA5: Molecular Evolutionary Genetics Analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods. Mol Biol Evol 28: 2731–2739.
  64. 64. Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, et al. (1997) Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res 25: 3389–3402.
  65. 65. Li L, Stoeckert CJ Jr, Roos DS (2003) OrthoMCL: identification of ortholog groups for eukaryotic genomes. Genome Res 13: 2178–2189.
  66. 66. Edgar RC (2004) MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res 32: 1792–1797.
  67. 67. Castresana J (2000) Selection of conserved blocks from multiple alignments for their use in phylogenetic analysis. Mol Biol Evol 17: 540–552.
  68. 68. Stamatakis A (2006) RAxML-VI-HPC: maximum likelihood-based phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics 22: 2688–2690.
  69. 69. Cintas NA, Koike ST, Bull CT (2000) Rappini bacterial blight declines with delayed replanting in the Salinas Valley of California; New Orleans, Louisiana. S15.