To compare the in vitro growth ability of the aldH-deficient mutant (ΔaldH) and the pat-deficient mutant (Δpat) from SPC9018 with that of SPC9018, P. cichorii strains were incubated in PY-medium and the optical density at 600 nm (OD600) of the bacterial suspensions was measured. The growth rate of both ΔaldH and Δpat was equal to that of SPC9018, showing that deletion of aldH and pat does not affect in vitro growth of the bacteria (Figure 1).

When inoculated with SPC9018, symptoms on leaves of eggplant and lettuce developed within one day post-inoculation and then progressed beyond the inoculated area within three days post-inoculation (Figure 2). The ΔaldH and the Δpat mutants lost their virulence on eggplants (Figure 2A). However, both mutants exhibited virulence on lettuce, similar to SPC9018 (Figure 2B). To confirm the involvement of aldH and pat in P. cichorii virulence against eggplant, the ΔaldH and Δpat mutants were complemented with plasmids pPc-aldH and pPc-pat carrying aldH and pat, respectively, from the SPC9018 genome. The transformants, ΔaldH(Pc-aldH) and Δpat(Pc-pat), exhibited virulence against eggplant similar to SPC9018 (Figure 2A). These results indicate involvement of aldH and pat in SPC9018 virulence on eggplant but not on lettuce.

The ΔaldH mutant grew vigorously in the inoculated area of eggplant leaves, and reached a maximum population size of 2.6 × 10⁷ cfu/cm² at 24 h post-inoculation, similar to SPC9018 and the complemented strain ΔaldH(Pc-aldH) (Figure 3). In contrast, the Δpat mutant grew slower compared to SPC9018 and reached a maximum population size of 1.9 × 10⁷ cfu/cm² at 36 h post-inoculation. The complemented strain Δpat(Pc-pat) grew similarly to SPC9018.

The deduced amino acid sequences based on the nucleotide sequences of aldH and pat showed that the amino acid sequences characteristic of type III effectors [18] were not observed in the N-terminal of the proteins encoded in aldH and pat. The PSORT prediction of deduced amino acid sequences of the proteins encoded in aldH and pat showed localization in the bacterial cytoplasm and the bacterial inner membrane, respectively. Though the PSORT prediction is limited to identify secreted proteins through the Sec pathway, it is thought that SPC9018 may not extracellularly secrete both proteins.

Results in this study indicate involvement of pat but not aldH in bacterial growth in eggplant leaves. Hojo et al. [6] indicate that vigorous growth of SPC9018 in eggplant leaves immediately after invasion may be involved in its virulence. This evidence leads us to consider that the protein encoded in pat might be implicated in the rapid growth of P. cichorii immediately after invasion into eggplant leaves, and may be associated with stressors or signals present in eggplant but not in lettuce. Wang et al. [12] reported that mutation of aldH in Agrobacterium tumefaciens resulted in early expression of the quorum sensing signal degrading enzyme, AttM. Furthermore, aldehyde dehydrogenase is reportedly a member of virulence regulons in V. cholerae [13]. It is thus thought that either the signal transduction through the protein encoded in aldH or the protein itself might be involved in functional regulation of virulence factors such as quorum sensing, which are involved in induction of PCD in eggplant but not in lettuce. Further analysis of aldH and pat will be required for elucidation of the virulence mechanism of SPC9018, as well as genetic and functional information regarding type III effectors.

HrpL, a member of the ECF family of alternative sigma factors [19], activates the hrp, hrc and type III effector genes of pseudomonads [6,10]. To analyze the dependency of aldH and pat expression on HrpL in P. cichorii, the expression of aldH and pat was examined by RT-PCR on samples taken at 8 h post-inoculation. In inoculated eggplant leaves, the expression of aldH and pat in the hrpL-deleted mutant, SPC9018-L [6] was similar to the parent strain SPC9018 (Figure 4). When RNA treated with DNase I was used as template in PCR reactions, no product was observed.

HrpL is thought to bind to a consensus bipartite cis element (hrp box) present in the promoter region of the hrp, hrc and type III effector genes [19,20,21]. The hrp box (GGAACC-N15-16-CCANNCA) exists in the promoter regions of hrpA, hrpF, hrpW, avrF, avrE and hrpJ in the genomic DNA of P. cichorii, and expression of these genes, is dependent on HrpL [6,10]. Absence of the hrp box in the promoter region of aldH and pat and results in this study indicate that SPC9018 expresses aldH and pat independently of HrpL.

aldH is also located in the genomic DNA of other pseudomonads such as P. viridiflava and P. syringae strains [10,11]. To compare the phylogenetic diversity between aldH, nucleotide sequences (Table S1) of aldH from P. cichorii, P. viridiflava and P. syringae were analyzed. In the phylogenetic tree based on the nucleotide sequence of aldH, all these species represent different clusters (Figure 5).

To analyze functional conservation of aldH among pseudomonads, the ΔaldH mutant was transformed with aldH genes originating from the P. viridiflava BS group strain Pv9504, P. syringae pv. tomato strain DC3000 and P. aeruginosa strain PAO1, creating ΔaldH(Pv-aldH), ΔaldH(Pst-aldH) and ΔaldH(Pa-aldH), respectively. All transformants exhibited virulence on eggplant, similar to ΔaldH(Pc-aldH) and SPC9018 (Figure 2A). The virulence of these transformants suggests that aldH is conserved amongst pseudomonads and that phylogenetic diversification of aldH may correspond to species diversification.

pat is located in the genome of P. cichorii strains and the S-PAI of P. viridiflava AS and BS group strains [10]. To analyze the functional conservation of pat between SPC9018 and Pv9504, the Δpat mutant was transformed with pat originating from P. viridiflava strain Pv9504. The transformant, Δpat(Pv-pat) exhibited virulence on eggplant similar to SPC9018 (Figure 2A). The virulence of this transformant suggests functional conservation of pat between P. cichorii and P. viridiflava BS and AS group strains harboring the S-PAI, supporting a common ancestry of pat between P. cichorii strains and the BS and AS group strains of P. viridiflava, as well as the hrp genes.

The composition of bacterial genomes can change rapidly and dramatically through a variety of processes, including horizontal gene transfer [24] which has been recognized as the universal event throughout bacterial evolution [14,25]. Acquiring horizontally transferred genes is an efficient way to alter the genotype of a bacterium, leading to the creation of a new phenotype or even a new species [26,27]. Although a single gene might have a low horizontal transfer index (HTI) purely by chance, it is unlikely that a large cluster of neighboring genes would all have low HTIs by chance. Such clusters are considered to be a single unit simultaneously inserted into the genome [28]. In particular, it has been suggested that a number of pathogenicity genes were horizontally transferred as large clusters, PAIs [29]. It is thus thought that a pathogenicity island in SPC9018 consists of the hrp and pat. The pathogenicity island with aldH is implicated in virulence of SPC9018 on eggplant but not on lettuce.

Virulence analysis showed that deletion of aldH and pat resulted in loss of SPC9018 virulence on eggplants but not on lettuce. When 31 species of Asteraceae plants (Table S2) including lettuce (Lactuca sativa) were inoculated with SPC9018, we could observe necrotic regions on leaves of all plant species at three days post-inoculation, indicating that SPC9018 is virulent on all Asteraceae plants. To analyze the involvement of hrp, aldH and pat in SPC9018 virulence on the Asteraceae species, we inoculated the Asteraceae plants with a hrcC-mutant, SPC9018-hrcC, and with the ΔaldH and Δpat mutants. The mutants, SPC9018-hrcC, ΔaldH and Δpat lost their virulence on 9, 16 and 6 species, respectively (Figure 6 and Table S2). Between them, all mutants lost their virulence on 5 species.

To determine the relationship between the involvement of hrp, aldH and pat in P. cichorii virulence on these plants and phylogeny among the Asteraceae species, phylogenetic trees were constructed based on the combined nucleotide sequences of ndhF and rbcL (Figure 6 and Table S2). The NJ phylogenetic tree placed 31 species in two major clusters (Figure 6). Involvement of hrp, aldH and pat in SPC9018 virulence on the respective species had no relationship with phylogeny of the Asteraceae species. It is thus thought that the implication of not only hrp but also aldH and pat in P. cichorii virulence might have arisen after species diversification of Asteraceae plants and be responsible for the virulence of P. cichorii towards respective species.

Araki et al. [10] and Hojo et al. [6] demonstrated that P. cichorii strains acquired the hrp genes through horizontal transfer from a common ancestor with the S-PAI of P. viridiflava and implicated the hrp in its virulence. Symptoms on P. cichorii-infected lettuce leaves are characterized by shiny, dark brown, firm necrotic spots [2,3]. P. cichorii also causes necrotic spots on eggplant distinct from the disease symptoms on lettuce leaves [6,7]. The development of disease symptoms on both lettuce leaves and eggplant leaves is closely associated with PCD in P. cichorii-infected tissues [5,7]. Virulence of P. cichorii on eggplant is hrp-dependent. On the other hand, the hrp are not directly implicated in induction of PCD in infected lettuce leaves and the disease development on lettuce leaves caused by the hrp-deficient mutants is delayed compared with that caused by the parent strain, since the hrp genes play a role in P. cichorii growth at the early infection stages in lettuce leaves [6]. From these evidences, Hojo et al. [6] demonstrate that the effect of the hrp on P. cichorii virulence differs between lettuce and eggplant. Results in this study suggested conservation of aldH amongst pseudomonads and supported a common ancestor of pat between P. cichorii and P. viridiflava strains harboring the S-PAI. Furthermore, results in this study implicate not only the hrp genes but also aldH and pat in the diversity of its virulence towards susceptible host plants. Therefore, it is hypothesized that P. cichorii might maintain the hrp, aldH and pat to establish its virulence on respective host plants.

The bacterial strains and plasmids used in this study are listed in Table S3. P. cichorii, P. viridiflava, P. syringae pv. tomato and P. aeruginosa strains were routinely grown in PY-medium (5% polypeptone, 2% yeast extract) at 30 °C. Escherichia coli strains were grown in LM medium [30] at 37 °C. The optimal density at 600 nm (OD600) of the bacterial suspensions was measured with Ultrospec 1100pro (GE Healthcare, Tokyo, Japan). Ampicillin (50 μg/mL, nakalai tesque, Kyoto, Japan), chloramphenicol (50 μg/mL, nakalai tesque), kanamycin (50 μg/mL, nakalai tesque) and tetracycline (30 μg/mL, nakalai tesque) were used in selective media. Populations of SPC9018 and mutants in planta were assayed in three independent experiments using PCSM plates [31] and PCSM plates containing appropriate antibiotics.

Isolation of genomic DNA, plasmid DNA manipulations, and PCR analysis were performed using standard techniques [6,32]. P. cichorii was transformed by electroporation as described by Hojo et al. [6]. Double-stranded DNA sequencing templates were prepared with GenElute™ Plasmid Miniprep Kits (SIGMA Chemical, St. Louis, MO, USA). Sequences were determined using an ABI Prism 3100-Avant Genetics Analyzer (Applied Biosystems, Tokyo, Japan). DNA sequence data were analyzed using the DNASIS-Mac software (Hitachi Software Engineering, Yokohama, Japan). Enzymes including restriction endonucleases (Takara, Ohtsu, Japan) and Ex-taq polymerase (Takara) were used according to the manufacturer’s instructions. The primers used in this study are listed in Table S4.

Plasmids designed to create deletion mutants using the kanamycin-resistant gene originating from pUCK191 [33] were based on pHSG398 (Takara) ligated sacB originating from pUCD800 [34]. The construction of the clones is described in detail in Supplementary Materials. The resulting plasmids, pD4-1-2KmSacB and pD5-1-2KmSacB, were electroporated into SPC9018 cells and the resultant kanamycin- and sucrose-resistant, and chloramphenicol-sensitive recombinants, an aldH-deficient mutant (ΔaldH) and a pat-deficient mutant (Δpat), were selected, respectively.

For creation of the hrcC-mutant, a plasmid, pHSG398sacBhrcCKm, was created. The construction of the clones is described in detail in Supplementary Materials. The plasmid pHSG398sacBhrcCKm was electroporated into SPC9018 cells and a resultant kanamycin and sucrose-resistant recombinant, SPC9018-hrcC, was selected.

For complementation of ΔaldH with aldH originating from pseudomonads, the plasmids, pPc-aldH, pPst-aldH and pPa-aldH, were created. A detailed cloning procedure is described in Supplementary Materials. The plasmids pPc-aldH, pPst-aldH and pPa-aldH containing aldH, originating from SPC9018, P. syringae pv. tomato strain DC3000 and P. aeruginosa strain PAO1, were transformed into ΔaldH competent cells and chloramphenicol-resistant transformants, ΔaldH(Pc-aldH), ΔaldH(Pst-aldH) and ΔaldH(Pa-aldH) were created, respectively. The plasmid pPv-aldH was also transformed into ΔaldH competent cells to create a tetracycline-resistant ΔaldH(Pv-aldH) transformant.

For complementation of Δpat with pat originating from SPC9018 and Pv9504, the plasmids, pPc-pat and pPv-pat, were created, respectively. A detailed cloning procedure is described in Supplementary Materials. The plasmids pPc-pat and pPv-pat including pat, originating from SPC9018 and Pv9504, were transformed into Δpat competent cells and a chloramphenicol-resistant Δpat(Pc-pat) transformant and a chloramphenicol-resistant Δpat(Pv-pat) transformant, were created, respectively.

To analyze the expression of aldH and pat by reverse transcription-PCR (RT-PCR), total RNA was isolated from five of each set of the bacteria-infiltrated area in eggplant leaves (0.6 g) 8 h after infiltration with with 50 μL of the bacterial suspension (1.0 × 10⁸ cfu/mL) of P. cichorii strains, and DNase I (Applied Biosystems, Tokyo, Japan) treatment was used to remove the genomic DNA from the RNA preparation [35]. The cDNA of aldH and pat was synthesized from total RNA (6 μg) using reverse transcription RAV-2 (Takara) utilizing primers SEMI-Back and ORF5-RT-Rv, respectively. The PCR was carried out with the primers: SEMI-Front and SEMI-Back to amplify a 500 bp amplicon specific to aldH; and ORF5-RT-Fw and ORF5-RT-Rv to amplify a 450 bp DNA fragment specific to pat. The cDNA of 16S rRNA was synthesized utilizing a primer, 16S-rRNA-Rv. The PCR was carried out with primers 16S-rRNA-Fw and 16S-rRNA-Rv to amplify a 448 bp amplicon specific to 16S rRNA.

A 1.7 kb DNA fragment containing aldH was amplified from the genomic DNA of P. cichorii strains (Table S2) and sequenced with aldH-Fw1 and aldH-Rv as primers.

The partial ndhF and rbcL fragments were amplified by PCR from the genomic DNA of 31 species of Asteraceae plants (Table S2) using the primers ndhF-11FW and ndhF-22RV for ndhF; and 1-1 and NN3-2 for rbcL. The nucleotide sequences of partial ndhF and rbcL DNA fragments were determined directly from the PCR fragments using ndhF-11FW, ndhF-22RV, ndhF-11RV, ndhF-22FW, 1-2FW, 1-2RV, 1-1 and NN3-2.

The nucleotide sequences of aldH (1,584 bp), and combined ndhF (1,996–2,023 bp) and rbcL (1244–1247 bp) were aligned and phylogenetic trees were constructed using ClustalW by the neighbor-joining (NJ) method [22,23]. Genetic distances were computed with Kimura’s two-parameter model [39]. The NJ phylogenetic tree was drawn by TreeView.

Eggplant (Solanum melongena L. cv. Senryo No. 2) and the plants listed in Table S2, including lettuce (Lactuca sativa L. cv. Success) were grown in pots containing a high-grade potting mix (Tsuchitaro; Sumitomo Forestry Co. Ltd., Tokyo, Japan) at 25 °C. Light (16 h/day) was supplied at 10,000 L × throughout the experimental period. Five-week-old test plants were inoculated by leaf-infiltration using a 1 mL disposable syringe with 1.0 × 10⁸ cfu/mL bacteria in a 20 μL volume. Following inoculation, plants were incubated for 3 days under 100% relative humidity at 25 °C and then grown at 25 °C. For all assays, inoculum concentrations were determined spectrophotometrically and confirmed by dilution plating. Plants were inspected for symptoms daily for five days after inoculation. We replicated virulence assays on tested species at five trials. Within each trial, 12 plants for each strain were treated, yielding 60 plants per strain.

Areas (1 cm²) inoculated with P. cichorii strains were excised from the eggplant leaves of five plants at 0, 12, 24 and 36 h post-inoculation and ground using a mortar and pestle. Samples (0.1 mL) of the original solution and 10-fold serial dilutions were spread onto three plates of selective agar media containing appropriate antibiotics. Colonies were counted after 2 days of incubation at 30 °C to estimate the population.

The nucleotide sequences determined in this study have been deposited in the DDBJ/GenBank international nucleotide sequence database under the accession numbers shown in Table S1 and Table S5.