The hrp genes of Pseudomonas cichorii are essential for pathogenicity on eggplant but not on lettuce.

Pseudomonas cichorii causes necrotic lesions in eggplant and rot in lettuce. Through transposon insertion into P. cichorii strain SPC9018 we produced two mutants, 4-57 and 2-99, that lost virulence on eggplant but not lettuce. Analyses showed that a transposon was inserted into the hrpG gene in 4-57 and the hrcT gene in 2-99. Nucleotide sequences of the hrp genes of SPC9018 are homologous to those of Pseudomonas viridiflava BS group strains. The pathogenicity of 4-57 on eggplant was restored by transformation with an hrpF operon, originating from either SPC9018 or the BS group member P. viridiflava strain 9504 (Pv9504). These data suggested the involvement of hrp genes in the pathogenicity of SPC9018 on eggplant, and functional conservation of hrpF operons between SPC9018 and Pv9504. Both the hrpS mutant and the hrpL mutant were unable to cause necrotic lesions on eggplant leaves but retained their pathogenicity against lettuce. These results suggest that the pathogenicity of P. cichorii is hrp-dependent in eggplant, but not in lettuce.

INTRODUCTION

Pseudomonas cichorii causes necrotic leaf spot on eggplants and rot on lettuce. Bacterial rot caused by the bacteria is called ‘varnish spot’ in California or ‘tar’ in Japan. The rot is characterized by shiny, dark-brown, firm necrotic spots on leaves underneath the second or the third outermost head leaves. Microscopic observations using lux-marked P. cichorii and an immunofluorescent antibody against the bacteria have shown that the bacteria first invade lettuce head leaves through stomata and then colonize intercellular spaces (Hikichi , 1998). After proliferation in intercellular spaces of the mesophyll, the bacteria spread throughout the whole leaf via the vascular bundle (Hikichi ). The symptoms appear and develop as the bacteria multiply and spread. Our previous studies have shown that de novo protein synthesis in lettuce leaves is required for the development of disease symptoms (Hikichi ; Kiba ). Moreover, the development of disease symptoms is closely associated with programmed cell death (PCD), following heterochromatin aggregation and laddering of genome DNA in the P. cichorii-infected lettuce cells (Kiba ). P. cichorii causes necrotic spot on eggplant, sweet pepper, celery and okra, distinct from the disease symptoms on lettuce. Kiba have shown that development of necrotic lesions following PCD on leaves of eggplants inoculated with P. cichorii is commonly associated with de novo protein synthesis, intracellular reactive oxygen species and caspase III-like protease activity.

In several Gram-negative phytopathogenic bacteria, the hrp genes are essential determinants for disease development on compatible hosts and for elicitation of the hypersensitive response (HR) on resistant plants (Alfano & Collmer, 1997). The hrp genes encode proteins in the type III secretion system (TTSS), which is believed to transport virulence proteins directly into the host cells. These proteins subsequently cause leakage of plant nutrients into the extracellular spaces of infected tissues and suppress host defences. Nine of the hrp genes have been renamed hrc (HR and conserved) to indicate that they encode conserved components that are also present in the type III secretion machinery of the animal pathogens Yersinia, Shigella and Salmonella (Bogdanove ). Recently Araki have reported that hrp genes exist in the genomic DNA of P. cichorii strain 83-1, but the roles of these genes in the pathogenicity of P. cichorii remain unclear.

In this study, we isolated two mutants that lost virulence on eggplants after transposon mutagenesis of P. cichorii strain SPC9018 (SPC9018). However, the mutants retained their virulence on lettuce. Molecular analysis revealed that the transposons were inserted in the hrpG and hrcT genes. Pathogenicity analysis using hrp mutants showed that the hrp genes of P. cichorii are essential for its ability to cause necrotic lesion symptoms on eggplant, but not to cause rot symptoms on lettuce.

METHODS

Bacterial strains, culture conditions and plasmids.

Bacterial strains and plasmids are listed in Table 1. P. cichorii strains were routinely grown in PS medium (Wakimoto ) at 30 °C. Escherichia coli strains were grown in LM medium (Hanahan, 1983) at 37 °C. Ampicillin (50 μg ml−1), kanamycin (50 μg ml−1) and tetracycline (30 μg ml−1) were used in selective media. Populations of SPC9018 and mutants in planta were assayed in three independent experiments using PCSM plates (Uematsu ) and PCSM plates containing kanamycin.

Table 1.

Strains and plasmids used in this study

Strain or plasmid	Relevant characteristics	Reference or source
E. coli strains
DH5α	recA1 endA1 gyrA96 thi-1 hsdR17supE44 Δ(lac)U169(φ80lacΔM15)	Takara
P. cichorii strains
SPC9018	Wild-type	Kiba et al. (2006a)
2-99	SPC9018 derivative, carries a EZ : : Tn <KAN-2>, Kmr	This study
4-57	SPC9018 derivative, carries a miniTn5lacZ2, Kmr	This study
SPC9018-L	hrpL-deleted mutant of SPC9018, Kmr	This study
SPC9018-S	hrpS-deleted mutant of SPC9018, Kmr	This study
4-57F	Transformant of 4-57 with phrpFoperon, Kmr, Tcr	This study
4-57FPV	Transformant of 4-57 with phrpFoperonPV, Kmr, Tcr	This study
P. viridiflava strain
9504	BS type	This study
Plasmids
pBluescript II	KS+, Apr	Stratagene
pUC118	Ampr	Takara
pHSG398	Cmr	Takara
pLAFR3	pLAFR1 containing HaeII fragment of pUC8	Staskawicz et al. (1987)
placZ2	P18Sfi derivative containing mini-Tn5lacZ2, Kmr	de Lorenzo et al. (1990)
pUCK191	pUC18 derivative containing Kmr from Tn903	Tsuge et al. (2001)
pUCD800	pUCD5 derivative containing sacB, Kmr	Gay et al. (1985)
p2-99	13.9 kb KpnI fragment containing the EZ : : Tn <KAN-2> in pUC118	This study
p4-57	5.6 kb KpnI fragment containing the miniTn5lacZ2 in pUC118 pGEM-T	This study
pP2-99	703 bp SphI- and XhoI-digested fragment of p2-99 in pUC118	This study
pP4-57	1.1 kb HindIII- and SmaI-digested fragment of p4-57 in pUC118	This study
pHOJO	30 kb fragment from SPC9018 genomic DNA in pBluescript II	This study
pKAJI	24.3 kb fragment from SPC9018 genomic DNA in pBluescript II	This study
pL3-4	870 bp PstI- and SacI-digested PCR fragment (3-4 fragment) in pHSG398	This study
phrpL	464 bp EcoRI- and SacI-digested PCR fragment (1-2 fragment) in pL3-4	This study
p398-LKm	1.4 kb blunt-ended fragment of pUCK191 containing KmR in phrpL	This study
p398-LKmsac	2.6 kb blunt-ended KpnI fragment containing sacB from pUCD800 ligated into blunt-ended EcoRI site of p398-LKm	This study
p398 S	4.7 kb SmaI- and EcoRI-digested fragment of pHOJO in pHSG398	This study
p398-SKm	1.4 kb blunt-ended KpnI fragment containing KmR from pUCK191 ligated into EcoRV site of p398-S	This study
p398-SKmsac	2.6 kb blunt-ended fragment contaning sacB from pUCD800 ligated into blunt-ended EcoRI site of p398-SKm	This study
phrpFoperon	3.6 kb PCR fragment containing the hrpF operon from SPC9018 genomic DNA	This study
phrpFoperonPV	4.1 kb PCR fragment containing the hrpF operon from Pv9504 genomic DNA	This study

General DNA manipulations.

Isolation of genomic DNA, plasmid DNA manipulations, PCR and Southern blot analyses were performed using standard techniques (Sambrook ). P. cichorii was transformed by electroporation as described by Allen . Double-stranded DNA sequencing templates were prepared with GenElute Plasmid Miniprep kits (Sigma Chemical). Sequences were determined using an Automated DNA Sequencer model 373 (Applied Biosystems). DNA sequence data were analysed using the dnasis-Mac software (Hitachi Software Engineering). Enzymes including restriction endonucleases (Takara) were used according to the manufacturer's instructions.

Transposon mutagenesis.

To create P. cichorii mutants with transposon insertions, 0.5 μg placZ2 (de Lorenzo ) containing mini-Tn5lacZ2 was used for electroporation of SPC9018 competent cells. The EZ : TN transposome-mediated insertion system (Epicentre) was also used (Tsujimoto ). The transposon EZ : : TN (0.1 μg) was mixed with an equal volume of 100 % glycerol and 2 μl EZ : TN transposase (1 unit μl−1), and incubated at 37 °C for 10 min. Aliquots (1 μl) were used for electroporation of P. cichorii. Transposition clones were selected by plating on PS medium containing kanamycin.

Transposon-inserted site in genomic DNA of mutants.

To determine whether a single transposon insertion had occurred in the genome of mutants 2-99 and 4-57 through the insertion by EZ : : Tn and mini-Tn5lacZ2, respectively, their genomic DNAs were isolated using the AquaPure Genomic DNA Isolation kit (Bio-Rad). DNA from 2-99 was digested with EcoRI, 4-57 DNA was digested with XhoI, and the DNA fragments were then separated by agarose gel electrophoresis and hybridized with KmR from EZ : : Tn (2-99) and KmR from pUCK191 (4-57) (Tsuge ).

The KpnI fragments from genomic DNA of the mutants were ligated into the plasmid vector pUC118 (Takara), and the resulting plasmids were transformed into E. coli DH5α (Takara). Kanamycin-resistant transformants were isolated. Each harboured plasmid p2-99 or p4-57, which contained 13.9 or 5.6 kb DNA fragments from genomic DNA of 2-99 and 4-57, respectively. The inserts carried on p2-99 and p4-57 were also sequenced.

Determination of the nucleotide sequence of the hrp cluster.

To create a genomic library of SPC9018, genomic DNA from SPC9018 was isolated and partially digested by Sau3AI. DNA fragments (10–20 kb) were collected by sucrose density-gradient centrifugation (Kanda ). The DNA fragments were ligated into the BamHI site of pBluescript II KS+ (Stratagene), and transformed into E. coli DH5α, to create a genomic library of SPC9018.

A 703 bp SphI- and XhoI-digested fragment of p2-99, and a 1.1 kb HindIII- and SmaI-digested fragment of p4-57 were ligated into pUC118 to create plasmids pP2-99 and pP4-57, respectively. The plasmids were used as templates in PCRs. Southern blot hybridization was performed using DIG-labelled DNA probes (Roche Molecular Biochemicals) to probe the hrp cluster from the SPC9018 genomic library. The DIG-labelled probes were PCR-amplified with M4 (5′-GTTTTCCCAGTCACGAC-3′) and RV (5′-CAGGAAACAGCTATGAC-3′) according to the manufacturer's protocol. The resultant positive transformant hrpG-posi harboured the plasmid pHOJO containing a 30 kb insert. The other positive transformant, hrcT-posi, harboured pKAJI containing a 24.3 kb insert. These were used to sequence the hrp cluster.

Complementation of 4-57 with hrpF operons.

A 3.6 kb fragment was PCR-amplified from the genomic DNA of SPC9018 with the following primers: 5′-GCTCTAGAGGGTCAACTGGGCTGGACGTTG-3′ (named Xba-FW-hrpFoperon) with an added XbaI site (underlined) and 5′-GCTCTAGATGCGCGTCGCGTTGAGAGTTCG-3′ (named Xba-RV-hrpFoperon) with an added XbaI site (underlined). PCR amplification was performed with 1 cycle of 94 °C for 2 min, 5 cycles of 94 °C for 1 min, 62 °C for 1 min, and 72 °C for 3 min, 20 cycles of 94 °C for 1 min, 65 °C for 1 min, and 72 °C for 3 min. The XbaI-digested fragment was ligated into the XbaI site of pLAFR3 (Staskawicz ), and phrpFoperon was created. A 4.1 kb fragment was PCR-amplified from the genomic DNA of Pseudomonas viridiflava strain 9504 (Pv9504), which belongs to the BS group, with the following primers: 5′-GCTCTAGACTCATGTTGACCCGTCGCAGTC-3′ (named Xba-PV-FW-hrpFoperon) with an added XbaI site (underlined) and 5′-GCTCTAGAGCATGTCGCGTTGGGAGTTCGC-3′ (named Xba-PV-RV-hrpFoperon) with an added XbaI site (underlined), based on the nucleotide sequences of P. viridiflava strains ME3.1b, RMX23.1a and RMX3.1b (Araki ). PCR amplification was performed with 1 cycle of 94 °C for 2 min, 5 cycles of 94 °C for 1 min, 60 °C for 1 min, and 72 °C for 3 min, 20 cycles of 94 °C for 1 min, 65 °C for 1 min, and 72 °C for 3 min. The XbaI-digested fragment was ligated into the XbaI site of pLAFR3, and phrpFoperonPV was created. phrpFoperon and phrpFoperonPV were transformed into 4-57 competent cells, and the tetracycline-resistant transformants 4-57F and 4-57FPV were created, respectively.

Creation of hrpL and hrpS mutants.

An 870 bp fragment (named 3-4) was PCR-amplified from the genomic DNA of SPC9018 with the following primers: 5′-CAGCCCTGCAGAACGCTAAC-3′ (named 4-Pst) and 5′-CGAGCTCTGAACAGTTTTGTCCC-3′ (named 3-Sac) with an added SacI site (underlined). PCR amplification was performed with 1 cycle of 94 °C for 2 min, 5 cycles of 94 °C for 1 min, 50 °C for 1 min, and 72 °C for 1 min, 20 cycles of 94 °C for 1 min, 55 °C for 1 min, and 72 °C for 1 min. The SacI- and PstI-digested 3-4 fragment was ligated into the SacI and PstI sites of pHSG398 (Takara), and pL3-4 was created. A 464 bp fragment (named 1-2) was PCR-amplified from the genomic DNA of SPC9018 with the following primers: 5′-GGAATTCGGGGCGTTCCACGCTTTC-3′ (named 1-RI) with an added EcoRI site (underlined) and 5′-GGGAGCTCGGTCACTGCATGCCTTTGACTTC-3′ (named 2-Sac) with an added SacI site (underlined). PCR amplification was performed with 1 cycle of 94 °C for 2 min, 5 cycles of 94 °C for 1 min, 50 °C for 1 min, and 72 °C for 30 s, 20 cycles of 94 °C for 1 min, 55 °C for 1 min, and 72 °C for 30 s. The EcoRI- and SacI-digested 1-2 fragment was ligated into the EcoRI and SacI sites of pL3-4, and phrpL was created. A 1.4 kb KpnI-digested fragment of pUCK191 containing KmR was blunt-ended by T4 DNA polymerase and ligated into the blunt-ended phrpL SacI site to create p398-LKm. A 2.6 kb PstI- and BamHI-digested DNA fragment containing sacB from pUCD800 (Gay ) was blunt-ended by T4 DNA polymerase and ligated into the blunt-ended p398-LKm EcoRI site to create p398-LKmsac. This plasmid was electroporated into SPC9018 cells and the resulting kanamycin- and sucrose-resistant recombinant, SPC9018-L, was selected. Southern blot analysis was performed to verify correct insertion of the 2.7 kb fragment containing KmR into the hrpL locus in the genetic backgrounds isolated (data not shown), and this showed that SPC9018-L was an hrpL-deficient mutant of SPC9018.

A 4.7 kb EcoRI- and SmaI-digested fragment of pHOJO was ligated into the EcoRI and SmaI sites of pHSG398, and p398-S was created. A 1.4 kb KpnI-digested fragment of pUCK191 containing KmR was blunt-ended by T4 DNA polymerase and inserted into the the p398-S EcoRV site to create p398-SKm. A 2.6 kb BamHI- and PstI-digested DNA fragment containing sacB from pUCD800 was blunt-ended by T4 DNA polymerase and ligated into the blunt-ended p398-SKm EcoI site to create p398-SKmsac. This plasmid was electroporated into SPC9018 cells and the resultant kanamycin- and sucrose-resistant recombinant, SPC9018-S, was selected. Southern blot analysis was performed to verify correct insertion of the 6.1 kb fragment containing KmR in the hrpS locus in the genetic backgrounds isolated (data not shown), showing that SPC9018-S was an hrpS-deficient mutant of SPC9018.

Virulence assays.

Eggplant (Solanum melongena L. cv. Senryo no. 2), lettuce (Lactuca sativa L. cv. Success), celery (Apium graveolens L. cv. Topseller), sweet pepper (Capsicum annuum cv. Shosuke) and okra plants (Abelmoschus esculentus cv. Gulliver) were grown in pots containing commercial soil (Tsuchitaro, Sumitomo Forestry) in a growth room at 25 °C. Light (16 h day−1) was supplied at 10 000 lux throughout the experimental period. Five-week-old test plants were inoculated by leaf infiltration using a 1 ml disposable syringe with 1.0×108 c.f.u. ml−1 bacteria in a 20 μl volume. For all assays, inoculum concentrations were determined spectrophotometrically and confirmed by dilution plating. Lettuce plants were coded and inspected for symptoms daily for 7 days after inoculation. Plants were rated on a zero-to-three disease index scale: 0, no symptoms; 1, discolouring; 2, browning; 3, collapse. Other plants were coded and inspected for symptoms daily for 7 days after inoculation. Plants were rated on a zero-to-three disease index scale: 0, no symptoms; 1, discolouring at inoculated sites; 2, necrosis at inoculated sites; 3, necrosis at the periphery of the inoculate sites. Within each trial, 12 plants of each strain were treated, yielding 60 plants per strain.

Bacterial population in planta.

Areas (1 cm2) inoculated with P. cichorii strains were excised from eggplant and lettuce leaves of five plants at 0, 1, 2 and 3 days after inoculation and ground using a mortar and pestle. Samples (0.1 ml) of the original solution and 10-fold serial dilutions thereof were spread onto three plates of selective agar media of PCSM (Uematsu ) for SPC9018, media containing 50 μg kanamycin ml−1 for SPC9018-L and 4-57, and media containing 50 μg kanamycin ml−1 and 30 μg tetracycline ml−1 for 4-57F. Colonies were counted after 2 days of incubation at 30 °C to estimate the population.

RESULTS

Selection of mutants losing virulence on eggplant

SPC9018 caused necrotic lesions (Fig. 1a) on eggplant and rot on lettuce (Fig. 1b) within 3 days after inoculation (Fig. 2). In total, 849 and 731 transposon mutants were derived from SPC9018 by random EZ : : Tn and mini-Tn5-insertion, respectively. We selected two mutants, 2-99 and 4-57, that were unable to cause necrotic lesions on eggplant within 7 days of inoculation (Fig. 1a). However, the mutants retained their ability to cause rot on lettuce leaves, although rot symptoms caused by the mutants were delayed compared to those caused by SPC9018 (Fig. 2). Other mutants retained similar virulence to SPC9018 against both eggplant and lettuce.

Fig. 1.

Necrotic lesions on eggplant leaves (a) and rot on lettuce leaves (b) after inoculation with P. cichorii strains. Five-week-old plants were inoculated by leaf infiltration with P. cichorii strains. a, SPC9018; b, 2-99; c, 4-57; d, SPC9018-S; e, SPC9018-L; f, 4-57F; g, 4-57FPV. Inoculates consisted of 1.0×108 c.f.u. ml−1 bacteria and distilled water (h) in a 20 μl volume. Plants were grown in a growth room at 25 °C (10 000 lux, 16 h per day) for 3 days.

Fig. 2.

Virulence of P. cichorii strains on lettuce. Values represent the mean±sd of five separate experiments. Asterisks denote values significantly different from the disease index of inoculation with P. cichorii strain SPC9018 (*P<0.05).

Cloning of the transposon-containing fragment from mutants

In Southern analysis a single transposon insertion was observed in the DNA of 2-99 and 4-57 (data not shown). No hybridization signal was observed in the DNA of parental strain SPC9018. The KpnI-digested 13.9 and 5.6 kb fragments containing the inserted region of transposons from the genomes of 2-99 and 4-57 were cloned into pUC118, and p2-99 and p4-57 were created, respectively. The nucleotide sequences of inserted fragments showed that the transposons were inserted into the hrcT gene in p2-99 and the hrpG gene in p4-57 (Fig. 3).

Fig. 3.

Assembly of ORFs of the DNA region that includes the hrp genes of P. cichorii strain SPC9018 genomic DNA.

Sequence analysis of the hrp cluster

The Sau3AI-digested 30.0 and 24.3 kb fragments containing hrpG and hrcC from the genome of SPC9018 were cloned into pBluescript II KS+, and pHOJO and pKAJI were created, respectively. Inserts of pHOJO and pKAJI were sequenced and ORFs were analysed based on the nucleotide sequence of hrp genes in 83-1. Analyses of nucleotide sequences of the 30.0 and 24.3 kb inserts showed that the DNA region including the hrp cluster of SPC9018 consisted of 49 ORFs. These ORFs were divided into four regions: A, B, C and D (Table 2, Fig. 3).

Table 2.

Homologues of open reading frames in A, B, C and D regions of pathogenicity islands of P. cichorii strain SPC9018

ORF		Homologue
Name	Start–stop	Bacterium	Gene	Locus tag	Product
A1	42–920	P. viridiflava strain RMX3.1b		PAI-R2-ORF5	LysR family transcriptional regulator
A2	1078–2661		aldH	PAI-R2-ORF6	Aldehyde dehydrogenase
A3	Complement (2868–3449)			PAI-R2-ORF7	Phosphinothricin N-acetyltransferase
A4	Complement (3449–3988)		hrpL	PAI-R2-ORF8	HrpL
A5	4268–5386		hrpJ	PAI-R2-ORF9	HrpJ
A6	5383–7488		hrcV	PAI-R2-ORF10	HrcV
A7	7499–8482		hrpQ	PAI-R2-ORF11	HrpQ
A8	8563–9837		hrcN	PAI-R2-ORF12	HrcN
A9	9824–10 288		hrpO	PAI-R2-ORF13	HrpO
A10	10 285–10 845		hrpP	PAI-R2-ORF14	HrpP
A11	10 847–11 911		hrcQ	PAI-R2-ORF15	HrcQ
A12	11 980–12 561		hrcR	PAI-R2-ORF16	HrcR
A13	12 576–12 836		hrcS	PAI-R2-ORF17	HrcS
A14	12 838–13 635		hrcT	PAI-R2-ORF18	HrcT
A15	13 660–14 745		hrcU	PAI-R2-ORF19	HrcU
B1	29 511–35 000		avrE	PAI-R2-ORF20	AvrE
B2	35 111–35 719			PAI-R2-ORF21	Aspartyl beta-hydroxylase
B3	Complement (35 966–36 295)			PAI-R2-ORF23	HrpW-specific chaperone
B4	Complement (36 297–37 982)		hrpW	PAI-R2-ORF24	HrpW
B5	Complement (38 129–38 527)		avrF	PAI-R2-ORF25	AvrF
B6	Complement (38 629–38 979)		hrpV	PAI-R2-ORF26	HrpV
B7	Complement (38 972–39 163)		hrpT	PAI-R2-ORF27	HrpT
B8	Complement (39 291–41 309)		hrcC	PAI-R2-ORF28	HrcC
B9	Complement (41 302–41 715)		hrpG	PAI-R2-ORF29	HrpG
B10	Complement (41 702–41 950)		hrpF	PAI-R2-ORF30	HrpF
B11	Complement (41 994–42 572)		hrpE	PAI-R2-ORF31	HrpE
B12	Complement (42 538–43 161)		hrpD	PAI-R2-ORF32	HrpD
B13	Complement (43 158–43 949)		hrcJ	PAI-R2-ORF33	HrcJ
B14	Complement (43 960–44 328)		hrpB	PAI-R2-ORF34	HrpB
B15	Complement (44 364–45 386)		hrpZ	PAI-R2-ORF35	HrpZ
B16	Complement (45 411–45 617)		hrpA	PAI-R2-ORF36	HrpA
B17	Complement (45 735–46 673)			PAI-R2-ORF37	HrpS/HrpR-like protein
C1	15 379–16 395	C. violaceum strain ATCC 12472		CV_1407	Dehydrogenase
C2	16 429–17 544		degt	CV_1406	Aspartate aminotransferase
C3	17 545–18 057			CV_1405	Unknown
C4	18 054–19 142			CV_1404	Aminotransferase
C5	19 123–20 493			CV_1403	Unknown
C6	20 490–21 188			CV_1402	Phosphoglycolate phosphatase
C7	21 191–22 303			CV_1401	Dihydrorhizobitoxine desaturase
C8	22 393–23 208			CV_1400	5′-Methylthioadenosine phosphorylase
C9	23 205–23 801			CV_1399	Unknown
C10	23 810–24 724			CV_1398	Unknown
C11	24 721–25 908			CV_1397	Beta-phosphoglucomutase
C12	25 915–26 571			CV_1396	ABC transporter
C13	Complement (26 678–28 246)	P. syringae pv. tomato strain DC3000/X. campestris pv. campestris strain ATCC 33913		PSPTO_3183/Xcc0038	Pirin/FND-dependent NADH azoreductor
C14	Complement (28 375–29 121)	P. putida strain F1		Pput_1855	Unknown
D1	47 215–48 021	P. aeruginosa strain PAO1		PA2204	Binding protein component of ABC
D2	48 090–48 830			PA2203	Amino acid permease
D3	48 833–49 519			PA2202	Amino acid permease

Fifteen ORFs were present in region A (14.7 kb). These encoded a LysR family transcriptional regulator, aldehyde dehydrogenase, phosphinothrin N-acetyltransferase, HrpL, HrpJ, HrcV, HrpQ, HrcN, HrpO, HrpP, HrcQ, HrcR, HrcS, HrcT and HrcU. Each ORF (Table 2, Fig. 3) was preceded by a ribosome-binding site. Region B (17.2 kb) included 17 ORFs encoding HrpS, HrpA, HrpZ, HrpB, HrcJ, HrpD, HrpE, HrpF, HrpG, HrcC, HrpT, HrpV, AvrF, HrpW, HrpW-specific chaperone, asparaginyl beta-hydroxylase and AvrE. Nucleotide sequences and alignments of these ORFs located in regions A and B showed homology to those of 83-1 and P. viridiflava BS-type strain RMX3.1b. Comparison to 83-1 sequences (Araki ) showed that the consensus sequence of the hrp box may be GGAACC-N15-16-CCANNCA, which was identified at 84 bp upstream of hrpA, 31 bp upstream of hrpF, 135 bp upstream of hrpW, 79 bp upstream of avrF and 61 bp upstream of hrpJ. Putative Rho-independent terminators were located between hrpS and hrpA and between hrpL and the phosphinothricin N-acetyltransferase gene, but no others were found.

Region C (14.8 kb) was located between region A and region B, and contained 14 ORFs (Table 2, Fig. 3). Nucleotide sequences of 12 ORFs, C1–C12, were homologous to those of CV_1407–CV_1396 of Chromobacterium violaceum strain ATCC 12472 (Brazilian National Genome Project Consortium, 2003). Furthermore, nucleotide sequences of C14 were homologous to those of Pput1855 of Pseudomonas putida strain F1. In comparison with region C of P. cichorii strain 83-1, one ORF, C13, existed in SPC9018 alone. The nucleotide sequence of C13 at position 94–825 showed homology to the PSPTO_3183-encoded pirin from Pseudomonas syringae pv. tomato strain DC3000 at position 31–762 (Buell ). The nucleotide sequence of C13 at 1196–1569 showed homology to the Xcc0038-encoded FND-dependent NADH azoreductor from Xanthomonas campestris pv. campestris strain ATCC 33913 at position 347–720 (da Silva ).

Three ORFs were present in region D (3.2 kb), encoding two amino acid permeases and the binding protein component of an ABC transporter, which showed homology to those of Pseudomonas aeruginosa strain PAO1 (Stover ) (Table 2, Fig. 3).

Complementation of 4-57 with the hrpF operon

To confirm the involvement of the TTSS in P. cichorii virulence against eggplant, the hrpT mutant with the transposon insertion 4-57 was transformed with a plasmid carrying the hrpF operons from the SPC9018 genome. The transformant 4-57F (Figs 1a and 2) showed virulence against both eggplant and lettuce plants, similar to SPC9018. Furthermore, 4-57 was transformed with phrpFoperonPV containing the hrpF operon from the Pv9504 genome, and the transformant 4-57FPV was created (Fig. 2). 4-57FPV showed virulence against both eggplant and lettuce plants. These results suggest that the hrpF operon is involved in the virulence of SPC9018 against eggplant, and that the hrpF operon from Pv9504 is functional in SPC9018.

Virulence of hrpL and hrpS mutants

To analyse the involvement of hrpS and hrpL in the pathogenicity of SPC9018, hrpS- and hrpL-deficient mutants were created. Both the hrpS- and the hrpL-deficient mutants lost their virulence on eggplant, in a manner similar to 2-99 and 4-57 (Fig. 1a). Both the hrpS mutant and the hrpL mutant retained their virulence on lettuce leaves, although rot symptoms caused by the mutants were delayed compared to those caused by SPC9018 (Figs 1b and2). All hrp mutants lost their virulence on celery, okra and sweet pepper.

Population of P. cichorii strains in eggplant and lettuce leaves

Populations of 4-57 and the hrpL-deficient mutant SPC9018-L showed little change after inoculation into eggplant leaves, remaining at 1.4×104–3.2×104 c.f.u. cm−2 at 3 days after inoculation. On the other hand, 1 day after inoculation the parent strain reached its maximum population size of 2.3×106 c.f.u. cm−2, and 4-57F reached its maximum of 3.4×106 c.f.u. cm−2 (Fig. 4a).

Fig. 4.

Population dynamics of P. cichorii strains in infiltrated eggplant leaves (a) and lettuce leaves (b). Values represent the mean±sd of five separate experiments. Asterisks denote values significantly different from the population of P. cichorii strain SPC9018 (*P<0.05).

SPC9018 and 4-57F grew vigorously in lettuce leaves, and reached 4.3×106 and 5.2×106 c.f.u. cm−2 respectively, by 1 day after inoculation. 4-57 and SPC9018-L grew more slowly in lettuce leaves than the parent strain, and reached population sizes of 2.4×106 and 1.7×106 c.f.u. cm−2, respectively, by 2 days after inoculation (Fig. 4b).

The parental and the mutant strains grew similarly in both PS medium and PCSM medium (data not shown).

DISCUSSION

Lettuce leaf tissues inoculated with P. cichorii show browning and collapse symptoms. P. cichorii also causes necrotic leaf spots and induces cell death in eggplant. Our previous studies showed that there may be differences not only in the induction kinetics and level of plant responses, but also in the infection-related responses between bacterial rot in lettuce leaves and necrotic leaf spots in eggplant leaves (Kiba , b). The results of the present study showed that the hrp-deficient mutants lost their ability to grow vigorously in eggplant leaves and also their virulence towards eggplant, indicating that the virulence of P. cichorii towards eggplant is dependent upon the hrp genes. On the other hand, the hrp-deficient mutants retained their virulence towards lettuce. Therefore, the virulence functions of P. cichorii in lettuce differ from those in eggplant.

The nucleotide sequences of the P. cichorii hrp genes and their genetic structure were homologous to those of P. viridiflava BS group strains, suggesting a common ancestor of hrp clusters between P. viridiflava BS group strains and P. cichorii strains. P. viridiflava harbours two structurally distinct and highly diverged pathogenicity island (PAI) paralogues, T-PAI and S-PAI. These paralogues are integrated at different chromosome locations in the genome of P. viridiflava AT group strains, and AS group and BS group strains, respectively (Araki ). Phylogenetic analysis showed that the time of the most recent common ancestor of T-PAI and S-PAI predates the split of P. viridiflava from other Pseudomonas species. This indicates that one of the PAIs cannot have originated as a recent duplication event of the other, and a recent horizontal gene transfer cannot explain the presence/absence polymorphism of S-PAI or T-PAI. Phylogenetic analysis using nucleotide sequences of gyrB and rpoD demonstrates that within the ‘P. syringae complex’, P. cichorii strains form an independent monophyletic cluster with two strains of P. syringae pv. syringae (Yamamoto ). Our preliminary phylogenetic analysis using nucleotide sequences of hrpL and hrpS/hrpA also showed that 12 strains of P. cichorii, including SPC9018 and 83-1, form a monophyletic cluster independent of P. viridiflava and P. syringae (data not shown). It was thus thought that the time of the most recent common ancestor between S-PAI of P. viridiflava and the hrp clusters of P. cichorii may predate the split of P. cichorii from other Pseudomonas species. The ORFs C1–C12 and C14 in region C may have been acquired subsequently in common with C. violaceum and P. putida, respectively. The existence of ORF C13 in SPC9018 but not 83-1 suggests diversity among P. cichorii strains.

There are distinct differences in virulence and host-specificity between AT group strains, and AS and BS group strains of P. viridiflava. Araki have suggested that these differences may be maintained by selection as alternative means of interacting with different hosts. The host range of P. cichorii differs from that of P. viridiflava, although both pathogens infect lettuce and okra. The hrpF operon from P. viridiflava BS group strain 9504 was able to complement the virulence of hrpG mutants on eggplants, suggesting functional conservation of hrpF operons between SPC9018 and Pv9504. Comparing SPC9018 and P. viridiflava strain RMX3.1b, the amino acid sequence identities of the TTSS-dependent effectors AvrE and HrpW were 33.6 and 61.8 %, respectively (data not shown). Furthermore, the hrp mutants of SPC9018 lost their ability to cause necrotic spots on not only eggplant but also celery, sweet pepper and okra. Therefore, after acquisition of the the DNA region that includes the hrp genes, variation of TTSS effector genes under valance selection may lead to the acquisition by P. cichorii of virulence against eggplant, celery, sweet pepper and okra. This would result in differences of virulence between P. cichorii and the P. viridiflava BS group. However, the roles of these candidates in bacterial virulence remain unclear.

The need for an hrp gene for virulence has been documented in both non-macerating plant pathogens and in macerating Erwinia spp. and Pectobacterium spp. In the soft-rot pathogens, hrpN mutants are weakly pathogenic on chicory (Bauer , 1995), whereas hrpN mutants are fully pathogenic on host plants (Cui ). Furthermore, plants infected with an hrcC mutant of Pectobacterium (Erwinia) carotovora subsp. carotovora show delayed symptom development (Rantakari ). Lehtimäki have reported that high basal-level expression of hrp-regulated genes in Pectobacterium (Erwinia) carotovora subsp. carotovora has a negative impact on disease progress in the susceptible host plant Arabidopsis thaliana. The macerating enzymic bombardment launched by the necrotrophic Erwinia spp. and Pectobacterium spp. is a robust method of invasion compared with biotrophic bacterial species, which depend fully on the hrp gene functions for virulence. Although biotrophic pathogens also carry additional virulence determinants such as genes for toxins and extracellular polysaccharides, they are nonvirulent without the TTSS, at least under laboratory conditions. Therefore, the effect of the TTSS differs between necrotrophic and biotrophic pathogens. P. cichorii induces PCD in eggplant leaves, leading to development of necrotic leaf spots (Kiba ). The hrp-deficient mutants lost their ability to grow vigorously in eggplant leaves and also their virulence on eggplants, suggesting that the bacteria have a biotrophic interaction with eggplants. SPC9018 grows vigorously in intercellular spaces after invasion through stomata in lettuce leaves. This leads to induction of PCD, resulting in induction of collapse and browning symptoms (Hikichi ; Kiba ). Although the hrp-deficient mutants of SPC9018 retained their virulence on lettuce plants, the mutants grew more slowly, and the appearance of disease symptoms on infected lettuce leaves was delayed compared with the wild-type strain. Our data suggest that the bacteria have a more necrotrophic interaction with lettuce, and the hrp cluster plays a role in virulence at the early stages of infection into lettuce leaves, although the hrp genes are not directly implicated in induction of PCD. Therefore, after introduction into P. cichorii by horizontal gene transfer, the putative TTSS-dependent effector proteins may hinder or delay the plant defence response, giving the bacteria time to multiply before inducing PCD in lettuce leaves.