Bacterial strains, plants and plasmids

Pseudomonas syringae pv. maculicola strain M2 (Rif^R) was a kind gift from Dr. Jeffrey L. Dangl (^{Ritter and Dangl, 1995}), and PsmMut8 was obtained in this work. E. coli S17-1 λpir (thi pro hsdR hsdM ΔrecA RP4-2traTc::Mu Km::Tn7) (^{de Lorenzo et al., 1990}) was obtained from Dr. Kate J. Wilson. E. coli DH5α competent cells (supE44ΔlacU169 (f80 lacZ DM15) hsdR17 recA1 endA1 gyrA96 thi-1 relA1) (^{Sambrook and Russell, 2001}) were used for cloning experiments. A SwaI restriction site was added into the SmaI site on pUIRM504 (^{Marsch-Moreno et al., 1998}) to form the plasmid pMDC505 (unpublished results); with this change, the transposable element pTn5cat (^{Marsch-Moreno et al., 1998}) was modified into pTn5cat1. King’s B medium (King et al., 1954), minimal medium M9 (^{Sambrook and Russell, 2001}) or M9CA (Difco) was used to culture P. s. maculicola strains and in the assay to determine the conditions for cat expression, with or without the additions described below. LB medium was used to culture the E. colistrains. Chloramphenicol, rifampicin and kanamycin were purchased from Serva or Sigma-Aldrich Chemicals.

Mutagenesis and mutant selection

Mutants of PsmM2 were generated using the transmid element pTn5cat1 according a published protocol (^{Marsch-Moreno et al., 1998}). E. coli S17-1 (pMDC505) was used to mobilize pTn5cat1 to PsmM2 by conjugation, and the bacteria were then spread onto M9 Rif₅₀ Km₅₀ plates. Mutants were screened for their ability to growth on M9 Cm₅₀ with plant extract. To obtain crude plant extract, mature rosette leaves from 3-week-old Arabidopsis thaliana plants were frozen in liquid nitrogen and ground into a powder, which was then centrifuged at 13,000 rpm for 10 to 20 min. The liquid phase was recovered and added to the growth medium as an effector of pathogenesis.

Assay for promoter strength

The promoter strength was evaluated as the cell density after the bacteria were grown in a medium containing chloramphenicol (^{Alvarez-Mejia et al., 2013}). The assays were performed in sterile 96-well polystyrene plates. First, 50 μL of a 0.04-OD₆₂₀ culture of mutant PsmMut8 in KB Km₅₀ was added to wells containing 200 μL of M9, M9Ca or KB medium supplemented or not with plant extract (2 μL/mL) or sucrose (5%); all media contained kanamycin (50 μg/mL) and chloramphenicol (150 μg/mL). The plates were incubated at 28 °C, and the cell density was measured at 0, 24 and 48 h using a Titertek Multiskan Plus (EFLAB, Joint Venture Company of Lab System and Flow Laboratories) with a 492-nm filter.

Pathogenesis assays

To test the ability of the mutants to induce disease in A. thaliana, 3-week-old plants were inoculated by infiltration with mutant or wild type PsmM2 cell suspensions (~20 μL per leaf). The cell suspensions were prepared by growing PsmMut8 or PsmM2 in 5 mL KB, incubated at 28 °C overnight with strong shaking to reach an 0.4 of OD₆₀₀. Then, 3 mL were centrifuged at 14,000 rpm for 2 min at 4 °C (rotor: Sorvall SS34). The pellet was washed two times with sterile water, and the cells were resuspended in 3 mL of sterile distilled water. Leaves were inoculated with the undiluted cell suspension or with a 1:10 dilution.

Cloning and sequencing

Total PsmMut8 DNA was purified using a previously described method (^{Chen and Kuo, 1993}). First, ten μg of DNA were completely digested using the restriction endonuclease EcoRI in a reaction volume of 50 μL. The enzyme was then inactivated at 65 °C for 20 min. Next, 1 μg of cut DNA was religated with T4 DNA ligase in a reaction volume of 50 μL at 28 °C for 4 h. The ligated DNA was then used to transform competent E. coli DH5α cells to become kanamycin resistant. To sequence the cloned chromosomal fragments, oligonucleotides 1212 (5′-GTGCCTGACTGCGTTA-3′; from the mob end), 1213 (5′-CCTTAGCTCCTGAAA-3′: from the cat end), 1658 (5′-GTTGACCTACGTCAACGCTGGC-3′), 2176 (5-GTGTCGAACACCGAAAG-3 to sequence hrpB), and 2149 (5-TCTGAAGAGTGGCGTTGGAAGC-3 to sequence hrpA) were used. Restriction endonucleases and T4 DNA ligase were purchased from New England BioLabs, Inc. or Invitrogen. Enzymes were used following the suppliers’ recommendations.

Bioinformatics analysis and alignment

hrpAZB operons from diverse Pseudomonas strains were retrieved from the GenBank database and used in our analysis (Table 1). Most of them had been used in a previous work (^{Inoue and Takikawa, 2006}). Nucleotide polymorphism analysis was conducted using DnaSP (^{Rozas et al., 2003}), and the sliding window analysis for hrpAZB operon was conducted using 25 nt in a window of 50 nt only for unique P. syringae strains. Bioinformatics analysis was performed using the BLASTn program (^{Altschul et al., 1990}; ^{Worley et al., 1998}), and alignments were performed using Clustal W and edited with BioEdit (^{Hall, 1999}); Pseudomonas viridiflava and Pseudomonas cichorii were included as outgroups.

Selection of PsmMut8

A collection of PsmM2 mutants harboring the pTn5cat1 transposon-based construct was screened for the induction of cat expression in M9 medium containing a plant extract (see the Materials and Methods section for details). A total of 14 candidates were identified by their ability to grow in M9 Km₅₀ Cm₁₅₀because the reporter gene cat was induced by the plant extract. All of these mutants were tested in pathogenesis assays by inoculating Arabidopsis plants. Mutant number 8 (PsmMut8) was selected because it was unable to infect and cause disease symptoms or hypersensitivity reaction (HR) in either A. thaliana or Brassica oleracea (Figure 1).

Figure 1 Pathogenesis and HR assays for PsmM2 and PsmMut8. A. Arabidopsis leaves were infected by PsmM2 but not by PsmMut8. B. HR assay in collard leaves; PsmM2 but not PsmMut8 was able to produce HR, similarly to hrpZ⁻ from Pseudomonas syringae pv. glycinea (Psg) and Pseudomonas syringae pv. tomatoDC3000 (PstDC3000). Phosphate buffer was used as a control 

Promoter expression detected in PsmMut8

The cat reporter gene in pTn5cat1 allows for the estimation of promoter expression under conditions that resemble those in the apoplast. The cell density in liquid media in the presence of chloramphenicol is associated with the resistance level to the antibiotic, suggesting that the measurement of cell density in the presence of chloramphenicol in different media (M9, M9Ca or KB) with or without the addition of plant extract or sucrose reflects the expression level of the detected promoter under these conditions (^{Alvarez-Mejia et al., 2013}). The cell density values of PsmMut8 growing in different media at 28 °C after 48 h are shown in Figure 2. The cell density was higher in M9 than in KB medium, suggesting that chloramphenicol resistance in response to the plant extract was increased in M9 but that casamino acids preclude the stimulatory effect of the plant extract. No different effects were observed in the assay with sucrose.

Figure 2 PsmMut8 was cultured at 28 °C for 48 h in M9, M9Ca or KB medium. All of the media contained chloramphenicol (Cm, 150 μg/mL), and some of the media were supplemented with plant extract (Ext) or sucrose (Suc) 

pTn5cat1 is inserted into a gene homologous to the hrpZ gene of Pseudomonas syringae

A 14-kb chromosomal fragment corresponding to the pTn5cat1 borders and their flanking genomic sequences were cloned, sequenced, and compared to the genomic information contained in GenBank. Both flanking sequences are homologous to the hrpZ gene from Pseudomonas syringae pv. tomato DC3000 (PstDC3000) (99% identity, six nucleotide substitutions over 1,110 bp, Figure 3A). Alignment with other sequences reported in GenBank for HrpZ proteins revealed two shared regions between PsmM2 and PstDC3000, including genomic locations 102–125, IGAGGGGGGIGGAGSGSGVGGGLS, and 229–244, SGVTSGGGLGSPVSDS.

Figure 3 A. The insertion of pTn5cat1 into the hrpZ gene. IR, inverted repeated; cat, chloramphenicol acetyltransferase; neo, neomycin phosphotransferase; ori_ColE1, replication origin type ColE1; mob, mobilization region from RP4. The PacI, PmeI and SwaI restriction sites are shown. The angled arrows indicate the position of the oligonucleotides for the following sequences: A. 1213, 5′-TTTCAGGAGCTAAGG-3′; B. 1212, 5′-GTGCCTGACTGCGTTA-3′; and C. 1638, 5′-CGTGGTTTGCAGTCGGTTT-3′. B. The genes detected around pTn5cat1 are similar to hrpS, hrpA, hrpB and hrpC (GenBank accession number AY325899) 

Additional sequences flanking pTn5cat1 are similar to the hrp genes of Pseudomonas syringaeDC3000

To further investigate the location of the interrupted gene in PsMut8, we sequenced the regions upstream and downstream of hrpZ. All of the generated sequences corresponded to previously identified genes encoding Hrp proteins: hrpS, hrpA, hrpZ, hrpB and hrpC(Figure 3B). The first and the last open reading frames (ORFs) were only partially sequenced. A putative hrp box (GGAACCGATTCGCAGGCTGCTGCCACCTA) was identified in the 5′ region of hrpA (^{Zwiesler-Vollick et al., 2002}), and a putative ribosome binding site (RBS) was identified within the hrpA gene. The 3′-UTRs of hrpA and hrpZ are predicted to fold into hairpin structures reminiscent of bacterial transcription terminators (TGAGTACCAAGCAATCACGCTGGTAAATCTTA and GCCCCCTCATCAGAGGGGGC, respectively). The presence of a putative RBS within the terminator suggests that the transcription of hrpZ proceeds independently of hrpA. To explore a possible conservation of the hrpAZB operon in different pathovars, including PsmM2 and PstDC3000, we conducted a phylogenetic analysis with 35 Pseudomonas syringae sequences; 2 different Pseudomonas species were included as outgroups. Our analysis was based on maximum likelihood estimations and the Kimura two-parameter substitution model with 1000 bootstraps. Our results showed that PsmM2 belongs to phylogroup II, as described by Inoue (^{Inoue and Takikawa 2006}), or group 5, as described by Guttman (^{Guttman et al., 2006}), and is closely related to the tomatopathovar, as well as to other maculicola strains (Figure 4). They also showed that nucleotide polymorphisms within the operon are particularly abundant in the hrpA gene and the 5 region of hrpZ, whereas polymorphisms are less abundant in the intergenic regions (Figure 5).

Figure 4 Phylogenetic distribution of the hrpABZ operon by the maximum likelihood method. PsmM2 is located in clade V, and P. s. tagetis is basal to clades IV and V 

Figure 5 Nucleotide polymorphism analysis (pi) for the hrpAZBoperon. Only sequences from Pseudomonas syringaepathovars were used. bp, base pair