Introduction

In all natural environments, plants are constantly exposed to a large number of microorganisms, including potential pathogens, and beneficial and saprophytic microorganisms. Plants have not acquired an immune system like that of animals. Therefore, they have developed monitoring systems that recognize potential invading pathogens and a wide range of immune responses to defend themselves (Hacquard et al. 2017; Jones and Dangl 2006). The first line of plant defense against invading pathogens is pathogen-associated molecular pattern (PAMP)-triggered immunity (PTI), which recognizes conserved molecules such as flagellin and elongation factor Tu (EF-Tu) of the invading pathogens using plant pattern-recognition receptors (PRRs), such as FLS2 and EFR, respectively (Zipfel 2008; Zipfel and Felix 2005; Zipfel et al. 2006). After recognition of invading pathogens with PRRs, plants activate a number of defense mechanisms including stomatal-based defense, rapid oxidative burst, restriction of nutrient transfer from the cytosol to the apoplastic space, the accumulation of antimicrobial compounds including phytoalexins, and the activation of hormone-mediated signaling pathways leading to defense responses (Ahuja et al. 2012; Bednarek 2012; Berens et al. 2017; Chen et al. 2010; Cowan 1999; Melotto et al. 2006, 2008; O’Brien et al. 2012; Sawinski et al. 2013; Wang et al. 2012). Stomatal-based defense to restrict the entry of bacterial pathogens by stomatal closure is one of the earliest responses in PTI (Melotto et al. 2008; Underwood et al. 2007). However, successful bacterial pathogens have evolved to acquire multiple virulence factors such as phytotoxins and type III secretion system (TTSS) effectors to overcome PTI and stomatal-based defense (Lozano-Durán et al. 2014; Melotto et al. 2017).

Pseudomonas syringae is a bacterial pathogen that causes economically important plant diseases (Mansfield et al. 2012; Xin and He 2013). It can induce a wide range of symptoms such as leaf blight, spots, galls, and wilt on different plants and can be classified into > 50 pathovars (pv.) based on their host plant specificities and the type of disease symptoms. Pseudomonas syringae pv. tomato strain DC3000 (Pst DC3000), which causes bacterial speck disease of tomato, has been used as a model pathogen to investigate the molecular basis of plant–pathogen interactions because of its genetic tractability and its pathogenicity on Arabidopsis (Ishiga et al. 2011; Katagiri et al. 2002; Xin and He 2013). It is predicted that Pst DC3000 has over 200 virulence-related genes based on the whole genome sequence, including genes related to TTSS effectors, the phytotoxin coronatine (COR), extracellular polysaccharides (EPS), siderophores, and tolerance to reactive oxygen species (ROS) (Buell et al. 2003; Xin and He 2013). A comprehensive search for TTSS effectors has revealed > 30 effector molecules in the Pst DC3000 genome (Cunnac et al. 2009). Recent studies have revealed that several effectors including HopM1, HopF2, HopX1, AvrB, HopZ1, and XopR play an important role in overcoming PTI and stomatal-based defense in Arabidopsis (Gimenez-Ibanez et al. 2014; Hurley et al. 2014; Jiang et al. 2013; Lozano-Durán et al. 2014; Wang et al. 2016; Zhou et al. 2015). In addition to these effectors, COR has also been reported to suppress PTI resulting in stomatal reopening during Pst DC3000 infection (Lee et al. 2013; Melotto et al. 2006, 2008; Wasternack 2017; Zheng et al. 2012). The Arabidopsis fls2 efr1 double mutant showed enhanced disease susceptibility to a COR-defective mutant because of the lack of stomatal-based defense (Zeng and He 2010). Further, when a COR-defective mutant was used to infiltrate the apoplastic space, thus bypassing stomatal-based defense, this mutant induced typical disease symptoms in Arabidopsis, indicating an important role for COR in overcoming stomatal-based defense (Melotto et al. 2006). Although there is a good understanding of the importance of TTSS effectors and COR in the virulence of Pst DC3000, the function of a large number of potential virulence factors encoded in the Pst DC3000 genome and their modes of action are not fully understood.

Pseudomonas syringae is considered to have two lifestyles, including an epiphytic phase on the surface of plant leaves and an endophytic phase in the apoplastic space (Melotto et al. 2008; Xin and He 2013). The epiphytic phase of P. syringae is one of the first steps in the infection process. Epiphytic fitness has been characterized as a virulence factor based on studies of P. syringae pv. syringae B728a (Pss B728a) (Yu et al. 2013, 2014). The exopolysaccharide alginate, a copolymer of mannuronic and guluronic acid and the major EPS of P. syringae has been reported to play a role during the epiphytic phase by protecting bacterial cells against external stresses such as oxidative and UV stresses (Keith et al. 2003; Laue et al. 2006; Schenk et al. 2006; Yu et al. 1999). Kidney bean inoculated with a Pss B728a algT mutant, deficient in the production of alginate, had reduced disease symptoms together with bacterial growth compared to the wild type (Yu et al. 1999). In addition, Pss B728a has a very pronounced epiphytic phase on plants, whereas Pst DC3000 does not colonize the leaf surface (Xin and He 2013). P. syringae is thought to be able to shift from the epiphytic phase to the endophytic phase during a successful infection cycle. One of the major virulence factors in P. syringae needed to make this shift includes effector proteins, which are delivered into the host through the TTSS to suppress PTI and to facilitate disease development (Brooks et al. 2005; Lindeberg et al. 2012; Nomura et al. 2005; Uppalapati et al. 2007). Not only type III effectors, but also phytotoxins function overcome PTI and stomatal-based defense during the transition from the epiphytic phase to the endophytic phase (Melotto et al. 2006; Uppalapati et al. 2007). However, we still have much to discover about the mechanisms of this bacterial pathogen transition because of the lack of information about the virulence factors related to infection.

In the present study, we demonstrated that the Pst DC3000 sigma factor AlgU regulates not only the expression of genes related to alginate biosynthesis, including algD, but also the expression of genes related to virulence, including TTSS effectors and the phytotoxin COR. Furthermore, the Pst DC3000 ΔalgU, but not the ΔalgD mutant showed reduced virulence during host plant infection, indicating that alginate may not have a role in virulence in Pst DC3000. Thus, our results suggest the importance of the downstream components of AlgU, such as COR, to overcome stomatal-based defense in the virulence of Pst DC3000.

Materials and methods

Plant materials and growth conditions

Arabidopsis thaliana seedlings and Solanum lycopersicum cv. Glamour plants were used for the pathogenicity assay. A. thaliana ecotype Colombia (Col-0) was used as the wild-type plant in this study. The fls2 efr1 line was obtained from Dr. Saijo (Tintor et al. 2013). Briefly, A. thaliana seeds were germinated and grown on 1/2 strength Murashige and Skoog (MS) medium (0.3% phytagel) with Gamborg vitamins (Sigma-Aldrich, St. Louis, MO, USA). A. thaliana plants were incubated in a growth chamber at 24 °C with a light intensity of 200 µE m−2 s−1 and a 12 h light/12 h dark photoperiod and used 2 weeks after germination for the pathogen assays. Tomato plants were used for pathogen assays 4 weeks after germination and subsequent growth at 24 °C with a light intensity of 200 µE m−2 s−1 and a 12 h light/12 h dark photoperiod.

Bacterial strains and growth conditions

Pseudomonas syringae pv. tomato strain DC3000 (Pst DC3000) was used as the pathogenic strain to inoculate tomato and Arabidopsis. Wild-type (WT) Pst DC3000; mutants ∆algU, ∆algD, and COR-deficient DB29; and an algU-complemented line (∆algU/algU) were grown at 28 °C on mannitol-glutamate (MG; Keane et al. 1970) or King’s B (KB; King et al. 1954) media containing 25 µg/ml of rifampicin. Before inoculation, bacteria were suspended in sterile distilled H2O, and the bacterial cell densities at 600 nm (OD600) were measured using a JASCO V-730 spectrophotometer (JASCO, Tokyo, Japan).

Generation of ∆algU and ∆algD mutants

The genetic regions containing algU and algD and the surrounding regions were amplified using sets of PCR primers (4224_1: 5′-CGCGCTGTATGTATCCACTT-3′ and 4224_2: 5′-GCCATATCGATGCTGGTCTT-3′ for algU, and 1243_1: 5′-CTGCTCAACCTTGCCTCTAC-3′ and 1243_2: 5′-AACCAGCTGGACTTGTCATC-3′ for algD) that were designed based on the registered sequence of Pst DC3000 (NC_004578) with PrimeStar HS DNA polymerase (TaKaRa, Otsu, Japan), then dA was added to the 3′ end of the PCR product with 10× A-attachment mix (TOYOBO, Osaka, Japan). The resultant DNA was inserted into the pGEM-T Easy vector (Promega, Madison, WI, USA). The recombinant plasmid DNA was then used to obtain pGEM-algU and pGEM-algD as templates, and inverse PCR was carried out using a set of primers (4224_3: 5′-cgggatccTCCTCTTCCTGGGTTAGCAT-3′; 4224_4: 5′-cgggatccGTTGTTGCAGGAATCCTGA-3′ for algU, and 1243_3: 5′- cgggatccCCAAATATGCTGATACGCAT-3′; 1243_4: 5′-cgggatccGAAGGTATCTGCTGGTAA-3′for algD; lowercase letters indicate the artificial sequence for the digestion by BamHI) to delete an open reading frame for algU and algD. Then, the PCR product and template DNA were digested with BamHI and DpnI. The resultant DNA was self-ligated with T4 DNA ligase (Ligation-Convenience kit, Nippon Gene, Tokyo, Japan). The algU- and algD-deleted DNA constructs were introduced into the EcoRI site of the mobilizable cloning vector pK18mobsacB (Schäfer et al. 1994). The resulting plasmids containing the DNA fragment lacking algU and algD were then used to transform E. coli S17-1. The deletion mutant was obtained by conjugation and homologous recombination according to the method previously reported (Shimizu et al. 2003). Transconjugants were selected on KB agar containing 30 µg/ml of kanamycin (Km) and 30 µg/ml of rifampicin (Rif). We generated mutants by incubating the transconjugants on a KB agar plate containing 25 µg/ml rifampicin and 10% sucrose. The specific deletions in the ∆algU and ∆algD mutants were confirmed by PCR using the primers 4224_1 and 4224_2 for ∆algU and 1443_1 and 1443_2 for ∆algD.

To confirm whether the altered phenotype of the ∆algU mutant originates from a corresponding mutation, an algU-complemented strain was generated. The algU-containing DNA fragment from pGEM-algU was transferred into the vector pDSK519 (Keen et al. 1988) at the SacI and SphI sites to generate pDSK-algU. The pDSK-algU construct was introduced into the ∆algU mutant by electrophoresis to generate the complemented strain ΔalgU/algU.

Growth curve assay

Pst DC3000 strains including the wild type, ∆algU mutant, and complemented line ∆algU/algU were grown at 28 °C for 24 h in KB broth. The strain suspensions were adjusted to an OD600 of 0.1 with fresh KB broth, and the bacterial growth dynamics were measured at OD600 for 24 h.

Spray- and flood-inoculation methods

Spray inoculations were used to observe disease symptoms on tomato leaves as described previously (Uppalapati et al. 2007). Briefly, plants were sprayed with a bacterial suspension (OD600  = 0.1) in sterile distilled water containing 0.025% Silwet L-77 (OSI Specialties, Danbury, CT, USA) until runoff. The plants were then incubated in growth chambers at ∼ 100% RH for the first 24 h, then at ∼ 70% RH for the rest of the experimental period. The inoculated plants were observed for 7 dpi for symptom development.

A flood-inoculation method was used to observe disease symptoms in Arabidopsis as described previously (Ishiga et al. 2011). Briefly, 40 ml of bacterial suspension made in sterile distilled H2O containing 0.025% Silwet L-77 was dispensed onto a plate containing 2-week-old Arabidopsis seedlings for uniform inoculation, then the plates were incubated for 2–3 min at room temperature. After the bacterial suspension was decanted from the plates, the plates were sealed with 3M Micropore 2.5 cm surgical tape (3M, St. Paul, MN, USA) and incubated at 24 °C with a light intensity of 200 µE m−2 s−1 and 12 h light/12 h dark.

To assess bacterial growth in tomato leaves and Arabidopsis seedlings, we measured the internal bacterial population at several times. For determining internal bacterial growth, the inoculated leaves (tomato) and seedlings (Arabidopsis) were collected, and the total mass of each was measured, then the leaves and seedlings were surface-sterilized with 5% H2O2 for 3 min, then washed three times with sterile distilled water. The plants were then homogenized in sterile distilled water, and the diluted samples were plated onto solid MG agar medium. The bacterial colony forming units (CFU) were normalized as CFU/mg using the total mass of the inoculated leaf and seedling samples. The bacterial population at 0 dpi was estimated using leaves harvested 1 hpi. The bacterial populations were evaluated in three independent experiments.

Real-time quantitative RT-PCR

For expression profiles of Pst DC3000 genes, bacteria were grown in KB broth for 24 h, then adjusted to OD600 of 0.2 with fresh MG broth and grown for 2 h. To analyze the expression profiles of Pst DC3000 genes during infection, we flood-inoculated Arabidopsis wild-type Col-0 plants with WT Pst DC3000, ΔalgU, or ΔalgU/algU at 5 × 106 CFU/ml for 3, 6, 12, 24, and 48 h, then the total RNAs including plant and bacterial RNAs were extracted from infected leaves and purified. Mixed plant–bacterial RNA populations were utilized for the expression profiles without bacterial RNA enrichment. Total RNA extraction and real-time quantitative RT-PCR (RT-qPCR) were done as described previously (Ishiga and Ichinose 2016). Total RNA was extracted using RNAiso Plus (TaKaRa) according to the manufacturer’s protocol. Two micrograms of total RNA was treated with gDNA Eraser (TaKaRa) to eliminate genomic DNA, and the DNase-treated RNA was reverse transcribed using the PrimeScript RT reagent Kit (TaKaRa). The cDNA (1:20) was then used for RT-qPCR using the primers shown in Supplementary Table S1 with SYBR Premix Ex Taq II (TaKaRa) on a Thermal Cycler Dice Real Time System (TaKaRa). Pst DC3000 oprI (outer membrane lipoprotein precursor), proC (pyrroline-5-carboxylate reductase), and rpoD (RNA polymerase sigma factor) were used as internal controls to normalize gene expression. The Arabidopsis UBIQUITIN EXTENSION PROTEIN 1 (UBQ1) was used as an internal control to normalize gene expression in Arabidopsis. The average CT values were calculated using the 2nd derivative maximum method, and triplicate samples were used to determine fold expression relative to the controls.

Stomatal assay

A modified method was used to assess stomatal response as described previously (Chitrakar and Melotto 2010). Briefly, Arabidopsis and tomato plants were grown for 4 weeks after germination as described before. Pst DC3000 strains including the wild type, the ∆algD mutant, the ∆algU mutant, DB29, and ΔalgU/algU were grown at 28 °C for 48 h on MG agar. The strains were adjusted to OD600 of 0.2 in sterile distilled water. Dip-inoculated leaves with or without COR (100 ng/ml; Sigma-Aldrich) were directly imaged using a Leica TCS SP8 confocal microscope equipped with a white light laser (Leica, Wetzlar, Germany). A reflected image of the leaf surface was obtained by illuminating the sample with 561 nm wavelength, and reflected light was detected through a 558–566 nm filter. The stomatal aperture (width) of at least 60 stomata was measured. The average and standard error for the stomatal aperture width were calculated. The stomatal apertures were evaluated in three independent experiments.

Results

Pst DC3000 AlgU is required for full virulence in tomato and Arabidopsis

To assess the importance of AlgU and AlgD in the virulence of Pst DC3000, we constructed an algU mutant (ΔalgU) and algD mutant (ΔalgD). In the growth curve assay for Pst DC3000 wild-type (WT), ΔalgU, and ΔalgD in KB and MG broth, the populations of the strains did not differ significantly (Supplementary Fig. 1a, b). Next, we challenged tomato and A. thaliana with Pst DC3000 WT, ΔalgU, and ΔalgD, and observed disease progression. Both WT Pst DC3000 and ΔalgD caused typical necrotic cell death lesions with severe chlorosis on tomato leaves at 7 dpi; however, ΔalgD-inoculated leaves developed milder chlorosis (Fig. 1a). On the other hand, tomato leaves inoculated with ΔalgU appeared healthy with reduced necrotic cell death and chlorosis compared with WT Pst DC3000 (Fig. 1a). In tomato leaves inoculated with ΔalgU, bacterial populations were significantly lower than for WT Pst DC3000 and ΔalgD (Fig. 1b).

Fig. 1
figure 1

Disease symptoms on tomato leaves and bacterial population dynamics in leaves after 4-week-old plants were sprayed with 1 × 108 colony forming units (CFU)/ml of the wild-type (WT) Pseudomonas syringae pv. tomato DC3000 (Pst DC3000), the ΔalgU mutant (ΔalgU), or the ΔalgD mutant (ΔalgD). Bacterial populations in leaves were estimated by dilution plating on selective medium as described in the methods. a Chlorosis and necrotic lesions on tomato leaves at 7 days post inoculation (dpi). b Population levels at 0, 2, 4, and 6 dpi. Vertical bars indicate standard error for three independent experiments. Asterisks indicate a significant difference from WT in a t test (**P < 0.01). (Color figure online)

Full size image

Pst DC3000 is pathogenic not only to tomato, but also to A. thaliana (Ishiga et al. 2011), so did the same inoculation tests using A. thaliana. Consistent with the results for tomato, WT Pst DC3000 caused typical water soaking lesions with severe chlorosis, whereas ΔalgU caused reduced disease symptoms with less chlorosis at 3 dpi (Fig. 2a). Interestingly, Arabidopsis plants inoculated with ΔalgD developed less severe symptoms (Fig. 2a), and the ΔalgU populations were significantly lower compared to those of WT Pst DC3000 and ΔalgD (Fig. 2b). Taken together, these results clearly indicate that AlgU has a critical role in the virulence of Pst DC3000, whereas AlgD may partially contribute to virulence in tomato and Arabidopsis.

Fig. 2
figure 2

Disease symptoms and bacterial population dynamics in Arabidopsis plants flood-inoculated with a suspension of 5 × 106 CFU/ml of the wild-type (WT) Pseudomonas syringae pv. tomato DC3000 (Pst DC3000), the ΔalgU mutant (ΔalgU), or the ΔalgD mutant (ΔalgD). 2-week-old Arabidopsis plants were flood-inoculated, and bacterial populations were estimated by dilution plating on selective medium as described in the methods. a Water-soaked lesions with chlorosis at 3 days post-inoculation (dpi). b Population levels at 0, 2, and 3 dpi. Vertical bars indicate the standard error for three independent experiments. Asterisks indicate a significant difference from WT in a t test (*P < 0.05; **P < 0.01). (Color figure online)

Full size image

AlgU regulates the expression of virulence genes in Pst DC3000 in broth

It has been reported that AlgU activates the transcription of target genes related to alginate synthesis, virulence, motility, and oxidative stress (Markel et al. 2016). We thus investigated the expression profiles of these genes in WT Pst DC3000, ΔalgU, and algU-complemented line (ΔalgU/algU) using RT-qPCR and gene-specific primer sets (Supplementary Table S1) Our results demonstrated that the expression of algU was higher in ΔalgU/algU than in WT Pst DC3000 (Fig. 3a), perhaps because of the higher copy number of the algU gene after complementation with pDSK519 plasmid. The expression of algD was induced in the WT Pst DC3000 and the ΔalgU/algU culture, but repressed in the ΔalgU culture (Fig. 3b), indicating that AlgU can positively regulate genes related to alginate biosynthesis.

Fig. 3
figure 3

Expression profiles of genes involved in the virulence of wild-type (WT) Pseudomonas syringae pv. tomato DC3000 (Pst DC3000), the ΔalgU mutant (ΔalgU), and algU-complemented line (ΔalgU/algU) in liquid mannitol-glutamate (MG) broth. WT Pst DC3000, ΔalgU, and ΔalgU/algU were grown in MG broth to an OD600 of 0.2. Gene expression of selected Pst DC3000 genes was normalized using housekeeping genes oprI, proC, and rpoD by real-time quantitative reverse transcription-polymerase chain reaction (RT-qPCR) with gene-specific primer sets (Supplementary Table S1). a algU, b algD, c hrpL, d cfl, e cmaB, f fliC, and g tuf. Vertical bars indicate the standard error for three biological replicates. Asterisks indicate a significant difference from WT in a t test (*P < 0.05; **P < 0.01). ND indicates not detected by RT-qPCR

Full size image

TTSS effectors and the phytotoxin COR play central roles in the virulence of Pst DC3000 by suppressing plant immunity and also facilitating disease development in host plants (Buell et al. 2003; Xin and He 2013). To investigate the importance of AlgU in the regulation of genes related to virulence, we analyzed the expression profiles of genes encoding transcriptional regulators of the TTSS effectors and the COR biosynthesis genes, including hrpL, cfl, and cmaB. The expression of hrpL, cfl, and cmaB was clearly suppressed in ΔalgU compared to their levels in WT Pst DC3000 and the ΔalgU/algU (Fig. 3c–e). These results suggest that AlgU may function in modulating the expression of the genes related to the TTSS and COR biosynthesis.

Virulence of the ΔalgU mutant was impaired in Arabidopsis (Fig. 2). Because PRRs such as FLS2 and EFR can recognize flagellin and EF-Tu, respectively (Zipfel 2008; Zipfel and Felix 2005; Zipfel et al. 2006), then the reduced virulence of the ΔalgU mutant might result from increased production of PAMPs, including flagellin and EF-Tu, rather than from the reduced expression of virulence genes. To rule out this possibility, we confirmed the expression profiles of fliC and tuf, encoding flagellin and EF-Tu, respectively. As shown in Fig. 3f, g, the expression of fliC and tuf did not differ significantly between WT Pst DC3000 and ΔalgU.

AlgU regulates the expression of virulence genes during Pst DC3000 infection of Arabidopsis

Since the virulence of ΔalgU was reduced in tomato and in Arabidopsis (Figs. 1, 2), we next determined the expression profiles of the Pst DC3000 genes involved in virulence during infection. In this assay, Arabidopsis wild-type Col-0 plants were flood-inoculated with WT Pst DC3000, ΔalgU, or the ΔalgU/algU, and at 3, 6, 12, 24, and 48 hpi, total RNAs (plant and bacterial) were purified for real-time RT-qPCR. The expression of the Pst DC3000 genes was not detected in the noninoculated Arabidopsis control plants (data not shown).

The expression of algD was significantly suppressed in Arabidopsis plants inoculated with ΔalgU in comparison with the WT Pst DC3000 (Fig. 4a, b). The expression of hrpL was induced in Arabidopsis inoculated with WT Pst DC3000 at 24 hpi; however, hrpL transcripts were differentially regulated during ΔalgU infection compared with WT Pst DC3000 (Fig. 4c). Interestingly, the expression of cfl and cmaB were suppressed in plants inoculated with ΔalgU at 3 and 6 hpi in comparison with the WT Pst DC3000 (Fig. 4d, e). These results suggest that AlgU can positively regulate the transcription of hrpL and COR-related genes during Pst DC3000 infection.

Fig. 4
figure 4

Expression profiles of genes involved in the virulence of wild-type (WT) Pseudomonas syringae pv. tomato DC3000 (Pst DC3000), the ΔalgU mutant (ΔalgU), or algU-complemented line (ΔalgU/algU) after inoculation of 2-week-old plants of Arabidopsis with a suspension of 5 × 106 colony forming units (CFU)/ml. Total RNA was extracted at 3, 6, 12, 24, and 48 h post inoculation (hpi) for use in real-time quantitative reverse transcription-polymerase chain reaction (RT-qPCR) with gene-specific primer sets. Expression of a algU, b algD, c hrpL, d cfl, e cmaB, f fliC, g oxyR, and h katE was normalized using housekeeping genes oprI, proC, and rpoD (Supplementary Table S1). Vertical bars indicate the standard error for three biological replicates. Asterisks indicate a significant difference from WT in a t test (*P < 0.05; **P < 0.01). ND indicates not detected by RT-qPCR

Full size image

The expression of fliC was induced at 12 hpi in plants inoculated with the WT Pst DC3000 and ΔalgU and at late stages of ΔalgU infection (Fig. 4f). The expression of katE, but not oxyR, was induced during late infection by WT Pst DC3000; however the transcript level of katE was lower in Arabidopsis plants inoculated with ΔalgU than after inoculation with WT Pst DC3000 (Fig. 4h), suggesting that AlgU can positively regulate the expression of oxidative stress tolerance genes.

The ΔalgU mutant displays full virulence on the Arabidopsis fls2 efr1 mutant

The interactions between P. syringae and Arabidopsis have been utilized to investigate the molecular basis of plant defense responses against pathogens and have revealed a number of key regulators for signaling pathways (Chisholm et al. 2006; Gimenez-Ibanez and Rathjen 2010; Ishiga et al. 2011; Macho and Zipfel 2014; Xin and He 2013). We next determined the expression profiles of the plant defense marker genes including AtPR1 and AtPR2 during infection. In this assay, Arabidopsis wild-type Col-0 plants were flood-inoculated with WT Pst DC3000 or ΔalgU for 6, 12, 24, and 48 h, and then the total RNAs including plant and bacterial RNAs were purified for real-time RT-qPCR. The expression of AtPR1 was induced in plants inoculated with WT Pst DC3000 and the ΔalgU mutant, however the expression levels of AtPR1 were greater in plants inoculated with the ΔalgU mutant at 24 hpi compared to those inoculated with WT Pst DC3000 (Supplementary Fig. S2a). Moreover, the expression levels of AtPR2 were greater in plants inoculated with the ΔalgU mutant at 48 hpi compared to those inoculated with WT Pst DC3000 (Supplementary Fig. S2b).

Plants can recognize the bacterial invasion by PRRs leading to PTI (Zipfel 2008; Zipfel et al. 2004, 2006). To investigate the virulence of ΔalgU on Arabidopsis mutants compromised in PTI, we utilized a PRRs mutant, fls2 efr1. Interestingly, the bacterial growth revealed enhanced susceptibility of the fls2 efr1 mutant to the ΔalgU mutant, whereas the populations of WT Pst DC3000 were comparable to those in the wild-type Col-0 (Fig. 5a, b). Moreover, the ΔalgU mutant multiplied to almost the same levels as WT Pst DC3000 (Fig. 5b), indicating that the ΔalgU mutant displays full virulence on the Arabidopsis fls2 efr1 mutant. These results suggest that the virulence factors regulated by AlgU can affect FLS2/EFR-dependent PTI during Pst DC3000 infection.

Fig. 5
figure 5

Disease symptoms on wild-type Arabidopsis Col-0 and fls2 efr1 mutant after flood-inoculation with 5 × 106 colony forming units (CFU)/ml of wild-type (WT) Pseudomonas syringae pv. tomato DC3000 (Pst DC3000) or the ΔalgU mutant (ΔalgU) or dipping into WT Pst DC3000, ΔalgD, ΔalgU, DB29, or algU-complemented line (ΔalgU/algU) with or without coronatine (COR). 2-week-old plants were flood-inoculated, and then the bacterial populations were measured as described in methods. a Water-soaked lesions with chlorosis at 3 dpi. b Bacterial populations in leaves estimated by dilution plating on selective media at 2 dpi. c Stomatal aperture width on intact Arabidopsis leaves 4 h after dip-inoculation with WT Pst DC3000, ΔalgU, or DB29. d Stomatal aperture width on intact Arabidopsis leaves 4 h after dip-inoculation with WT Pst DC3000, ΔalgD, ΔalgU, DB29, or algU-complemented line (ΔalgU/algU) with or without COR (100 ng/ml). e Stomatal aperture width in intact tomato leaves 4 h after dip-inoculation with WT Pst DC3000, ΔalgD, ΔalgU, DB29, and ΔalgU/algU with or without COR (100 ng/ml). In all bar graphs, vertical bars indicate the standard error for three biological replicates. Asterisks indicate a significant difference from WT in a t test (**P < 0.01)

Full size image

The stomatal-based defense mechanism in plants is responsible for closing stomata in response to the perception of PAMPs of invading pathogens (Melotto et al. 2017). In turn, bacterial COR and TTSS effectors can prevent this PAMP-triggered stomatal closure (Melotto et al. 2017). To investigate PTI, especially stomatal-based defense, we examined the stomatal response after inoculation with WT Pst DC3000, the ΔalgU mutant, and the COR-deficient mutant DB29 in Arabidopsis. WT Pst DC3000 was able to cause the stomata to reopen, whereas stomatal closure of Arabidopsis was triggered after inoculation with DB29 and ΔalgU at 4 hpi; however, more stomata closed when triggered by ΔalgU than by DB29 (Fig. 5c). Interestingly, bacterium-triggered stomatal closure was not observed on fls2 efr1 mutant plants inoculated with WT Pst DC3000, ΔalgU, or DB29 (Fig. 5c). To further clarify the role of AlgU-regulated COR, we next examined the stomatal response after inoculation with WT Pst DC3000, the ΔalgD mutant, the ΔalgU mutant, DB29, or the algU-complemented line (ΔalgU/algU) with or without exogenous application of COR in Arabidopsis and tomato. WT Pst DC3000 and the ΔalgD mutant were able to reopen stomata, whereas bacterium-triggered stomatal closure was observed in ΔalgU- and DB29-inoculated Arabidopsis at 4 hpi (Fig. 5d). Interestingly, the application of COR completely restored the phenotype of ΔalgU and DB29 (Fig. 5d). Consistent with these results in Arabidopsis, WT Pst DC3000 and the ΔalgD mutant were able to reopen stomata, whereas bacterium-triggered stomatal closure was observed on ΔalgU- and DB29-inoculated tomato at 4 hpi (Fig. 5e), and the application of COR was able to cause reopening of stomata in tomato inoculated with ΔalgU and DB29 at 4 hpi. These results clearly suggest that AlgU-regulated COR production can have an important role in overcoming PTI, especially stomatal-based defense in the early virulence of Pst DC3000.

Discussion

AlgU, one of the sigma factors that regulates genes responsive to the environment, also regulates the expression of genes related to alginate biosynthesis enzymes, such as AlgD (Okkotsu et al. 2014). The EPS alginate produced by Pseudomonas spp. contributes to biofilm formation and resistance to various external stresses (Keith et al. 2003). In the present mutant analysis of AlgU and AlgD, we found that the ΔalgU mutant clearly showed reduced disease symptoms as well as reduced bacterial populations in host tomato and Arabidopsis, suggesting that AlgU plays a critical role in the virulence of Pst DC3000 (Figs. 1, 2). The host plants inoculated with ΔalgD mutant showed milder disease symptoms, suggesting that AlgD may partially contribute to virulence. Gene expression profiles indicated that algD was differentially expressed in the ΔalgU mutant background, suggesting that AlgU plays a role in regulating gene expression of alginate biosynthesis. Based on the gene expression profiles in Figs. 3c, 4c, it is tempting to speculate that AlgU can regulate the expression of hrpL during Pst DC3000 infection and that AlgU-regulated TTSS effectors may have roles in the virulence of Pst DC3000. Our results were consistent with a previous study describing interactions between tomato and Pst DC3000 (Markel et al. 2016). Markel et al. (2016) characterized the phenotype of a ΔalgU mutant and reported that AlgU, but not AlgD, plays an important role in virulence. On the basis of their transcriptome analysis (RNA-seq) to characterize the AlgU regulon, they suggested that AlgU regulates the expression of TTSS effectors, resulting in modulation of virulence. Thus, AlgU can function as a virulence factor by regulating the expression of genes related to TTSS effectors.

In addition to genes related to TTSS effectors, we also demonstrated that the expression of genes related to the virulence of Pst DC3000 were differentially regulated in the ΔalgU mutant background. Interestingly, the expression of genes related to biosynthesis of the phytotoxin COR, such as cfl and cmaB were clearly suppressed in the ΔalgU mutant background compared to WT Pst DC3000 and algU-complemented line (Fig. 3d, e). Moreover, the expression of these genes was suppressed in the early stage of ΔalgU mutant infection (Fig. 4d, e). The expression of genes related to COR is activated by the sigma factor, HrpL, and CorR, and COR contributes to the virulence of Pst DC3000 by suppressing the host defense response (Sreedharan et al. 2006; Uppalapati et al. 2007). Furthermore, COR can suppress PTI, especially stomatal-based defense in the early stage of Pst DC3000 infection in Arabidopsis and tomato (Lee et al. 2013; Melotto et al. 2006, 2008; Wasternack 2017; Zheng et al. 2012). Based on the gene expression profiles of cfl and cmaB (Fig. 4d, e), AlgU apparently can regulate the production of COR and contribute to virulence by suppressing stomatal-based defense in the early stage of infection. Consistent with the hypothesis, we demonstrated that the virulence phenotype of the ΔalgU mutant was identical to the COR-deficient mutant, DB29, with respect to stomatal-based defense in tomato (Fig. 5e). Interestingly, the application of COR was able to restore the ability of the ΔalgU mutant to reopen stomata in Arabidopsis and tomato (Fig. 5d, e). We also showed the ΔalgU mutant displayed full virulence on the Arabidopsis fls2 efr1 mutant, which is compromised in PTI and stomatal-based defense (Fig. 5c). Arabidopsis double mutant fls2 efr1 is more susceptible to a COR-defective mutant because it lacks the stomatal-based defense (Zeng and He 2010). Together, these results suggest that AlgU contributes to the virulence of Pst DC3000 by regulating COR production to overcome PTI, especially early stomatal-based defense.

Our stomatal response assay demonstrated that DB29 and the ΔalgU mutant induced stomatal closure compared to the WT Pst DC3000 in tomato and Arabidopsis (Fig. 5c–e). However, ΔalgU-triggered stomatal closure was greater than that triggered by DB29 in Arabidopsis (Fig. 5c). It is important to note that the differences between tomato and Arabidopsis in ΔalgU-triggered stomatal closure compared to that triggered by DB29 might be explained by the differential responses of the hosts to COR; purified COR induces chlorosis on leaves of tomato, but not on Arabidopsis leaves (Bent et al. 1992). We also previously demonstrated that purified COR suppresses the expression of genes related to chloroplast-localized ROS-detoxifying enzymes in tomato, but not in Arabidopsis (Ishiga et al. 2012). Therefore, further characterization of responses to COR in tomato and in Arabidopsis is needed to understand the molecular mechanisms of COR in the virulence of Pst DC3000.

Expression profiles in this study revealed a complex effect by AlgU on expression of virulence genes including algD, hrpL, cfl, cmaB, oxyR, and katE. Expression profiles of algD were clearly correlated with that of algU not only in vitro (Fig. 3b), but also during infection by Pst DC3000 (Fig. 4b), suggesting that AlgU can directly regulate the expression of algD. On the other hand, the expression profile of other virulence genes differed from that of algU (Figs. 3, 4). The different expression profiles among virulence genes in ΔalgU mutant background might be explained by complex regulation of virulence genes by multiple transcriptional regulators. Previous studies demonstrated that CorR, a positive regulator for coronafacic acid (CFA) and coronamic acid (CMA), which are intermediates in the COR biosynthesis, bind the promoter region of hrpL and regulate the expression (Sreedharan et al. 2006). Therefore, further identification of complex gene regulatory networks that function to modulate the virulence during Pst DC3000 infection is needed to understand the mechanisms by which AlgU and other regulators, such as HrpL and OxyR function in the virulence of Pst DC3000.

EPS production in planta is correlated with virulence in numerous phytopathogenic bacteria including Ralstonia solanacearum, Erwinia stewartii, and Xanthomonas campestris (Dolph et al. 1988; Kao et al. 1992; Katzen et al. 1998; Saile et al. 1997). Moreover, several studies have shown that alginate plays an important role in the virulence of P. syringae. In bean leaves inoculated with an alginate-defective mutant of P. syringae pv. syringae, disease symptoms and bacterial populations were significantly reduced compared to the wild-type (Yu et al. 1999). In the present study, the ΔalgD mutant induced milder disease symptoms than the WT Pst DC3000 did in tomato and Arabidopsis (Figs. 1, 2), suggesting that alginate may partially contribute to virulence in all P. syringae pathovars. Gene expression analysis of algD in the interaction between Pst DC3000 and tomato showed that the promoter activity of algD was only detected in leaves showing necrotic cell death (Keith et al. 2003). Similarly, our previous studies also showed that the expression of algD was induced in the late stage of Pst DC3000 infection in Arabidopsis (Ishiga and Ichinose 2016). These results suggest that alginate may play a role in the virulence of Pst DC3000, especially in modulating disease-associated necrotic cell death. Further characterization of alginate and other EPS is needed to understand the interactions between the pathogenic bacteria and plants.

Light is known to function as a key factor regulating virulence, including the motility of Pst DC3000 (Río-Álvarez et al. 2014). Several studies demonstrated that motility is a critical virulence factor during P. syringae infection (Clarke et al. 2016; Nogales et al. 2015; Shimizu et al. 2003). Light inhibits the motility of Pst DC3000 by regulating the expression of genes related to flagellum-dependent motility, such as fliC (Río-Álvarez et al. 2014). Based on gene expression profiles, the expression of fliC was clearly induced in the dark (Fig. 4f), but transcript levels of fliC in the WT Pst DC3000 and the ΔalgU mutant were not affected (12 hpi), suggesting that AlgU does not regulate motility in the virulence of Pst DC3000.

OxyR, a key transcription factor related to the virulence and oxidative stress tolerance of Pst DC3000, regulates the expression of genes related to H2O2-detoxifying enzymes, such as the kat genes (Ishiga and Ichinose 2016). Based on gene expression profiles, the expression of oxyR was suppressed in the ΔalgU mutant at 6 hpi (Fig. 4g). The expression of katE, which is downstream of oxyR, was suppressed at 6, 24, and 48 hpi (Fig. 4h), suggesting that AlgU regulates not only the genes related to virulence factors including TTSS effectors and COR, but also the genes related to oxidative stress tolerance through the regulation of OxyR. ROS also play a key role in the defense responses of plants. A study utilizing the Arabidopsis NADPH oxidase rbohD mutant showed that the ROS burst via RbohD plays an important role in PTI (Kadota et al. 2014). Thus, it is tempting to speculate that not only type III effectors and COR, but also OxyR and oxidative stress tolerance mechanisms play an important role in avoiding PTI in the early infection stage in Arabidopsis. Therefore, further identification of regulators that function to modulate the virulence factors during infection and characterization of their modes of action in conjunction with AlgU are needed to understand the mechanisms of AlgU and other regulators in plant–bacterial pathogen interactions.