The Pseudomonas syringae effector protein HopZ1a suppresses effector-triggered immunity

Authors

  • Alberto P. Macho,

    1. Instituto de Hortofruticultura Subtropical y Mediterranea, Universidad de Málaga-Consejo Superior de Investigaciones Científicas (IHSM-UMA-CSIC), Depto Biología Celular, Genética y Fisiología, Campus de Teatinos, Málaga E-29071, Spain
    Search for more papers by this author
  • Carlos M. Guevara,

    1. Instituto de Hortofruticultura Subtropical y Mediterranea, Universidad de Málaga-Consejo Superior de Investigaciones Científicas (IHSM-UMA-CSIC), Depto Biología Celular, Genética y Fisiología, Campus de Teatinos, Málaga E-29071, Spain
    Search for more papers by this author
  • Pablo Tornero,

    1. Instituto de Biología Molecular y Celular de Plantas (Universidad Politécnica de Valencia – CSIC) Avda de los Naranjos s/n. Valencia E-46022, Spain
    Search for more papers by this author
  • Javier Ruiz-Albert,

    1. Instituto de Hortofruticultura Subtropical y Mediterranea, Universidad de Málaga-Consejo Superior de Investigaciones Científicas (IHSM-UMA-CSIC), Depto Biología Celular, Genética y Fisiología, Campus de Teatinos, Málaga E-29071, Spain
    Search for more papers by this author
  • Carmen R. Beuzón

    1. Instituto de Hortofruticultura Subtropical y Mediterranea, Universidad de Málaga-Consejo Superior de Investigaciones Científicas (IHSM-UMA-CSIC), Depto Biología Celular, Genética y Fisiología, Campus de Teatinos, Málaga E-29071, Spain
    Search for more papers by this author

Author for correspondence:
Carmen R. Beuzón
Tel: + 34 952 131959
Email: cbl@uma.es

Summary

  • The Pseudomonas syringae pv syringae type III effector HopZ1a is a member of the HopZ effector family of cysteine-proteases that triggers immunity in Arabidopsis. This immunity is dependent on HopZ1a cysteine-protease activity, and independent of known resistance genes. We have previously shown that HopZ1a-triggered immunity is partially additive to that triggered by AvrRpt2. These partially additive effects could be caused by at least two mechanisms: their signalling pathways share a common element(s), or one effector interferes with the response triggered by the other.
  • Here, we investigate the molecular basis for the partially additive effect displayed by AvrRpt2- and HopZ1a-triggered immunities, by analysing competitive indices, hypersensitive response and symptom induction, PR-1 accumulation, expression of PR genes, and systemic acquired resistance (SAR) induction.
  • Partially additive effects between these defence responses require HopZ1a cysteine-protease activity, and also take place between HopZ1a and AvrRps4 or AvrRpm1-triggered responses. We establish that HopZ1a-triggered immunity is independent of salicylic acid (SA), EDS1, jasmonic acid (JA) and ethylene (ET)-dependent pathways, and show that HopZ1a suppresses the induction of PR-1 and PR-5 associated with P. syringae pv tomato (Pto)-triggered effector-triggered immunity (ETI)-like defences, AvrRpt2-triggered immunity, and Pto or Pto (avrRpt2) activation of SAR, and that suppression requires HopZ1a cysteine-protease activity.
  • Our results indicate that HopZ1a triggers an unusual resistance independent of known pathways and suppresses SA and EDS1-dependent resistance.

Introduction

Many Gram-negative pathogenic bacteria secrete proteins directly inside the host cell, using a type III secretion system (T3SS). These proteins, called type III effectors (T3Es), modulate diverse processes inside the host, including the suppression of defence responses triggered after pathogen recognition (Gohre & Robatzek, 2008). Plants have evolved specific mechanisms to recognize effector activities directly or indirectly, usually triggering a drastic response accompanied by programmed cell death, called the hypersensitive response (HR), which restricts pathogen growth (named ETI for effector-triggered immunity; Chisholm et al., 2006). Effectors triggering strong ETI have traditionally been known as avirulence factors (Mansfield, 2009). Pathogens have also evolved effector activities to suppress ETI, cell death and other HR-associated phenomena, allowing pathogen growth and disease development (Jones & Dangl, 2006). Effector recognition by the plant is mediated by the products of resistance genes (R genes). R gene-mediated defences are usually associated with the accumulation of salicylic acid (SA; Glazebrook, 2005), although SA-independent defence pathways, such as the one dependent on EDS1 (Enhanced Disease Susceptibility-1; Parker et al., 1996), also contribute to the HR (Bartsch et al., 2006). EDS1 and SA have recently been described to act redundantly to regulate R-gene-mediated signalling (Venugopal et al., 2009). Virulence activity has been demonstrated for several effectors in the absence of their respective R proteins (Kearney & Staskawicz, 1990; Ritter & Dangl, 1995; Jackson et al., 1999; Chen et al., 2000; Kim et al., 2005a; Sohn et al., 2009). Likewise, virulence activity has been reported for AvrRpm1 while simultaneously determining ETI, although, in this situation, its contribution to virulence was offset by its activation of plant defences (Kim et al., 2009).

Some T3Es may undergo diversification to avoid plant recognition (Ma et al., 2006). Some strains from Yersinia spp. and Pseudomonas syringae secrete effector proteins that are members of the YopJ/HopZ family of cysteine-proteases (Shao et al., 2002), which constitute a fine example of effector diversification in response to selective pressure exerted by plant defences (Ma et al., 2006). HopZ1a, from P. syringae pv syringae A2 strain, is the most similar version to that encoded by the ancestral allele, and triggers plant defences in different species, such as Arabidopsis thaliana (Arabidopsis), Nicotiana benthamiana or Glycine max (soybean) (Ma et al., 2006; Zhou et al., 2009). Other effectors from the HopZ family have also been shown to trigger plant defences in different hosts (Deng et al., 2003; Ma et al., 2006; Vinatzer et al., 2006; Zhou et al., 2009).

Several studies have recently analysed the basis of HopZ1a recognition by plant defence mechanisms (Lewis et al., 2008; Macho et al., 2009; Zhou et al., 2009). In Arabidopsis and soybean, cysteine-protease activity and myristoylation and subsequent localization to cellular membranes are both required for the defence response triggered by HopZ1a (Lewis et al., 2008; Zhou et al., 2009). HopZ1a-triggered defence in Arabidopsis is independent of the thoroughly characterized R proteins RPM1, RPS2, RPS4 or RPS5 (Lewis et al., 2008). A report from our team has shown HopZ1a-triggered defence response to be partially additive to that triggered by AvrRpt2 (Macho et al., 2009), a well-studied effector that cleaves the plant defence-related protein RIN4, triggering the HR through the activation of the resistance protein RPS2 (Axtell & Staskawicz, 2003; Mackey et al., 2003). AvrRpt2 and HopZ1a-triggered immunity could be partially additive as a result of at least two different general mechanisms: their signalling pathways functionally overlap, sharing a common element(s); or one of the effectors interferes with the defence responses triggered by the other. As an example of the latter mechanism, transgenic expression of the effector HopF2 has been recently shown to interfere with AvrRpt2-mediated cleavage of RIN4, thus compromising AvrRpt2-triggered ETI (Wilton et al., 2010). Although HopZ2, another member of the HopZ-effector family, has been shown to promote pathogen growth in Arabidopsis (Lewis et al., 2008), no virulence activity has been reported for HopZ1a to date.

In this work, we show that the cysteine-protease activity of HopZ1a is required to determine a partially additive effect on bacterial growth for HopZ1a- and AvrRpt2-triggered defence responses. Likewise, we show that the defence responses triggered by HopZ1a and AvrRps4, or HopZ1a and AvrRpm1, are also partially additive. We also establish that HopZ1a-triggered immunity is independent of SA- and EDS1-dependent pathways, and rule out the possibility that other defence-related hormones like jasmonic acid (JA) or ethylene (ET) may be involved. Furthermore, we show that HopZ1a suppresses both weak ETI-like defences triggered by P. syringae pv tomato DC3000 (Pto) and AvrRpt2-triggered immunity, and show that this suppression is dependent on the catalytic residue necessary for HopZ1a cysteine-protease activity. Finally, we demonstrate that HopZ1a also suppresses systemic acquired resistance (SAR) triggered by either virulent or avirulent bacteria. We propose interference with AvrRpt2, AvrRps4 and AvrRpm1-triggered immunity to be caused by suppression of ETI-associated defences resulting from HopZ1a virulence activity.

Materials and Methods

Bacterial strains and growth conditions

Bacteria were grown O/N (overnight) at 37°C for Escherichia coli DH5α (Hanahan, 1983), and either 48 h at 19°C (only for macroscopic HR assays) or O/N at 28°C, for P. syringae strains, in LB (Luria-Bertani) supplemented with kanamycin (50 μg ml−1 for E. coli DH5α; 15 μg ml−1 for P. syringae strains), gentamycin (10 μg ml−1), or cycloheximide (2 μg ml−1), as appropriate. Strains Pto DC3000 wt (Cuppels, 1986), Pto DC3000 ΔhrcC (referred as T3SS-deficient mutant in the text; Mudgett & Staskawicz, 1999) and P. syringae pv phaseolicola 1448a (Teverson, 1991) were used for plant inoculations. All the chemicals, antibiotics, oligonucleotides and reagents were purchased from Sigma, unless otherwise stated.

Plant material

Unless otherwise stated, A. thaliana (Arabidopsis) plants accession Columbia wild-type (Col-0), and mutant derivatives were grown in growth chambers with 8 h light : 16 h dark cycles at 21°C. The following mutants were used: rpm1-1 (Grant et al., 1995), rps2 (SALK_087581), rps4 (SAIL_519-B09), ndr1-1 (Century et al., 1995), eds1-2 (Col-0) (Parker et al., 1996), NahG (Lawton et al., 1995), jin1 and jin4 (Berger et al., 1996), ein2-5 (Alonso et al., 1999), 35S:ERF1 (Berrocal-Lobo et al., 2002), pad4-1 (Jirage et al., 1999), rar1-21 (Tornero et al., 2002), sgt1b-1 (Azevedo et al., 2002), win3-1 (Lee et al., 2007), fmo1-1 (Bartsch et al., 2006), nho1-1 (Lu et al., 2001), hsp90.2-1 (Hubert et al., 2003), sid2-1 (Wildermuth et al., 2001), and npr1-1 (Cao et al., 1994). The mentioned lines were provided by the Nottingham Arabidopsis Stock Centre (NASC; http://www.arabidopsis.info) or by generous colleagues. coi1-1 mutant seeds (F2; Xie et al., 1998) were grown in MS (Duchefa, Haarlem, the Netherlands) agar plates with 0.5% 10 mM methyl-jasmonate stock solution (MeJa; diluted in ethanol; Duchefa) for 15 d, and seedlings nonresponsive to MeJa (homozygous coi1 mutants) were transferred to soil substrate for subsequent experiments. Corresponding wild-type (wt) control plants were grown in MS agar plates with 0.5% ethanol before transference to soil substrate. Tomato plants (Solanum lycopersicum) cv MoneyMaker, were grown in growth chambers with 16 h light : 8 h dark cycles at 22–28°C.

Bacterial inoculations

Competitive index assays to measure growth attenuation were performed as described for Arabidopsis plants (Macho et al., 2007). Briefly, 4- to 5-wk-old plants were inoculated with a 5 × 104 colony-forming unit (cfu) ml−1 mixed bacterial suspension, containing equal cfu of wt and effector-expressing strains, using a blunt syringe. Serial dilutions of the inoculum were plated onto LB agar and LB agar with kanamycin to confirm dose and relative proportion between the strains, which should be close to one. At 2 or 4 d post-inoculation (dpi), three 10-mm-diameter leaf discs were homogenized by mechanical disruption into 1 ml of 10 mM MgCl2. Then, bacteria were enumerated by plating serial dilutions onto LB agar with cycloheximide, and LB agar with kanamycin and cycloheximide, to differentiate the strains within the mixed infection. Bacterial enumeration was carried out in the dilution displaying between 50 and 500 colonies per plate. Similar experimental conditions were used for single inoculations, inoculating with single instead of mixed bacterial suspensions.

For measuring SAR, plants were initially inoculated with either 10 mM MgCl2 (mock), Pto or Pto-expressing effectors at 5 × 105 cfu ml−1. After 2 d, secondary leaves were infiltrated with Pto at 5 × 104 cfu ml−1. Growth of Pto was measured 4 dpi in secondary leaves, as already described.

For macroscopic HR assays, fully expanded leaves of 4- to 5-wk-old plants were inoculated with a 5 × 107 cfu ml−1 bacterial suspension using a blunt syringe, and symptoms were documented at 20 or 24 h post-inoculation (hpi). Three independent experiments were carried out with each strain, with similar results. A minimum of 30 leaves was infiltrated per strain and plant genotype. The same dose was used in inoculations for RNA extractions, while for protein extractions the concentration was 106 cfu ml−1.

For symptom visualization, 3- to 4-wk-old plants were sprayed with a bacterial suspension containing 5 × 107 cfu ml−1 in 10 mM MgCl2 and 0.02% Silwet-L77 (Crompton Europe Ltd, Evesham, UK). Plants were then covered for 24 h to keep humidity high, and symptoms were documented at 4 and 6 dpi.

Competitive index assays

Competitive indices (CIs) correspond to the effector-expressing strain-to-wt output ratio divided by their input ratio, within a mixed infection (Macho et al., 2009). For cancelled-out index (COI) calculations, the output ratio between the strain expressing two effectors and the strain expressing one effector was divided by their input ratio (Beuzón & Holden, 2001; Beuzón et al., 2001). Competitive indices shown are the mean of three replicates showing typical results of at least two independent experiments. Error bars represent the standard error. Each CI was analysed using a homoscedastic and two-tailed Student’s t-test and the null hypothesis: mean index is not significantly different from 1.0, or from other CI value when otherwise specified (P-values < 0.05).

Plasmids

Plasmids generated in this work, as well as parent vectors, are listed in Table 1. DNA fragments for cloning were PCR-amplified using Pfu polymerase (Promega; Madison, WI, USA); the corresponding primers are detailed as Supporting Information (Table S1). The gene hopZ1aPsy was cloned from P. syringae pv syringae strain 7B40 (ICMP13516), and confirmed by sequencing to be 100% identical to the originally described effector from the P. syringae pv syringae A2 strain (Sundin et al., 2004). A boiled preparation of P. syringae pv syringae 7B40 cells was used as template PCR amplification. A fragment containing the coding sequence of the FLAG epitope followed by a stop codon was amplified from pAME9 and cloned into pAMEX to generate pAMEFLAG. The gene avrRps4, without the stop codon, was amplified from P. syringae pv phaseolicola strain 1448a and cloned into pAMEFLAG, generating a plasmid that expresses the AvrRps4 effector protein fused to a FLAG epitope. Genomic DNA of P. syringae pv phaseolicola strain 1448a was extracted using Jet Flex Extraction Kit (Genomed, Löhne, Germany), and used as template for PCR amplification. Plasmid pVSP61::avrRpm1 (Bisgrove et al., 1994) was used as template for PCR amplification of avrRpm1 fragments. The fragment, including the hopZ1a ORF (Open Reading Frame), was cloned with blunt ends into the EcoRV restriction site of pAMEX. pAME8 was digested with HindIII and blunt-ended with Klenow fragment (GE HealthCare, Little Chalfont, UK) to act as recipient of blunt-ended hopZ1a, avrRpm1 and avrRps4 fragments. A fragment containing the gentamycin resistance cassette was excised from pMGm with KpnI and cloned into the KpnI restriction site of pAME23 to generate pAME23Gm. Generation of the plasmid expressing the HopZ1a catalytic mutant is detailed in Methods S1. All effectors are expressed from the nptII promoter. All constructs were verified by sequencing.

Table 1.   Plasmids used in this study
NameParent vectorPromoterExpressed effectorsResistanceaReference
  1. aAmp, Km and Gm indicate resistance to ampicilin, kanamycin, and gentamycin, respectively.

pBBR1-MCS4pBBR1-MCSlacZNoneAmpKovach et al. (1995)
pMGmN/AaacC1NoneGmMurillo et al. (1994)
pAMEXpBBR1-MCS4nptIINoneAmp, KmMacho et al. (2009)
pAME9pAME5nptIINoneAmp, KmMacho et al. (2009)
pAME8pAME4nptIIAvrRpt2Amp, KmMacho et al. (2009)
pAME23pAME8nptIIHopZ1a-6His + AvrRpt2Amp, KmMacho et al. (2009)
pAME23GmpAME23nptIIHopZ1a-6His + AvrRpt2Amp, Km, GmThis work
pAMEFLAGpAMEXnptIINoneAmp, KmThis work
pAMEF6pAMEFLAGnptIIAvrRps4-FLAGAmp, KmThis work
pAME26pAMEF6nptIIHopZ1a-6His + AvrRps4-FLAGAmp, KmThis work
pAME27pAMEXnptIIHopZ1aC216AAmp, KmThis work
pAME28pAME8nptIIAvrRps4 + AvrRpt2Amp, KmThis work
pAME30pAMEXnptIIHopZ1aAmp, KmThis work
pAME33pAME8nptIIHopZ1a + AvrRpt2Amp, KmThis work
pAME34pAME8nptIIHopZ1aC216A + AvrRpt2Amp, KmThis work
pAME36pAME8nptIIAvrRpm1 + AvrRpt2Amp, KmThis work
pAME37pAME30nptIIHopZ1a + AvrRpm1Amp, KmThis work
pAME41pAMEXnptIIAvrRpm1Amp, KmThis work

Protein and RNA

Protein and RNA extractions, western blot and quantitative real-time PCR (qRT-PCR) assays were carried out using standard protocols, and are detailed in Methods S1. Serum against PR-1 protein, as described in Wang et al. (2005), was used for western blot hybridizations.

Results

Interference or functional overlap between HopZ1a- and AvrRpt2-triggered immunity in Arabidopsis requires HopZ1a cysteine protease activity

We previously showed that the defence response against HopZ1a in Arabidopsis is partially additive to that triggered by AvrRpt2 (Macho et al., 2009). We established this using CI in mixed infections of Pto DC3000 co-inoculated with Pto expressing either HopZ1a or AvrRpt2, or both. In this experimental setting, co-inoculated strains have been shown to grow as they would in individual infections (Macho et al., 2007). Since the CI is calculated as the test strain-to-wt output ratio divided by their input ratio, a CI significantly smaller than one indicates a growth defect for the strain being tested. HopZ1a- and AvrRpt2-triggered responses were found to be partially additive, since the CI of a strain expressing both effectors simultaneously was lower than the CIs of each of the strains expressing the effectors individually, but not as low as would be expected if their defence responses were completely independent (i.e. CI expected to be equal to the product of both CIs) (Fig. 1a left panel; Macho et al., 2009). Analysing further the molecular mechanisms behind this result, we have confirmed that the amount of hopZ1a and avrRpt2 transcripts accumulated when expressed individually is similar to the amounts accumulated when they are encoded within the polycistron from which both are expressed simultaneously (data not shown). Thus, we can rule out differences in expression as the reason for HopZ1a- and AvrRpt2-triggered responses not causing fully additive growth attenuation. We further confirmed these responses as partially additive by directly comparing growth attenuation determined by either HopZ1a or AvrRpt2 expression with that determined by the expression of both, using a previously published modification of the CI, the COI (Beuzón & Holden, 2001; Beuzón et al., 2001). We co-inoculated the strain expressing both effectors with either of the strains expressing the effectors individually, and calculated the COI as the ratio between both strains in the output divided by the ratio in the input. In this manner, it is possible to reliably quantify the reduction of growth determined by one of the effectors in the presence of the defence triggered by the other, by cancelling out the contribution of the effector that is expressed by both strains (Beuzón et al., 2001). COI directly measures the differences in growth between Pto expressing one of the effectors and Pto expressing both effectors. Since this comparison is carried out within the same plant, it allows for greater accuracy. The COI of Pto (avrRpt2 hopZ1a) in mixed infection with Pto (hopZ1a) was significantly different from one, but also significantly higher than the CI of Pto (avrRpt2) (Fig. 1a). This indicates that the defence response triggered by AvrRpt2 when coexpressed with HopZ1a is not as efficient in restricting bacterial growth as it is when AvrRpt2 is expressed alone, further showing that although these responses are additive, they are not fully additive. The same conclusion can be reached from the analysis of the reciprocal COI, in which Pto expressing both effectors was co-inoculated with Pto expressing HopZ1a (Fig. 1a). However, it could still be argued that the degree of attenuation of Pto (avrRpt2 hopZ1a) is below the threshold of detection of these assays, thus hindering identification of these responses as fully additive. Bacterial counts on individual infections 4  dpi showed a 104-fold increase for Pto, a 100-fold increase for Pto (hopZ1a), and a 50-fold increase for Pto (avrRpt2 hopZ1a), whereas nonpathogenic control strains (i.e. Pto T3SS-deficient mutant and P. syringae pv phaseolicola 1448a) showed considerably smaller increases in growth (Fig. 1b). Therefore, CI and COI assays would have allowed detection of a stronger attenuation caused by fully additive responses. Thus, the partially additive restriction of growth determined in Arabidopsis by coexpression of HopZ1a and AvrRpt2 is likely to be caused by one of these two general mechanisms: their signalling cascades functionally overlap by sharing a common element(s) or target(s), or one of the effectors interferes with the defence response triggered by the other. A straightforward mechanism for the latter alternative would be that the virulence activity of one of the effectors partially suppresses the defence response triggered by the other, in a manner similar to that described for HopF2 (Wilton et al., 2010). However, it cannot be formally ruled out that the defence response triggered by one of the effectors may somehow antagonize that triggered by the other.

Figure 1.

 HopZ1a and AvrRpt2-triggered responses display a partially additive effect that is dependent of HopZ1a cysteine-protease activity. (a) Graphical representation of competitive indices (CIs) and cancelled-out indices (COIs) resulting from mixed infections of Pseudomonas syringae pv tomato DC3000 wild-type (Pto) or derivatives expressing the indicated effectors. Inoculations were carried out in either wild-type (wt, left and central panels) or rps2 plants (right panel). Index values correspond to the mean of three samples and error bars represent standard error. Mean values marked with the same letter were not significantly different from each other as established by Student’s t-test (< 0.05). (b) Bacterial growth following single inoculations of the indicated strains in wild-type plants. Samples were taken immediately after inoculation, 0 d post-inoculation (dpi; light grey bars) or 4 dpi (dark grey bars). Values correspond to the mean of three samples and error bars represent standard error. Mean values marked with the same letter were not significantly different from each other as established by Student’s t-test (< 0.05). Experiments were repeated at least twice with similar results.

A mutation in a cysteine residue of HopZ1a that abolishes protease activity in vitro has already been shown to be necessary for HopZ1a-triggered immunity (Lewis et al., 2008). We generated a mutant gene encoding a HopZ1a protein carrying a substitution in this cysteine residue (HopZ1aC216A), and assayed its effect on bacterial growth when expressed alone or simultaneously with AvrRpt2. In keeping with the results of Lewis et al. (2008), HopZ1aC216A did not trigger immunity since its expression in Pto did not affect bacterial growth (CI not significantly different to one; Fig. 1a). A strain simultaneously expressing AvrRpt2 and HopZ1aC216A showed an attenuation of growth similar to that caused by the expression of AvrRpt2 alone (Fig. 1a). This result would be expected if the defence responses triggered by HopZ1a and AvrRpt2 shared a common element or were antagonistic, since in the absence of HopZ1a-triggered immunity, the AvrRpt2-triggered response would not be affected. However, this result would also be consistent with one of the effectors suppressing the defence response triggered by the other. If this was the case, we could rule out AvrRpt2 as the interfering effector, since expression of AvrRpt2 does not alter HopZ1a-triggered restriction of growth in rps2 plants (Fig. 1a). In this plant genotype, AvrRpt2 does not trigger resistance, but should still display virulence activity.

Interference or functional overlap also takes place between HopZ1a and AvrRps4- or AvrRpm1-triggered immunity

We used AvrRps4Pph1448a, similar to previously characterized AvrRps4Ppi151 (Hinsch & Staskawicz, 1996), and confirmed that it also triggers ETI in Arabidopsis, restricting bacterial growth in a RPS4-dependent manner when expressed from virulent bacteria (Fig. 2a). Pto expressing HopZ1a and AvrRps4 from a polycistron displayed an attenuation of growth close to 120-fold (Fig. 2a), significantly different from that determined by expression of AvrRps4 or HopZ1a alone, but far from the 1500-fold attenuation (equal to the product of the attenuations caused by these effectors) expected if the responses were completely independent. A similar result was obtained for Pto expressing HopZ1a and AvrRpm1 from a polycistron, which displayed a growth attenuation close to 500-fold (Fig. 2b), significantly lower than those determined by either effector expressed individually, but still far from the expected 104-fold if these responses were fully additive. Thus, as happens with AvrRpt2-triggered immunity, either HopZ1a also suppresses AvrRps4 and AvrRpm1 defence responses, or the defence response it triggers in Arabidopsis shares a signalling component with both these responses, or antagonizes them both. We further validated the potential for CI and COI assays to detect interactions between different ETI by analysing the responses of AvrRpm1 and AvrRpt2, or AvrRps4 and AvrRpt2 in a similar manner. AvrRpt2- and AvrRps4-triggered responses displayed a partially additive effect on bacterial growth (Fig. S1a), in keeping with their signalling cascades overlapping in EDS1 (Feys et al., 2001), whereas AvrRpt2 displayed full interference with AvrRpm1-triggered immunity (Fig. S1b), as previously described (Ritter & Dangl, 1995).

Figure 2.

 HopZ1a- and AvrRps4- or AvrRpm1-triggered responses display a partially additive effect. Graphical representation of competitive indices (CIs) resulting from mixed infections of Pseudomonas syringae pv tomato (Pto) with derivatives expressing the indicated effectors. Inoculations were carried out in wild-type (wt) RPS4 plants (a, left and central panels) or rps4 mutant plants (a, right panel), or wt RPM1 plants (b, left and central panels) or rpm1 mutant plants (b, right panel). Index values shown in (a) and (b) correspond to the mean of three samples and error bars represent the standard error. Mean values marked with the same letter were not significantly different from each other as established by Student’s t-test (< 0.05). Experiments were repeated twice with similar results.

HopZ1a-triggered immunity is independent of SA and other pathways leading to the HR

To test the possibility of the HopZ1a-triggered immune response sharing signalling components with those triggered by AvrRpt2, AvrRps4 or AvrRpm1, we analysed growth of Pto (hopZ1a) vs that of Pto, in plants carrying mutations in genes important for the R gene-mediated defences triggered by these three effectors. NDR1 (Century et al., 1995) is required for most coiled-coil (CC)-nucleotide binding site (NBS)-leucine rich repeat (LRR) types of R proteins (including RPS2; Aarts et al., 1998), EDS1 (Parker et al., 1996) is essential for most Toll-interleukin1-like (TIR)-NBS-LRR type R proteins (including RPS4, Aarts et al., 1998), and RAR1 (Muskett et al., 2002; Tornero et al., 2002) is a co-chaperone that contributes to the function of many R proteins (including RPS2, RPS4 and RPM1; Tornero et al., 2002). If one of these genes were important for HopZ1a-triggered immunity, growth of Pto (hopZ1a) would be closer to growth of Pto in mutant plants than in wt (i.e. CI closer to one). However, none of the CIs obtained in ndr1, eds1 or rar1 were closer to one (Fig. 3a). Furthermore, the CI of Pto (hopZ1a) in plants carrying the NahG transgene, encoding a salicylate hydroxylase that converts SA to catechol, which accumulate very low concentrations of SA and present enhanced susceptibility to infection with Pto (Delaney et al., 1994), was not significantly closer to one (Fig. 3a). In fact, the CI values in these mutants were even lower than the CI in wt plants. However, detailed examination of cfu values (Fig. S2) showed that growth of Pto is increased in these plants, in keeping with previous reports (Century et al., 1995; Feys et al., 2001), whereas growth of Pto (hopZ1a) is not different from its growth in wt type plants, supporting the idea that the HopZ1a-triggered response is independent of these elements. Similar results were obtained for eds1 and NahG plants when the assays were carried out at earlier stages of the infection process (Fig. 3b). Additionally, we also found that HopZ1a-triggered immunity is not compromised to any degree in plants carrying mutations in other components that contribute to SA biosynthesis, SA-downstream signalling, R-protein stability, SA-independent defence, or nonhost resistance (Table 2).

Figure 3.

 Effect of different defence mutants on the responses triggered by effectors HopZ1a, AvrRpt2, AvrRps4, and AvrRpm1. Graphical representation of competitive indices (CIs) resulting from mixed infections of Pseudomonas syringae pv tomato (Pto) with either Pto (hopZ1a) (a, b), Pto (avrRpt2) (c), Pto (avrRps4) (d) or Pto (avrRpm1) (e). Inoculations were carried out in the indicated genotypes, and samples were taken either at 2 (b) or 4 d post-inoculation (dpi; a, c–e). Index values correspond to the mean of three samples and error bars represent the standard error. Mean values marked with the same letter were not significantly different from each other as established by Student’s t-test (< 0.05). Experiments were repeated three times with similar results. Col-0, Columbia wild-type.

Table 2.   Competitive index (CI) values obtained in different plant genotypes co-inoculated with Pseudomonas syringae pv tomato (Pto) and Pto (hopZ1a)
Plant genotypeReferenceCISE
Col-0 0.0210.0012
pad4PAD4, Jirage et al. (1999)0.0170.0064
ndr1 pad40.0140.0095
ndr1 rar1 sgt1bSGT1, Azevedo et al. (2002)0.00680.0025
hsp90.2Hubert et al. (2003)0.00630.0017
win3Lee et al. (2007)0.00370.0004
fmo1Bartsch et al. (2006)0.00710.00035
nho1Lu et al. (2001)0.00490.0005
sid2Wildermuth et al. (2001)0.00460.0011
npr1Cao et al. (1994)0.00240.001
Solanum lycopersicum cv MoneyMaker
 Wild-type 0.890.17

When the CIs for Pto (avrRpt2) and Pto (avrRps4) were determined in ndr1, rar1, eds1 and NahG plants, results reflected the diminished ability of these plants to restrict growth of one or both the effector-expressing strains (Fig. 3c,d). In agreement with previous reports (Aarts et al., 1998), AvrRpt2-triggered immunity was abolished in ndr1 mutant plants and NahG plants (CIs not significantly different from one), while it was partially compromised in rar1 mutant plants (Fig. 3c). We also found a small, though statistically significant, reduction of AvrRpt2-triggered immunity in eds1 mutants (Fig. 3c). AvrRpt2-RPS2-mediated resistance has been reported as not requiring EDS1 for defence signalling (Aarts et al., 1998), although it has also been reported as weakly compromised in eds1 plants (Feys et al., 2001). CI assays display an increased sensitivity compared with standard growth assays (Macho et al., 2007) and have thus allowed us to establish a statistically significant contribution of EDS1 to AvrRpt2-triggered immunity that has been missed by other assays (Fig. 3c). In keeping with previous reports (Aarts et al., 1998), eds1 plants have completely lost their ability to restrict growth of Pto (avrRps4) (CI not significantly different from one), while, remarkably, resistance to these bacteria in NahG plants was only slightly compromised (Fig. 3d). AvrRps4-triggered immunity was significantly, though not completely, reduced in rar1 plants, and not affected in ndr1 plants (Fig. 3d). As expected, AvrRpm1-triggered immunity was not affected in either eds1 or NahG plants (Fig. 3e).

We also carried out spray inoculations to monitor virulence-associated symptoms. When sprayed over wt plants, Pto induced chlorosis by 4–6 dpi (Fig. 4a). Mutants in eds1 sprayed with Pto showed more severe symptoms, although not as dramatic as NahG plants at that time point (Fig. 4a). However, after spray inoculation with Pto (hopZ1a), none of the tested plant genotypes showed any virulence-associated symptoms (Fig. 4a), illustrating that resistance is not compromised in either eds1 or NahG plants. Pto (avrRpt2), sprayed as a control, triggered efficient resistance in wt plants, as indicated by the absence of disease symptoms, and this resistance was lost in NahG plants (Fig. 4a).

Figure 4.

 HopZ1a-triggered immunity is not dependent on salicylic acid (SA) or EDS1. (a) Virulence symptoms caused by spray-inoculated Pseudomonas syringae pv tomato (Pto), Pto (hopZ1a), or Pto (avrRpt2). Wild-type, NahG and eds1 Arabidopsis plants were sprayed with bacterial suspensions containing 5 × 107 cfu ml−1, and were photographed 6 d post-inoculation (dpi). Images are representative of 15 inoculated plants per strain and plant genotype. (b) Hypersensitive response (HR) to hand-infiltration with bacterial suspensions containing 5 × 107 cfu ml−1. Photographs were taken 24 hpi, and leaves that showed clear loss of turgor are labelled ‘HR’. Images are representative of > 30 inoculated leaves per strain and plant genotype. Experiments were repeated at least twice with similar results. Col-0, Columbia wild-type.

In agreement with previous reports (Lewis et al., 2008), when Pto (hopZ1a) was inoculated directly into the leaf at a high dose (5 × 107 cfu ml−1), a clear HR was detected by 20–24 hpi, which was absent in leaves infiltrated with Pto expressing HopZ1aC216A (Fig. 4b). In keeping with CI and spraying assays, the HR triggered by expression of HopZ1a was not abolished in NahG or eds1 plants. Control leaves inoculated with Pto or Pto (avrRpt2) produced the expected results (Fig. 4b).

These results indicate that the defence response triggered by HopZ1a is completely independent of SA-dependent pathways and does not share any of the tested signalling components with those triggered by AvrRpt2, AvrRps4 or AvrRpm1. Therefore, the partial additive effect on bacterial growth established for these responses is most likely the result of interference between the effectors, either by HopZ1a suppressing AvrRtp2-, AvrRps4- and AvrRpm1-triggered immunity, or HopZ1a-triggered defence being somehow antagonistic to the responses triggered against these three effectors.

HopZ1a-triggered immunity is independent of jasmonic acid and ethylene

Jasmonic acid and ET are plant hormones that play an essential role in plant immunity against different pathogens and insects. Although recent research has provided us with a more complex scenario (van Wees et al., 2000; De Vos et al., 2006; Mur et al., 2006; Truman et al., 2007), JA and ET-dependent pathways have been classically described as antagonistic to SA (Glazebrook, 2005). Thus, we sought to determine if JA or ET were involved in the plant defence against HopZ1a.

We compared growth of Pto (hopZ1a) vs Pto in plants impaired in JA responses, such as jin1 and jin4 mutants (Berger et al., 1996), JA perception, such as a coi1 mutant (Feys et al., 1994), or ET perception, such as an ein2 mutant (Guzman & Ecker, 1990), and transgenic plants overexpressing ERF1 (Ethylene Response Factor-1; Berrocal-Lobo et al., 2002), and wt plants. The CI values obtained from jin4, ein2 or 35S::ERF1 were not different from that obtained in wt plants, although values from jin1 and coi1 mutant plants were significantly closer to one (Fig. 5a). A CI closer to one indicates that growth of Pto (hopZ1a) is closer to that of Pto in these mutants than in wt plants. However, detailed examination of cfu values (Fig. S3) showed that CI values were closer to one because, in keeping with previous reports (Feys et al., 1994; Nickstadt et al., 2004), growth of Pto is reduced in these plants, and not because growth restriction of Pto (hopZ1a) is compromised, as would be expected if these genes were involved in HopZ1a-triggered immunity. Bacterial enumerations and CI calculations in these plants were also carried out at earlier stages of infection (2 dpi), with equivalent results (data not shown). A similar result was observed when growth of Pto (avrRpt2) was compared with that of Pto in these mutants (Figs S4, S5).

Figure 5.

 HopZ1a-triggered immunity is not dependent on jasmonic acid (JA) or ethylene (ET). (a) Graphical representation of competitive indices (CIs) resulting from mixed infections of Pseudomonas syringae pv tomato (Pto) together with Pto (hopZ1a). Inoculations were carried out in the indicated genotypes, and samples were taken 4 d post-inoculation (dpi). Index values correspond to the mean of three samples and error bars represent the standard error. Mean values marked with the same letter were not significantly different from each other as established by Student’s t-test (< 0.05). (b) Hypersensitive response (HR) against Pto (hopZ1a). Arabidopsis plants with the indicated genotypes were infiltrated with a bacterial suspension containing 5 × 107 cfu ml−1. Photographs were taken 24 h post-inoculation (hpi), and leaves that showed clear necrosis are labelled ‘HR’. coi1 plants are shown with their wild-type control, as these plants were grown in different conditions (see the Materials and Methods section). Images are representative of > 30 inoculated leaves per strain and plant genotype. (c) Virulence symptoms caused by spray inoculation with bacterial suspensions containing 5 × 107 cfu ml−1 of Pto (hopZ1a). Plants were photographed 6 dpi. Images are representative of 15 inoculated plants per strain and plant genotype. Experiments were repeated at least twice with similar results. Col-0, Columbia wild-type.

We also used hand-infiltration at high doses of inoculum to test for a contribution of JA or ET to the macroscopic HR triggered by Pto (hopZ1a) by 20–24 hpi. Supporting the results obtained using growth assays, none of the plant genotypes altered HopZ1a-triggered HR (Fig. 5b). These results were further confirmed by spray inoculation with Pto (hopZ1a) and Pto, of jin1 and wt plants, after which (4–6 dpi) no virulence-associated chlorosis could be observed in either plant background, supporting the theory that HopZ1a-triggered immunity is not compromised in jin1 (Fig. 5c). Control experiments carried out for both assays with Pto (avrRpt2) produced the expected results and are shown in Fig. S4. Altogether, our results indicate that HopZ1a-triggered immunity does not depend on JA or ET signalling pathways.

The fact that HopZ1a-mediated restriction of growth is not compromised in any of the mutants analysed raised the formal caveat of its having a general toxic effect for bacterial growth within the plant, unrelated to ETI. However, growth of Pto (hopZ1a) in S. lycopersicum (tomato) plants, where this strain does not trigger an HR (data not shown), was not statistically different from growth of Pto (Table 2).

HopZ1a suppresses AvrRpt2-triggered defence responses

Having ruled out common elements between the signalling cascades, or antagonistic defence pathways, as the reasons behind the partially additive effect displayed by HopZ1a and AvrRpt2-, AvrRps4- or AvrRpm1-triggered immunities, we tested the alternative hypothesis of HopZ1a suppressing ETI. We first analysed accumulation of PR-1 in the presence or absence of HopZ1a in different infections. Time points for the analysis were established after a time course experiment, the results of which are included as Fig. S6. Pto triggered a reproducible PR-1 accumulation by 48 hpi, which was not observed in a T3SS-deficient mutant derivative (Fig. 6a). However, PR-1 accumulation by that time point was clearly reduced when leaves were inoculated with Pto (hopZ1a) (Fig. 6a). When Pto expressed the HopZ1a mutant protein lacking the cysteine-protease activity (HopZ1aC216a), the amount of accumulated PR-1 was not different from that observed following inoculation with Pto (Fig. 6a). Thus, HopZ1a suppresses Pto-triggered PR-1 accumulation and this suppression requires its cysteine-protease activity. Suppression of Pto-triggered PR-1 accumulation has also been reported for AvrRpt2, although this activity has only been detected in an rps2 mutant (Chen et al., 2004). We confirmed AvrRpt2-suppression of Pto-triggered PR-1 accumulation in rps2 plants using our experimental settings (Fig. S7).

Figure 6.

 HopZ1a suppresses effector-triggered immunity. (a) Western blot showing PR-1 accumulation in wt plants inoculated with Pseudomonas syringae pv tomato (Pto) T3SS− or Pto expressing the indicated effectors. Ten micrograms of total protein were loaded per sample, and Coomassie staining is shown as loading control. Results shown are representative of three independent replicates. (b–d) Quantitative-real time-PCR (qRT-PCR) analysis of PR-1, PR-5 and PDF1.2 gene transcripts, monitored 8 h after inoculation of wild-type Arabidopsis plants with Pto, Pto (avrRpt2) or Pto (hopZ1a). Bars represent the mean induction of gene transcripts normalized to the housekeeping actin gene for three replicates. Error bars represent the standard error. Relative expression values were divided by the expression measured in an untreated control to obtain fold increase values. The results shown are representative of two independent experiments.

In plants carrying the RPS2 gene, AvrRpt2 triggers earlier accumulation of PR-1 at 24 hpi. When hopZ1a was expressed simultaneously with avrRpt2, PR-1 accumulation was clearly reduced (Fig. 6a), indicating that HopZ1a is also capable of suppressing AvrRpt2-triggered PR-1 accumulation, thus supporting the theory that it interferes with AvrRpt2-triggered ETI. The quantitative data corresponding to these analyses are shown in Fig. S8.

To test if HopZ1a-mediated suppression of PR-1 accumulation was carried out at a transcriptional level, we performed qRT-PCR in plants inoculated with Pto, Pto (avrRpt2), or Pto (hopZ1a). As expected, Pto caused an increase of PR-1 transcripts of > 20-fold compared with basal expression levels, while Pto (avrRpt2) caused a > 200-fold increase (Fig. 6b). However, the amount of PR-1 transcripts detected in leaves infected with Pto (hopZ1a) was clearly reduced compared with those infected with Pto (Fig. 6b), a result consistent with suppression of PR-1 accumulation by HopZ1a (Fig. 6a) taking place at the transcriptional level.

We also analysed transcription of other defence-related molecular markers, carrying out qRT-PCR to detect transcript abundances of PR-5 and PDF1.2. While PR-5 transcription is, as with PR-1, usually associated with defence against biotrophic pathogens (including SA-dependent pathways) and SAR induction (Ryals et al., 1996), PDF1.2 transcription is a marker for ET and JA defence responses, usually considered antagonistic to SA-dependent pathways (Epple et al., 1997). Transcription of PR-5 followed a similar pattern to that observed for PR-1 (Fig. 6c), thus further supporting HopZ1a activity of suppression of Pto-triggered defences. Transcription of PDF1.2 showed an opposite pattern to that observed for PR-1 and PR-5 (Fig. 6d), with lower expression in leaves infected with Pto (avrRpt2) than in those infected with Pto, and higher than both in leaves infected with Pto (hopZ1a). Since we have established that HopZ1a-triggered immunity does not require the ET or JA pathways (Fig. 5), induction of PDF1.2 expression could be either the result of HopZ1a suppression of the antagonistic SA pathway, or a sign that HopZ1a directs activation of the ET or JA pathways.

HopZ1a suppresses SAR

Both virulent and avirulent bacteria trigger SAR (Cameron et al., 1994). To determine if, in addition to suppressing local resistance, HopZ1a could suppress activation of SAR by Pto or Pto (avrRpt2), we treated primary leaves by infiltrating them with 10 mM MgCl2 (mock), Pto or Pto (avrRpt2), and different strains either expressing or not expressing HopZ1a or its mutant derivative. Two days after infiltration in primary leaves, secondary leaves were inoculated with Pto, and growth was monitored 4 dpi. Fig. 7a) shows that pre-inoculation with Pto or Pto (avrRpt2) triggered a clear SAR, as seen by a restriction of bacterial growth in secondary leaves. However, when primary leaves were inoculated with Pto (hopZ1a) or Pto (avrRpt2 hopZ1a), bacterial growth in secondary leaves was not restricted (Fig. 7a), indicating that HopZ1a can suppress the SAR induced by both Pto and Pto (avrRpt2). When HopZ1aC216A was expressed instead of HopZ1a, bacterial growth in secondary leaves was restricted (Fig. 7a), indicating that suppression of SAR by HopZ1a requires its cysteine-protease activity.

Figure 7.

 HopZ1a suppresses systemic-acquired resistance (SAR). (a) Growth of Pseudomonas syringae pv tomato (Pto) in secondary leaves of 2 d after preinoculation of primary leaves. Primary leaves of Arabidopsis accession Columbia wild-type (Col-0) plants were infiltrated with 10 mM MgCl2 (mock), or bacterial suspensions containing 5 × 105 cfu ml−1 of Pto, or Pto expressing the indicated effectors. After 2 d, secondary leaves were infiltrated with 5 × 104 cfu ml−1 of Pto, and growth was determined 4 d post-inoculation (dpi). Values correspond to the mean of three samples and error bars represent standard error. Asterisks indicate results that are significantly different from growth in plants preinoculated with 10 mM MgCl2 (mock) (< 0.05). Experiments were repeated with similar results. (b) Western blot showing PR-1 accumulation in secondary leaves of plants in which primary leaves were preinoculated 2 d before with either Pto or Pto expressing the indicated effectors. Ten micrograms of total protein were loaded per sample, and Coomassie staining is shown as loading control.

To determine if HopZ1a suppression of SAR correlated with suppression of PR-1 accumulation in systemic tissue, we carried out western blot analysis for immuno-detection of PR-1 on secondary leaves 2 d after inoculating primary leaves with 10 mM MgCl2 (mock), Pto, Pto (avrRpt2), Pto (hopZ1a) or Pto (avrRpt2 hopZ1a), as well as with Pto and Pto (avrRpt2) expressing HopZ1aC216A. In keeping with the SAR detected in growth assays, inoculation of primary leaves with Pto or Pto (avrRpt2) determined the accumulation of PR1 in secondary leaves (Fig. 7b). However, expression of HopZ1a in either Pto or Pto (avrRpt2) determined a complete absence of PR-1 in secondary leaves (Fig. 7b), thus indicating that HopZ1a suppression of SAR correlates with its suppression of PR-1 accumulation in systemic tissues. This suppression is also dependent on HopZ1a cysteine-protease activity since expression of HopZ1aC216A allowed accumulation of PR-1 in secondary leaves, even if the amount accumulated in Pto expressing the mutant was not as high as that accumulated after inoculation with Pto.

Discussion

In this work, we used CI assays to analyse and quantify the additive effects that different effector-triggered responses have on restricting bacterial growth, establishing that HopZ1a interferes with AvrRpt2-, AvrRps4- and AvrRpm1-triggered immunities. Using the appropriate experimental settings, in which co-inoculated strains do not interfere with each other, experimental errors are minimized in CI assays, allowing a very accurate direct comparison between growths of the strains that determines an increase in the sensitivity in comparison with standard growth assays (Macho et al., 2007). CIs have been a key asset in allowing detection of the functional interactions between HopZ1a and different effector-triggered immunities. In addition, CIs and other assays have allowed us to rule out partially additive effects between these effector-triggered responses being caused by functional overlap between their signalling cascades or by dependency on antagonistic pathways, establishing HopZ1a-triggered immunity to be independent of SA-, JA- or ET-dependent pathways, as well as EDS1-mediated SA-independent pathways.

Defence against HopZ1a is independent of SA and other pathways leading to the HR

Establishing differences in how a given plant genotype affects bacterial growth restriction associated with a particular effector-triggered response can be difficult, particularly since the effect of the plant genotype on growth of the wt bacteria needs to be taken into account, and many defence mutants do have an impact on wt bacterial growth (Feys et al., 2001). The independence of HopZ1a-triggered immunity from both SA-dependent and partially SA-independent (mediated by EDS1; Bartsch et al., 2006; Venugopal et al., 2009) defence pathways established using CI assays (Fig. 3a,b) is fully supported by the results obtained using HR and disease symptom assays (Fig. 4a, Table 2).

EDS1 and NDR1 operate essential signalling pathways in different R gene-dependent defences (Aarts et al., 1998). Our results indicate that recognition of HopZ1a relies on different signalling elements not yet described or associated with R gene-dependent resistance (Figs 3, 4, and Table 2). Our results also show that HopZ1a detection does not depend on RAR1, SGT1 or HSP90 (Table 2), contrary to what has been described for previously characterized R genes (reviewed in Shirasu, 2009). Even defences triggered by AvrRpm1, which do not require EDS1 or SA to restrict bacterial growth (Fig. 3e), have been described to depend on RAR1, NDR1 and HSP90 (Tornero et al., 2002; Hubert et al., 2003).

Our experiments have also allowed quantification of the previously described implications of different defence genes in AvrRpt2- and AvrRps4-triggered responses, including a small but significant implication of EDS1 in AvrRpt2-triggered immunity. Previous studies have reported EDS1 as either weakly contributing (Feys et al., 2001) or not contributing (Aarts et al., 1998; Venugopal et al., 2009) to AvrRpt2-triggered immunity. The slight increase in Pto (avrRpt2) growth in eds1 plants may be hard to establish and can easily go undetected by standard growth assays. Interestingly, Venugopal et al. (2009) recently showed that EDS1 protein acts redundantly with SA signalling in contributing to AvrRpt2-triggered resistance and other CC-NBS-LRR R protein-mediated defence signalling. However, they could only detect an EDS1 contribution when combining eds1 and sid2 mutations. Since we can indeed detect the impact of eds1 in AvrRpt2-triggered resistance, but not in HopZ1a-triggered resistance, a similar situation is unlikely to take place in this case.

HopZ1a interferes with ETI

We show that HopZ1a suppresses Pto-triggered induction of PR-1. Since PR-1 accumulation is not observed in leaves inoculated with a T3SS-mutant derivative, this accumulation could be taking place through activation of a weak ETI-component of basal defences resulting from inoculation with a virulent pathogen like Pto (Gohre & Robatzek, 2008). Pathogen-associated molecular patterns (PAMPs) are very conserved molecules that trigger immunity against nonhost pathogens, upon recognition by pattern-recognition receptor proteins (PRR proteins; Nurnberger et al., 2004; Zipfel, 2008). Many T3SS effectors have been reported to suppress PAMP-Triggered Immunity (PTI)-associated responses (Espinosa & Alfano, 2004; Abramovitch et al., 2006); however, suppression of ETI as a mechanism of effector-mediated promotion of virulence seems to be an emerging theme for T3SS effectors (Jones & Dangl, 2006; Guo et al., 2009). We show that HopZ1a can suppress the following: ETI determined by different effectors through different and mostly independent signalling cascades; weak ETI triggered by Pto; and SAR. However, we cannot rule out that HopZ1a could also be suppressing Pto-triggered PTI, as is the case for other effectors (Guo et al., 2009). This suggests either that HopZ1a can act on different plant targets, thus affecting different signalling pathways, or that its virulence target is a common element acting either upstream or downstream of those typically defined as signalling from the R proteins.

Effector virulence activity is usually undetectable within the context of their R-mediated recognition in the local site of infection. AvrRpm1 triggers plant defences in the absence of its cognate R-protein, RPM1, through RPS2-mediated recognition (Kim et al., 2009), but has been proposed to carry out its virulence activity in this context, although its impact on bacterial growth may be cancelled out by RPS2-mediated restriction of growth (Kim et al., 2009). Similarly, a HopZ1a-mediated promotion of growth that would reveal its virulence activity is not detected in the context of HopZ1a-triggered immunity, which importantly restricts bacterial growth; however it is indirectly shown by its interference with AvRpt2-, AvrRps4- and AvrRpm1-triggered immunities in bacterial growth assays, and by its suppression of SAR (discussed later in this section). This contribution to virulence in the context of HopZ1a-triggered immunity is further supported by the results obtained when analysing common molecular markers for defence responses. However, in the local site of infection, this is only possible because of the fact that HopZ1a-triggered immunity represents a rather unusual class of ETI that does not dependent on the cognate ETI signalling pathways. Plant backgrounds lacking recognition of HopZ1a may show a direct contribution to bacterial growth if, as observed for other effectors (e.g. AvrRpt2, AvrRps4 or AvrRpm1; Chen et al., 2000; Kim et al., 2005b; Sohn et al., 2009), the virulence target(s) of HopZ1a was different from its R protein. Our results are consistent with two models. While HopZ1a could actively deplete the ETI (exemplified by PR1 expression; Fig. 6), it could also act by promoting the JA pathway (exemplified by PDF1.2 expression; Fig. 6). However, the former model is more likely, since we did not detect any impact on growth of Pto (hopZ1a) in any of the JA or ET mutants tested (Fig. S3). As with several P. syringae effectors, HopZ1a is myristoylated and targeted to the plant membrane (Lewis et al., 2008). The plasma membrane is a common location for defence-related proteins (e.g. Boyes et al., 1998; Mackey et al., 2002; Axtell & Staskawicz, 2003; Coppinger et al., 2004), and thus HopZ1a target(s) could be defence-related proteins associated with the membrane. A possible candidate to be targeted by HopZ1a is the membrane-associated protein RIN4, which is also targeted by other effectors (Mackey et al., 2002; Axtell & Staskawicz, 2003; Wilton et al., 2010). Interestingly, one of these effectors, HopF2, interferes with AvrRpt2-mediated cleavage of RIN4, abolishing AvrRpt2-triggered immunity (Wilton et al., 2010). However, RIN4 could not be HopZ1a’s main or only target, since Pto-triggered induction of PR-1 has been shown to be independent on RIN4 (Lim & Kunkel, 2004), and AvrRps4-triggerd immunity has not been shown to require RIN4. Furthermore, HopZ1a interference with AvrRpt2- or AvrRpm1-triggered responses is only partial, whereas these responses fully depend on RIN4. In any event, RIN4 could not be the target guarded by the R-protein responsible for HopZ1a-triggered immunity, since this response does not require RIN4 (Lewis et al., 2008).

The degree of resistance triggered by HopZ1a is similar to that triggered by AvrRpt2 (Fig. 1a). However, their respective effects on SAR are strikingly different (Fig. 7). HopZ1a is able – even in a wt and therefore resistant background – to produce a strong virulence effect in a systemic leaf, an effect not accomplished by either of the control strains (including Pto (avrRpt2) and Pto (hopZ1aC216A)).

In other words, the virulence effect of HopZ1a could have evolved to affect not only the local site of infection but also the signal that protects systemic leaves of the already infected plant. While Pseudomonas spp. do not travel inside the plant, a usual mechanism of infection involves aerosols produced by the splash of raindrops (Goto, 1992). Therefore, the presence of HopZ1a in Pseudomonas spp. could represent an evolutionary advantage for the dispersion of the pathogen.

Acknowledgements

We are grateful to T. Duarte for technical assistance. We also wish to thank A. Dobón, T. Rosas-Díaz and R. Lozano-Durán for their help with some of the experiments. The work was supported by project grants and from the Ministerio de Educación y Ciencia (BIO2006-00673), and Ministerio de Ciencia e Innovación (BIO2009-11516) (Spain) to C.R.B. The work was co-funded by Fondos Europeos de Desarrollo Regional (FEDER).

Ancillary

Advertisement