Layered basal defenses underlie non-host resistance of Arabidopsis to Pseudomonas syringae pv. phaseolicola

Authors

  • Jong Hyun Ham,

    1. Department of Horticulture and Crop Science, The Ohio State University,
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Min Gab Kim,

    1. Department of Horticulture and Crop Science, The Ohio State University,
    2. Department of Plant Cellular and Molecular Biology, The Ohio State University,
    3. Program in Plant Molecular Biology and Biotechnology, Program in Molecular Cellular and Developmental Biology, Rm. 306C Kottman Hall, The Ohio State University, Columbus, OH 43210, USA
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Sang Yeol Lee,

    1. Program in Plant Molecular Biology and Biotechnology, Program in Molecular Cellular and Developmental Biology, Rm. 306C Kottman Hall, The Ohio State University, Columbus, OH 43210, USA
    Search for more papers by this author
  • David Mackey

    Corresponding author
    1. Department of Horticulture and Crop Science, The Ohio State University,
    2. Department of Plant Cellular and Molecular Biology, The Ohio State University,
    3. Environmental Biotechnology National Core Research Center, Gyeongsang National University, Jinju 660-701, Korea, and
      (fax +614 292 7162; e-mail mackey.86@osu.edu).
    Search for more papers by this author

(fax +614 292 7162; e-mail mackey.86@osu.edu).

Summary

Arabidopsis is a non-host for Pseudomonas syringae pv. phaseolicola NPS3121 (Pph), a bacterial pathogen of bean. Pph does not induce a hypersensitive response in Arabidopsis. Here we show that Arabidopsis instead resists Pph with multi-layered basal defense. Our approach was: (i) to identify defense readouts induced by Pph; (ii) to determine whether mutations in known Arabidopsis defense genes disrupt Pph-induced defense signaling; (iii) to determine whether heterologous type III effectors from pathogens of Arabidopsis suppress Pph-induced defense signaling, and (iv) to ascertain how basal defenses contribute to resistance against Pph by individually or multiply disrupting defense signaling pathways with mutations and heterologous type III effectors. We demonstrate that Pph elicits a minimum of three basal defense-signaling pathways in Arabidopsis. These pathways have unique readouts, including PR-1 protein accumulation and morphologically distinct types of callose deposition. Further, they require distinct defense genes, including PMR4, RAR1, SID2, NPR1, and PAD4. Finally, they are suppressed differentially by heterologous type III effectors, including AvrRpm1 and HopM1. Pph growth is enhanced only when multiple defense pathways are disrupted. For example, mutation of NPR1 or SID2 combined with the action of AvrRpm1 and HopM1 renders Arabidopsis highly susceptible to Pph. Thus, non-host resistance of Arabidopsis to Pph is based on multiple, individually effective layers of basal defense.

Introduction

Plants have an innate immune system that copes with persistent challenges from a diverse array of parasites (Jones and Dangl, 2006). Plants are resistant to most potentially pathogenic microbes and, consequently, disease is the exception rather than the rule (Thordal-Christensen, 2003). Resistance often results because the plant recognizes a threatening organism and activates defense signaling pathways that produce an effective defense response (Jones and Takemoto, 2004). Conversely, susceptibility often results when a pathogen has the capacity to colonize the host and suppress its defenses (Jones and Dangl, 2006).

Recognition of pathogen-associated elicitors by plants is a prerequisite for the induction of defense responses. Microbe-associated molecular patterns (MAMPs) are found within proteinaceous or non-proteinaceous molecules that often provide critical functions for pathogen viability (Zipfel and Felix, 2005). Such molecules are evolutionarily stable targets suitable for direct defense perception. Examples of molecules containing MAMPs include flagellin, lipopolysaccharides, and elongation factor Tu. Gram-negative plant pathogenic bacteria secrete type III effectors directly into host cells via the type III secretion system (TTSS). Although intended to promote virulence, these effectors also can be recognized by defense receptors. Type III effectors are sometimes recognized indirectly via induced molecular patterns in host molecules (Mackey and McFall, 2006). Thus, plants can directly and indirectly recognize a variety of pathogen-associated molecules and induce defenses.

Pseudomonas syringae pv. tomato strain DC3000 (Pto) translocates at least 33, and perhaps as many as 50, type III effectors (Chang et al., 2005; Lindeberg et al., 2006), which are collectively required for pathogenicity on susceptible hosts (Collmer et al., 2000). One explanation for the low virulence of TTSS-deficient bacteria is that they fail to suppress MAMP-induced defense responses (Chisholm et al., 2006). Indeed, one function of type III effectors is suppression of MAMP signaling (Hauck et al., 2003; He et al., 2006; Kim et al., 2005b; de Torres et al., 2006). Additional functions of effectors include suppression of a variety of host defenses. Some prevent the hypersensitive response (HR), either by suppressing the generalized programmed cell death machinery (Abramovitch et al., 2003; Bretz et al., 2003; Espinosa et al., 2003; Jamir et al., 2004) or by suppressing specific defense signaling pathways (Kim et al., 2005a; Mackey et al., 2003; Reuber and Ausubel, 1996; Ritter and Dangl, 1996; Tsiamis et al., 2000). Other effectors suppress basal defenses, including modification of the plant cell wall (DebRoy et al., 2004; Hauck et al., 2003; Keshavarzi et al., 2004; Kim et al., 2005b). Counter to the defense-suppressing activities of type III effectors, plants have evolved R-proteins that detect type III effectors and thus promote defenses. Considering that plants have numerous R-proteins and phytopathogenic bacteria deliver large numbers of effectors, the complexity of the overall interaction becomes immediately apparent.

AvrRpm1 and HopM1 are type III effectors with the capacity to suppress callose deposition, indicative of cell-wall-associated defense (DebRoy et al., 2004; Kim et al., 2005b). AvrRpm1, an effector originally identified in strains of P. syringae pv. maculicola (Pma), is required for full virulence of Pma on Arabidopsis (Ritter and Dangl, 1995; Rohmer et al., 2003). HopM1 is part of the conserved effector locus (CEL) of Pto, a region of DNA that is required for full virulence of Pto on tomato and Arabidopsis (Alfano et al., 2000; DebRoy et al., 2004). HopM1 suppresses cell-wall-associated defenses by inducing degradation of AtMIN7, a guanine exchange factor important for vesicle trafficking to the cell surface (Nomura et al., 2006).

Plant pathogens have often co-evolved with a narrow host range. The ability to suppress defenses in host plants contributes to compatibility. When a pathogen is non-virulent on all varieties of a species of plant, the interaction is referred to as non-host (Thordal-Christensen, 2003). The lack of co-evolution in non-host interactions produces an unadapted pathogen that often induces multiple host defense responses. Considering these responses as independent ‘layers’ of defense provides an explanation for the typical durability of non-host resistance relative to cultivar-level resistance against a co-evolved pathogen. This point was elegantly illustrated for non-host resistance of Arabidopsis to unadapted powdery mildew pathogens (Lipka et al., 2005; Stein et al., 2006). Resistance depends on at least two layers of host defense. The first, which depends on the PEN genes, prevents fungal penetration at the site of spore germination. The second, which requires the EDS1–SAG101–PAD4 signaling complex (Wiermer et al., 2005), mediates resistance to fungi that have successfully penetrated. When both layers of defense are inactive, more fungal development occurs than when either layer alone is disrupted. Thus, each layer is independently effective, and together they contribute to full resistance of Arabidopsis to non-host powdery mildew pathogens.

Non-host resistance against Gram-negative bacteria has been divided into two classes, based on the type of response induced in the host (Mysore and Ryu, 2004). Type 2 non-host resistance is associated with the HR, a widely studied response that includes the rapid death of host cells exposed to the pathogen. The less well studied, but more common, type 1 non-host resistance occurs in the absence of a HR. We hypothesized that basal defenses occurring in the absence of a HR can underlie type 1 non-host resistance.

We sought to determine the contribution of basal defenses to resistance against Gram-negative phytopathogenic bacteria. For several reasons, we opted to study the resistance of Arabidopsis to P. syringae pv. phaseolicola strain NPS3121 (Pph). Firstly, we wished to study basal defenses in the absence of the HR, which can obscure other defense responses. Pph is a model pathogen that causes halo blight in bean (Lindgren et al., 1986). Arabidopsis demonstrates type 1 non-host resistance against Pph. Secondly, we sought an interaction in which resistance does not result from the lack of a niche, and Arabidopsis is a susceptible host for several other species of P. syringae. Thirdly, we wanted to study heterologous type III effectors from species of P. syringae that are pathogenic on Arabidopsis. When expressed in Pph, such effectors are delivered by the TTSS of Pph, permitting tests of their biological activities. Fourthly, the PphArabidopsis interaction has been a model system for several studies of non-host resistance (He et al., 2006; Lu et al., 2001; Tao et al., 2003; van Wees and Glazebrook, 2003).

We sought to determine whether disrupting basal defenses would increase the susceptibility of Arabidopsis to Pph. We identified multiple Pph-induced defense responses. Pph (TTSS–), a mutant strain lacking the TTSS, induces small callose deposits and weak accumulation of PR-1. The responses to Pph (TTSS–) are similar to those induced by purified MAMPs. Wild-type Pph, with a functional TTSS, induces morphologically distinct big callose deposits and strong accumulation of PR-1, in addition to small callose deposits. Using Arabidopsis plants with mutations in known defense genes (PMR4, RAR1, SID2, NPR1, and PAD4), we define a minimum of three defense signaling pathways activated by Pph. We studied the effect of heterologous type III effectors, from bacteria pathogenic to Arabidopsis, on Pph-induced defenses. The Arabidopsis pathogen Pto, when co-inoculated with Pph, dominantly suppresses defense responses induced by Pph and promotes the growth of Pph. These activities require Pto to have a functional TTSS, indicating the involvement of type III effectors. Individual heterologous effectors also suppress Pph-induced defenses. AvrRpm1, expressed from a transgene in the plant, suppresses both types of callose deposition. HopM1, expressed from a plasmid in Pph, suppresses both types of callose deposition and additionally suppresses the induction of PR-1. By combining the expression of AvrRpm1 and/or HopM1 with mutations in host defense genes we disrupt multiple pathways leading to basal defenses and enhance the susceptibility of Arabidopsis to Pph by 3–4 logs. Thus, resistance of Arabidopsis to Pph depends on multiple layers of basal defense.

Results

Establishing readouts for Pph-induced basal defenses in Arabidopsis

Because non-host resistance is so common, effective, and durable, we hypothesized that it probably involves activation of multiple, distinct defense pathways. In turn, we reasoned that virulent plant pathogens must negate the function of multiple distinct defense pathways to be successful. To test these ideas we had to first identify defense pathways activated by a non-host pathogen that are not activated by a related, virulent pathogen. We have compared responses of Arabidopsis to Pph and Pto (Figure 1) and have identified three distinct defense readouts specifically activated by Pph.

Figure 1.

  Active defense responses of Col-0 wild-type Arabidopsis to Pseudomonas syringae pv. phaseolicola NPS3121 (Pph).
(a) PR-1 accumulation patterns at 48 h after infiltration. PR-1 was detected by Western blot analysis using anti-PR-1 antibodies.
(b) Callose deposition patterns induced by different bacteria. Deposits of callose greater than 20 μm in diameter are classified as big, those less than 20 μm in diameter are classified as small. The average and SE of the number of callose deposits per mm2 are from five experiments for Pto and Pph and three experiments for Pto (TTSS–) and Pph (TTSS–) (TTSS, type III secretion system). Scale bar indicates 0.2 mm.
(c) Graph of the frequency versus size distribution of callose deposits induced by Pph and Pph (TTSS–).
(d) Distributions of callose (top) and dead cells (bottom) in the same leaf area following challenge with Pph and co-staining with aniline blue and trypan blue. Scale bars indicate 0.2 mm.
(e) Magnified view of a big callose deposit (top) that coincides with the borders of an individual mesophyll cell (bottom). Sample stained as in (d). Scale bars indicate 0.1 mm.

The first response we measured was accumulation of PR-1 protein, a classic marker of salicylic acid (SA)-dependent plant defense (Figure 1a). We found that Pph strongly induced accumulation of PR-1 while Pto did not. To test the contribution of type III effectors and MAMPs, we compared induction of PR-1 by Pph and a Pph mutant deficient in type III secretion. Pph (TTSS–) induced significantly less PR-1 than wild-type Pph, but did induce a low level similar to that induced by Pto (TTSS–) or a purified MAMP (Figure 1a and data not shown). Thus, MAMPs presented by the TTSS– mutant make a relatively minor contribution to induction of PR-1 by Pph and the majority of PR-1 induced by Pph requires a functional TTSS.

The second defense readout we evaluated is callose accumulation (Figure 1b). Papillae are defense-associated cell-wall thickenings that incorporate toxic and structural compounds (Bestwick et al., 1995). Callose is a structural component of papillae that can be visualized by fluorescence microscopy following aniline blue staining (Adam and Somerville, 1996). We found that callose was strongly induced by Pph, but not by Pto. Further, although Pph and Pph (TTSS–) both induce callose deposits, the size distribution of the deposits was distinct (Figure 1c). Pph (TTSS–) induced small deposits that are indistinguishable from those induced by Pto (TTSS–) or purified MAMPs (data not shown). Thus, MAMPs from Pph are the likely elicitors of these ‘small callose’ deposits. Only Pph with a functional TTSS induced numerous callose deposits exceeding 20 μm in diameter, which we have called ‘big callose’.

We wondered if big callose coincided with single cells or if it occurred in between two or more cells. We also wondered if big callose might be associated with dead plant cells. To test these possibilities, leaves challenged with Pph were simultaneously stained with trypan blue (to identify dead cells) and aniline blue (to detect callose) (Figure 1d,e). Although Pph does not induce a macroscopic HR, it does induce some dead cells. The dead cells occurred predominantly around the vasculature (data not shown). It is unclear whether dead cells induced by Pph result from a defense-associated response (e.g. a ‘weak’ HR) or a virulence-associated response (e.g. necrosis). Notably, the location of big callose did not coincide with that of dead cells (Figure 1d). Higher-magnification images revealed that big callose deposits correspond with single, viable cells (Figure 1e).

Arabidopsis defense mutants variably affect Pph-induced defenses

Before we could test the role of individual defense pathways in non-host resistance against Pph, we needed tools to further characterize and ultimately disrupt those pathways. Our first approach was to use Arabidopsis plants with mutations in defense genes to assess the contribution of those genes to Pph-induced defense readouts (Figure 2). We tested mutants with known roles in SA signaling, R-protein function, and cell-wall-based defense.

Figure 2.

  Disruption of Pseudomonas syringae pv. phaseolicola NPS3121 (Pph)-induced defense responses in known defense mutants.
(a) Accumulation of PR-1 induced by Pph or Pph (TTSS–) (TTSS, type III secretion system). Western blot analysis is from samples taken at 24, 48, and 72 h post-infiltration. The first five rows are from one experiment and the last two rows (Col-0 and pad4) are from a separate experiment. The results are collectively representative of at least three independent experiments.
(b) Callose deposition induced by Pph. The average and SE of the number of callose deposits per mm2 are from three independent experiments.

Salicylic acid frequently accumulates downstream of pathogen recognition and upstream of defense responses, including accumulation of PR-1. Therefore, we tested plants with mutations in genes involved in SA signaling, namely SID2, NPR1, and PAD4. SID2 is an isochorismate synthase that is required for defense-associated production of SA (Wildermuth et al., 2001). In sid2 plants, Pph induced minimal PR-1 (Figure 2a). NPR1 contributes to SA-induced expression of defense genes, including PR-1. Salicylic acid stimulates translocation of NPR1 from the cytoplasm to the nucleus (Mou et al., 2003), where it is thought to act as a transcriptional co-factor (Fan and Dong, 2002). Pph induced a reduced quantity of PR-1 in npr1 plants (Figure 2a). PAD4 is a component of the EDS1–SAG101–PAD4 signaling complex, which, among other things, contributes to amplification of the accumulation of SA (Falk et al., 1999; Feys et al., 2005; Jirage et al., 1999). As in npr1 plants, Pph induced a reduced quantity of PR-1 in pad4 plants (Figure 2a). Contrary to their involvement in PR-1 induction, SID2, NPR1, and PAD4 were not required for the induction of small or big callose by Pph (Figure 2b) or for the induction of small callose by Pph (TTSS–) (data not shown). Notably, the increase in big callose deposits observed in pad4 plants is probably independent of SA as no similar increase is observed in sid2 plants. Taken together, the data indicate that SA signaling is not required for Pph-induced callose production, but is important for Pph-induced expression of PR-1.

Another gene tested for its contribution to Pph-induced defense readouts was RAR1. RAR1 functions as a co-chaperone required for the accumulation and function of numerous R-proteins (Holt et al., 2005; Muskett et al., 2002; Tornero et al., 2002). Pph induced about 5% as many big callose deposits in rar1 plants as in wild-type plants (Figure 2b). RAR1 was not required for the induction of small callose by Pph (Figure 2b) or by Pph (TTSS–) (data not shown). The increase in small callose deposits in rar1 plants is consistent with a reported increase in flagellin-induced callose in this mutant background (Shang et al., 2006). Notably, RAR1 was not required for effector-induced accumulation of PR-1 (Figure 2a). Thus, of those defense responses we have examined, only big callose is induced via a pathway requiring RAR1.

A final gene tested for its contribution to Pph-induced defense readouts was PMR4. PMR4 is a callose synthase responsible for defense-associated callose deposition (Jacobs et al., 2003; Nishimura et al., 2003). Not surprisingly, PMR4 was required for the induction of both small and big callose by Pph (Figure 2b) and small callose by Pph (TTSS–) (data not shown). Meanwhile, mutation of PMR4 caused more rapid and stronger induction of PR-1 by Pph and Pph (TTSS–) (Figure 2a). This result is reminiscent of previous studies in which pmr4 showed enhanced expression of PR-1 (Vogel and Somerville, 2000) and other SA-dependent defense-associated genes (Nishimura et al., 2003) in response to a fungal pathogen, Erysiphe cichoracearum.

Collectively, these data indicate that multiple elicitors from Pph stimulate multiple defense signaling pathways in Arabidopsis. We have established readouts for three Pph-induced pathways: (i) a pathway induced by Pph (TTSS–) that leads to cell-wall fortifications exemplified by small callose, (ii) a RAR1-dependent pathway induced by wild-type Pph that leads to cell-wall fortifications exemplified by big callose, and (iii) a SA-dependent pathway induced by wild-type Pph that leads to defense gene expression exemplified by accumulation of PR-1 protein (Figure 7).

Figure 7.

  Model of signaling network for Pseudomonas syringae pv. phaseolicola NPS3121 (Pph)-induced basal defenses.
Elicitors from Pph activate microbe-associated molecular pattern receptors and R-proteins (blue). Defense proteins (purple) transduce signals to readouts of defense responses (green). Type III effectors (red) suppress signaling pathways and possibly carry out additional unknown functions.

Pto suppresses Pph-induced defenses and enhances Pph growth

In Arabidopsis, Pph induces defenses that are not induced by Pto (Figure 1). We conducted co-inoculation experiments with Pph and Pto to determine whether defense induction by Pph or defense suppression by Pto is dominant and how the overall defense response relates to bacterial growth.

First, we investigated the effect of Pto on the induction of defense readouts by Pph (Figure 3a,b). In these experiments, we co-inoculated leaves with a 1:1 ratio of two different bacteria at high concentration [1 × 108 colony-forming units (CFU) ml−1 of each]. Wild-type Pto suppressed each of the Pph-induced defense responses. Pto (TTSS–) failed to suppress any of these responses, indicating that type III effectors from Pto are required for suppression of Pph-induced defenses.

Figure 3.

  Co-inoculated Pseudomonas syringae pv. tomato strain DC3000 (Pto) dominantly suppresses P. syringae pv. phaseolicola NPS3121 (Pph)-induced defenses and enhances growth of Pph. Both activities of Pto are dependent on its type III secretion system (TTSS).
(a) Accumulation of PR-1 induced by Pph or Pph (TTSS–) infiltrated alone (H2O) or with Pto, Pto (TTSS–), or Pto (ΔCEL) (CEL, conserved effector locus). Western blot analysis is from samples taken 48 h after infiltration. Data are representative of three or more independent experiments.
(b) Big and small callose induced when Pto, Pto (TTSS–), or Pto (ΔCEL) are infiltrated alone (H2O) or with Pph or Pph (TTSS–). The average and SE of the number of callose deposits per mm2 is from at least two independent experiments.
(c) Growth of Pph or Pph (TTSS–) infiltrated alone (H2O) or with Pto (TTSS–), Pto (ΔCEL), or Pto. Data are representative of five independent experiments and error bars show SD. For all experiments in this figure, bacteria were co-infiltrated at a 1:1 ratio.

We wondered whether suppression of Pph-induced defenses by Pto would permit enhanced growth of Pph (Figure 3c). By using differential antibiotic selection, the growth of Pph and Pto was measured independently following their co-inoculation into leaves. In this experiment, a lower inoculum of 1 × 106 CFU ml−1 for each bacterium was used to allow the bacteria to proliferate. At this bacterial density, a HR assay produces a non-confluent response. Thus, only a fraction of cells contacted by Pph were also contacted by Pto, and vice versa. Nonetheless, Pto enhanced by 2 logs the growth of either Pph or Pph (TTSS–). The robust growth of Pto was unaffected by the co-inoculation of Pph (data not shown). Pto (TTSS–), which failed to suppress Pph-induced defenses, also failed to enhance Pph growth. Thus, suppression of Pph-induced defenses by Pto correlates with enhanced growth of Pph.

Though the TTSS from Pto was required for suppression of Pph-induced defenses, it remained unclear which type III effectors contribute to defense suppression. DebRoy et al., showed that HopM1 and AvrE each suppress basal defense signaling (DebRoy et al., 2004). These functionally redundant type III effectors are part of the CEL in the Pto genome, which is required for full virulence (Alfano et al., 2000). Thus, we tested whether Pto lacking the CEL (ΔCEL) was altered in its ability to suppress Pph-induced defenses or to enhance the growth of Pph. Like wild-type Pto, Pto (ΔCEL) suppressed Pph-induced big callose and PR-1 protein accumulation (Figure 3a,b). Unlike Pto, Pto (ΔCEL) induced small callose whether inoculated alone or co-inoculated with Pph. Though its growth was enhanced, Pph grew to a lower level when co-inoculated with Pto (ΔCEL) than with wild-type Pto. Defenses associated with the induction of small callose by Pto (ΔCEL) may account for the decreased growth of Pph.

Heterologous type III effectors suppress Pph-induced defenses

When Pph is co-inoculated with another bacterium (as in Figure 3) multiple heterologous effectors are simultaneously delivered into the cells of the host. To reduce this complexity, we have studied heterologous type III effectors individually. Because effectors from compatible pathogens are known to suppress plant defenses, we chose to test the role of individual type III effectors from pathogens of Arabidopsis in the suppression of Pph-induced defenses (Figure 4). HopM1 and AvrRpm1 were selected because each suppresses basal defenses in Arabidopsis (DebRoy et al., 2004; Kim et al., 2005b) and contributes to the growth of virulent bacteria (DebRoy et al., 2004; Ritter and Dangl, 1995; Rohmer et al., 2003). We studied HopM1, from DC3000, by expressing it in Pph from a plasmid. We studied AvrRpm1 by inducibly expressing it in Arabidopsis from a transgene (Dex:AvrRpm1) (Aoyama and Chua, 1997; Mackey et al., 2002). RPM1 is an R protein of Arabidopsis that induces a HR in response to AvrRpm1 (Grant et al., 1995). All the Dex:AvrRpm1 plants used in this study carry a non-functional mutant allele of RPM1 (Bisgrove et al., 1994) so that the virulence function(s) of AvrRpm1 can be studied in the absence of RPM1 activation.

Figure 4.

  HopM1 and AvrRpm1 suppress Pseudomonas syringae pv. phaseolicola NPS3121 (Pph)-induced defenses.
(a) Big and small callose deposition induced by Pph (TTSS–) (TTSS, type III secretion system), Pph, or Pph (HopM1) in rpm1/rps2 and in Dex:AvrRpm1 in rpm1/rps2. The average and SE of the number of callose deposits per mm2 is representative of five independent experiments.
(b) Accumulation of PR-1 induced by Pph (TTSS–), Pph, or Pph (HopM1) in Col-0 or Dex:AvrRpm1 in rpm1/rps2. Western blot analysis is from samples taken 48 h after infiltration. Data are representative of four experiments.

Pph lacks AvrRpm1. Pph does not induce a RPM1-dependent HR, and the HR is not suppressed because Pph carrying a plasmid expressing AvrRpm1 does induce the HR. Pph also lacks a functional HopM1. In the sequenced strain of Pph (1448A; Joardar et al., 2005) hopM1 has a frameshift mutation at codon 545, which is predicted to disrupt the function of HopM1 (Nomura et al., 2006). We have sequenced the portion of hopM1 from NPS3121 (the strain used in this study), which includes the mutation and eventual stop codon, and found it to be identical to the corresponding region of hopM1 from strain 1448A (data not shown).

Pph-induced callose deposition was suppressed by both HopM1 and AvrRpm1 (Figure 4a). AvrRpm1 is known to suppress MAMP-induced signaling (Kim et al., 2005b), so it is not surprising that, when expressed in Arabidopsis, AvrRpm1 blocked the production of small callose induced by Pph (TTSS–). AvrRpm1 also suppressed the production of small and big callose in response to wild-type Pph. Similarly, HopM1, when expressed by Pph, suppressed the production of small and big callose induced by Pph. Thus, AvrRpm1 and HopM1 may suppress defense signaling after convergence of the signaling pathways leading to small and big callose.

Interestingly, HopM1 partially suppressed the accumulation of PR-1 induced by Pph (Figure 4b). The partial suppression of accumulation of PR-1 by HopM1, in combination with mutations in NPR1 or PAD4, was sufficient to reduce the amount of PR-1 in these mutant backgrounds to undetectable levels (data not shown). Thus, in addition to its ability to suppress callose production, HopM1 also suppresses SA-dependent accumulation of PR-1. Unlike HopM1, AvrRpm1 did not significantly suppress Pph-induced accumulation of PR-1 (Figure 4b).

Type III secretion systems can deliver heterologous effectors. Thus, we supposed that Pph (HopM1) delivers both HopM1 and the normal complement of Pph effectors into plant cells. We conducted two experiments to verify those assumptions. Firstly, Pph and Pph (HopM1) were equally competent at eliciting a HR in tobacco (data not shown). As this HR is induced by type III effectors from Pph, HopM1 apparently does not prevent their delivery. Secondly, in a co-inoculation of Pph and Pph (HopM1), HopM1 suppressed small and big callose deposition (data not shown). Thus, HopM1 functions inside plant cells to suppress defenses induced by Pph.

Disrupting Pph-induced defense signaling increases the susceptibility of Arabidopsis to Pph

We hypothesized that basal defenses might underlie the resistance of Arabidopsis to Pph. Thus, suppressing basal defenses might permit growth of Pph. Pto suppressed Pph-induced defense responses and also enhanced the growth of Pph (Figure 3). However, Pto delivers many type III effectors likely to alter the host in a variety of ways. Thus we sought to challenge our hypothesis in a more refined manner. By disrupting host defenses with plant mutations and individual, heterologous type III effectors, we have enhanced the growth of Pph in Arabidopsis. The data indicate that basal defenses are critical to non-host resistance of Arabidopsis against Pph.

We first tested plants with mutations in SID2, NPR1, PAD4, RAR1, and PMR4. Although these mutations disrupt individual, Pph-induced defense responses (Figure 2), each mutant was able to fully restrict the growth of infiltrated Pph (Figures 5 and 6, data not shown, and van Wees and Glazebrook, 2003). Similarly, wild-type plants maintained full resistance against Pph (HopM1) (Figures 5 and 6). Thus, the host defenses disrupted by these mutations or by HopM1 are individually dispensable for resistance to Pph.

Figure 5.

  Non-host resistance of Arabidopsis to Pseudomonas syringae pv. phaseolicola NPS3121 (Pph) is disrupted by the action of HopM1 and multiple defense mutations.
Pph (TTSS–) (TTSS, type III secretion system), Pph, or Pph (HopM1) were infiltrated into the indicated plants. Growth of Pph (HopM1) was enhanced in pmr4/pad4 and in pmr4/npr1. Data are representative of three out of four independent experiments and error bars represent SD.

Figure 6.

  Non-host resistance of Arabidopsis to Pseudomonas syringae pv. phaseolicola NPS3121 (Pph) is disrupted by the action of AvrRpm1.
The growth of Pph (TTSS–) (TTSS, type III secretion system) and Pph is enhanced in plants expressing AvrRpm1. Bacterial expression of HopM1 further promotes the growth of Pph. Mutation of (a) NPR1 and (b) SID2 in combination with in planta expression of AvrRpm1 permits even more growth of Pph (HopM1). All data are from day 4 and are representative of five different experiments; error bars represent SD.

Remarkably, disruption of multiple defense pathways significantly enhanced the growth of Pph in Arabidopsis. The growth of Pph (HopM1) was enhanced by more than 3 logs in Arabidopsis with mutations in both PMR4 and PAD4 and more than 2 logs in pmr4/npr1 (Figure 5). Single mutant pmr4, npr1, and pad4 plants were resistant to Pph (HopM1). Double mutant pmr4/npr1 and pmr4/pad4 plants were resistant to both Pph (TTSS–) and wild-type Pph. Pph (HopM1) induced disease symptoms in pmr4/npr1 and pmr4/pad4 (data not shown); however, the growth of Pph (HopM1) and the associated symptoms were both less than those observed with wild-type Pto in wild-type Col-0. Pto typically grew about 1 or 2 logs more and produced more severe disease symptoms in Col-0 than did Pph (HopM1) in pmr4/pad4 (data not shown). Thus, mutation of PMR4 and NPR1 or PAD4 in combination with the action of HopM1 permits Pph to grow and cause disease, though not to the same extent as a virulent pathogen.

We also tested the effect of a second bacterial effector, AvrRpm1, on the virulence of Pph (Figure 6). The growth of Pph was enhanced by about 1 log in Dex:AvrRpm1 plants relative to control plants. The growth of Pph was enhanced by about 2–3 logs in Dex:AvrRpm1 in npr1 or Dex:AvrRpm1 in sid2, indicating that AvrRpm1 provides a virulence function(s) complementary to the disruption of SA signaling in these mutant backgrounds. The growth of Pph (HopM1) was enhanced by about 2–3 logs in Dex:AvrRpm1, indicating that HopM1 provided a virulence function(s) complementary to the function(s) of AvrRpm1. Remarkably, combining the expression of AvrRpm1 and HopM1 with mutation of NPR1 or SID2 permitted still higher growth of Pph. In fact, Pph (HopM1) in Dex:AvrRpm1 in npr1 or Dex:AvrRpm1 in sid2 grew to levels comparable to those achieved by Pto in wild-type Col-0 (data not shown). Thus, by manipulating multiple layers of basal defense, we have made Arabidopsis highly susceptible to Pph.

Discussion

The overall importance of basal defenses in non-host resistance has remained unclear. It was recently shown that suppressing MAMP signaling increases the growth of Pph in Arabidopsis (He et al., 2006; de Torres et al., 2006). However, these reports failed to explain why bacterial growth still remained much lower than that of virulent strains of P. syringae. Our results demonstrate that wild-type Pph induces defenses additional to those stimulated by the MAMPs of Pph (TTSS–). When MAMP- and TTSS-dependent defenses are suppressed, Arabidopsis becomes increasingly susceptible to Pph. Thus, non-host resistance against Pph results from multiple pathways that induce individually effective basal defenses.

Using callose deposition and PR-1 accumulation as defense readouts, we have demonstrated that Pph induces multiple defense signaling pathways in Arabidopsis (Figure 7). The signaling pathways are discriminated by: (i) the eliciting bacterial molecules, (ii) the readouts induced, (iii) the requirement for host defense genes, and (iv) the sensitivity to suppression by heterologous type III effectors. By combining mutations and heterologous type III effectors to simultaneously disrupt multiple Pph-induced defense pathways, we demonstrated that these pathways collectively contribute to non-host resistance of Arabidopsis to Pph.

Pph produces multiple elicitors of defense in Arabidopsis (Figure 7). Pph (TTSS–) induces small callose deposits and weak expression of PR-1. These responses are independent of type III effector delivery and are indistinguishable, at our level of analysis, from responses induced by purified MAMPs, such as flg22. We therefore hypothesize that these responses are induced by MAMPs from Pph. Consistent with this, AvrRpm1, a known suppressor of MAMP signaling, suppresses the induction of small callose by Pph. Wild-type Pph, with a functional TTSS, induces additional defense responses. We hypothesize that type III effectors from Pph induce big callose deposits and strong expression of PR-1. Alternatively, Pph may elicit defenses that are dependent on the TTSS itself, rather than type III effectors.

Plant genetic requirements for Pph-induced defense responses

The defense pathways elicited by Pph have distinct genetic requirements (Figure 7). Among these could be the function of ‘weak’ R-proteins that do not induce a HR (Gassmann, 2005). The pathway leading to big callose requires RAR1, which is known to be required for the function of numerous NB-LRR-type R-proteins, including some involved in basal defense (Holt et al., 2005; Muskett et al., 2002; Shirasu et al., 1999; Tornero et al., 2002). Thus, we hypothesize that Arabidopsis encodes a RAR1-dependent R-protein required for the induction of big callose. The pathways leading to small and big callose converge downstream of RAR1, as each pathway requires the callose synthase PMR4. Thus, big callose deposits may result from elevated flux through the converged pathway when both MAMP receptors and the putative R-protein are activated.

The Pph-induced pathway leading to strong accumulation of PR-1 does not require RAR1. We hypothesize that Arabidopsis encodes a RAR1-independent R-protein that triggers the strong induction of PR-1. The role of SA further distinguishes the pathways leading to callose from the pathway leading to strong PR-1 accumulation. Mutations that disrupt SA production (sid2) (Wildermuth et al., 2001), SA amplification (pad4) (Glazebrook et al., 1997), or SA-induced transcription (npr1) (Fan and Dong, 2002; Mou et al., 2003) disrupt the strong induction of accumulation of PR-1. On the contrary, SA signaling is not required for the induction of small or big callose.

Pph (TTSS–) induces more accumulation of PR-1 in pmr4. This enhancement is reminiscent of enhanced defense gene expression induced by powdery mildew in pmr4, which was speculated to result from a more intimate interaction of the fungus and the host (Nishimura et al., 2003; Vogel and Somerville, 2000). Similarly, papillae lacking callose may permit more efficient activation of MAMP receptors by Pph. Resistance of pmr4 to powdery mildew is dependent on SA signaling (Nishimura et al., 2003). Similarly, disruption of SA signaling contributes to the enhanced growth of Pph in pmr4 plants.

Suppression of Pph-induced defense responses by heterologous type III effectors

The defense-suppressing activity of virulence factors from Pto is dominant over the defense-inducing activities of Pph. When co-inoculated with Pph, Pto suppresses Pph-induced defenses and enhances the growth of Pph. Type III effectors from Pto are required for both activities. Pto has effectors with redundant defense-suppressing activities. Both HopM1 (from within the CEL) and effectors other than HopM1 [that are delivered from Pto (ΔCEL)] suppress callose and strong PR-1 accumulation induced by Pph. HopM1 also suppresses defenses elicited by other effectors from Pto (DebRoy et al., 2004) and these defenses induced by Pto (ΔCEL) probably account for its inability to enhance Pph growth as well as Pto. Recent work has demonstrated that MAMPs from Pph induce rapid defense responses, including MAPK signaling (He et al., 2006). Individual effectors from Pto, specifically AvrPto and AvrPtoB, suppress those defense pathways and enhance the growth of Pph in Arabidopsis (He et al., 2006; de Torres et al., 2006). Taken together with our results, it is clear that a variety of type III effectors from Pto suppress defense pathways active against Pph. These results indicate that both non-host and pathogenic bacteria elicit similar defense responses and the lack of pathogenicity of non-host bacteria results from an inability to suppress those responses.

To simplify the contribution of heterologous virulence factors, we studied individual type III effectors from bacterial pathogens of Arabidopsis. Both AvrRpm1 and HopM1 suppress the induction of both small and big callose by Pph. To suppress both defense readouts, the effectors probably function downstream from where the pathways converge. The recent observation that HopM1 suppresses vesicle traffic to the plasma membrane provides a mechanistic model for how HopM1 suppresses both types of cell-wall modifications (Nomura et al., 2006). Differences apparently exist between the function of AvrRpm1 and HopM1. HopM1 partially suppresses the accumulation of PR-1 induced by Pph while AvrRpm1 does not. This activity of HopM1 could also result from its inhibition of vesicle traffic (Wang et al., 2005). Despite its inability to suppress accumulation of PR-1, AvrRpm1 alone enhances the growth of Pph, which HopM1 does not. It is important to note that AvrRpm1 is over-expressed in planta and HopM1 is expressed within Pph from a multicopy plasmid. Thus, the timing and level of expression of these effectors could alter their normal function. Nonetheless, transgenically expressed AvrRpm1 and bacterially expressed HopM1 each have an activity lacked by the other.

Layering of non-host resistance

Non-host resistance is widely thought to result from multiple barriers that must be overcome by a pathogen (Thordal-Christensen, 2003). Our results indicate that multiple barriers can exist within the classification of basal defenses. Although individual mutants or heterologous type III effectors suppress various Pph-induced defense readouts, they promote limited Pph growth in the case of AvrRpm1 or not at all in other cases. High-level growth of Pph is only observed when multiple type III effectors and/or host mutations are combined. The layered nature of non-host resistance against Pph is most apparent in plants expressing AvrRpm1. Expression of AvrRpm1 can weakly enhance the growth of Pph. When combined with AvrRpm1, expression of HopM1 or mutation of NPR1 or SID2 adds incrementally to the growth of Pph. Finally, the simultaneous action of AvrRpm1, HopM1, and mutation of NPR1 or SID2 permits growth of Pph comparable to that of Pto in wild-type Arabidopsis.

Peart et al. (2002) classified non-host resistance of Nicotiana benthamiana as SGT1 dependent and SGT1 independent. They suggested that SGT1-dependent non-host resistance might share a similar defense mechanism with R-mediated resistance, whereas SGT1-independent non-host resistance may involve pre-existing barriers or result from a lack of appropriate host and pathogenicity factors. Our findings add further to this distinction. Non-host resistance may result from multiple layers of defense that combine R-protein-mediated resistance and one or more additional layers of defense. Thus, the majority of non-host resistance may rely in part or whole on R-mediated resistance. This supposition is supported by significant overlap between defense responses associated with non-host and gene-for-gene resistance (Navarro et al., 2004; Tao et al., 2003).

Experimental procedures

Plants

Arabidopsis thaliana ecotype Col-0 was used as the wild-type background in this work. The mutants used in this study were as follows: npr1-1 changes amino acid 334, in the third ankyrin-repeat consensus sequence, from histidine to tyrosine (Cao et al., 1997); pad4-1 has a premature stop at codon 359 (Jirage et al., 1999); pmr4-1 has a premature stop at codon 687 (Nishimura et al., 2003); rar1-21 has a premature stop at codon 52 (Tornero et al., 2002); and sid2-2 has a deletion spanning codons 439 to 455 within the predicted chorismate-binding domain (Wildermuth et al., 2001). Plants were grown at 23°C for 8 h of light and 16°C for 16 h of darkness.

DEX:AvrRpm1 in rpm1/rps2, DEX::AvrRpm1 in rpm1/npr1, and DEX:AvrRpm1 in rpm1/sid2 were generated by crossing the DEX:AvrRpm1/rpm1 (Mackey et al., 2002) with rps2-101C, npr1-1, and sid2, respectively. These transgenic lines were induced by spraying with 20 μm dexamethasone (Sigma-Aldrich, http://www.sigmaaldrich.com/) containing 0.005% Silwet L-77 (Lehle Seeds, http://www.arabidopsis.com/) 24 h prior to the infiltration of bacteria.

Bacteria

Pseudomonas syringae pv. tomato strain DC3000 (Pto) and P. syringae pv. phaseolicola strain NPS3121 (Pph) are a pathogen and a non-host pathogen of Arabidopsis, respectively. The TTSS-deficient strain of Pto has a mutation in hrcC and the TTSS-deficient strain of Pph (strain NPS4007) is a hrpE::Tn5 derivative of Pph NPS3121 (Rahme et al., 1991). The ΔCEL mutant of Pto is strain CUCPB5115 (Alfano et al., 2000). The plasmid pORF43 (DebRoy et al., 2004), which carries HopM1 and the chaperone ShcM (Badel et al., 2003), was transformed into Pph by electroporation to generate Pph (HopM1).

For analyses of PR-1 accumulation and callose deposition, 108 CFU ml−1 of bacteria (OD600 = 0.2) were infiltrated with a needle-less 1 cm3 syringe into leaves of 5-week-old plants. For experiments with mixed bacteria, the infiltrated suspension contained 108 CFU ml−1 of each bacterium, so that the total concentration of bacteria was 2 × 108 CFU ml−1. In control experiments, bacteria were infiltrated into Col-0 at 106 CFU ml−1, at which concentration Pto does not induce tissue collapse within 48 h. At this reduced concentration, Pph still induced accumulation of PR-1 and Pto did not. Thus, the lack of PR-1 protein in response to Pto does not result from tissue collapse.

Growth curve assays were conducted by syringe-infiltrating 106 CFU ml−1 of bacteria in 10 mm MgCl2 into leaves of 5- to 6-week-old plants. After the infiltrated leaves were no longer visibly wet with infiltrate (about 4 h), the plants were kept in 100% humidity (domes on) for the remainder of the experiment. Later, bacteria were titered by grinding leaf disks to homogeneity in 10 mm MgCl2, plating serial dilutions, and counting colonies.

For co-inoculation growth curves, kanamycin selection was used to specifically determine the growth of Pph and Pph (TTSS–). Pph carried pVSP61, an empty vector conferring kanamycin resistance, and Pph (TTSS–) has the kanamycin resistance gene associated with Tn5 inserted in hrpE. Pto and the hrcC mutant of Pto were resistant to tetracycline because they carried the vector pBBR1MCS-3 (Kovach et al., 1994). For co-inoculation growth curves, the infiltrated suspension contained 106 CFU ml−1 of each bacterium, so that the total concentration of bacteria was 2 × 106 CFU ml−1. The growth of individual bacteria was monitored by titering and plating on kanamycin or tetracycline selection. Following growth in planta and titering, the veracity of kanamycin- and tetracycline-resistant colonies was confirmed as Pph and Pto, respectively, by virulence assays.

Protein

Soluble protein preparation and Western blot analysis were performed following a previously described method (Kim et al., 2005b). Briefly, about 3 cm2 of leaf tissue from three leaves was ground in 100 μl of grinding buffer [20 mm 2-amino-2-(hydroxymethyl)-1,3-propanediol (Tris)–HCl (pH = 7.5), 150 mm NaCl, 1 mm EDTA, 1% Triton X-100, 0.1% SDS, 5 mm DTT, and plant protease inhibitor cocktail (Sigma-Aldrich)]. Following centrifugation at 20 000 g for 10 min at 4°C, the soluble supernatant was recovered and protein concentration was determined by the Bio-Rad protein assay (Bio-Rad, http://www.bio-rad.com/). Samples were resolved on 12% SDS-PAGE gels (Mini-PROTEAN, Bio-Rad) and transferred to polyvinylidene difluoride membrane (Millipore, http://www.millipore.com/). Western blots were carried out by standard methods using anti-PR-1 sera (Kliebenstein et al., 1999) at a dilution of 1:5000.

Staining for callose and dead plant cells

For callose staining, 4- to 5-week-old leaves were syringe-infiltrated with 108 CFU ml−1 of each bacterium or 10 mm MgCl2. At 12–16 h after infiltration, whole leaves were collected and stained with aniline blue as previously described (Hauck et al., 2003) and callose deposition was examined with a Nikon Eclipse 80i epifluorescence microscope (http://www.nikon.com/).

For simultaneous staining of dead plant cells and callose, leaves challenged with 108 CFU ml−1 of Pph were collected at 12–16 h after infiltration and consecutively stained with trypan blue (Koch and Slusarenko, 1990) and aniline blue. In brief, leaves were immersed in the trypan blue staining solution [about 0.083 mg trypan blue in a mixture of 1 volume of lactophenol alcohol (a 1:1:1:1 volume mix of glycerol, saturated phenol, lactophenol, and deionized water) and 2 volumes of 95% ethanol (EtOH)] and incubated for 5 min at 95°C and then for an additional 4 h at room temperature. After the removal of the trypan blue staining solution, the leaf samples were incubated for 30 min at 65°C and subsequently washed with fresh lactophenol alcohol for 4 h and with 50% EtOH for 24 h at room temperature. Then the leaf samples were destained with the chloral hydrate solution, which was made by dissolving 250 g of chloral hydrate in 100 ml of distilled water. Finally, those leaves were stained with aniline blue for callose staining, submerged in 50% glycerol, and mounted on slide glasses. The same leaf areas were observed for both dead cells and callose depositions using two different light sources, a white light lamp (for trypan blue) and a mercury lamp (for aniline blue).

Acknowledgements

This work was funded by grants from the NSF (MCB-0315673) and the Ohio Agricultural Research & Development Center of The Ohio State University. MGK was partially supported by a grant from the MOST/KOSEF to the Environmental Biotechnology National Core Research Center (grant #: R15-2003-012-01001-0), Korea. We thank Jian-Min Zhou for Pph and Pph (TTSS–), Sheng Yang He for pORF43, Dan Kliebenstein for anti-PR-1 antibodies, and Marc Nishimura and Shauna Somerville for pmr4/npr1 and pmr4/pad4 seeds. We appreciated helpful comments on the manuscript from David Coplin, Jeff Dangl, and Aidan McFall.

Ancillary