A new allele of the coronatine-insensitive locus (COI1) was isolated in a screen for Arabidopsis thaliana mutants with enhanced resistance to the bacterial pathogen Pseudomonas syringae. This mutant, designated coi1-20, exhibits robust resistance to several P. syringae isolates but remains susceptible to the virulent pathogens Erisyphe and cauliflower mosaic virus. Resistance to P. syringae strain PstDC3000 in coi1-20 plants is correlated with hyperactivation of PR-1 expression and accumulation of elevated levels of salicylic acid (SA) following infection, suggesting that the SA-mediated defense response pathway is sensitized in this mutant. Restriction of growth of PstDC3000 in coi1-20 leaves is partially dependent on NPR1 and fully dependent on SA, indicating that SA-mediated defenses are required for restriction of PstDC3000 growth in coi1-20 plants. Surprisingly, despite high levels of PstDC3000 growth in coi1-20 plants carrying the salicylate hydroxylase (nahG) transgene, these plants do not exhibit disease symptoms. Thus resistance to P. syringae in coi1-20 plants is conferred by two different mechanisms: (i) restriction of pathogen growth via activation of the SA-dependent defense pathway; and (ii) an SA-independent inability to develop disease symptoms. These findings are consistent with the hypotheses that the P. syringae phytotoxin coronatine acts to promote virulence by inhibiting host defense responses and by promoting lesion formation.
In response to microbial attack, plants activate a complex series of general defense responses that are believed to inhibit colonization of plant tissue by micro-organisms (Heath, 2000). These inducible defenses include a rapid oxidative burst; accumulation of elevated levels of the endogenous signaling compounds salicylic acid (SA) and jasmonic acid (JA); induction of several pathogenesis-related (PR) genes; and production of antimicrobial phytoalexins and lytic enzymes (Felix et al., 1999; Glazebrook et al., 1997; Hammond-Kosack and Jones, 1996; Lamb et al., 1989). These defenses are also often induced in response to infection by virulent pathogens. However, in these interactions induction of defenses responses occurs at a relatively late stage of infection and, although this does not prevent disease development, appears to be important in limiting pathogen aggressiveness (Glazebrook et al., 1997; Jakobek et al., 1993). Thus it is likely that successful plant pathogens evade or actively inhibit induction of host defenses to facilitate colonization of plant tissues. Presumably the ability to avoid detection, or to suppress activation of defense responses normally induced upon microbial attack, are traits that may distinguish successful plant pathogens from non-pathogenic organisms (Alfano and Collmer, 1996; Felix et al., 1999; Lamb et al., 1989). However, the pathogen virulence factors that are involved in these processes, and how they function to modify host defense mechanisms, are not well understood (Alfano and Collmer, 1996).
One example of a virulence factor that may promote parasitism by inhibiting host defenses is the bacterial phytotoxin coronatine (COR). COR is produced by several strains of the bacterial pathogen Pseudomonas syringae and has been demonstrated to contribute to virulence of P. syringae (Bender et al., 1999; Moore et al., 1989). Pseudomonas syringae mutants that do not produce COR exhibit a reduction in both growth and disease symptom development on several plant species, including Arabidopsis thaliana (Bender et al., 1987; Budde and Ullrich, 2000; Mittal and Davis, 1995). The mode of action of COR in contributing to pathogen virulence is not well understood. However, based on both structural similarities and its effects on plant tissue, COR has been proposed to function as a molecular mimic of methyl jasmonate (MeJA), an endogenous plant hormone involved in defense signaling (Bender et al., 1999; Feys et al., 1994; Reymond and Farmer, 1998). Both MeJA and COR induce similar responses in plants, including inhibition of root elongation in A. thaliana, accumulation of anthocyanin, production of ethylene, and leaf senescence (Bender et al., 1999).
In a series of experiments aimed at addressing the role of COR in virulence of P. syringae, Mittal and Davis (1995) found that a mutant of P. syringae pv. tomato strain DC3000 impaired in COR biosynthesis exhibited reduced growth and symptom production on A. thaliana. They also demonstrated that this reduced virulence was correlated with an enhanced ability to elicit expression of defense-related genes. These findings led Mittal and Davis (1995) to hypothesize that COR is important in early stages of infection by P. syringae, and may be involved in suppression of general host defenses that normally serve to inhibit colonization of plant tissue. However, the molecular basis for how COR modulates host defenses has not been explored.
The identification and characterization of A. thaliana mutants that are insensitive to COR is likely to provide insight into the molecular basis of COR virulence activity (Feys et al., 1994). A series of COR-insensitive (coi) mutants was isolated on the basis that they exhibited normal root elongation in the presence of 50 µM COR. Interestingly, the coi mutants defined a single locus, COI1, and were also shown to be insensitive to MeJA (Feys et al., 1994). The coi1-1 mutant has been extensively studied and has been shown to exhibit several additional phenotypes, including a defect in pollen development that renders it male sterile and resistance to the bacterial pathogen P. syringae pv. atropurpurea (Feys et al., 1994). Further evidence that the JA-signaling pathway is impaired in the coi1-1 mutant stems from studies showing that mutant plants do not express several JA-responsive genes upon application of MeJA (Benedetti et al., 1995; Benedetti et al., 1998; Feys et al., 1994), or after infection with the fungal pathogen Alternaria (Penninckx et al., 1996; Thomma et al., 1998). The COI1 gene has been isolated and encodes a predicted protein that contains an F-box motif and a series of leucine-rich repeats, suggesting that the COI1 protein may be involved in targeting proteins for polyubiquination and degradation (Xie et al., 1998). The mechanism(s) underlying resistance to P. syringae in the coi1-1 mutant has not been investigated. Presumably this resistance is due, at least in part, to the fact that mutant plants are insensitive to the virulence factor COR. Investigation of the molecular basis of resistance in COI1 mutant plants would provide valuable insight into the mechanism of COR function, and could also provide clues as to how a mutant that is defective in a host defense response-signaling pathway confers pathogen resistance.
To identify plant genes that govern susceptibility to virulent pathogens, we carried out a screen for A. thaliana mutants with enhanced disease resistance to P. syringae (Boch et al., 1998; G. Kalinowski, M. Verbsky and B. Kunkel, unpublished results). One of the mutants isolated in this screen defines a new allele of COI1 and was thus designated coi1-20. To better understand the mechanism(s) underlying resistance in the coi1-20 mutant, we investigated the molecular basis of resistance in this line. Here we demonstrate that resistance to P. syringae strain PstDC3000 in the coi1-20 mutant is correlated with the hyper-activation of SA synthesis and SA-mediated defense responses upon infection. These results suggest that the SA-signaling pathway is sensitized in the coi1-20 mutant such that defenses are rapidly induced in response to infection by the virulent PstDC3000 strain. We also show that coi1-20 plants carrying mutations in the SA-signaling pathway allow high levels of PstDC3000 growth but, unexpectedly, do not develop disease symptoms. These findings suggest that the enhanced resistance phenotype of the coi1 mutant is mediated via two different mechanisms: (i) sensitization of an SA-dependent defense-signaling pathway responsible for restricting pathogen growth; and (ii) an SA-independent impairment in disease symptom formation.
To identify genes that govern susceptibility to P. syringae, we conducted a screen for A. thaliana mutants that exhibit reduced disease symptoms upon infection with the P. syringae pv. tomato strain PstDC3000. Approximately 17 500 M2 plants derived from ethylmethane sulfonate (EMS)-mutagenized seeds of A. thaliana ecotype Columbia (Col-0) were screened, and 32 mutant lines that exhibited enhanced resistance in the M3 generation were isolated. One mutant, designated coi1-20, exhibited strong resistance to PstDC3000 (Figure 1), and was chosen for further analysis. As this mutant was male sterile it was propagated by fertilizing mutant flowers with pollen from wild-type plants, and was maintained as a heterozygous line. Several additional mutants exhibiting enhanced disease resistance that have been isolated in this screen are described elsewhere (Boch et al., 1998; G. Kalinowski, M. Verbsky and B. Kunkel, unpublished results).
Genetic analysis of coi1-20
To determine the genetic basis of the disease resistance phenotype in the coi1-20 mutant, we crossed the mutant to wild-type Col-0 plants. As summarized in Table 1, the F1 progeny from this cross were susceptible to PstDC3000 and were self-fertile, indicating that both enhanced resistance to PstDC3000 and male sterility in coi1-20 plants are caused by a recessive mutation(s). The F1 plants were allowed to self-pollinate, the resulting F2 progeny were assayed for disease resistance and fertility, and resistance to PstDC3000 was observed to segregate as a recessive, single-gene trait (Table 1). Male sterility co-segregated with disease resistance in these plants. In over 200 coi1-20 F2 progeny scored from a cross between coi1-20 and COI1 plants (Table 1 and data not shown) we did not detect any recombinant plants in which resistance to PstDC3000 and the male sterile phenotype were separated. We also noted a third phenotype, an upright growth habit in which the rosette leaves were held in a more vertical position, such that the angle between the petiole and the main axis of the plant was smaller than observed in wild-type plants, co-segregated with the enhanced resistance and male sterile phenotypes among the F2 progeny. This upright phenotype appeared to become more pronounced after PstDC3000 inoculation. These results indicate that the three phenotypes associated with the coi1-20 mutation are conferred by a defect at a single locus or by mutations at tightly linked loci.
Plants were inoculated by dipping in bacterial suspensions of P. syringae strain PstDC3000 containing the surfactant Silwet L-77, and were scored 4–5 days after inoculation. R, resistant plants exhibiting no disease symptoms; S, susceptible plants with disease symptoms consisting of individual water-soaked lesions and chlorosis.
Data presented for second back-cross of coi1-20 to wild-type Col-0 (COI1).
All resistant plants were sterile; all susceptible plants were fertile.
This unusual combination of male sterility and enhanced disease resistance phenotypes is reminiscent of those reported for coronatine-insensitive (coi1) mutants (Feys et al., 1994). In addition to exhibiting male sterility and enhanced resistance to P. syringae, coi1 mutants display a marked insensitivity to both the bacterial phytotoxin COR and the plant hormone MeJA (Feys et al., 1994). To determine whether resistance in our A. thaliana mutant is due to a mutation at the COI1 locus, we conducted genetic mapping experiments, direct phenotypic comparisons and complementation analysis with the coi1-1 mutant isolated by Feys et al. (1994).
The COI1 gene is located on the south arm of chromosome 2 between the visible markers as and cer8 (Xie et al., 1998). To determine the map position of our mutation we crossed the coi1-20 mutant to wild-type plants from the ecotype Landsberg erecta (Ler;Table 1). F2 progeny from this cross were used to map coi1-20 relative to a molecular marker (m429) known to be tightly linked to COI1 (Konieczny and Ausubel, 1993; http://www.arabidopsis.org/maps/CAPS_Chr2.html). In our mapping population of 30 F2 homozygous coi1-20 mutant plants (Table 1 and data not shown) we found no recombinants that separated the coi1-20 mutation from m429, demonstrating that coi1-20 maps to the COI1 region of chromosome 2.
Feys et al. (1994) previously reported that the coi1-1 mutant is resistant to infection by the plant pathogen P. syringae pv. atropurpurea. To determine if the coi1-1 and coi1-20 mutants exhibit similar levels of resistance to our P. syringae isolate, we subjected self-progeny from coi1-1 and coi1-20 heterozygous plants to infection with PstDC3000 and monitored appearance of disease symptoms. Both the coi1-1 and coi1-20 segregating families gave rise to approximately 25% resistant individuals that exhibited few or no visible disease symptoms (Table 1). The degree of resistance exhibited by the two mutants was indistinguishable.
The sensitivity of the coi1-20 mutant to COR and MeJA was assayed and compared to that of the coi1-1 mutant by germinating segregating families of the mutants on Murashige and Skoog (MS) (Murashige and Skoog, 1962) plates containing 1 µm COR or 10 µm MeJA. Both segregating populations gave rise to approximately 25% COR- or MeJA-insensitive seedlings as measured by root elongation 6 days following germination (Feys et al., 1994; see Experimental procedures). These results indicate that, like coi1-1, coi1-20 exhibits resistance to both COR and MeJA.
Allelism tests between coi1-1 and coi1-20 revealed that the two mutants do not complement one another, indicating that the lines carry mutations at the same locus (Table 1). Thus, given our findings that: (i) the coi1-1 and coi1-20 mutations map to the same position on chromosome 2; (ii) both mutant lines exhibit resistance to PstDC3000 and are insensitive to COR and MeJA; and (iii) the two mutants fail to complement one another, we conclude that the mutant we isolated based on its enhanced disease resistance phenotype defines a new allele of COI1. Thus we have assigned this mutant the designation coi1-20. The molecular nature of the coi1-20 allele is not presently known. However, based on our observation that the COR- and MeJA-insensitive phenotypes of coi1-20 are indistinguishable from those of coi1-1– a mutation that results in a truncated protein that appears to abolish JA-mediated signaling (Feys et al., 1994; Xie et al., 1998) – it is likely that coi1-20 also represents a strong loss of function allele.
coi1-20 confers resistance to multiple strains of P. syringae
To determine whether enhanced resistance conferred by the coi1-20 mutation is specific to PstDC3000, we inoculated coi1-20 plants with two additional virulent strains of P. syringae and observed that the mutant plants failed to develop disease symptoms following inoculation with either P. syringae pv. tomato strain Pst3455 or P. syringae pv. maculicola strain Psm m4 (data not shown). These observations are consistent with the finding by Feys et al. (1994) that coi1-1 is resistant to P. syringae, and suggest that resistance in the coi1-20 line is not due to gain of a novel capacity to specifically detect infection by PstDC3000, but rather appears to be conferred by enhanced resistance to several different virulent strains of P. syringae.
To determine if resistance in coi1-20 plants was associated with restricted growth of PstDC3000 within the plant, growth of the pathogen in coi1-20 leaf tissue was monitored over several days. As shown in Figure 1(b), growth of PstDC3000 was significantly reduced in coi1-20 plants, obtaining a final concentration of only 104−105 cfu cm−2. This was in marked contrast to the higher levels of bacterial growth observed in wild-type, susceptible siblings of coi1-20 where PstDC3000 reached a final concentration of approximately 106 cfu cm−2(Figure 1b).
coi1-20 plants also exhibited normal macroscopic tissue collapse indicative of the hypersensitive response (HR) when inoculated with high levels of PstDC3000 expressing the avirulence genes avrRpm1 or avrB (data not shown). Thus the coi1-20 mutation does not impair the basic ability of A. thaliana plants to mount an HR when challenged by an avirulent P. syringae strain. These results also demonstrate that RPM1-mediated pathogen recognition is functional in coi1-20 mutant plants.
Coi1-20 plants do not exhibit enhanced resistance to virulent fungal or viral pathogens.
To determine whether the coi1-20 mutant exhibited enhanced resistance against other virulent pathogens of A. thaliana, we examined its response to the fungal powdery mildew pathogen Erisyphe cichoracearum UCSC1, and the viral pathogen cauliflower mosaic virus (CaMV) (Adam and Somerville, 1996; Leisner and Howell, 1992; Melcher, 1989). Segregating F2 plants from a cross of coi1-20 to Col-0 were challenged with virulent E. cichoracearum UCSC1, and scored for disease symptoms 6 and 10 days following inoculation. As is summarized in Table 2, 100% of the plants exhibited extensive white, powdery fungal growth on leaf surfaces (Figure 1c). In two of these experiments, individual plants were allowed to flower following scoring of disease symptoms, in order to provide positive identification of coi1-20 homozygotes based on the male sterile phenotype. As expected, approximately 25% of the F2 plants were coi1-20 mutants (Table 2). Thus both coi1-20 homozygotes and their wild-type siblings appeared to be equally susceptible to E. cichoracearum.
Table 2. Inoculation of coi1-20 plants with virulent Erisyphe and CaMV isolates
Plant genotypes were determined by allowing self progeny from coi1-20/COI1 heterozygous plants to flower and scoring for sterility. The segregating families were comprised of approximately 25% coi1-20 and 75% wild-type, fertile individuals (coi1-20/COI1 and COI1/COI1; indicated as COI1/__).
coi1-20/COI1 seg, self progeny from coi1-20/COI1 heterozygous plants. The COI1 genotype was not determined in these experiments, and could not be obtained for CaMV-infected plants as they died prior to flowering.
Disease phenotype: +, plant exhibited typical disease symptoms; –, no disease symptoms evident. Disease phenotypes were determined by visual observation. For E. cichoracearum-infected plants, diseased plants were characterized by dense fungal growth and abundant sporulation. For CaMV-infected plants, diseased plants exhibited mosaic symptoms on the rosette leaves.
Segregating F2 plants from a cross of coi1-20 to Col-0 were challenged with the virulent CaMV isolate CM1841 and scored for disease symptoms daily from the point of inoculation until 42 days post-inoculation (dpi). No delay in disease progression was observed in any of the segregating F2 plants relative to the wild-type controls, and all the F2 individuals essentially exhibited typical systemic symptoms of viral pathogenesis (Table 2). If the coi1-20 plants did exhibit altered responses to CaMV, we would have expected to see approximately 25% of the plants exhibiting either more or less severe symptoms after infection. These results indicate that, although the coi1-20 mutant is strongly resistant to virulent strains of P. syringae, it exhibits no enhanced resistance to the virulent fungal or viral pathogens tested. Norman-Setterblad et al. (2000) have recently reported that coi1-1 plants are extremely susceptible to Erwinia carotovora, a bacterial pathogen not known to produce COR. Thus resistance in coi1 plants may be limited to bacterial pathogens that rely on COR as a key virulence factor.
coi1-20 resistance to PstDC3000 is correlated with rapid induction of PR-1
To investigate the molecular basis of enhanced resistance to P. syringae in coi1 mutants, we monitored expression levels of PR-1, a pathogenesis-related gene often used as a marker for SA-dependent host defense responses (Baker et al., 1997; Lamb et al., 1989), in infected coi1-20 plants. coi1-20 plants and their wild-type siblings were infiltrated with PstDC3000 and assayed for PR-1 expression at various times after infection. As is shown in Figure 2, the level of expression of PR-1 was undetectable in leaves of mature coi1-20 plants harvested immediately after infiltration with PstDC3000 (time 0, see Experimental procedures). Thus coi1-20 does not constitutively express PR-1 and does not belong to the constitutive expresser of PR (cpr) class of disease resistance mutants (Figure 2a). However, coi1-20 mutant plants responded to PstDC3000 infection by rapidly and strongly inducing PR-1 expression by 12 h post inoculation (Figure 2a). In contrast, expression of PR-1 was barely detectable in the susceptible, wild-type plants, even at 48 h post inoculation with PstDC3000. The rapid induction of PR-1 in coi1-20 plants is consistent with the enhanced resistance to PstDC3000 exhibited by the mutant. These results also suggest that coi1-20 plants are sensitized to infection with PstDC3000 and respond by rapidly inducing defense responses following pathogen attack. These results, taken in conjunction with the observations that coi1 mutants are insensitive to COR and that P. syringae mutants that do not produce COR exhibit reduced virulence on A. thaliana (Mittal and Davis, 1995; D. Brooks, A. Kloek and B. Kunkel, unpublished results), are consistent with the hypothesis that COR promotes virulence of P. syringae by inhibiting induction of host defense response upon pathogen attack (Mittal and Davis, 1995), and suggest that COR may function by modulating the SA-dependent defense pathway.
Rapid induction of PR-1 expression in coi1-20 plants on infection with PstDC3000 is dependent on SA signaling
As PR-1 belongs to the systemic acquired resistance (SAR) class of defense-related genes that are induced in response to elevated SA levels in the plant, we examined the role of SA and SA-dependent signaling pathways in resistance exhibited by the coi1-20 mutant. To generate coi1-20 lines that are impaired for SA signaling, coi1-20 was crossed to an A. thaliana Ler transgenic line carrying the nahG gene (Bowling et al., 1994) and to a Col-0 line carrying a mutation at the NPR1 locus (Cao et al., 1994). The nahG transgenic line carries a bacterial salicylate hydroxylase transgene which converts SA to the inactive product catechol, thus preventing accumulation of SA (Delaney et al., 1994; Gaffney et al., 1993). The npr1-1 mutant line carries a mutation that disrupts SA-dependent activation of defense-related transcripts (Cao et al., 1994). Both lines are extremely susceptible to infection with P. syringae (Cao et al., 1994; Delaney et al., 1994).
PR-1 expression in coi1-20, coi1-20 npr1-1 and coi1-20 nahG plants was monitored by RNA blot analysis 12 h after inoculation with PstDC3000. coi1-20 plants strongly expressed PR-1 12 h post-inoculation, whereas the coi1-20 npr1-1 and coi1-20 nahG plants failed to express detectable PR-1 message (Figure 2b). These results indicate that the rapid induction of PR-1 expression (and presumably other defense responses) in coi1-20 plants infected with PstDC3000 is dependent on both SA and NPR1.
Analysis of the requirement for SA in coi1-20 mediated resistance to PstDC3000
To assess the contribution of SA-mediated defenses to resistance to PstDC3000 in coi1-20, development of disease symptoms and pathogen growth on coi1-20 nahG and coi1-20 npr1-1 plants and their nahG (COI1/COI1 nahG or coi1/COI1 nahG) or npr1 (COI1/COI1 npr1 or coi1/COI1 npr1) siblings were monitored following inoculation with PstDC3000. As illustrated in Figure 3(a,b), both double-mutant lines appeared to be resistant to PstDC3000, and exhibited only slight chlorosis and few or no individual water-soaked lesions. In contrast, npr1-1 and nahG plants were extremely susceptible to PstDC3000, and developed severe chlorosis and numerous water-soaked lesions.
To determine if the phenotypic resistance was associated with restriction of pathogen growth in these lines, PstDC3000 was infiltrated into leaves of coi1-20 nahG, coi1-20 npr1-1, and their corresponding wild-type siblings (COI1/COI1 npr1 or coi1/COI1 npr1 and COI1/COI1 nahG or coi1/COI1 nahG), and bacterial growth monitored over a 4 day period. Surprisingly, coi1-20 nahG plants allowed very high levels of bacterial growth that were essentially indistinguishable from those observed in nahG plants (Figure 4a). The coi1-20 npr1-1 plants also supported high levels of bacterial growth, obtaining a level similar to that normally observed in wild-type plants (Figure 4b), but that was intermediate to levels observed in coi1-20 and npr1-1 plants (Figure 4c). Thus the restriction of PstDC3000 growth in coi1-20 mutant plants appears to be completely dependent on SA and partially dependent on NPR1.
These high levels of bacterial growth were unexpected, given the lack of disease symptoms on coi1-20 nahG and coi1-20 npr1 plants infected with PstDC3000 (Figure 3a,b). These results suggest that resistance to PstDC3000 in coi1-20 may involve two distinct mechanisms; an SA-dependent mechanism that limits pathogen growth, and an SA-independent mechanism that affects symptom development. However, the fact that the coi1 nahG line was generated in a mixed genetic background (e.g. by crossing Col-0 coi1-20 × Ler nahG) should be taken into consideration when interpreting the results of the bacterial growth experiments for the coi1 nahG plants.
coi1-20 plants accumulate elevated levels of SA in response to infection with PstDC3000
The repression of coi1-mediated restriction of bacterial growth by the nahG transgene suggests that an SA-dependent signaling pathway is critical in mediating some aspects of resistance to PstDC3000 in the coi1-20 mutant, and that this pathway has become either sensitized or partially de-repressed in the mutant. We envision two possible ways in which SA signaling could be altered in the coi1-20 plants. One hypothesis is that coi1-20 mutants are more sensitive to SA, such that even a small increase in SA levels (e.g. on infection with PstDC3000) results in unusually rapid and strong induction of SA-dependent defense responses. Alternatively, the coi1-20 mutant could produce elevated levels of SA, either constitutively or upon infection with PstDC3000.
To address the possibility that coi1-20 mutant plants might exhibit increased sensitivity to SA or similar chemical inducers of defense responses, we examined their response to benzothiadiazol (BTH), a synthetic analog of SA that has been shown to induce the SAR signal transduction pathway in A. thaliana (Lawton et al., 1996). We used RNA blot experiments to monitor the induction of PR-1 expression in coi1-20 and wild-type plants following application of BTH. Both coi1-20 and wild-type plants exhibited identical induction kinetics and responsiveness to a range of doses of BTH (50, 100 and 200 µm; data not shown). These results suggest that coi1-20 mutant plants do not exhibit a significant increase in sensitivity to BTH, which suggests that sensitivity to SA is also likely to be unaltered.
To test the second hypothesis, we monitored SA levels in the coi1-20 mutant. As shown in Figure 5, in uninoculated plants the mean levels of both free SA and SA glucoside (SAG) were higher in mutant tissue than in wild-type plants (Figure 5a,b; 0 h time points in Figure 5c,d). However, although we observed elevated levels of SA and SAG in uninoculated coi1-20 plants in four independent experiments (Figure 5 and data not shown), these levels were not significantly higher than those observed in wild-type plants (P > 0.05 in unpaired t-tests for all experiments). In contrast, within 24 h following inoculation with PstDC3000, both SA and SAG levels increased to significantly higher levels in coi1-20 plants than in wild-type plants (Figure 5c,d). These results are consistent with our observations that the SA-signaling pathway is sensitized in the coi1-20 mutant (Figure 2), and suggest that this occurs, at least in part, at the level of SA synthesis.
Our observation that mean SA levels were slightly elevated in uninoculated coi1-20 plants may indicate that the levels of SA and SAG in uninoculated coi1 plants are more highly variable than observed in wild-type plants. Although this may result in slightly higher than normal amounts of SA in uninoculated plants, these levels are apparently insufficient to induce constitutive expression of PR-1 (Figure 2). However, slightly elevated levels of SA in coi1-20 plants could serve to potentiate activation of the defense response-signaling pathway on infection with PstDC3000 (Shirasu et al., 1997).
Other A. thaliana mutants defective in JA production or signaling are susceptible to PstDC3000
To investigate whether the enhanced resistance to PstDC3000 is specific for coi1 mutants, or alternatively whether this phenotype is associated with a general defect in JA signaling, we assayed resistance to PstDC3000 in two other JA-related mutants, the fad3-2 fad7-2 fad8 triple mutant (McConn and Browse, 1996), and jar1-1 (Staswick et al., 1992). The fad3-2 fad7-2 fad8 line is deficient in several fatty acid desaturase activities and is incapable of producing linolenic acid, the lipid precursor of JA, and is therefore unable to accumulate JA (McConn and Browse, 1996). Like the coi1 mutants, fad3-2 fad7-2 fad8 is male sterile. However, this mutant is able to respond to JA (McConn et al., 1997), and is presumably sensitive to COR as well. The jar1-1 mutant, which was isolated in a screen for mutants exhibiting decreased sensitivity to JA (Staswick et al., 1992), is deficient in JA-induced expression of a subset of vegetative storage proteins, but is male fertile, and retains a significant degree of sensitivity to JA and COR when scored in root elongation assays (Staswick et al., 1992; V. Joardar, G. Kalinowski and B. Kunkel, unpublished results). The jar1-1 mutant has been reported to be susceptible to infection by PstDC3000 (Pieterse et al., 1998). However, studies investigating resistance to P. syringae in the fad triple mutant have not been reported.
The fad3-2 fad7-2 fad8 and jar1-1 mutants were infected with PstDC3000 and scored for disease symptoms. Unlike the coi1-20 mutant, both the fad3-2 fad7-2 fad8 and jar1-1 plants were fully susceptible to infection by PstDC3000, and exhibited numerous water-soaked lesions and extensive chlorosis (Figure 3c,d). Thus among the JA-signaling mutants tested, resistance to PstDC3000 appears to be unique to COI1 mutants.
In a screen for A. thaliana mutants with enhanced disease resistance we isolated a new allele of COI1, coi1-20. Consistent with previous finding by Feys et al. (1994), this mutant exhibits strong resistance to P. syringae, consisting of the complete absence of visible disease symptoms and restriction of bacterial growth within mutant leaf tissue (Figure 1a,b). coi1-20 mutant plants do not exhibit any significant alteration in their responses to the virulent fungal pathogen E. cichoracearum(Figure 1c; Table 2) and the viral pathogen CaMV, suggesting that the enhanced resistance exhibited by the coi1-20 mutant could be limited to phytopathogenic pseudomonads. These findings are interesting given that in other studies the coi1-1 mutant exhibited increased susceptibility to several fungal plant pathogens including Pythium sp. (Vijayan et al., 1998), Alternaria brassicicola (Thomma et al., 1998), and Botrytis cinerea (Thomma et al., 1998), as well as to Erwinea carotovora (Norman-Setterblad et al., 2000), a macerating bacterial pathogen. Alternaria brassicicola, B. cinerea and Pythium sp. are not usually capable of causing disease on healthy A. thaliana plants, and resistance against all of these pathogens is thought to be controlled by JA-mediated defense responses (Dong, 1998; Staswick et al., 1992). Therefore it is not surprising that the coi1-1 mutant, with its insensitivity to JA, exhibits increased susceptibility to these pathogens. However, it is somewhat counter-intuitive that coi1 mutants exhibit strong resistance to P. syringae, a pathogen known to be controlled by SA-dependent defense responses (Dong, 1998). This finding prompted us to examine more closely the role of SA-dependent signaling in coi1-20 resistance to PstDC3000.
We have demonstrated that resistance to PstDC3000 in coi1-20 is correlated with enhanced induction of PR-1 expression on infection with PstDC3000 (Figure 2a). Although coi1-20 plants do not constitutively express PR-1, expression of this gene is much more rapidly and strongly induced following infection with PstDC3000 than is observed in wild-type plants. These findings suggest that in the coi1-20 mutant, the SA-mediated defense pathway is sensitized to respond to infection by PstDC3000. Other A. thaliana disease resistance mutants that exhibit enhanced activation of defense responses upon pathogen invasion have been reported (Frye and Innes, 1998; Vogel and Somerville, 2000), but none that we are aware of exhibits as rapid and strong a response as that shown by coi1-20.
As further evidence that SA-mediated defenses are required for coi1-20 resistance to PstDC3000, we found that coi1-20 nahG and coi1-20 npr1-1 double mutants do not express PR-1 in response to PstDC3000 (Figure 2b) and support elevated levels of pathogen growth compared to coi1-20 plants (Figure 4). In fact, both coi1-20 nahG plants and nahG plants allow indistinguishably high levels of pathogen growth, suggesting that the ability to restrict growth of PstDC3000 is fully dependent on SA. In contrast, coi1-20 npr1-1 plants allowed intermediate levels of pathogen growth (less than npr1-1 but more than the coi1-20 mutant), suggesting that SA signaling activates other defense pathways in addition to those regulated by NPR1. Evidence for NPR1-independent defense responses has also been reported in studies of cpr6, ssi1 and acd6 (Clarke et al., 1998; Rate et al., 1999; Shah et al., 1999). However, although the npr1-1 mutant appears to completely lack NPR1 function (Cao et al., 1994); X. Dong, personal communication), the fact that npr1-1 is a mis-sense mutation (Cao et al., 1997), rather than a true null allele, must be taken into consideration when interpreting these results.
The rapid and strong activation of defense responses in coi1-20 is correlated with a significantly greater increase in SA levels on infection by PstDC3000 than is observed in wild-type plants (Figure 5). These findings support our conclusion that resistance to PstDC3000 in coi1 plants is due, in part, to enhanced signaling through the SA-dependent defense pathway, and further suggest that this is mediated through increased SA synthesis. Based on these results, we hypothesize that COI1 is involved in modulation of SA signaling in response to infection by PstDC3000.
In addition to playing a role in resistance to P. syringae, we found that COI1 also affects symptom development. Despite the extensive bacterial growth supported by the coi1-20 nahG and coi1-20 npr1-1 double mutants, neither line develops significant disease symptoms (Figure 3a,b). These results indicate that the wild-type COI1 allele is required for development of visible signs of disease regardless of the levels of pathogen growth, and are consistent with the hypothesis that the plant host itself contributes actively to the production of disease symptoms (Greenberg et al., 2000; Morel and Dangl, 1997). In the light of these findings, resistance to PstDC3000 in coi1-20 appears to be conferred by two different mechanisms: (i) restriction of pathogen growth via an SA-dependent defense mechanism; and (ii) an SA-independent inability to develop disease symptoms.
The finding that the coi1-20 nahG plants can support extremely high levels of bacteria without showing any visible disease symptoms, a phenotype often referred to as tolerance, is quite remarkable. Tolerance to bacterial pathogens in A. thaliana has been reported in only a few cases (Bent et al., 1992; Buell and Somerville, 1995; Buell and Somerville, 1997; Tsuji et al., 1991). For example, on infection with virulent strains of P. syringae, the ethylene-insensitive ein2 mutant supports high levels of bacterial growth, yet it exhibits only mild disease symptoms (Bent et al., 1992). The mechanism underlying tolerance in these plants is not understood. However, the fact that both ein2 and coi1 plants exhibit reduced symptoms and are impaired in JA-responsive gene induction (Penninckx et al., 1996) suggests that both the ethylene- and JA-signaling pathways are involved in disease symptom development.
The molecular basis of P. syringae-induced disease lesion formation is not well understood. Several reports in the literature indicate that the P. syringae phytotoxin COR contributes to the formation of lesions on tomato, A. thaliana and soybeans (Bender et al., 1987; Budde and Ullrich, 2000; Mittal and Davis, 1995). This hypothesis is strengthened by our findings that COR-insensitive coi1-20 nahG and coi1-20 npr1 plants show no significant signs of disease despite high levels of pathogen growth. Ethylene has also been demonstrated to be important in later stages of P. syringae disease development, where it promotes lesion expansion and chlorosis (Bent et al., 1992; Lund, 1998).
Our finding that the coi1-20 mutant exhibits elevated SA signaling after infection with P. syringae is consistent with the hypothesis that the JA-signaling pathway negatively regulates induction of SA-dependent defenses. However, we found that other A. thaliana mutants defective in either JA production (fad3-2 fad7-2 fad8) or perception (jar1-1) exhibited wild-type susceptibility to PstDC3000 (Figure 3c,d), suggesting that resistance to P. syringae is not a common feature of all mutants that affect JA signaling. Rather, this may be specific to coi1 mutants, which are unique in that they appear to be completely impaired in perception of MeJA and COR. Thus we hypothesize that the enhanced resistance phenotype of coi1 plants is due primarily to the fact that these mutants are insensitive to COR.
It is likely that the enhanced resistance phenotype of the coi1 mutants provides insight into the role of COR in promoting virulence of P. syringae. Our data – taken in conjunction with observations that P. syringae mutants that do not produce COR exhibit reduced virulence on A. thaliana (Mittal and Davis, 1995; D. Brooks, A. Kloek and B. Kunkel, unpublished results) – are consistent with the hypothesis that COR promotes virulence of P. syringae by inhibiting host defense responses (Budde and Ullrich, 2000; Mittal and Davis, 1995). We propose that COR accomplishes this by acting in a COI1-dependent manner to interfere with SA signaling. The ability to inhibit or delay activation of SA-dependent host defense responses is likely to be an important trait of pathogenic bacteria, as this may provide a window of opportunity during which the pathogen can colonize the host tissue. The molecular mechanisms through which COR achieves this are not understood, and are currently under investigation.
Bacterial strains and plasmids
The bacterial pathogen strains Pseudomonas syringae pv. tomato (Pst DC3000, Pst 3455) and P. syringae pv. maculicola m4 have been described previously (Debener et al., 1991; Whalen et al., 1991). Pseudomonas syringae strains were cultured at 28°C in King's B medium (King et al., 1954) containing 50 µg ml−1 rifampicin. The fungal pathogen Erysiphe cichoracearum UCSC1 was maintained as described previously (Vogel and Somerville, 2000). The viral pathogen CaMV strain CM1841 (Gardner et al., 1981) was maintained on turnip (Brasssica rapa‘Just Right’) or in lyophilized turnip tissue stored at 4°C. (Schoelz et al., 1986).
Plant material, growth conditions and inoculation procedures
The coi1-20 mutant was isolated from a population of M2 plants derived from EMS-mutagenized seed of a Columbia (Col-0)/Nossen (No-0) hybrid A. thaliana line carrying the rps2-201C mutation (Boch et al., 1998). The coi1-20 plants used in the experiments described here were derived from self-fertilized seed harvested from a heterozygous F2coi1-20/COI1 individual generated from a second back-cross of coi1-20 rps2-201C to Col-0 rps2-201C (Kunkel et al., 1993) or from a second back-cross of coi1-20 rps2-201C to wild-type Col-0 plants. As the presence or absence of the rps2 mutation in the coi1-20 plants is not relevant to the experiments described in this work, for the sake of clarity we have chosen not to refer to the RPS2 allele present in each plant line. As coi1 plants are male sterile, coi1-20 mutant lines were maintained as heterozygotes.
The Ler nahG transgenic line (Bowling et al., 1994) and the Col-0 npr1-1 mutant (Cao et al., 1994) were obtained from Scott Bowling and Xinnian Dong (Duke University). The coi1-1 mutant line (Feys et al., 1994) was obtained from John Turner (University of East Anglia). The jar1-1 (Staswick et al., 1992) and the fad 3-2 fad7-2 fad8 (McConn and Browse, 1996) mutant lines were obtained from Paul Staswick (University of Nebraska) and John Browse (Washington State University), respectively. Arabidopsis thaliana plants were grown from seed in growth chambers under an 8 h photoperiod at 24°C.
Pseudomonas syringae inoculation of plants was carried out by dipping entire leaf rosettes of 3–5-week-old plants into bacterial suspensions of 2–4 × 108 colony forming units (cfu) ml−1 containing 0.02% of the surfactant Silwet L-77 (OSi Specialties Inc., Danbury, CT, USA) and 10 mm MgCl2 as previously described (Kunkel et al., 1993). Bacterial growth within leaf tissue was monitored as described by Whalen et al. (1991).
Arabidopsis thaliana plants were inoculated with E. cichoracearum using settling towers as previously described (Vogel and Somerville, 2000). CaMV virions were partially purified from infected turnip leaves according to Schoelz et al. (1986) to concentrate the inoculum. The inoculum was applied to glass rods, then rubbed onto A. thaliana leaves that had been lightly dusted with carborundum.
For bacterial growth experiments in coi1-20 plants, 15–20 progeny derived from self-fertilized seed harvested from a heterozygous coi1-20/COI1 individual were infiltrated with PstDC3000 suspensions of 1 × 105 cfu ml−1. Data were collected from each individual inoculated plant over a 4-day period. coi1-20 homozygous plants were identified based on their resistance and male-sterile phenotypes, and data from individual coi1-20 plants were then pooled. Data from the phenotypically wild-type plants (coi1-20/COI1 and COI1/COI1) were also pooled and presented as controls in these experiments.
Genetic analysis and mapping
The coi1-20 rps2-201C mutant was crossed to both Col-0 rps2-201C (Kunkel et al., 1993) and wild-type Col-0 to determine the genetic basis of resistance. Data from the cross to Col-0 are presented in Table 1. The F2 progeny from the cross of the coi1-20 rps2-201C mutant to Col-0 rps2-201C segregated in a ratio of 97 resistant plants to 29 susceptible plants (χ23 : 1= 0.27; P > 0.5). Complementation tests between coi1-1 (Feys et al., 1994) and coi1-20 were conducted by fertilizing homozygous coi1-20 plants with pollen from heterozygous coi1-1/COI1 plants. Non-complementation between coi1-1 and coi1-20 was established by the appearance of male-sterile, PstDC3000-resistant F1 progeny (coi1-20/coi1-1) occurring in a 1 : 1 ratio with fertile, susceptible (coi1-20/COI1) F1 plants (Table 1). The phenotypically wild-type F1 plants were allowed to self-pollinate, and gave rise to approximately 25% male-sterile, resistant progeny.
Genetic linkage analysis using co-dominant cleaved amplified polymorphic DNA sequences (CAPS) (Konieczny and Ausubel, 1993) was performed utilizing progeny from the cross between coi1-20 and Ler plants (Table 1). Genomic DNA was isolated from leaf tissue of F2 plants from this cross according to the procedure of Tai and Tanksley (1990), with modifications as described by Kunkel et al. (1993).
Coronatine, methyl jasmonate and BTH assays
The sensitivity of coi1-20 mutants to COR and MeJA was assayed by germinating segregating families of the mutants on Murashige and Skoog (MS) (Murashige and Skoog, 1962) plates containing 1 µm coronatine or 10 µm MeJA (Feys et al., 1994). Homozygous coi1-20 plants were identified by their root-elongation phenotype 6 days following germination on either COR or MeJA MS plates. Putative coi1-20 mutants were verified by transferring the seedlings to soil, growing them to maturity, and scoring for male sterility.
Arabidopsis thaliana plants were sprayed to run-off with 50, 100 or 200 µm benzo(1,2,3)thiadiazole-7-cabothioic acid (BTH, Novartis Inc., Research Triangle Park, NC, USA) dissolved in distilled H2O (Lawton et al., 1996). BTH-treated plants were grown in a separate growth chamber, and tissue samples were harvested and frozen in liquid nitrogen at various times following treatment. Tissue samples were stored at −80°C until coi1-20 mutants could be identified by scoring for male sterility. coi1-20 and wild-type (coi1-20/COI1 and COI1/COI1) tissues were pooled and saved for RNA isolation and RNA blot analyses.
Generation of coi1-20 npr1-1 and coi1-20 nahG double mutants
coi1-20 npr1-1 and coi1-20 nahG double mutants were generated using homozygous coi1-20 mutant plants as the female parent in crosses with npr1-1 and nahG plants. F2 lines homozygous for npr1-1 or nahG and heterozygous for coi1-20 were identified by screening F3 progeny using seedling assays. npr1 homozygous plants become bleached when grown in the presence of SA (Cao et al., 1994; Cao et al., 1998). This assay facilitated the identification of coi1/COI1 npr1/npr1 lines that gave rise to approximately 25% bleached, MeJA-insensitive seedlings when germinated on MS plates containing 125 nm SA, 10 µm MeJA. coi1/COI1 nahG/nahG lines were identified as giving rise to approximately 25% kanamycin-resistant, MeJA-insensitive seedlings on MS plates containing 50 µg ml−1 kanamycin, 10 µm MeJA, utilizing the linked kanamycin resistance marker to select for the presence of the nahG transgene (Bowling et al., 1994).
RNA isolation and Northern analysis
Total RNA was isolated from A. thaliana leaf tissue using the RNeasy Plant RNA isolation kit (Qiagen, Chatsworth, CA, USA). RNA gel-blot analysis was carried out according to Sambrook et al. (1989). Total RNA (2–3 µg) was loaded in each lane. Hybridization probes were prepared using the Prime-it II kit (Stratagene, La Jolla, CA, USA). The A. thaliana cDNA corresponding to the PR-1 gene was used as a probe (Uknes et al., 1992). As a loading standard, a 3.7 kb EcoRI fragment from the 10 kb genomic region carrying A. thaliana rRNA genes was used (Vongs et al., 1993). The RNA blots were analyzed using X-ray film or a phosphorimager (Molecular Dynamics, Sunnyvale, CA, USA).
For RNA analysis of inoculated coi1-20 plants, 15–20 progeny derived from self-fertilized seed harvested from a heterozygous coi1-20/COI1 individual were infiltrated with PstDC3000 suspensions of 1 × 105 cfu ml−1. Tissue from individual plants was harvested at various times post-infiltration and flash-frozen in liquid nitrogen. Tissue samples were stored at −80°C until coi1-20 homozygous plants could be identified based on their male sterile phenotypes. Tissue from at least three homozygous coi1-20 mutants was pooled for each time point and used for RNA isolation. Tissue from phenotypically wild-type plants (coi1-20/COI1 and COI1/COI1) tissues were also pooled and subject to RNA blot analyses.
Progeny derived from self-fertilized seed harvested from a heterozygous coi1-20/COI1 individual were inoculated with PstDC3000 by dipping into suspensions of 2–4 × 108 cfu ml−1 containing 0.02% Silwet L-77. Tissue from individual plants was harvested just prior to inoculation (time 0) and at 12 and 24 h post-infection. Plants were scored for disease symptoms at 4 dpi, then placed in the greenhouse to flower. Tissue samples were stored at −80°C until coi1-20 homozygous plants could be verified by scoring for male sterility. Tissue from the resistant, male sterile plants was combined to generate three or four pools of coi1-20 tissue for each time point. Tissue from phenotypically wild-type siblings (coi1-20/COI1 and COI1/COI1) was combined to generate three or four pools of COI1/__ tissue for each time point. SA and SAG were extracted, and levels assayed as described by Bowling et al. (1994).
We thank Julia Lifitz for technical assistance and Zhongying Chen, Jean Greenberg, Grant Kalinowski and Rodolfo Zentella for helpful discussion and comments on the manuscript. We are grateful to Carol Bender (Oklahoma State University) for the generous gift of coronatine, to Novartis, Inc. for providing BTH, to John Turner (University of East Anglia), Paul Staswick (University of Nebraska) and John Browse (Washington State University) for seed from the coi1-1, jar1-1 and fad triple mutant lines, respectively, and to Xinnian Dong (Duke University) for providing the npr1-1 mutant and nahG transgenic line. A.P.K. was a DOE-Energy Biosciences Research Fellow of the Life Sciences Research Foundation. This work was supported by an NIH grant (GM52536) awarded to B.N.K and an NSF grant (MCB9723952) awarded to D.F.K.