Loss of a Callose Synthase Results in Salicylic Acid-Dependent Disease Resistance

See allHide authors and affiliations

Science  15 Aug 2003:
Vol. 301, Issue 5635, pp. 969-972
DOI: 10.1126/science.1086716


Plants attacked by pathogens rapidly deposit callose, a β-1,3-glucan, at wound sites. Traditionally, this deposition is thought to reinforce the cell wall and is regarded as a defense response. Surprisingly, here we found that powdery mildew resistant 4 (pmr4), a mutant lacking pathogen-induced callose, became resistant to pathogens, rather than more susceptible. This resistance was due to mutation of a callose synthase, resulting in a loss of the induced callose response. Double-mutant analysis indicated that blocking the salicylic acid (SA) defense signaling pathway was sufficient to restore susceptibility to pmr4 mutants. Thus, callose or callose synthase negatively regulates the SA pathway.

Plants defend themselves from pathogens with a variety of chemical and physical defenses (1). As first reported in 1863 by de Bary, the most prominent physical defense is the rapid synthesis of callose, an amorphous, high-molecular-weight β-1,3-glucan (2). During fungal infections, callose is deposited in cell wall appositions (papillae) that form beneath infection sites and are thought to provide a physical barrier to penetration (3). One of the collection of Arabidopsis mutants resistant to the powdery mildew Erysiphe cichoracearum, pmr4, is resistant to other biotrophic pathogens (E. orontii and Peronospora parasitica), and resistance appears to act after the pathogen has penetrated the plant cell wall (4). Also, pmr4 produces dramatically less callose in response to powdery mildew infection or wounding.

We mapped the PMR4 gene to the top arm of chromosome 4 (4), a region that contains a glucan synthase-like gene (GSL5 = CalS12, At4g03550). Both GSL5 mRNA and callose accumulate in a mitogen-activated protein kinase mutant (mpk4) that constitutively expresses systemic acquired resistance (5). A GSL5 cDNA can partially complement a callose synthase–deficient yeast mutant (5). We tested a heterozygous line with a transferred DNA (T-DNA) insertion in GSL5 and found that it segregated for powdery mildew resistance (6). Nine additional ethylmethane sulfonate alleles of pmr4, isolated in separate screens, were sequenced to confirm the identity of pmr4 (table S1) (6). Eight had a mutation that introduced a stop codon within GSL5. The exception, pmr4-2, had a splicing mutation at the second intron-exon junction (table S1). Thus, PMR4 encodes a callose synthase responsible for producing callose in response to biotic and abiotic stresses (see below), and the other eleven Arabidopsis callose synthases cannot compensate for loss of PMR4.

To verify that infection and wound callose were missing from pmr4 plants, we visualized callose using a callose-specific monoclonal antibody (6). The localization of callose to papillae and wound sites was qualitatively similar to our previous results with aniline blue staining (4). Callose was present in wild-type cells adjacent to wound sites, but almost completely absent at pmr4-1 wound sites (Fig. 1, A and B). Callose in wild-type papillae appeared in rings surrounding the point of penetration by the powdery mildew pathogen (Fig. 1C). Penetration sites in pmr4-1 plants had either low levels of callose or lacked callose altogether.

Fig. 1.

Callose deposition at papillae and wound sites in wild type and pmr4-1. (A and B) Callose deposition (green fluorescence) at 24 hours after wounding in (A) Col-0, wild-type and (B) pmr4-1. Scale bar, 50 μm. (C and D) Callose deposition around penetration sites (ps) formed in response to powdery mildew penetration 3 days after inoculation in Col-0 (C) and pmr4-1 (D). Penetration sites in (D) pmr4-1 have greatly reduced amounts of callose but are identifiable by the presence of haustoria. Callose was detected with a monoclonal antibody. Plant cell walls, powdery mildew hyphae (h), and conidia (c) were stained with propidium iodide (red fluorescence). Penetration sites (ps) are indicated in (C) and (D) by arrows. Small callose deposits (arrowheads) seen in (C) and (D) are at the junctions of stomatal guard cells. Scale bar, 10 μm. (E and F) Transmission electron micrographs of papillae, 1 day after inoculation with powdery mildew in Col-0 (E) and pmr4 (F). Other labels are w, host cell wall; p, papilla; hc, host cell. Scale bar, 0.5 μm.

Because callose is a major component of papillae, we used transmission electron microscopy to determine whether the overall structure of papillae in pmr4 plants was grossly different from that in wild type (6). Even in the absence of callose, pmr4-1 plants were able to form papillae (Fig. 1F). Furthermore, these papillae were not significantly different from papillae formed in wild-type plants in size or shape. The main ultrastructural difference was a prominent fibrillar network in the pmr4-1 papilla (Fig. 1, E and F). Thus, papillae formation occurs independently of callose deposition.

Arabidopsis normally produces callose after treatment with the cellulose synthase inhibitor dichlobenyl (7), treatment with chitin oligomers, or inoculation with the barley pathogen Blumeria graminis f.sp. hordei (6). pmr4 plants do not produce callose after any of these treatments (8). Callose production, aside from stress-related callose, was unaffected in pmr4 plants. Callose deposition was normal in pmr4-1 in plasmodesmata and at the junctions of stomatal guard cells (Fig. 1, C and D), as well as in pollen tubes, cell plates, and the vasculature (8), sites where callose accumulates during the course of development (9). Thus, PMR4 appears to be the main biosynthetic enzyme responsible for the callose response to biotic, abiotic, and chemical stresses, but not for callose deposition that is part of normal plant development.

The pmr4 mutants develop at the same rate as wild-type and do not outwardly appear to be stressed (4). This lack of severe pleiotrophic phenotypes might be expected as callose is not a major component of unstressed plant cell walls (9). Consistent with this, no constitutive changes in the cell walls were detected by Fourier transform infrared spectrometry when wild-type and pmr4-1 plants were compared (10). Thus, disease resistance was unlikely to develop as an indirect result of a nonspecific stress or as a consequence of abnormal cell wall formation.

When infected with powdery mildew, the pmr4 mutants developed lesions reminiscent of hypersensitive cell death. All of the pmr4 mutants developed lesions at infection sites that were not observed in the compatible wild-type interaction (Fig. 2, A and B). Dead cells were visible microscopically after trypan blue staining as early as 3 days after inoculation with powdery mildew (Fig. 2, C and D) (6). These early lesions were associated with fungal growth and appeared under both penetration sites and conidia. The lesions became visible to the eye about 10 to 14 days after inoculation and did not spread beyond infection sites.

Fig. 2.

pmr4 plants develop lesions. (A) Single leaves 10 days after inoculation with powdery mildew. Col-0 (left) does not develop lesions. The two pmr4-1 leaves, center and right, demonstrate the range of lesion sizes. (B) Higher magnification of right-hand leaf from (A). Small lesions (dc) are present in all pmr4 plants after infection. Scale bar, 100 μm. (C and D) At higher magnification, lesions can be observed in pmr4-1 (D) but not Col-0 (C) under conidia (c) and hyphae (h) 4 days after inoculation, following trypan blue staining for dead cells (dc). Scale bar, 25 μm.

Because resistance and hypersensitive cell death are often regulated by salicylic acid (SA) or ethylene and jasmonate signaling, we crossed the pmr4-1 mutant with a set of mutants having defects in components of these pathways (6, 11). The resistance phenotypes of pmr4-1 Etr1-1 and pmr4-1 coi1-1 double mutants were indistinguishable from pmr4-1 plants, which indicated that the ethylene and jasmonate signaling pathways did not contribute to resistance in pmr4 plants (Fig. 3A).

Fig. 3.

pmr4 resistance acts via the SA pathway but not the ethylene and jasmonate pathways. (A) Double-mutant analysis of the SA signal transduction pathway (pad4, NahG, npr1) and of the ethylene and jasmonate signal transduction pathways (Etr1, coi1) in pmr4 resistance. (B) Leaves stained with trypan blue. Trypan blue stains fungal propagules (p) and dead host-plant cells (dc). (C) Aniline blue staining for callose. Callose-rich papillae are visible in the wild-type Col-0 (wt). Leaf hairs (t) autofluoresce. Scale bar, 50 μm. In (A), (B), and (C), plants were photographed 10 days after inoculation with powdery mildew.

In contrast, mutations in the SA pathway were capable of restoring susceptibility to pmr4-1. Both pmr4-1 NahG plants and pmr4-1 pad4-1 plants supported wild-type levels of fungal growth and asexual reproduction (Fig. 3, A and B). The restoration of susceptibility was associated with decreased lesion formation but was not accompanied by restoration of callose in the papillae (Fig. 3C). Thus, callose is not required as a compatibility factor for fungal growth. Because blocking SA signaling suppressed the lesion phenotype, cell death was likely to be a consequence of SA pathway activation and not a direct result of the loss of callose in pmr4-1 mutants. The pmr4-1 npr1-1 plants supported an intermediate level of fungal growth. Thus, pmr4-based resistance probably acts through both NPR1-dependent and NPR1-independent branches of the SA pathway (11, 12).

The double-mutant analysis implicated the SA pathway as a mechanism of pmr4-based resistance. To better understand the effects of the pmr4 mutation on downstream targets of SA signaling, we compared the gene expression of wild-type and pmr4-1 plants using full-genome microarrays (6). Four biological replicates of powdery mildew infected and uninfected plants were examined to identify any constitutive or induced expression changes in pmr4-1 relative to wild-type. The data were filtered for consistency of expression among replicates, and genes showing at least a two-fold change in expression were retained (13). Of the 24,000 genes on the microarray, 685 genes passed these filters (table S2) and fell into two main clusters; genes induced by and genes repressed by pathogen attack (Fig. 4). A common pattern was that SA and pathogen-responsive genes were up-regulated in pmr4-1 plants and that this up-regulation further increased after infection (Fig. 4). These hyperinduced genes included a variety of known SA and pathogen-inducible genes. Common marker genes for SA-dependent systemic acquired resistance, including members of the PR-1, PR-2, PR-3, and PR-5 families, were all up-regulated in pmr4 plants (Table 1), as were the SA biosynthesis genes isochorismate synthase and a MATE (multidrug and toxin extrusion) transporter (14, 15). Of the 50 most highly up-regulated genes, more than half fell into SA- and pathogen-induced groups. Expression of the plant defensins, PDF1.2a and PDF1.2b, was down-regulated in pmr4-1 plants, which was consistent with reports that activation of the SA pathway is associated with repression of the ethylene and jasmonate pathways (16). Thus, it appears that the basis for the resistance observed in pmr4 plants was an enhanced activation of the SA signal transduction pathway.

Fig. 4.

pmr4-1 has a hyperinduced SA response to powdery mildew. Hierarchical display of ratio values (relative to uninfected Col-0) of 685 genes with altered expression in pmr4-1 mutants. Red indicates up-regulated genes; green, down-regulated genes; and yellow, genes with no change in expression. Infected samples were taken 3 days after inoculation. CI, infected Col-0; CU, uninfected Col-0; PI infected pmr4-1; PU, uninfected pmr4-1.

Table 1.

Examples of pathogen- and SA-induced genes that are hyperinduced or repressed in pmr4-1 after infection. Numbers indicate average fold induction relative to uninfected Col-0. CI, infected Col-0; PI, infected pmr4-1; PU, uninfected pmr4-1.

Locus Annotation CI PI PU
Pathogen- or SA-induced
At2g14610 PR-1—like 6.3 11.7 2.4
At3g57260 PR-2 7.9 13.4 2.3
At2g43570 Endochitinase isolog 11.2 25.2 3.1
At1g75040 Thaumatin-like (PR-5) 10.0 18.7 2.4
At4g34135 Salicylate-induced glucosyltransferase 4.0 9.0 1.9
At1g33960 AIG1 (Pseudomonas inducible) 5.8 22.3 2.6
At1g65690 HIN1-like (harpin-inducible in tobacco) 6.9 13.2 2.4
At4g10500 SRG1-like (stress-induced ascorbate ox) 11.7 26.0 3.1
At2g29350 Tropinone reductase (alkaloid biosynthesis) SA biosynthesis 7.2 12.6 2.0
At1g74710 Isochorismate synthase (EDS16 = SID2) 3.9 6.3 1.5
At4g39030 MATE-like transporter (EDS5 = SID1) SA repressed 2.9 4.2 1.4
At5g44420 PDF1.2a 1.1 0.3 0.5
At2g26020 PDF1.2b 1.0 0.4 0.4

One hypothesis to explain the pmr4 resistance phenotype is that callose could be an induced defense response (5, 9, 17), which then acts to limit further defense responses. This negative feedback would contain potentially damaging defense responses, similar to the feedback regulation of cell death by LSD1, which acts both upstream and downstream of SA (18). In the pmr4 mutants, feedback regulation is removed, which derepresses defense responses and leads to enhanced resistance and cell death. In any case, lack of either callose or the PMR4 protein is capable of triggering the SA pathway in unexpected ways. The commonly held view that callose contributes to resistance by physically limiting fungal penetration needs to be reevaluated in light of these results.

The cell wall has recently been linked to biotic stress signaling by a cellulose synthase (CESA) mutant, cev1, which is resistant to fungal pathogens and constitutively activated for defense pathways (19, 20). The rat1 mutant, resistant to Agrobacterium tumefaciens, carries a defect in a cell-surface arabinogalactan protein and has elevated PR1 mRNA levels relative to wild type (21, 22), reminiscent of our observations with pmr4. These mutants and the properties of the pmr4 mutant suggest a role for cell wall components in reporting on the status of the extracellular environment to infected or wounded cells.

Supporting Online Material

Materials and Methods

Tables S1 to S3

References and Notes

View Abstract

Navigate This Article