Report

Plant Pathogen Recognition Mediated by Promoter Activation of the Pepper Bs3 Resistance Gene

See allHide authors and affiliations

Science  26 Oct 2007:
Vol. 318, Issue 5850, pp. 645-648
DOI: 10.1126/science.1144958

Abstract

Plant disease resistance (R) proteins recognize matching pathogen avirulence proteins. Alleles of the pepper R gene Bs3 mediate recognition of the Xanthomonas campestris pv. vesicatoria (Xcv) type III effector protein AvrBs3 and its deletion derivative AvrBs3Δrep16. Pepper Bs3 and its allelic variant Bs3-E encode flavin monooxygenases with a previously unknown structure and are transcriptionally activated by the Xcv effector proteins AvrBs3 and AvrBs3Δrep16, respectively. We found that recognition specificity resides in the Bs3 and Bs3-E promoters and is determined by binding of AvrBs3 or AvrBs3Δrep16 to a defined promoter region. Our data suggest a recognition mechanism in which the Avr protein binds and activates the promoter of the cognate R gene.

Resistance (R) proteins, a class of plant immune receptors that mediate recognition of pathogen-derived avirulence (Avr) proteins, are a well-studied facet of the plant defense system (1). The bacterial plant pathogen Xanthomonas campestris pv. vesicatoria (Xcv) uses a type III secretion (T3S) system to inject an arsenal of about 20 effector proteins into the host cytoplasm that collectively promote virulence (2). R protein–mediated defense in response to Xcv effector proteins is typically accompanied by a programmed cell death response referred to as the hypersensitive response (HR).

One Avr protein that R proteins recognize is AvrBs3, a member of a Xanthomonas family of highly conserved proteins (3). The central region of AvrBs3 consists of 17.5 tandem near-perfect 34–amino acid repeat units that determine avirulence specificity (4). AvrBs3 also contains nuclear localization signals (NLSs) and an acidic transcriptional activation domain (AD) (5, 6), similar to eukaryotic transcription factors, and induces host gene transcription (7). Mutations in the NLS or AD of AvrBs3 abolish pathogen recognition by the matching pepper R gene Bs3 (5, 8), which suggests that recognition involves the transcriptional activation of host genes.

Previously we identified bacterial artificial chromosome (BAC) clones derived from the pepper (Capsicum annuum) cultivar Early California Wonder 30R (ECW-30R) that cover the Bs3 gene (9). For complementation-based identification, fragments of a Bs3-containing BAC (9) were cloned into a plant transformation vector and were delivered into Nicotiana benthamiana leaves via Agrobacterium tumefaciens–mediated transient transformation. Two nonidentical clones carrying the same coding sequence triggered an HR in N. benthamiana when cotransformed with avrBs3. A genomic DNA fragment containing only the predicted coding sequence and ∼1 kb of sequence upstream of the ATG mediated AvrBs3 recognition, confirming that this gene is Bs3 (Fig. 1A).

Fig. 1.

(A) Recognition specificity of the Bs3 allele from ECW-30R. The Bs3 gene and/or avr genes were expressed transiently in N. benthamiana leaves via A. tumefaciens (OD600 = 0.8). Dashed lines mark the inoculated areas. Four days after infiltration, the leaves were cleared to visualize the HR (dark areas). (B) Bs3-E and/or avr genes were transiently expressed in N. benthamiana leaves. (C) The relationship between domain structure and activity of AvrBs3, AvrBs3 derivatives, and AvrBs4. Plus and minus signs indicate presence or absence of the HR in N. benthamiana upon coexpression of the pepper Bs3 or Bs3-E allele, respectively. For details, see Fig. 1A. White- and gray-boxed areas in the central part of the protein represent the repeat region of AvrBs3 and AvrBs4, respectively. AD refers to the C-terminal acidic transcriptional activation domain. (D) Gene structure of the ECW-30R Bs3 and the ECW Bs3-E alleles. Exons, introns, untranslated regions, and promoter regions are displayed to scale as white, black, gray, and hatched boxes, respectively. The length of these elements (in base pairs) is indicated within the boxes. Differences between the Bs3 alleles are marked in boldface. A 13-bp insertion in the Bs3-E promoter relative to the Bs3 promoter is underlined. Nucleotide positions of the promoter and exon 3 polymorphisms are relative to the transcriptional and translational start sites, respectively. Amino acids encoded by the polymorphic region in exon 3 (E, Glu; L, Leu; F, Phe) are depicted above and below the nucleotide sequences.

AvrBs3 mutants lacking the AD (AvrBs3ΔAD) or repeat units 11 to 14 (AvrBs3Δrep16) did not trigger HR in pepper Bs3 plants (4, 5) and also failed to trigger HR in N. benthamiana when coexpressed with the cloned Bs3 gene (Fig. 1A). AvrBs4, which is 97% identical to AvrBs3 but is not recognized by pepper Bs3 genotypes (10), also did not trigger HR in N. benthamiana when coexpressed with Bs3 (Fig. 1A). Therefore, Bs3 mediates specific recognition of wild-type AvrBs3 in both pepper and N. benthamiana, but not when AvrBs3 lacks the AD or repeat units 11 to 14; nor does Bs3 mediate recognition of the AvrBs3-like AvrBs4 protein (Fig. 1C).

The Bs3 gene has three exons and two introns (Fig. 1D), is 342 amino acids long (fig. S1), and is homologous to flavin-dependent mono-oxygenases (FMOs) (fig. S2) (11). Bs3 is most closely related to FMOs of the Arabidopsis YUCCA family (fig. S3) but lacks a stretch of ∼70 amino acids present in all related FMOs (fig. S4).

The AvrBs3 derivative AvrBs3Δrep16, which lacks repeat units 11 to 14, triggers HR in the pepper cultivar ECW but not in the near-isogenic Bs3-resistant cultivar ECW-30R (4). We transformed N. benthamiana with the ECW Bs3 allele (termed Bs3-E) including ∼1 kb of the promoter and showed that it mediated recognition of AvrBs3Δrep16 but not AvrBs3 (Fig. 1B). Furthermore, AvrBs3Δrep16 lacking the C-terminal AD did not trigger HR when coexpressed with Bs3-E (Fig. 1B), and Bs3-E did not mediate recognition of AvrBs4. Thus, Bs3 and Bs3-E represent functional alleles with distinct recognition specificities (Fig. 1C). The coding sequences of the two Bs3 alleles differ by a single nucleotide conferring a nonsynonymous change in exon 3, resulting in a leucinephenylalanine difference (Fig. 1D and fig. S1). The promoter regions also differed by a 13–base pair (bp) insertion in Bs3-E compared to Bs3, at position –50 relative to the transcription start site.

We fused the Bs3 promoter to the Bs3-E coding sequence and vice versa, then cotransformed N. benthamiana with these chimeras in combination with avrBs3, avrBs3Δrep16, or the corresponding AD mutant derivatives. The Bs3 promoter fused to the Bs3-E coding sequence mediated exclusively AvrBs3 recognition, whereas the reciprocal chimera (Bs3-E promoter fused to the Bs3 coding sequence) mediated exclusively recognition of AvrBs3Δrep16 (Fig. 2). Thus, the promoter and not the coding region determines recognition specificity of the pepper Bs3 alleles.

Fig. 2.

Chimeras containing the promoter (arrow) of the Bs3 allele (white) and the coding region (box) of the Bs3-E allele (black) or the reciprocal combination (right side of the leaf) were expressed together with avrBs3, avrBs3Δrep16, and derivatives as indicated. Asterisks mark areas in which only A. tumefaciens delivering the chimeric constructs was infiltrated. Dashed lines mark the inoculated areas. Four days after inoculation, leaves were cleared to visualize the HR (dark areas).

Semiquantitative reverse transcription polymerase chain reaction (RT-PCR) revealed strongly increased Bs3 transcript levels in pepper ECW-30R Bs3 plants upon infection with avrBs3-expressing, but not avrBs3Δrep16-or avrBs4-expressing, Xcv strains (Fig. 3). Likewise, Bs3-E levels in ECW Bs3-E plants increased upon infection with avrBs3Δrep16-expressing Xcv strains, but not when infected with avrBs3-or avrBs4-expressing Xcv strains. AD-mutant derivatives of avrBs3 and avrBs3Δrep16 did not induce accumulation of Bs3 or Bs3-E mRNA. Expression patterns were unaltered in the presence of the translation inhibitor cycloheximide (fig. S5), which indicates that accumulation of the Bs3 and Bs3-E transcripts was independent of de novo protein synthesis. Agrobacterium-mediated transient coexpression of avrBs3 and a Bs3-GFP (green fluorescent protein) fusion construct under the control of the Bs3 promoter caused GFP emission, whereas delivery of Bs3-GFP on its own did not result in GFP emission (fig. S6). Together these data indicate that AvrBs3 and AvrBs3Δrep16 induce transcription of the respective R genes Bs3 and Bs3-E, and that the subsequent accumulation of these R proteins triggers HR. In agreement with this result, constitutive expression of Bs3 or Bs3-E under the cauliflower mosaic virus 35S promoter triggered an avr-independent HR (fig. S7). We identified Bs3 mutants with single amino acid replacements that were not compromised in protein stability but no longer triggered HR when expressed in N. benthamiana (fig. S8), indicating that the enzymatic activity of Bs3 is crucial to its function as a cell death inducer.

Fig. 3.

Semiquantitative RT-PCR on cDNA of non-infected and Xcv-infected pepper ECW-30R (Bs3) and ECW (Bs3-E) leaves 24 hours after infection. The avrBs3-like genes that are expressed in the given Xcv strains are indicated in parentheses. Elongation factor 1α (EF1α) was amplified as a control.

Electrophoretic mobility shift assays (EMSAs) with GST-AvrBs3 fusion protein and biotin-labeled Bs3 and Bs3-E promoter fragments (Fig. 4A) showed that AvrBs3 bound to both Bs3- and Bs3-E–derived promoter fragments containing the polymorphism, although affinity appeared higher for the Bs3-derived fragment (Fig. 4B). Competition assays with labeled Bs3-derived promoter fragments and nonlabeled Bs3- and Bs3-E–derived promoter fragments, and vice versa, confirmed that AvrBs3 binds with high affinity to the Bs3 promoter fragment and with low affinity to the Bs3-E promoter fragment (Fig. 4C). In contrast, AvrBs3 did not bind to a DNA fragment from a nonpolymorphic region of the Bs3 promoter (Fig. 4B). Furthermore, EMSA studies showed that both AvrBs3 and AvrBs3Δrep16 have a higher affinity for the Bs3 promoter than for the Bs3-E promoter (Fig. 4 and fig. S9). Therefore, promoter binding per se of AvrBs3 or AvrBs3Δrep16 is not the basis for promoter activation specificity.

Fig. 4.

(A) Probes derived from Bs3 and Bs3-E promoter sequences used in EMSAs. Numbering is relative to the transcriptional start site. The 13-bp insertion in the Bs3-E promoter is indicated in boldface. Positions of biotin-labeled DNA fragments are indicated by lines above and below the promoter sequences. Probes I and II correspond to Bs3 and Bs3-E promoters, respectively, whereas probe III corresponds to an identical region in both promoters. (B) EMSA with AvrBs3 and Bs3-or Bs3-E–derived probes in a 6% non-denaturing polyacrylamide gel. Protein amounts are in fmol. Positions of the bound and free probe are indicated on the left. (C) EMSA competition experiment between AvrBs3 and different amounts (in fmol) of a nonlabeled competitor probe. (D) Chromatin immunoprecipitation was conducted with AvrBs3-specific antibodies on extracts from ECW-30R (Bs3) and ECW (Bs3-E) plants that were infected with Xcv wild-type (WT avrBs3) strains or an isogenic type III secretion-deficient Xcv mutant strain (ΔhrcV avrBs3). Leaves were harvested 12 hours after inoculation. Semiquantitative PCR with 32, 34, and 36 cycles was conducted before immunoprecipitation (input) or on immunoprecipitated material (IP). ECW-30R (Bs3) and ECW (Bs3-E) derived PCR products differ in size because of a 13-bp insertion in the Bs3-E promoter.

We performed chromatin immunoprecipitation assays by infiltrating pepper ECW-30R (Bs3) and ECW (Bs3-E) leaves either with avrBs3-expressing Xcv wild-type strains or with an isogenic hrcV mutant strain. HrcV is a conserved protein of the core T3S system with mutants incapable of delivering T3S effector proteins (12). After immunoprecipitation with an antibody to AvrBs3 (13), enrichment of the Bs3 but not the Bs3-E promoter region was detected by semiquantitative PCR (Fig. 4D). This demonstrates that Xcv-delivered AvrBs3 binds to the Bs3 promoter in vivo with higher affinity than to the Bs3-E promoter. Given that Bs3 promoter enrichment was detected in leaf material inoculated with the wild type but not with the hrcV mutant strain, we conclude that the Bs3 promoter is bound before cell lysis.

We also infected the pepper cultivar ECW-123R containing the R genes Bs1, Bs2, and Bs3 with xanthomonads delivering either the structurally unrelated AvrBs1, AvrBs2, or AvrBs3 protein or none of these Avr proteins. RT-PCR showed that the Bs3-derived transcripts were detectable only upon infection with avrBs3-expressing Xcv strains (fig. S10). Therefore, Bs3 is not transcriptionally activated in the course of the Bs1-or Bs2-mediated HR.

Isolation of the pepper Bs3 gene uncovered a mechanistically novel type of recognition mechanism and a structurally novel type of R protein that shares homology to FMOs. Recently, FMO1, an Arabidopsis protein that is sequence-related to Bs3 (fig. S2), was shown to be involved in pathogen defense (1416). Thus, FMO1 and Bs3 may have similar functions. However, FMO1 is transcriptionally induced by a variety of stimuli including virulent and avirulent microbial pathogens (14, 16, 17). In contrast, Bs3 was not induced by virulent Xcv strains (Fig. 3), nor by resistance reactions mediated by the pepper R genes Bs1 and Bs2 (fig. S10). Moreover, 35S-driven Bs3 alleles triggered an HR reaction (fig. S7), whereas a 35S-driven FMO1 gene mediates broad-spectrum resistance but not HR (14, 15). Thus, Arabidopsis FMO1 and pepper Bs3 differ with respect to their transcriptional regulation and function.

Our results show that the bacterial effector protein AvrBs3 binds to and activates the promoter of the matching pepper R gene Bs3. Analysis of host genes that are up-regulated by AvrBs3 (“upa” genes) in a compatible Xcv-pepper interaction (7, 18) led to the identification of the upa-box (TATATAAACCN2-3CC), a conserved DNA element that was shown to be bound by AvrBs3 and that is also present in the Bs3 promoter (Fig. 1D) (18). This suggests that binding of AvrBs3 to the upa-box is crucial for activation of corresponding promoters. However, binding of an AvrBs3-like protein does not necessarily result in promoter activation, because AvrBs3Δrep16 bound with higher affinity to the Bs3 than to the Bs3-E promoter (fig. S9) but only activated the Bs3-E and not the Bs3 promoter (Fig. 3). Because AvrBs3Δrep16 and AvrBs3 differ in their structure, we postulate that upon DNA binding, their functional domains (e.g., AD) are exposed at different promoter locations, which may define whether AvrBs3Δrep16 and AvrBs3 are able to activate a given promoter. Additionally, given that the Bs3 promoter determines recognition specificity, the Bs3 promoter might be coevolving to maintain compatibility with rapidly changing AvrBs3-like proteins, similar to that seen in the NB-LRR proteins (19, 20).

We consider it likely that not only AvrBs3 but also other AvrBs3 homologs bind to and activate promoters of matching R genes. The recently isolated rice R gene Xa27, which mediates recognition of the AvrBs3-like AvrXa27 protein from Xanthomonas oryzae pv. oryzae (21), is transcriptionally induced by AvrXa27, and thus it is tempting to speculate that the Xa27 promoter is a direct target of AvrXa27. However, whether AvrXa27 acts directly at the Xa27 promoter remains to be clarified.

Supporting Online Material

www.sciencemag.org/cgi/content/full/318/5850/645/DC1

Materials and Methods

Figs. S1 to S10

References

References and Notes

View Abstract

Navigate This Article