SEARCH

SEARCH BY CITATION

Keywords:

  • broad-range resistance;
  • expression-level polymorphism (ELP);
  • R gene;
  • gene expression marker;
  • callose;
  • blackleg

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Here, we describe the rapid cloning of a plant gene, Leptosphaeria maculans 3 (RLM3Col), which encodes a putative Toll interleukin-1 receptor-nucleotide binding (TIR-NB) class protein, which is involved in defence against the fungal pathogen L. maculans and against three other necrotrophic fungi. We have, through microarray-based case control bulk segregant comparisons of transcriptomes in pools of Col-0 × An-1 progeny, identified the absence of a locus that causes susceptibility in An-1. The significance of this locus on chromosome 4 for L. maculans resistance was supported by PCR-based mapping, and denoted resistance to RLM3Col. Differential susceptible phenotypes in four independent T-DNA insertion lines support the hypothesis that At4g16990 is required for RLM3Col function. The mutants in RLM3Col also exhibited an enhanced susceptibility to Botrytis cinerea, Alternaria brassicicola and Alternaria brassicae. Complementations of An-1 and T-DNA mutants using overexpression of a short transcript lacking the NB-ARC domain, or a genomic clone, restored resistance to all necrotrophic fungi. The elevated expression of RLM3Col on B. cinerea-susceptible mutants further suggested convergence in signalling and gene regulation between defence against B. cinerea and L. maculans. In the case of L. maculans, RLM3Col is required for efficient callose deposition downstream of RLM1Col.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Leptosphaeria maculans, a devastating fungal pathogen on Brassica oil crops worldwide, is an emerging model for hemibiotrophic fungi because of the genetic and genomic tools available, or under development (Gout et al., 2006; Rouxel and Balesdent, 2005), and because of its well-characterized interactions (Balesdent et al., 2005). Until recently, gene-for-gene type resistance in Arabidopsis has only been reported for biotrophic pathogens (Glazebrook, 2005). Despite its primarily necrotrophic nature, L. maculans does, however, display a gene-for-gene relationship with both Brassica napus and Arabidopsis (Delourme et al., 2006; Staal et al., 2006). Transgressive segregation in crosses between the Arabidopsis accessions Ler-0 and Col-0 revealed two evolutionary connected resistance loci, RLM1Col and RLM2Ler (Staal et al., 2006). The RLM1Col locus comprises a cluster of homologous Toll interleukin-1 receptor-nucleotide binding-leucine-rich repeat (TIR-NB-LRR) encoding genes, where At1g64070 plays a major role in L. maculans resistance. The TIR-NB-LRR (TNL) family of plant disease resistance genes consists of protein domains of pre-eukaryotic evolutionary origin, which are of central importance for cell-death regulation, and for innate immunity, in both plants and animals, thereby indicating conserved mechanisms (Staal and Dixelius, 2007). Several L. maculans resistance genes in B. napus (LepR1, LmR1, CLmR1, Rlm1, Rlm3, Rlm7 and Rlm9) are mapped to genomic loci that correspond to the chromosome segment on Arabidopsis chromosome 1, which harbours RLM1Col (Delourme et al., 2004; Mayerhofer et al., 2005; Parkin et al., 2005). The sequence of an R gene analogue (RGA) marker associated with the recessive Brassica juncea derived resistance rjlm2 in B. napus also shows homology with the TIR-NB-LRR subfamily TNL-H, of which RLM1Col is a member (Meyers et al., 2003; Saal and Struss, 2005). However, none of the nine currently known AvrLm genes identified from Brassica–L. maculans interactions correspond with Arabidopsis RLM1Col resistance (Staal et al., 2006).

Another feature of the Arabidopsis–L. maculans pathosystem is the large extent of naturally occurring resistance (Bohman et al., 2004). Natural genetic variation has traditionally been utilized to identify plant genes of interest through segregant analysis, which in many cases is laborious and time consuming. In addition, there is a problem in providing evidence for the functional significance of a specific gene, especially for genes with a low penetrance or where the genetics are complex (Salvi and Tuberosa, 2005; Weigel and Nordborg, 2005). The ability to fine-map and identify genes of particular interest has been greatly enhanced by the generation of high-density polymorphism maps (Jander et al., 2002), and the almost global coverage of T-DNA insertion mutants in Arabidopsis (Alonso et al., 2003). Novel tools in genomics provide us with new alternative strategies for high-density mapping via single-feature polymorphisms (SFPs) on oligonucleotide microarrays (Borevitz et al., 2003). Co-segregation analysis of SFPs with a mutant phenotype has enabled rapid identification of genes such as ERECTA (Borevitz et al., 2003) and HKT1 (Gong et al., 2004). To further demonstrate the power of bulk segregant analysis on high-density microarrays, even complex quantitative trait locus (QTL) analysis can be accomplished via the eXtreme Array Mapping method (Wolyn et al., 2004). Alternative gene cloning approaches are cDNA-AFLP or subtractive hybridization, which rely heavily on decreased complexity via bulked segregant populations. In some cases, a transcriptional difference has led to the direct discovery of a mutated gene (Kang et al., 2003), and expression levels in well-defined comparisons have also been used to identify novel genes of a defined function (Leonhardt et al., 2004).

The aim of the study was to explore previously found natural variation in the interaction between Arabidopsis and Lmaculans (Bohman et al., 2004), in order to identify additional resistance components. We used a microarray strategy: hybridizing cDNA from pools of susceptible resistant An-1 × Col-0 F3 progeny. The role of the new resistance to Leptosphaeria maculans 3 (RLM3Col) gene, encoding a putative TIR-NB protein, was further confirmed by genetic complementation and transcriptional studies, besides pathogen specificity assessments.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Case–control comparisons of resistant and susceptible progeny

In order to identify the factor important for resistance in Col-0 (our gene of interest), but absent in An-1, we needed to exclude all other accession-dependent variations in gene expression. This was accomplished by microarray comparisons on pools of susceptible (‘case’) and resistant (‘control’) Col-0 × An-1 F3 progeny of individuals that had been classified as resistant or susceptible in F2. To generate reliable comparison material, healthy uninoculated leaves were harvested from individuals where a susceptible phenotype had been established for the susceptible F3 lines. At the same time, healthy uninoculated leaves from F3 lines of resistant F2 parents, lacking segregation of resistance, were harvested as control material. In order to avoid environmental bias between the two populations, the F3 lines were grown intermixed.

Pooling of progeny decreased the unlinked differential expression between the samples (Figure S1). The material for each array was generated at different harvesting times and from different inoculation events, to ensure that the variation between the experiments was kept high in order to identify the most significant and consistent differences. Genes more highly expressed in the resistant samples showed a clear genetic bias towards a region on chromosome 4, where four genes (At4g13100, At4g16990, At4g19530 and At4g23290) showed consistently greater than twofold differential expression in all arrays (Table S1). Genes highly expressed in the susceptible samples, on the other hand, did not show any genetic bias. Consequently, the overlap between the arrays was greater among the genes highly expressed in the resistant samples, than in the genes highly expressed in the susceptible samples (Figure 1a,b). Linkage of susceptibility to the marker SNP102 at chromosome 4 supports the hypothesis that the locus (RLM3Col), with a high density of genes with elevated expression in the resistant samples, confers resistance to L. maculans (Table S2). The genes highly expressed in the susceptible samples are interpreted as environmentally induced, as a consequence of compatible interaction. Among the genes highly expressed in the susceptible samples, ACC (1-aminocyclopropane-1-carboxylic acid) synthase and isochorismate synthase indicate enhanced salicylic acid (SA) and ethylene (ET) synthesis after compatible interactions together with the β-1,3-glucanase gene PR2 (Table S3). The only gene consistently highly expressed among the susceptible samples is WRKY60.

image

Figure 1.  Number of greater than twofold differentially expressed genes for each array, and the level of overlap. V1 represents a comparison between pools of Col-0 × An-1 progeny with at least 10 individual F2 parents, V2 comprises pools of F3 progeny with at least 15 individual F2 parents, and V3 shows a comparison between the parental accessions Col-0 and An-1. Numbers within the Venn diagrams represent the number of genes that are upregulated in one array, in two arrays or in all three arrays out of 21 121 CATMA cDNA clones printed on each array. (a) Distribution and overlap of genes with greater than twofold higher expression in the resistant pool; for a list of overlapping genes, see Table S1. (b) Distribution and overlap of genes with greater than twofold higher expression in the susceptible pool; for a list of upregulated genes in the susceptible pool, see Table S3.

Download figure to PowerPoint

Evaluation of RLM3Col candidate genes

Six genes (At4g03500, At4g13100, At4g16990, At4g19530, At4g19520 and At4g23290) on chromosome 4, as defined by both genetic mapping and cDNA arrays, were evaluated with RT-PCR (Figure 2a). Two genes failed to be amplified from An-1 cDNA, and one of these genes, At4g16990, an RPP5-like gene, was the candidate gene located closest to marker SNP102. The second gene that failed to be amplified, At4g23290, which is a kinase, could be excluded as an RLM3Col candidate based on genetic mapping. At4g16990 showed a low constitutive expression, which was strongly induced between 1 and 7 days post inoculation in Col-0, but which was completely absent in An-1 (Figure 2b). The gene induction was pathogen-independent, as it could not be differentiated from the mock-inoculated (wounded) material. The transcript length of the upregulated RPP5-like candidate gene from Col-0 corresponds to ∼1 kb, indicating an alternative splice variant of At4g16990 (Figure 2b). PCR amplification from genomic DNA of Col-0 and An-1 of the Complete Arabidopsis Transcriptome MicroArray (CATMA) gene sequence tag (GST) 96112-E12, which spans part of exon 3 and exon 4 of At4g16990, showed that this particular sequence was absent in An-1. At4g16990 GST amplification of genomic DNA from resistant and susceptible Col-0 × An-1 F3 lines showed a perfect co-segregation between the presence of At4g16990 and resistance (40 positive bands out of 40 resistant F3 lines, and with all amplifications out of 35 susceptible F3 lines being negative). PCR failed to amplify target sequences in a large (>4.7 kb) region at At4g16990 from An-1 genomic DNA, which implies that the absence of At4g16990 from An-1 (rlm3An-1) is caused by a large indel (insertion/deletion) polymorphism.

image

Figure 2.  RT-PCR analysis of differentially expressed genes within the RLM3 locus, and transcriptional analysis of At1g16990. (a) Reverse transcriptase PCR on unique gene sequences (cDNA) with over twofold higher expression in the resistant pool, in at least two arrays, and on average with greater than twofold higher expression in all three arrays. A 100-bp Generuler ladder (Fermentas, http://www.fermentas.com) was used as a size control, and PCR was repeated twice with similar results. Lanes: 1, At4g03500 (Col-0); 2, At4g03500 (An-1); 3, At4g13100 (Col-0); 4, At4g13100 (An-1); 5, At4g16990 (Col-0); 6, At4g16990 (An-1); 7, At4g19530 (Col-0); 8, At4g19530 (An-1); 9, At4g19520 (Col-0); 10, At4g19520 (An-1); 11, At4g23290 (Col-0); and 12, At4g23290 (An-1). (b) Northern blot analysis of RLM3 on the resistant accession Col-0 and the susceptible accession An-1 inoculated with Leptosphaeria maculans. A total of 7 μg of RNA was loaded in each lane. The membrane was hybridized with an At4g16990-specific CATMA probe (96112-E12); only the ∼1-kb transcript was detected using this specific probe. To verify equal loading, rRNA bands in the gel were stained with ethidium bromide and visualized under UV light. Abbreviations: hw, hours post water inoculation; hi, hours post inoculation; dw, days post water inoculation; di, days post inoculation. Time-point 0 is represented by untreated samples. Results were repeated twice.

Download figure to PowerPoint

Insertion mutants in candidate genes identified an alternative transcript as RLM3Col

Varying degrees of susceptibility to L. maculans were identified in T-DNA mutants of At4g16990 (Figure 3a, Figure S2a). Col-0 displays a completely resistant phenotype, whereas two RLM3Col mutants, gabi_235E02 (rlm3-1) and gabi_491E04 (rlm3-2) (Rosso et al., 2003), harbouring an insertion in the promoter, 535-bp upstream of the translation start of At4g16990, and in the first intron, respectively, revealed a clear L. maculans susceptibility (Figure 3b–d). The salk mutant 048620 (rlm3-3), with an insertion in the promoter 905-bp upstream of the translation start of At4g16990, also showed a susceptible phenotype (Figure S2c). Salk_067449 (rlm3-4), on the other hand, where the gene is truncated in exon 4, the second last exon of splice form 1, exhibited a moderately susceptible phenotype (Figure S2d). The two gabi mutants (rlm3-1 and rlm3-2) revealed a slightly higher degree of L. maculans susceptibility, compared with rlm3-3, when assessed by qPCR, indicating that rlm3-3 may have retained a small degree of RLM3 resistance (data not shown). In addition, salk_146865 (rlm3-5), with an insert downstream of the shortest gene model (At4g16990.1) in the intron between exons 6 and 7, showed a completely resistant phenotype (Figure S2e). No signs of susceptibility could be seen in insertion mutants in other candidate genes (At4g03500, At4g13100, At4g19530, At4g19520 and At4g23290) in the RLM3Col locus (data not shown). At the time of established susceptibility in the T-DNA mutants, An-1 showed a severely diseased phenotype, and most leaves were already dead (Figure 3e, Figure S2f).

image

Figure 3.  Disease phenotypes in response to Leptosphaeria maculans, Alternaria brassicicola, Alternaria brassicae and Botrytis cinerea in Col-0, An-1 and RLM3 T-DNA mutants. (a) T-DNA insertions in At4g16990 (RLM3): orientation and location in relation to exons and predicted protein domains. The gene model shows the predicted intron/exon composition (black boxes = exons) of the At4g16990.1 locus, and the numbers represent the nucleotide base coordinates in the genomic gene model. The green box indicates the TIR domain, the yellow box indicates the NB-ARC domain and the purple box indicates a sequence with homology to the RCC1 family. The mutation in gabi_235E02 (rlm3-1) is located in the promoter and is inserted in the opposite direction to RLM3 (blue arrowhead, susceptible phenotype). gabi_491E04 (rlm3-2) is inserted in the same direction as RLM3 and is located in the intron between exon 1 and 2 (red arrowhead, susceptible phenotype). Water-inoculated control plants did not display any disease-like phenotypes on any of the genotypes assessed. Responses on various genotypes to four fungal pathogens. (b–e) L. maculans responses at 17 dpi. At 21 dpi, the inoculated leaves of rlm3-1 were completely dead. The experiment was repeated five times, with 10–20 plants per genotype. (f–i) Assessment of the genotypes with A. brassicicola at 10 dpi resulted in a susceptible ‘spotted’ phenotype on rlm3-1, rlm3-2 and An-1, whereas Col-0 remained resistant. The experiment included 20 plants per genotype, and was repeated once. (j–m) B. cinerea inoculation on corresponding plant material resulted in moderately susceptible phenotypes on rlm3-1 and rlm3-2 at 14 dpi. The experiment included 20 plants per genotype, and was repeated once. (n–r) A. brassicae inoculation on the genotypes also resulted in similar patterns of susceptibility as B. cinerea at 14 dpi. The experiment included 20 plants per genotype, and was repeated once.

Download figure to PowerPoint

Tests of allelism in F1 progeny from crosses between An-1 and T-DNA mutants (rlm3-3, rlm3-4 and rlm3-5), and Col-0 as a control, confirmed that the susceptible phenotypes observed in the mutants are allelic to the recessive susceptibility trait in An-1 (Figure S2g-k). Furthermore, F1 progeny between An-1 × rlm3-3 and An-1 × rlm3-4, revealed that only the promoter mutant (rlm3-3) resulted in more susceptible progeny (Figure S2h), whereas An-1 × rlm3-4 retained a moderately susceptible phenotype, which appeared resistant at 14 days post-inoculation (dpi) (Figure S2g), and started to display a susceptible phenotype after 17 dpi. In conclusion, At4g16990.1 is responsible for the resistance response to L. maculans, and the results indicate that the C-terminal part of the protein encoded by this transcript only has moderate influence on RLM3Col function.

Isolation of cDNA, with a primer approximately 250-bp downstream of rlm3-4, identified a transcript with a length corresponding to the approximately 1-kb transcript (744 bp) that was observed to be upregulated (Figure 2b). Subsequent sequencing revealed that the transcript represented an alternative splice form, where exon 2 had been spliced away (Figure 4). The novel 744-bp coding sequence from the alternative transcript has lost its predicted NB-ARC domain, compared with the At4g16990.1 TIR-NB protein, and only encodes an N-terminal TIR domain and an amino acid sequence of unknown function (TIR-X).

image

Figure 4.  RT-PCR analysis of RLM3Col and RLM3TIR-X transcripts present in different genotypes. The gel picture shows a normal splice form (approximately 1.4 kb) with exon 2 (NB-ARC) and an alternative splice form (approximately 700 bp) without exon 2 (NB-ARC). Analysis of RLM3 expression from exon 1 to exon 4 using 5′-AAGATGTCCGCCACTCATTAGT-3′ (forward) and 5′-CATAGATAGCTCAACTGATG-3′ (reverse).

Download figure to PowerPoint

RLM3Col plays a role in resistance against several necrotrophic fungi

An-1 shows enhanced susceptibility to a range of pathogens, such as Botrytis cinerea, Alternaria brassicicola and Alternaria brassicae (Figures S2p,u,z and 3i,m,q). To analyse whether RLM3Col affected resistance against those pathogens, the susceptible and resistant F3 lines deriving from the An-1 and Col-0 cross were evaluated to clarify their responses. All L. maculans-susceptible lines showed susceptibility to the other three pathogens. Similarly, the same resistant F3 lines (25 of 40 lines) that showed a 3:1 segregation of resistance against L. maculans also showed the same pattern of segregation against A. brassicicola, A. brassicae and B. cinerea.

In order to obtain additional data on rlm3An-1, An-1 was crossed with the susceptible rlm1Ler rlm2Col double mutant of two Lmaculans-susceptible genotypes (Staal et al., 2006). The F1 progeny of this cross were found to be resistant to L. maculans, demonstrating that the susceptibility in An-1 does not result from the absence of RLM1 or RLM2 resistance. Importantly, it also demonstrates that rlm1Ler rlm2Col harbours a functional RLM3 gene, and that RLM3 alone cannot confer a gene-for-gene type of resistance. Fifty F2 plants from this cross were further evaluated against each of the four fungal pathogens. A 3:1 segregation of resistance against A. brassicicola, A. brassiciae and B. cinerea was observed, further supporting the earlier finding that RLM3 is a single dominant gene governing resistance to these necrotrophic pathogens. No epistasis (hypersusceptible plants) could be found in the F2 population after inoculation with L. maculans, indicating that RLM1/RLM2 and RLM3 resistances reside in the same pathway. These results support the hypothesis that resistance to all four pathogens depends on the same gene, and that RLM3Col is dominant. Inoculations with A. brassicicola revealed a moderate to clear susceptibility on the T-DNA mutants rlm3-1, rlm3-2 and rlm3-3, and a moderately susceptible phenotype on rlm3-4 (Figures S2l–o and 3f–i). A clear impairment of resistance in rlm3-1 and rlm3-2 to both B. cinerea and A. brassicae was found (Figure 3j–q), whereas rlm3-3 and rlm3-4 show a more moderate susceptibility (Figure S2r,s,w,x). Analogous to L. maculans, rlm3-5 displayed a fully resistant phenotype against A. brassicicola, A. brassicae and B. cinerea (Figure S2o,t,y).

RLM3 does not show any clear effects on jasmonic acid (JA)/ET or camalexin-dependent responses, which have been shown to be important for all of the pathogens that cause a susceptible phenotype on rlm3 material. This indicates that RLM3 influences a different set of required resistance responses. Further support for this hypothesis is the assessment of the Botrytis-induced kinase 1 (bik1) mutant, which shows enhanced susceptibility to both B. cinerea and A. brassicicola as a result of reduced JA/ET responses (Veronese et al., 2006), but remained resistant as wild type when challenged by L. maculans, suggesting that RLM3Col is a BIK1-independent resistance response. In order to further evaluate the link between the different pathogens that are able to infect rlm3 material, B. cinerea-susceptible (bos) mutants were evaluated. Interestingly, the bos1 mutant encoding a mutant R2R3MYB (Mengiste et al., 2003) shows a moderately susceptible phenotype, whereas bos3 (Veronese et al., 2004) displays a clearly susceptible phenotype to L. maculans (Figure S3). Taken together, these observations suggest a common resistance pathway for both pathogens, separate from bos2 that is resistant to L. maculans. RLM3Col expression is not downregulated in bos1 and bos3 mutants, compared with wild type, after inoculation with L. maculans (Figure 6), but is significantly upregulated (< 0.0004 and < 0.03, respectively), which further points to downstream converging points in the signalling pathway(s).

image

Figure 6.  Expression of At4g16990 in different genotypes, determined by quantitative real-time PCR. The forward primer was designed with the 5′ part on exon 3 and the 3′ part on exon 4, and the reversed primer spanned exon 4. The data was assessed relative to the expression of polyubiquitin 10 (At4g05320). Col-0 was set to a value of 100 for relative comparisons with other genotypes. Each genotype was assessed in three biological replicates of samples comprising nine or 10 plants. Each sample was amplified in triplicate. Bars indicate standard deviation between the different biological replications. Leaf samples were taken 2 days post inoculation with Leptosphaeria maculans.

Download figure to PowerPoint

In contrast with the responses towards necrotrophic fungi, assessments of rlm3-3, rlm3-4 and rlm3-5 did not reveal any enhanced susceptibility against Albugo candida, Hyaloperonospora parasitica or Phytophthora brassicae isolates (Table S4). A quantitative evaluation of different avr genes in a Pseudomonas syringae DC3000 background on Col-0, An-1 and the RLM3 T-DNA mutants revealed a reduced susceptibility, expressed as bacterial growth in all rlm3 genotypes evaluated at 3 dpi in the absence of avr genes (Figure S4). Under the given conditions, P. syringae showed over 100-fold growth after 3 days on wild-type Col-0 in the absence of R-avr gene interaction, whereas the growth was more moderate (in the 10-fold range) in An-1 and rlm3 mutants in the Col-0 background. This reduced bacterial growth is the opposite to the enhanced growth of the necrotrophic fungal pathogens evaluated. The presence of an R-avr interaction led to a drastic decrease or to the complete abolition of bacterial growth in all materials, indicating that the loss of RLM3Col does not impair any of the R-gene dependent responses towards P. syringae. This observation could imply a role of RLM3Col in directing basal defence responses towards pathways effective against necrotrophic fungi, and separate from pathways functioning against P. syringae. A differential response on An-1 compared with the five T-DNA insertion lines for the avrPphB gene was also found, but growth of this strain was as moderate as for the avr-absent strains (in the 10-fold range), suggesting that An-1 may lack RPS5 or may mount a less efficient defence response downstream of this gene.

Genomic RLM3Col complements An-1 and RLM3 T-DNA mutants

Genomic complementation of An-1, and the T-DNA mutants rlm3-1 and rlm3-2, using the TAC clone JAtY64O13, containing the complete genomic clone of RLM3 and its promoter (RLM3::RLM3Col), generated 15 T1 plants in the An-1 background and 18 T1 plants in the rlm3-2 background. Transformation of rlm3-1 did not generate any T1 plants. All surviving T1 plants demonstrated a restored resistance against L. maculans. In total, 40 Basta-resistant T2 plants, deriving from individual 3:1 segregation plates of RLM3::RLM3Col in the An-1 and rlm3-2 background, respectively, were assessed against all four necrotrophic fungal pathogens. All the plants were found to be resistant (Figure 5). This was further validated by inoculating the material with L. maculans expressing gdp::GUS (data not shown). In congruence, quantitative real-time PCR analysis on RLM3::RLM3Col-complemented An-1 and rlm3-2 genotypes revealed a significant increase (< 0.003) of At4g16990 expression, as early as 2 dpi (Figure 6).

image

Figure 5.  Comparison of disease phenotypes to Leptosphaeria maculans, Alternaria brassicicola, Alternaria brassicae and Botrytis cinerea in complemented lines harbouring RLM3::RLM3Col compared with wild-type genotypes at 21 dpi. The experiments were repeated twice, with 20 plants pere genotype. (a–d) L. maculans responses. (e–h) A. brassicicola responses. (i–l) B. cinerea responses. (m–p) A. brassicae responses.

Download figure to PowerPoint

Overexpression of an RLM3Col transcript enhances resistance responses associated with RLM1

Complementation using the 744-bp alternative Col-0 transcript of RLM3Col (At4g16990: exon1 + exon3 + exon4) driven by the 35S promoter (hereafter 35S::RLM3TIR-X) was also analysed. As An-1 displayed a clearer disease phenotype than any of the T-DNA insertion mutants in the Col-0 background, An-1 was used for further evaluation of the biological role of the TIR-X-encoded RLM3TIR-X. The majority of the An-1 35S::RLM3TIR-X plants exhibited spontaneous disease-like lesions on older leaves in the absence of a pathogen, a phenomenon observed after overexpression of other resistance genes (Gurr and Rushton, 2005; Xiao et al., 2005). The induction of cell death by overexpression of RLM3TIR-X was, however unexpected, as the NB-ARC/NOD domain is associated with programmed cell death (Jaroszewski et al., 2000). However, four out of the 41 T1 lines evaluated had a delayed local lesion development, compared with wild-type An-1, suggesting a resistant phenotype to L. maculans (Figure S5a). In agreement with our observations, recent analysis by overexpression of the TIR-NB domain of a truncated RPP1A gene revealed spontaneous (elicitor-independent) activation of hypersensitive response type cell death, and was sufficient to confer resistance to H. parasitica Cala2 (Weaver et al., 2006). Further assessment of systemic lesions after inoculation with GFP-tagged L. maculans showed that the progression of L. maculans growth was absent or severely reduced (Figure S5b,c). Quantitative real-time PCR experiments confirmed that a resistant An-1 35S::RLM3TIR-X T3 line had, on average, 46% less fungal biomass compared with plants in the An-1 background at 14 dpi (Figure S6).

Complementary assessments of 35S::RLM3TIR-X T2 offspring from an L. maculans resistant T1 plant in the An-1 background against A. brassicicola, A. brassicae and B. cinerea also revealed a partial restoration of resistance to those three pathogens (Figure S5d–f). In addition, evaluation of L. maculans growth in 24 rlm3-4 T1 plants complemented with 35S::RLM3TIR-X revealed enhanced susceptibility (57% increase in fungal biomass, at 19 dpi) in 19 individuals, and completely restored resistance (65% decrease in fungal biomass, at 19 dpi) in five individuals (Figure S6g; data not shown). The lines showing clear enhancement of the susceptible phenotype were found to completely lack RLM3Col expression because of co-suppression (Figure S7). This could also explain the low frequency of complemented An-1 plants, as the surviving plants either have low expression of the gene or represent lines where the expression is silenced. A role of alternative R-gene transcripts in signalling has been proposed, based on the observation that several R genes require their alternative splice forms to remain functional (Jordan et al., 2002). Whether truncated R-gene forms encode adaptor molecules, which are important in defence signalling, remains to be shown. To address this question, RLM3TIR-X was overexpressed in the rlm1Ler rlm2Col background, lacking the TIR-NB-LRR RLM1Col and RLM2Ler loci (Staal et al., 2006). A restoration of resistance in four out of 12 rlm1Ler rlm2Col35S::RLM3TIR-X T1 plants demonstrates that overexpressed RLM3TIR-X activates RLM1Col-dependent responses (Figure S5h), and is thus likely to influence signalling downstream of RLM1Col. 35S::RLM3TIR-X complementation of rlm3-3 also generated completely restored lines (Figure S5i), indicating that the partial restoration observed in An-1 is caused by additional susceptibility genes, possibly in lacking other RLM3 transcripts.

RLM3Col affects callose deposition

A role for RLM1Col in callose induction for resistance to L. maculans has recently been demonstrated (Staal et al., 2006). In order to establish whether RLM3Col influenced this RLM1Col response, callose deposition in An-1, Col-0 and the RLM3 T-DNA mutants was evaluated after challenge with L. maculans. The susceptible genotypes all revealed a decreased callose deposition compared with Col-0 (Figure 7). Surprisingly, the resistant mutant rlm3-5 also displayed reduced callose depositions (data not shown). Genomic complemented An-1 and rlm3-2 exhibit restored callose deposition levels, compared with wild type, whereas the former genotype showed a fourfold increase compared with wild type An-1. Likewise, callose deposition in the partially restored An-1 35S::RLM3TIR-X line was significantly higher (< 0.000005) than that observed in non-transformed An-1 (data not shown). Taken together, the accelerated cell death in many lines harbouring an overexpressed RLM3TIR-X gene and the modified callose deposition suggest that RLM3Col activates responses downstream of RLM1Col, but upstream of callose induction.

image

Figure 7.  Box plots illustrating the distribution of callose deposition densities in different Arabidopsis genotypes 2 days post L. maculans inoculation. An-1, rlm3-1 and rlm3-2 lines show significantly lower levels of callose compared with Col-0 or genomic RLM3::RLM3Col complemented lines. The black bars represent average values, the grey boxes represent the range of the first quartiles over and under the averages (50% of measured values), and the error bars represents the range of the two outer quartiles. Callose deposition densities (area above signal threshold/complete area) were determined by APS assess (L. Lamari, Univ. of Manitoba, Canada). The average density in Col-0 is set to 100. Data represent measurements on between nine and 23 individual samples per genotype.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Rapid identification of a gene using bulked segregant transcriptome comparisons

Microarrays have previously been reported as high-throughput tools for haplotyping the SFPs of accessions on labelled genomic DNA, which could be used for linkage disequilibrium mapping (Borevitz and Nordborg, 2003; Borevitz et al., 2003). The use of expression-level polymorphisms (ELPs) as gene expression markers (GEMs) has recently been demonstrated to be another method to generate high-density microarray-based markers, but is less robust than SFPs (West et al., 2006). In theory, however, GEMs do provide a biased output that favours the observation of differences of biological interest by choice of comparison material, which in turn facilitates faster gene identification. Here, we have described a rapid and efficient method to identify a gene that differs between two Arabidopsis accessions. By collecting material at variable time points for each replicate, we kept the variance for environment-dependent expressions high, and were thus able to exclude some confounding effects that otherwise could have complicated the transcription profile mapping procedure. The use of cDNA microarrays in a ‘case–control’ population design set-up for gene cloning has some major advantages, compared with cDNA-AFLP strategies, such as the full-genome information and the immediate identification of candidate genes. There are, however, some issues to consider compared with genetic mapping. A microarray will fail to identify the mutations responsible for a phenotype if the mutation is a minor base-pair substitution that changes the properties of the gene product. This may, however, be of minor concern when it comes to natural variation between Arabidopsis accessions, as the ratio between base-pair mismatches caused by indels compared with substitutions is 51:1 (Britten et al., 2003). A reassuring finding in our study that partially addresses this issue is the observation of transcriptional differences in genes genetically linked to our gene of interest. A similar strong genetic bias in transcription profiles was also recently reported in Ler-2 × Cvi-1 near isogenic lines (Juenger et al., 2006). Moreover, the expression of the gene of interest may well be lacking under the chosen conditions. The CATMA cDNA arrays used have, however, been shown to have high sensitivity, and are able to detect low copy transcripts to a greater extent than oligonucleotide-based arrays (Allemeersch et al., 2005), which make them particularly useful for this kind of study.

Functional characteristics of RLM3Col

The short form of RLM3Col (RLM3TIR-X) encodes a protein with a TIR domain and a sequence of undefined function. At4g16990 has been defined as a TIR-NB protein (Meyers et al., 2002), but the putative NB-ARC domain has been spliced away from this particular transcript. The domain of undefined function consists of 74 amino acids (EDLRNWRSEAEMLAEWIHLALGVSILLNIRSDGTTILKHLSYNRSMAQQAKIWWYENLERVCKKYNICGIDSST). This X domain, encoded by the end of exon 1, and exons 3 and 4 (Figure 3a), possesses homology with a region close to the C-terminal of the regulator of chromosome condensation (RCC1) protein family, which regulates the cell cycle by binding to histones (Nemergut et al., 2001; Renault et al., 2001). Interestingly, the recently identified TirA TIR domain-encoding gene, required for specialized immune cell function in slime mould, is also an RCC1-like protein (Chen et al., 2007). The C-terminal sequence of RLM3TIR-X harbours putative high-motility group (HMG) and whey acidic protein (WAP) domains. The HMG domain is normally found in low molecular weight chromatin proteins, and binds to DNA. Epigenetic regulation of resistance responses by modulation of histones may be an important mechanism. The histone deacetylase 19 gene is, for example, required for JA-dependent defence responses against A. brassicicola (Zhou et al., 2005). Previous functional analysis of the RCC1 family identified the N-terminal part as responsible for binding to the nuclear Ras-like GTPase Ran, whereas the C-terminal end of this family still has unknown function. Interaction of a RanGAP protein with the N-terminal of the NB-LRR protein Rx has also been shown to be required for PVX resistance (Sacco et al., 2007). In L. maculans defence, the functional requirement of the Rac GTPase activating protein (RacGap), encoded by At2g27440, has been observed (Bohman, 2001), which may share functional links to the interactions between Ran and the RCC1 family. Other deletion studies of RPS2 and N genes with CC-NB-LRR and TIR-NB-LRR motifs, respectively, imply that the CC-NB and TIR-NB genes retain some functions of the native R gene, which may require some specific modulations of its C-terminal domain to become active (Dinesh-Kumar et al., 2000; Tao et al., 2000). Alternative splicing is an important and common post-transcriptional regulatory mechanism in Arabidopsis, particularly among stress-response genes (Ner-Gaon et al., 2004; Wang and Brendel, 2006). Alternative splice forms of TIR-NB-LRR proteins have been suggested to be able to act as R-protein-specific downstream signalling components (Meyers et al., 2002), which have also been shown experimentally in the case of RPS4 (Jordan et al., 2002; Zhang and Gassmann, 2007; Zhang and Gassmann, 2003). Our observations in this study imply that RLM3 is another candidate that appears to act as a critical downstream component, but in this case for a resistance pathway shared between several necrotrophic fungal pathogens. However, a further assessment of additional constructs, harbouring various cDNAs driven by the RLM3 promoter, in a susceptible background, is required to conclusively resolve the function of alternative RLM3 transcripts.

RLM3Col influence on resistance against a wide array of pathogens

An-1 is the only accession found hitherto that displays a general susceptibility to a wide range of pathogens, such as A. candida, H. parasitica and P. brassicae, besides fungal necrotrophs. This observation indicates that this particular accession has lost components of fundamental importance for plant innate immunity. Surprisingly, RLM3Col activity does to some extent influence Col-0 resistance to A. brassicicola, A. brassicae and B. cinerea. Additional information on the Arabidopsis–A. brassicae interaction is unfortunately lacking. Resistance to A. brassicicola and B. cinerea has not been shown to be mediated by a single R gene, in contrast to L. maculans, and no enhanced disease severity from either pathogen has been observed on Arabidopsis gene signalling mutants. A common denominator between L. maculans and A. brassicicola is the susceptible phenotype on pmr4, a mutant in callose synthase (Staal et al., 2006; Ton and Mauch-Mani, 2004). Callose deposition is, at least in the case of L. maculans responses, in both Arabidopsis and Brassica species, dependent on R genes, as genotypes lacking R genes display greatly reduced callose deposition levels (Yu et al., 2005; Staal et al., 2006; Kaliff et al., 2007; Figure 7). In contrast to Arabidopsis–L. maculans resistance, which is independent of SA, JA and ET signalling (Bohman et al., 2004), resistance to A. brassicicola depends on JA (Schilmiller et al., 2007; Thomma et al., 1998). Resistance to B. cinerea, on the other hand, requires both JA and ET signalling (Thomma et al., 1998, 1999a). SA signalling does not play a major role in defence against B. cinerea, but it does contribute to some extent (Ferrari et al., 2003). However, oligogalacturonides released from plant cell walls increased resistance to B. cinerea in Arabidopsis, independently of SA, JA or ET, but required camalexin induction (Ferrari et al., 2007).

Camalexin is another common important factor, which is active against both Lmaculans (Bohman et al., 2004; Staal et al., 2006) and A. brassicicola (Thomma et al., 1999b; Zhou et al., 1999), and against some strains of B. cinerea (Denby et al., 2004). The role of camalexin in response to B. cinerea seems complex, as the phytoalexin-deficient mutant pad3 shows resistance (Ferrari et al., 2003). RLM3Col has no effect on either JA/ET-induced responses or camalexin induction in response to L. maculans (Bohman et al., 2004; data not shown), which implies that RLM3Col is required for a different set of resistance components.

The responses of B. cinerea and L. maculans also resemble each other in other aspects. The most intriguing difference among the bos mutants is the enhanced cell death and necrosis found in bos3 (Veronese et al., 2004). L. maculans also grows to various extents in cell-death mutants like ran1-1 and acd1-20, and in the respiratory burst NADPH-oxidase mutants, such as rbohD, rbohF and rbohDF (Bohman et al., 2004; Kaliff et al., 2007), despite the presence of RLM1Col. Ubiquitination of BOS1 has been proposed to be involved in the B. cinerea interaction, but also to act in a role of mediating signals governed by JA or reactive oxygen intermediates (Mengiste et al., 2003). Similarly, several ubiquitin-related components have now been shown to be of importance in defence against L. maculans (Bohman et al., 2002; M. Kaliff, C. Dixelius, unpublished data). Among other factors, ABA was found to be essential for the Arabidopsis–B. cinerea interaction by expression profiling (AbuQamar et al., 2006). In addition to the impact on callose deposition, ABA also plays an important role in defence against L. maculans, where ABA induces a hitherto unknown resistance factor, which can partly complement the absence of RLM1Col (Kaliff et al., 2007). Furthermore, WRKY transcription factors are required for resistance to B. cinerea and A. brassicicola, a response linked to the downregulation of PDF1.2 (Zheng et al., 2006), whereas the role of WRKY trancription factors in defence against L. maculans is so far unknown.

The previously identified L. maculans R gene RLM1Col and its putative paralogue RLM2Ler belong to the TIR-NB-LRR class (Staal et al., 2006), and it is interesting that the third resistance gene (RLM3) in this pathosystem is also related to the same group of genes. In contrast to the TNL-H RLM1Col and rjlm2 L. maculans resistance genes, the RLM3Col gene is more closely related to the TNL-E subfamily (Meyers et al., 2003). The potential functions of a truncated product of an R gene can be numerous, and the TIR domain alone could have signalling properties in addition to its well-established role as a protein–protein interaction domain.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Plant material

Uninoculated leaves of 35 F3 lines from susceptible F2, and 40 F3 lines from resistant F2, Col-0 × An-1 progeny were used for microarray comparisons and genetic mapping. L. maculans-susceptible individuals, derived from resistant F2 plants, were also used in PCR-based mapping. The parental accessions Col-0 and An-1 were used for one microarray comparison, RT-PCR and northern blot expression profiling of candidate genes. Plant material for each microarray comparison originated from a unique set of material, in order to obtain three independent biological replicates. The Arabidopsis material was cultured as described previously (Bohman et al., 2004). For evaluation of candidate genes, T-DNA insertion mutants gabi_491E04 (rlm3-1), gabi_235E02 (rlm3-2), salk_004453, salk_024404, salk_035263, salk_067449 (rlm3-4), salk_048620 (rlm3-3), WiscDsLox377-380N18, salk_054464, salk_021741, salk_112083, salk_135546 and salk_146865 (rlm3-5), all in the Col-0 background (Alonso et al., 2003; Rosso et al., 2003), were screened for susceptibility. The rlm1Ler rlm2Col line (Staal et al., 2006) was used for comparative functional studies of RLM3. For comparison with B. cinerea, bik1 (Veronese et al., 2006) and bos mutants (Mengiste et al., 2003; Veronese et al., 2004) were assessed.

Plant inoculations

Responses to L. maculans (M1, PHW1245 and Leroy) were determined as previously described (Bohman et al., 2004). Control leaf material for northern blot comparisons was inoculated with water (similar to the fungal leaf-puncture method). Plants were screened for L. maculans susceptibility until 21 dpi. Col-0, An-1 and selected T-DNA mutants were also inoculated with additional plant pathogens. B. cinerea (MUCL30158), A. brassicicola (MUCL20297) and A. brassicae (CBS980:2) were applied as described previously (Thomma et al., 1998), and plants were examined until 14 dpi. The degree of infection was further assessed by lactophenol trypan blue staining (Koch and Slusarenko, 1990). P. brassicae (former: porri) isolates d and HH were applied as agar plugs (Roetschi et al., 2001). Plants were dipped in P. syringae DC3000 solution harbouring either empty vector or avr genes, according to the method described by Tornero and Dangl (2001).

Experimental design

Each of the three microarray experiments consisted of a unique case–control comparison set: version 1 (V1) comprised lines from 10 resistant and 10 susceptible F2 parents, V2 comprised lines from lines of 15 F2 parents of each phenotype, and V3 comprised lines from the parental accessions Col-0 (resistant) and An-1 (susceptible). Each pool consisted of over 25 individual plants (2–3 individuals per F3 line), irrespective of the number of parental genotypes. Material for experiment V1 was collected immediately when the L. maculans-susceptible phenotype had become established (14 dpi). Material for V2 was harvested at 19 dpi, when the first signs of systemic spreading of the fungus were observed. V3 material was harvested at an intermediate stage (17 dpi).

Microarray hybridizations

Differential gene expression in green leaves between pools of L. maculans F3 plants, classified as either resistant or susceptible after inoculations, were assessed. RNA was isolated using the RNeasy kit (Qiagen, http://www.qiagen.com). The cDNA synthesis, slide hybridization and washings were performed as described in the protocols from the Meyerowitz laboratory, Caltech 15 (http://www.its.caltech.edu/~plantlab/html/protocols.html). Resistant material was labelled with Cy5, and susceptible material was labelled with Cy3, in V1 and V2 arrays, whereas V3 was dye swapped. Labelled cDNA was hybridized to CATMA 21-k cDNA arrays (http://www.catma.org). The CATMA arrays were analysed with genepix 4000B (MDS Analytical Technologies, http://www.moleculardevices.com).

Analysis of expression data

Array-by-array examinations were made by the arbitrary twofold expression difference determination, after filtering (removal of bad spots) and normalizations (adjustments of data distribution) with the kth package (http://www.biotech.kth.se/molbio/microarray/dataanalysis/index.html) in R (R Foundation for Statistical Computing, http://www.r-project.org/foundation) in accordance with the settings presented in the kth package tutorial (http://biotech.kth.se/molbio/microarray/usersguide/UsersGuide.htm). Simultaneous analysis on all three arrays was performed with a B-test within the same program package to identify significantly differentially expressed genes under the same filtering and normalization conditions.

Mapping of susceptibility traits

In addition to the transcriptional comparisons, we used SNP (SGCSNP41, SGCSNP24, SGCSNP64, SGCSNP102, SGCSNP58 and CER433105), CAPS (g4539), SSLP (CIW7) and microsatellite markers (MSAT4.39, MSAT4.25, MSAT4.18 and MSAT4.37) on chromosome 4, in order to decrease the number of potential gene candidates.

Candidate gene evaluation

Candidate genes were evaluated with reverse transcriptase PCR on Col-0 and An-1 cDNA generated with the SMART cDNA synthesis kit (Clontech, http://www.clontech.com). The primers used in RT-PCR were as follows: At4g03500-FOR, 5′-TCTTTTCCCCACTTGGGAAT-3′; At4g03500-REV, 5′-TTTTTGTTGTCACCAGCAGC-3′; At4g13100-FOR, 5′-GCCGGAGCTCAGTGACTTTA-3′; At4g13100-REV, 5′-TAGCCGGAGCCTTCCTTCT-3′; At4g16990-FOR, 5′-TTGGCACTAGGCGTGTCTATA-3′; At4g16990-REV, 5′-TCGTCCCGCAATACGATATT-3′; At4g19530-FOR, 5′-AGGTGCTCAAGTGTGGGTTAA-3′; At4g19530-REV, 5′-CTCCCATTTTGAGATTGTGC-3′; At4g19520-FOR, 5′-AATTTCGCTTCGTCCAAAGG-3′; At4g19520-REV, 5′-TCACGGGAAAGTTGGTCATA-3′; At4g23290-FOR, 5′-TGTCCTAATCAGGCAGAAGCT-3′; At4g23290-REV, 5′-TTGCCATCTTTCTTGGTTGTG-3′. Genomic DNA from Col-0, An-1 and susceptible and resistant F3 individuals were isolated for mapping. PCR and sequencing was performed as previously described (Staal et al., 2006).

Plant transformation

The TAC clone JAtY64O13 containing At4g16990 in pYLTAC17 (47.2 kb, spanning At4g16970–At4g17080) was used for the genomic complementation. None of the other genes are predicted to be involved in pathogen responses, and SNP data from MSQT and InDel data, via SFP, showed no obvious differences between Col-0 and An-1 in any other genes in the clone, other than those found in At4g16990. The short alternative transcript of At4g16990.1 was isolated with the At4g16990.1 5′ untranslated region (5′-UTR)-specific primer (5′-ATCTTCTTTCTATAGTTTCCATG+attB1-3′) and the 3′-CATMA GST (96112-E12) transcribed sequence (5′-TTTTTTGGTACACAATTGGGTTTG+attB2-3′). The PCR product was sequenced, cloned into the pDONR221/Zeo vector (Invitrogen, http://www.invitrogen.com), and was then recombined into the Gateway-compatible pGWB2 binary vector downstream of the 35S promoter (T. Nakagawa, Shimane University, Izumo, Japan). All constructs were transferred into Agrobacterium tumefaciens C58 by freeze–thaw transformation. Arabidopsis transformation was performed by floral dip (Clough and Bent, 1998), and seeds were selected on 50 μg ml−1 kanamycin, and 25 μg ml−1 Basta for the JatY transformed materials. Retransformed kanamycin-resistant T-DNA mutants were selected on 20 μg ml−1 hygromycin.

Transgenic L. maculans

The coding sequence of EGFP (Clontech) under control of the GPDA promoter (Farman and Oliver, 1992) was introduced into a binary vector harbouring an HPH gene cassette, and was used in Agrobacterium tumefaciens-mediated transformation of the L. maculans isolate Leroy (Gardiner and Howlett, 2004). A fluorescence microscope (Leica MZFLIII; Leica, http://www.leica.com) with 480/440-nm excitation and 510-nm barrier filters was used for GFP observations.

A GPDA::GUS transgenic L. maculans PHW1245 isolate was also included in the analysis, and GUS staining was performed as previously described (Weigel and Glazebrook, 2002).

Quantitative real-time PCR

Quantitative measurements of L. maculans growth (Brouwer et al., 2003) were performed on Col-0, rlm3-1, rlm3-2, rlm3-3, rlm3-4, rlm3-4, 35S::RLM3TIR-X, RLM3::RLM3Col and rlm3An-1 genotypes using the following primers: 5′-CTTATCGGATTTCTCTATGTTTGGC-3′ and 5′-GAGCTCCTGTTTATTTAACTTGTACATACC-3′ for Arabidopsis, and 5′-GGTGTTGGGTGTTTGTTCCAC-3′ and 5′-GGCTGCCAATTGTTTCAAGG-3′ for L. maculans. The PCR conditions were 2 min at 50°C and 10 min at 95°C, followed by 40 cycles of 15 sec at 95°C, and 50 sec at 64°C, followed by a dissociation curve analysis. Total plant RNA was isolated using the Aurum Total RNA Mini Kit (Bio-Rad, http://www.bio-rad.com) as described by the manufacturer. For cDNA syntheses, the iScript cDNA Synthesis Kit (Bio-Rad) was used with 500 ng of total RNA, as described by the manufacturer. The forward primer was designed with the 5′ part on exon 3 and the 3′ part on exon 4 (5′-GAGCTACAATCGAAGCATGG-3′), and the reverse primer was designed on exon 4 (5′-TCAGTTGAGCTATCTATGCCACA-3′), of At4g16990, and were assessed relative to the expression of polyubiquitin 10 At4g05320 (5′-GGCCTTGTATAATCCCTGATGAATAAG-3′ and 5′-AAAGAGATAACAGGAACGGAAACA TAGT-3′) (Czechowski et al., 2005). For the expression analysis, PCR conditions of 2 min at 50°C and 10 min at 95°C, followed by 40 cycles of 15 sec at 95°C, and 60 sec at 60°C were used, followed by a dissociation-curve analysis. All amplifications were performed in an ABI 7000 thermocycler (Applied Biosystems, http://www.appliedbiosystems.com) with 3 μm primers, 5 ng template DNA or cDNA per 25-μl reaction, together with 12.5 μl Power SYBR® Green PCR Master mix (Applied Biosystems). The quantitative real-time PCR results were statistically evaluated using the Student’s t-test.

RT-PCR

RT-PCR was performed using the same cDNA as for the qPCR. For the analysis the following PCR conditions were used: 5 min at 95°C, followed by 35 cycles of 40 sec at 95°C, 40 sec at 55°C and 1.5 min at 72°C, with 10 min at 72°C for the final extension. Two different primer pairs were used: 5′-AAGATG TCCGCCACTCATTAGT-3′ and 5′-CATAGATAGCTCAACTGATG-3′, as well as 5′-AAGATGTCCGCCACTCATTAGT-3′ and 5′-GCAAGAAGTACAATATATGTGG-3′, spanning exons 1–4, but amplifying two different sized fragments, approximately 700 and 600 bp, respectively.

Callose staining and quantification

Leaves were collected at 2 dpi and were stained with aniline blue, and image analysis was performed as previously described in Kaliff et al. (2007). The staining was performed on 9–23 biological replicates of each genotype. The mean and median values for each inoculation site were used as estimators of callose deposition for each leaf sample, and were compared using a one-tailed t-test.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

We thank E. Holub for screenings of rlm3 material against H. parasitica and A. candida isolates, B. Staskawicz for providing P. syringae with avr genes, F. Mauch for providing P. brassicae isolates, S. Bohman for the Col-0 × An-1 material, T. Mengiste for the bik1 and bos mutants and J. Jones for providing the JatY clone. We also want to thank R. Hopkins for language corrections. This work was supported by the Wallenberg Foundation, the Swedish Research Council (VR) and the two graduate research schools FGB and IMOP.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Table S1. Differentially expressed genes with higher expression in the resistant pool. Genes with >2-fold higher expression in resistant pool found in more than 1 array are listed. Differential expression is expressed as M value (log2(Resistant/Susceptible)). V1 constitutes array comparison of pools of 10 individual resistant and 10 susceptible F3 lines deriving from individual F2 parents. V2 comprises array comparison of pools of 15 resistant and susceptible F3 lines, and comparison between the parental accessions Col-0 and An-1 from V3. V3 was dye-swapped. Shaded entries are the most significant in a B test when arrays V1-V3 were analysed together. Entries in italic are below threshold (M>1) and bold indicate very strong up-regulations in the resistant pool.

Table S2. Genetic mapping of the RLM3 locus. Markers close to candidate genes At4g13100 and At4g23290 show less linkage to the susceptible phenotype than the marker close to At4g16990. Mapping was performed on Col-0 x An-1 F3 lines.

Table S3. List of genes that have significantly higher expression in the susceptible pool over all arrays (p<0.05) and at on average >2-fold difference. No significant over-representation of any specific locus could be seen in this group of genes, suggesting that environmental differences rather than genotypic differences are the main contributor to differential expression. Shaded AGI genes are located at chromosome 4. The gene that had a consistent over twofold higher expression in susceptible pool is marked with an “X”. The genes are presented in order of B-test level of significance.

Table S4. Responses to fungi, oomycetes and bacteria on Col-0, An-1 and RLM3 KO mutants (rlm3-3, rlm3-4 and rlm3-5). Experiment repeated 1-5 times using 10-40 plants/assessment.

Figure S1. Reduction of transcriptional complexity, illustrated by number of differentially expressed (>2-fold difference) genes found in comparisons between parental accessions, comparisons between pools of 10 susceptible and resistant F3 lines, and comparisons between pools of 15 susceptible and resistant F3 lines. The decrease in complexity is interpreted to be dependent on the number of individual F3 lines and hence number of the close recombination events.

Figure S2. Disease phenotypes in response to L. maculans, A. brassicicola, A. brassicae and B. cinerea in Col-0, An-1 and RLM3 T-DNA mutants. (a) T-DNA insertions in At4g16990 (RLM3), orientation and location in relation to exons and predicted protein domains. The gene model shows predicted intron/exon composition (black boxes = exons) of the At4g16990.1 locus and numbers represent nucleotide base coordinates in the genomic gene model. A green box indicates TIR-domain, yellow NB-ARC and purple a sequence with homology to the RCC1 family. The mutation in salk_048620 (rlm3-3) is located in the promoter and inserted in the opposite direction to RLM3 (red arrowhead, susceptible phenotype). The salk_067449 (rlm3-4) mutation is also in the opposite direction and located in the end of exon 4 (blue arrowhead, intermediate phenotype). Salk_146865 (rlm3-5) is inserted in the same direction as RLM3 and located in the intron between exon 5 and 6 (black arrowhead, resistant phenotype). Water-inoculated control plants did not display any disease-like phenotypes on any of the genotypes assessed. Responses on various genotypes to four fungal pathogens. (b-f) L. maculans responses at 17 dpi. Experiment repeated six times, 10-20 plants/genotype. (g-k) Genotypes were further evaluated for allelism (F1) by crosses to An-1 and challenged to L. maculans. Responses at 14 dpi showed allelism between rlm3An and rlm3-3 or rlm3-4. 10 plants/genotype, repeated once and confirmed once with A. brassicicola. (l-p) Screening of the genotypes with A. brassicicola 10 dpi resulted in a susceptible “spotted” phenotype on rlm3-3 and An-1, whereas rlm3-4, rlm3-5 and Col-0 remained resistant. 20 plants/genotype, repeated twice. (q-u) B. cinerea inoculation on corresponding plant material 14 dpi resulted in moderate susceptible phenotypes on rlm3-3 and rlm3-4. 20 plants/genotype, repeated twice. (v-z) A. brassicae inoculation on the genotypes also resulted in the same pattern of susceptibility at 14 dpi. 20 plants/genotype, repeated twice.

Figure S3. Disease phenotypes in response to L. maculans on B. cinerea susceptible (bos) mutants, a) bos1, b) bos2 and c) bos3. xx dpi/ xx plants/genotype, repeated three times.

Figure S4. Quantitative evaluation of Pseudomonas syringae growth. Col-0, An-1 and T-DNA mutants of At4g16990 were dipped in P. syringae DC3000 and strains harbouring different avr genes in a DC3000 background. (a) Quantification at 0 dpi. (b) Quantification at 3 dpi. The standard deviation from four individual samples is shown. The experiment was repeated once.

Figure S5. Comparison of phenotypes between complemented with 35S::RLM3TIR-X and wild-type genotypes. (a) T1 plants harbouring 35S::RLM3TIR-X (the ~1 kb alternative Col-0 transcript) in An-1 background showing delayed local lesion development (left) compared to An-1 (right), 16 dpi of L. maculans. Phenotype confirmed twice in T2 and T3, 30 plants/line. (b) Necrotic lesions on systemic leaves after inoculation with a GFP-tagged L. maculans were similar between (left) transformed (T2) and (right) non-transformed An-1 plants. Observations on 30 plants/genotype. c) UV-fluorescence analysis revealed that L. maculans growth was severely reduced or absent in the transgenic line (left), compared to non-transformed control (right). Observations on 30 plants/genotype. Evaluation of L. maculans resistant T2 lines to three additional necrotrophic fungi. (d) A. brassicicola at 10 dpi showed enhanced resistance (left) compared to An-1 (right), 20 plants/genotype, repeated once in T3. (e) B. cinerea at 14 dpi showed enhanced resistance (left) compared to An-1 (right), 20 plants/genotype, repeated once in T3. (f) A. brassicae at 14 dpi showed enhanced resistance (left) compared to wild-type An-1 (right), 20 plants/genotype, repeated once in T3. (g) rlm3-4 (middle) showing susceptibility to L. maculans, was either completely restored (left) or showed an enhanced susceptible phenotype due to co-suppression (right) in T1 at 17 dpi when complemented with 35S::RLM3. Lines with clearly enhanced susceptible or resistant phenotypes were confirmed in T2, 30 plants/line.

(h) Evaluations of over-expression of RLM3TIR-X in the rlm1Lerrlm2Col background showed enhanced resistance in transgenic plants (left) compared to background control (right) at 16 dpi. Resistant phenotype confirmed in T2, 20 plants/line. (i) T1 plants of the At4g16990 promoter mutant rlm3-3 showed complete complementation (left), compared to wild-type (right), 21 dpi of L. maculans. Resistant phenotype confirmed in T2, 20 plants/line.

Figure S6. Quantitative measurements of L. maculans fungal biomass in different genotypes An-1 (rlm3An-1) (susceptible) and An-1 complemented RLM3::RLM3Col (resistant) and 35S::RLM3TIR-X (resistant) compared to rlm3-1 and rlm3-2 and rlm3-2 complemented (resistant), expressed as pg L. maculans nuclear DNA relative to ng Arabidopsis nuclear DNA. Each value derives from samples comprising 12 plants from three individual biological experiments. Leaf material was collected 14 dpi and bars indicate standard deviation between different biological replications.

Figure S7. RT-PCR analysis of RLM3 transcripts present in different genotypes. The gel picture shows normal splice form (~1.4 kb with exon2 (NB-ARC)) and alternative (~600 bp) without exon 2 (NB-ARC). Analysis of RLM3 expression from exon 1 to exon 4 using 5’-AAGATG TCCGCCACTCATTAGT-3′(forward) and 5’-GCAAGAAGTACAATATATGTGG-3’ (reverse) primers.

Please note: Blackwell Publishing are not responsible for the content or functionality of any supplementary materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

FilenameFormatSizeDescription
TPJ_3503_sm_Fig S2.tif27232KSupporting info item
TPJ_3503_sm_fig1.pdf29KSupporting info item
TPJ_3503_sm_fig3.tif17985KSupporting info item
TPJ_3503_sm_fig4a.jpg3889KSupporting info item
TPJ_3503_sm_fig4b.jpg3194KSupporting info item
TPJ_3503_sm_fig5.jpg3416KSupporting info item
TPJ_3503_sm_fig6.pdf176KSupporting info item
TPJ_3503_sm_fig7.pdf47KSupporting info item
TPJ_3503_sm_table1.doc112KSupporting info item
TPJ_3503_sm_table2.doc108KSupporting info item
TPJ_3503_sm_table3.doc160KSupporting info item
TPJ_3503_sm_table4.doc112KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.