LepR3, found in the Brassica napus cv ‘Surpass 400’, provides race-specific resistance to the fungal pathogen Leptosphaeria maculans, which was overcome after great devastation in Australia in 2004. We investigated the LepR3 locus to identify the genetic basis of this resistance interaction.
We employed a map-based cloning strategy, exploiting collinearity with the Arabidopsis thaliana and Brassica rapa genomes to enrich the map and locate a candidate gene. We also investigated the interaction of LepR3 with the L. maculans avirulence gene AvrLm1 using transgenics.
LepR3 was found to encode a receptor-like protein (RLP). We also demonstrated that avirulence towards LepR3 is conferred by AvrLm1, which is responsible for both the Rlm1 and LepR3-dependent resistance responses in B. napus.
LepR3 is the first functional B. napus disease resistance gene to be cloned. AvrLm1's interaction with two independent resistance loci, Rlm1 and LepR3, highlights the need to consider redundant phenotypes in ‘gene-for-gene’ interactions and offers an explanation as to why LepR3 was overcome so rapidly in parts of Australia.
Brassica oilseed crops (canola/oilseed rape) are the world's third leading source of vegetable oil with Brassica napus L. (AACC, n =19), Brassica rapa (AA, n =10) and Brassica juncea (AABB, n =18) being the main Brassica oilseed species cultivated worldwide (Singh, 2006). Over the past four decades blackleg disease (stem canker) has become a major limitation for canola production (Fitt et al., 2006). The Dothideomycete fungus Leptosphaeria maculans is the causal agent of the disease. The disease cycle initiates with spores that are released from infected plant debris. The spores germinate on cotyledons and leaves of young plants producing germ tubes that enter through stomata, leading to the subsequent intercellular growth of the hyphae. Except for the formation of lesions on the host leaves, L. maculans remains asymptomatic through most of its life cycle within the adult plant until it forms a canker at the crown of the stem, which results in lodging and significant yield loss (Howlett et al., 2001). The most efficient approach for controlling blackleg disease is deployment of plant resistance genes (Delourme et al., 2006).
There have been many genetically characterized Brassica genes described for resistance to L. maculans, though, as yet, none of these genes have been cloned. The majority of these genes have been mapped to the Brassica A genome, including Rlm1 on chromosome A07 and LepR3 on A10 (Ansan-Melayah et al., 1998; Balesdent et al., 2001, 2002; Delourme et al., 2004, 2006; Mayerhofer et al., 2005; Yu et al., 2005, 2008; Long et al., 2011; Raman et al., 2012). At least three R-genes are carried in the B genome (Chèvre et al., 1996, 1997; Balesdent et al., 2002; Christianson et al., 2006; Kutcher et al., 2010) and none have yet been reported for the C genome. These genes convey resistance to L. maculans isolates in a race-specific manner. The corresponding avirulence (Avr) genes are often located in genetic clusters in the L. maculans genome (Balesdent et al., 2002) and several have been mapped or cloned (Cozijnsen et al., 2000; Gout et al., 2006; Fudal et al., 2007; Parlange et al., 2009; Ghanbarnia et al., 2012). The recent release of both the B. rapa (AA) genome (Wang et al., 2011) and the L. maculans genome (Rouxel et al., 2011) sequences should help facilitate a greater understanding of this host-pathogen system.
The L. maculans resistance (LepR) genes, LepR1, LepR2 and LepR3 have been introgressed into B. napus through a resynthesis of B. napus from a B. rapa ssp. sylvestris × B. oleracea var. alboglabra interspecies cross (Crouch et al., 1994; Buzza & Easton, 2002). LepR1 and LepR2 were identified and mapped to linkage groups A02 and A10, respectively (Yu et al., 2005). The third major blackleg resistance gene, LepR3, is present in the B. napus cultivar ‘Surpass 400’ and has been mapped to linkage group A10, at a distance of 11.7 cM from the LepR2 locus (Yu et al., 2008). LepR3-initiated resistance in ‘Surpass 400’ is associated with a hypersensitive response (Li et al., 2004, 2007), a feature of many, although not all, plant resistance gene (R-gene)-mediated defense responses (Hammond-Kosack & Jones, 1996). LepR3 was first described as a single, dominant gene from field-based studies (Li & Cowling, 2003), although several recent reports suggest the presence of two independent R-genes in ‘Surpass 400’ (Van de Wouw et al., 2009; Long et al., 2011). It was demonstrated that two independent L. maculans Avr genes, AvrLm1 and AvrLmS, trigger defence responses in ‘Surpass 400’, thus it was inferred that ‘Surpass 400’ contained both Rlm1 and a second gene referred to as ‘RlmS’ (Van de Wouw et al., 2009).
Plant resistance responses to pathogens fall under two general categories based on the pathogen molecules that trigger the responses; ‘pattern-triggered immunity’ (PTI) where slowly-evolving pathogen-associated molecular patterns (PAMPS) trigger basal defense responses, or ‘effector-triggered immunity’ (ETI), in which specific pathogen effectors, targeted to disrupt PTI, either directly or indirectly trigger specific R-genes (Jones & Dangl, 2006). Though PTI and ETI are often described as separate pathways, the responses likely function as a coordinated network (Katagiri & Tsuda, 2010). A common feature of most plant R-genes is the presence of leucine rich repeat (LRR) motifs that play a major role in recognition of pathogen effectors by facilitating protein–protein interactions (McDowell & Woffenden, 2003).
The majority of the plant R-proteins are predicted to be located intracellularly; however, there are several examples of extra-cytoplasmic LRR (eLRR)-containing R-proteins that are anchored to the plasma membrane via a transmembrane (TM) domain (Kruijt et al., 2005; Yang et al., 2012). One group of well-characterized eLRR R-genes are the tomato Cf genes that confer resistance against Cladosporium fulvum, the causal agent of tomato leaf mold disease. Cf genes encode a group of receptor-like proteins with recognition specificity for different avirulence proteins encoded by C. fulvum (Wulff et al., 2009). Hyphae of C. fulvum enter the intercellular space through stomata and are confined to the intercellular space (Thomma et al., 2005), which is analogous to the infection and growth of L. maculans hyphae within the plant host tissues. While three of the L. maculans effectors – AvrLm1, AvrLm4-7 and AvrLm6 – have been cloned (Gout et al., 2006; Fudal et al., 2007; Parlange et al., 2009) their function and subcellular location in the host remains to be determined. This paper provides insight into the molecular recognition of L. maculans by its host. Here we report high-resolution mapping of the LepR3 locus on linkage group A10 of B. napus, the investigation of collinearity between the LepR3 region of B. napus, B. rapa and Arabidopsis thaliana, and the cloning of the LepR3 gene, which encodes a receptor-like protein. We also demonstrate that the L. maculans avirulence gene AvrLm1 confers avirulence to both LepR3 and Rlm1. We believe this to be the first published report of the cloning of an R-gene from the important oilseed crop species Brassica napus.
Materials and Methods
L. maculans isolates and inoculum
Two L. maculans isolates characterized for their virulence on ‘Surpass 400’; ‘S005’ (avrLm1, avrLmS) and ‘P042’ (AvrLm1, AvrLmS) (Van de Wouw et al., 2009) and progeny of a cross between the two; ‘3R5’ (AvrLm1, avrLmS), ‘3C3’ (AvrLm1, AvrLmS) and ‘3R11’ (avrLm1, avrLmS) were kindly provided by Angela van de Wouw, University of Melbourne (Australia). Additional ‘AvrLm1’ isolates ‘WA74’, ‘87-41’ and ‘2354’, as well as ‘avrLm1’ isolate ‘99-53’ were sourced from the Rimmer Collection, AAFC Saskatoon. Inoculum was produced at a concentration of 2 × 107 pycnidiospores ml−1 as described in Yu et al. (2005). To confirm AvrLm1 genotypes of the isolates 9 μl of inoculum was added to 1 μl 10× TE (100mM Tris-Cl pH8, 10mM EDTA pH8), 2.5% Tween20, incubated at 95°C for 5 min then used as template for PCR with AvrLm1 F & R primers (see the Supporting Information, Table S1).
B. napus lines and populations
First backcross (BC1) and third backcross (BC3) generations of B. napus, segregating for LepR3, produced via crosses between the susceptible spring-type double-haploid line ‘Topas DH16516’ and the resistant variety ‘Surpass 400’ backcrossed to the susceptible parent were used for the mapping study. A total of 446 plants (97 BC1 and 349 BC3) were screened for resistance to the isolate ‘3R5’ (AvrLm1, avrLmS). All screening included ‘Topas DH16516’ (no resistance) and ‘Surpass 400’ (LepR3, RlmS) controls. The resistant B. napus varieties ‘Columbus’ (Rlm1, Rlm3) (Balesdent et al., 2006), ‘Quinta DH24288’ (Rlm1, Rlm3) (Kutcher et al., 2010) and ‘Quantum’ (Rlm3) (R. Kutcher, pers. comm.) were also used for pathotyping of isolates.
Determination of resistance phenotype
Seedlings were germinated in 96-cell trays in a controlled growth chamber (20°C, 16 h days (light intensity c. 450 μmol m−2 s−1 at bench level) and 18°C, 8 h nights). After 7 d the cotyledons were inoculated with a spore suspension of L. maculans (2 × 105 pycnidiospores per wound site, eight plants per test) and were rated 14 d postinfection. The resistance phenotype of the seedlings was rated using a continuous 0–9 scale modified after (Koch et al., 1991) as shown in Fig. S1. A score of 0–4 was classified as ‘resistant’, 5 as ‘intermediate’ and 6–9 as ‘susceptible’.
Molecular marker characterization of mapping populations
The approximate position of LepR3 was first determined using 87 BC3 individuals phenotyped for resistance using the L. maculans isolate ‘3R5’ (AvrLm1, avrLmS) and surveyed using several microsatellite markers from the region of chromosome A10 identified previously as containing the LepR3 locus (Yu et al., 2008). All microsatellite markers used in this work (prefixed with ‘sN’ or ‘sR’) were developed by and obtained from A. G. Sharpe and D. J. Lydiate, AAFC Saskatoon, Canada. After determining a wide interval for the gene the remaining BC1 and BC3 seedlings used in the fine mapping population were surveyed with the microsatellite markers sN8502 and sR9355, flanking the LepR3 region. Individuals displaying a recombination event between the two microsatellite markers were reserved for the mapping study. After rating for disease at 14 d post-inoculation, the infected cotyledons were removed to prevent spread of the disease. Four leaf discs totaling c. 20 mg (fresh weight) were collected from young leaves and freeze-dried for DNA extraction by cetyltrimethylammonium bromide (CTAB) method scaled down from (Murray & Thompson, 1980) for 96-well format. The initial genotypes for these ‘recombinants’ and their resistance phenotypes were confirmed by examining the segregation of resistance from BC1S1 that is self-progeny from the recombinant BC1 individuals. The PCR reactions for all markers were performed as described by Hughes et al. (2003). Genotyping was performed using a MegaBACE capillary sequencer (GE Health, Mississauga, ON, Canada).
Targeted marker creation
Further resolution was added to both maps by the targeted creation of markers that took advantage of small insertion/deletion events (‘indels’) between the parental lines. These markers were produced after first identifying the syntenic position of the microsatellite markers relative to the A. thaliana genome (Arabidopsis Genome Initiative, 2000; http://www.arabidopsis.org) by blastn (Altschul et al., 1997). After determining the collinear intervals spanning the target genes, the Brassica rapa Genome Project database was searched via blastn (http://www.brassica-rapa.org) for bacterial artificial chromosomes (BACs) matching the collinear intervals, using either microsatellite clone sequence or A. thaliana genes that fell within the collinear spans. Primers were then designed using these BAC sequences and used to amplify 1300–1400 bp fragments from both ‘Topas DH16516’ and ‘Surpass 400’ parents. These fragments were cloned using TOPO TA Cloning Kits (Invitrogen) and screened for correct-sized inserts via PCR. For each fragment three or four plasmids were purified (QIAprep Spin Miniprep Kit; Qiagen), sequenced (Sharpe Lab, PBI, Saskatoon, SK, Canada), aligned using Vector NTI software (Invitrogen) and polymorphic indels identified. New fluorescently-labeled primers were designed flanking the polymorphic sequence and used for further genotypic characterization of the mapping population. All indel markers (prefixed with ‘Ind’) designed during this study are listed in Table S1. The PCR reactions and genotyping were performed as for the microsatellite markers (above) except for the marker ‘Ind10-12’ for which the following program was used: 95°C, 3 min; (95°C, 30 s; 54°C, 30 s; 72°C, 45 s) × 35; 72°C, 5 min; 12°C hold.
Interval spacing was determined manually from recombination frequencies in the recombinant subpopulation (cMBC = (x/n)*100, where x = recombination events and n = total population size). Linkage between markers and the LepR3 locus was verified using joinmap 4 software (Stam, 1993) using a LOD threshold > 3 and the Kosambi mapping function (Kosambi, 1944). The marker positions were later aligned to the complete B. rapa genome sequence (www.brassicadb.org) following its public release.
Identification, cloning and transformation of candidate gene
A B. rapa BAC spanning the LepR3 interval was annotated for gene content using fgenesh software (www.softberry.com) trained to ‘Dicot plants (Arabidopsis)’ to predict gene content. The protein translations were then used to search for homologous A. thaliana genes via blastp, allowing a candidate gene to be identified. The candidate gene was cloned by first using Phusion DNA Polymerase (New England Biolabs, Whitby, ON, Canada) and primers anchored within the flanking genes to amplify a fragment of c. 7 kb spanning the candidate gene locus from both parental lines. After subcloning and partial sequencing, primers ‘GW-BnRLP F and R’ were designed with Gateway attB1 and attB2 recombination tags, respectively (Table S1), and used to amplify a nested fragment of 4.25 kb, containing the full candidate gene locus from ‘Surpass 400’, including an additional 1144 bp upstream and 535 bp downstream of the coding region to capture the native promoter and termination sequences, respectively. A corresponding fragment was also amplified from ‘Topas DH16516’ using primers ‘GW-BnRLP F and Rb’ (Table S1). The fragments from both parents were cloned into the Gateway (Invitrogen) entry vector pDONR-Zeo and sequenced. The resistant ‘Surpass 400’ parent allele was transferred to the Gateway-compatible plant transformation vector pMDC123, which contains the BastaR gene for selection in planta (Curtis & Grossniklaus, 2003). The construct was transferred to Agrobacterium tumefaciens strain ‘GV3101 pMP90’ and used to transform ‘Topas DH16516’ essentially as described by De Block et al. (1989).
Sequences of the ‘Topas DH16516’ lepR3 and ‘Surpass 400’ LepR3 alleles were deposited in GenBank (JX880109 and JX880110, respectively).
Analysis of B. napus transformants and LepR3 transcript
Transformants that tested positive for insertion of the candidate sequence via PCR were self-pollinated to produce T1 seeds, which were tested for segregation of the resistance phenotype using the L. maculans isolate ‘3R5’ as described above. A T1 seed lot showing 3R : 1S segregation consistent with a single insertion event was selected for further phenotypic analysis and production of homozygous T2 seed. Total RNA was isolated from cotyledons using a PureLink RNA Mini Kit (Ambion) and a near-full length copy of the LepR3 transcript was amplified from both ‘Surpass 400’ and a resistant T1 plant by reverse transcription using Superscript III One-Step RT-PCR System (Invitrogen) and the gene-specific primers ‘RTR3C F & R’, while ‘RT-Top Fc & Rc’ primers were used to amplify the ‘Topas DH16516’ transcript (Table S1). A GeneRacer kit (Invitrogen) was used for 5′ RACE analysis, using the gene-specific primer ‘RTR3-5′b’ (Table S1). Identification of functional domains within the LEPR3 protein was performed using InterProScan (http://www.ebi.ac.uk/Tools).
Analysis of LepR3-transgenic B. napus with AvrLm1-transgenic L. maculans
A Gateway-compatible fungal transformation vector was created by modifying the vector pPK2 (Covert et al., 2001) using the Gateway Vector Conversion Reagent System (Invitrogen). A 2221 bp genomic fragment spanning the L. maculans AvrLm1 locus, including 1131 bp upstream and 423 bp downstream of the coding region, was amplified from the AvrLm1 isolate ‘WA74’ using Phusion DNA polymerase (New England Biolabs) and the Gateway attB-tagged primers ‘AvrLm1G F and R’ (Table S1), transferred to the Gateway entry vector pDONR-Zeo and then to the GW-pPK2 fungal transformation vector. After confirmation of the insert by sequencing, the GW-pPK2-AvrLm1G construct was transferred into the A. tumefaciens strain ‘AGL1 pTiBo542’ and used for transformation of pycnidospores from the L. maculans isolate ‘3R11’ (avrLm1), as described by (Utermark & Karlovsky, 2008). Selection of transformed colonies was achieved after 7 d incubation on Czapek Dox medium supplemented with 120 μg ml−1 Hygromycin B. A pycnidiospore suspension produced from the resulting ‘3R11 : AvrLm1G’ transformant, along with the untransformed ‘3R11’, were used for inoculation of ‘Topas DH16516’, ‘Columbus’, ‘Quinta DH24288’ (Rlm1, Rlm3), ‘Quantum’ and ‘Surpass 400’ controls as well as three homozygous T2 lines produced from LepR3-transgenic ‘NLA8’ T1 plants, as described above.
Mapping of the LepR3 resistance locus
Mapping of the LepR3 locus was conducted using the isolate ‘3R5’ which had previously been pathotyped as ‘AvrLm1, avrLmS’ (A. Van de Wouw, pers. comm.). ‘3R5’ was virulent on the susceptible B. napus parental line ‘Topas DH16516’ and avirulent on the resistant parent ‘Surpass 400’ (Table 1). The initial mapping of resistance using the first 87 ‘Topas DH16516’ × ‘Surpass 400’ BC3 plants correlated strongly, though not perfectly, with the original map position described for LepR3 on chromosome A10 (Yu et al., 2008). The presence of one ‘recombinant’ individual displaying a recombination event between the microsatellite marker sN2428R and LepR3 suggested that the gene was located outside of the genetic interval previously described. Assaying the initial population with additional microsatellite markers located in the LepR3 region of A10 identified four markers (sR2428R, sN1982, sR1443 and sR0685) that mapped inside the interval defined by the flanking markers sN8502 and sR9355. Mapping of the locus with the full population of 446 individuals gave a 1 : 1 segregation ratio (212R : 234S; χ2 (1 : 1) = 1.09, P =0.297), indicating that resistance to ‘3R5’ is conferred by a single dominant gene. All six microsatellite markers spanning the map interval also conformed to the expected 1 : 1 ratio (χ2 (1 : 1) = 0.002–0.56, P =0.453–0.962). Sixteen individuals carrying recombination events between the flanking microsatellites (sN8502 and sR9355) were detected. This resulted in a genetic distance of 3.4 cM for the target region and a revised position for the LepR3 locus in a 0.8 cM interval between the microsatellite markers sR1443 and sR0685 (Fig. 1), c. 2.1 cM below the previously described location (Yu et al., 2008).
Table 1. Phenotypic reaction of Leptosphaeria maculans isolates on Brassica napus controls
Investigation of collinearity and targeted marker creation
The six microsatellite markers and their respective genomic clones were used for the analysis of collinearity between the LepR3 interval of B. napus and the A. thaliana genome. The sequences of five of the six microsatellite clones (except sR9355, which has no significant homology to A. thaliana) matched homologous sequences in a 695 kb region of chromosome 5 of A. thaliana (At5). The five matching markers were collinear in the two species and the relative spacing was conserved (not shown). The specific interval of B. napus chromosome A10 in which LepR3 resided was found to be syntenic to At5, between the genes At5g12150 (sR1443) and At5g11700 (sR0685). Further definition of the LepR3 region was obtained by developing genetic markers from B. rapa BACs that had homologous sequences in the syntenic region of At5. Several sequenced B. rapa BACs within or adjacent to the LepR3 interval were identified based on their homology to collinear A. thaliana genes. These BACs were used to produce seven targeted indel markers (prefixed ‘Ind10’, Table S1) that were placed within the 3.4 cM LepR3 interval and again demonstrated a conserved order with respect to the corresponding homologous sequences in A. thaliana (not shown). These new markers were used to refine the LepR3 map and define the locus to an interval of 0.4 cM (Fig. 1), syntenic to the At5g11880 – At5g11900 interval of At5 (Fig. 2).
Identification of candidate gene
Annotation of the B. rapa BAC ‘KBr080E24’ containing the chromosome A10 region matching the LepR3 interval revealed the presence of a 1890 bp predicted receptor-like protein (RLP)-coding gene that was nonsyntenic to the A. thaliana interval, with its best match (e-value of 5e−66) being AtRLP32 (At3g05650). Upon the release of the B. rapa genome (Wang et al., 2011) the gene was annotated as ‘Bra008930’ (Fig. 2). No other candidate genes were detected within the target interval. Amplification and sequencing of the candidate locus from B. napus revealed a predicted gene of 2853 bp (one exon) from the susceptible ‘Topas DH16516’ and 2556 bp (one exon) from the resistant cv ‘Surpass 400’. The B. napus homologues appeared to be larger than the predicted B. rapa gene (1890 bp). While the ‘Topas DH16516’ and ‘Surpass 400’ genes showed conserved homology over most of their lengths they also contained large indel variations: an indel in the ‘Topas DH16516’ 5′ region, two indels in the ‘Surpass 400’ LRR region and a large indel 3′ of the ‘Surpass 400’ CDS (Fig. 3), along with numerous amino acid variations within the CDS (Fig. S2).
Analysis of LepR3 transcript
Amplification of a near-full length LepR3 transcript from ‘Surpass 400’ cotyledon total cDNA produced two fragments: one with the expected size of 2431 bp (LepR3 candidate) as well as a second fragment of only 2262 bp. The second, smaller fragment was identical to the first except for the removal of 169 bp from within the CDS (Δ386..536). Rapid amplification of cDNA ends (RACE) analysis of the LepR3 transcript allowed for the determination of a 37 bp 5′ UTR, though the 3′ UTR was not successfully amplified. A transcript of the susceptible (lepR3) open reading frame (ORF) was also amplified from ‘Topas DH16516’ cDNA using unique primers, demonstrating that the LepR3 and lepR3 alleles are both expressed in their respective cultivars.
The predicted full-length transcript encodes one exon producing a protein 851 amino acids in length. InterProScan analysis of the predicted LEPR3 protein revealed a typical receptor-like protein motif structure (Fritz-Laylin et al., 2005): a signal peptide located at the N-terminal (residues 1–26) followed by an extensive eLRR region (34–731), a GXXXG-type transmembrane motif and terminating with a 24 amino acid cytoplasmic C-terminal region. The smaller, partially deleted fragment contained a premature stop codon after only 123 amino acids, meaning the resulting peptide would only contain a small portion of the eLRR domain.
Analysis of LepR3 transformants
A LepR3 genomic clone, driven by its native promoter, was transferred to the susceptible B. napus cv ‘Topas DH16516’. A total of 13 transformants (T0) were screened with the ‘Ind10-12’ marker positioned within the LepR3 CDS (Table S1) to confirm the presence of an insertion event, revealing five positive and eight negative transformants. Of the five positive plants, only one (‘NLA8’) displayed a balanced heterozygous-type polymorphism for ‘Ind10-12’ while the other four plants displayed polymorphisms heavily favoring the ‘Surpass 400’ allele, suggesting multiple insertion events (data not shown). After allowing self-pollination, eight T1 seedlings produced from each of four mature T0 plants (two negative lines ‘NLA6’ and ‘NLA7’ and two positive lines ‘NLA8’ and ‘NLA9’) were tested with ‘3R5’, the same L. maculans isolate used to map the LepR3 locus (data not shown). Both negative lines proved to be 100% susceptible, equal to the ‘Topas DH1516’ susceptible control, ‘NLA8’ displayed six resistant and two susceptible T1 individuals (Fig. 4) and ‘NLA9’ showed 100% resistance. To test the segregation ratio for resistance in the ‘NLA8’ T1 progeny, an additional 40 T1 individuals were screened with ‘3R5’, revealing a segregation for resistance (38R to 10S) conforming to a 3 : 1 ratio (χ2 (3 : 1) = 0.44, P =0.501), as expected from the segregation of a single dominant gene, suggesting a single insertion event and further confirming cloning of the LepR3 gene (Table 1). Several resistant ‘NLA8’ T1 individuals were self-pollinated and the resulting T2 seed was screened with ‘3R5’ to select LepR3-homozygous lines (data not shown).
Both the near full-length LepR3 transcript and the second, partially deleted LepR3 transcript (first detected from ‘Surpass 400’ cDNA) were also detected in the resistant, LepR3-transformed ‘NLA8’ T1 plants. However, no LepR3 transcript, partial or otherwise, was detected in the ‘Topas DH16516’ susceptible control, showing it to likely be a product of the LepR3 gene and not a related transcript from the B. napus genome (data not shown).
The L. maculans effector AvrLm1 confers avirulence to LepR3
Resistance of LepR3-containing B. napus cv ‘Surpass 400’ to L. maculans isolates carrying AvrLm1 had previously been attributed to the presence of Rlm1 in this cultivar. When several L. maculans isolates were inoculated on the LepR3-transgenic ‘NLA8’ T1 lines, all six of the isolates that contained AvrLm1 were avirulent on the LepR3-transgenic material while the three isolates lacking AvrLm1 were virulent. This suggested that AvrLm1 conferred avirulence to LepR3. The interaction of the L. maculans isolates with the control B. napus lines and ‘NLA8’ T1 plants is summarized in Table 1. The functionality of AvrLm1 in these isolates was confirmed by their avirulence on ‘Columbus’ (Rlm1, Rlm3) and virulence on ‘Quantum’ (Rlm3) and the AvrLm1 genotypes were also confirmed via PCR (data not shown). One isolate, ‘99-53’, was shown to be virulent towards Rlm1, yet still induced a resistance response on ‘Surpass 400’, and was therefore designated as ‘avrLm1, AvrLmS’, as described by Van de Wouw et al. (2009). This isolate was also virulent on the LepR3 transgenic material. We were unable to determine the ‘AvrLmS’ status of the AvrLm1 isolates ‘WA74’, ‘87-41’ or ‘2354’ at the time of scoring as the phenotype was masked by the isolates’ reaction to LepR3.
In total, 88 T1 individuals were screened with ‘AvrLm1’ isolates, producing a total segregation of 68 resistant to 20 susceptible plants (χ2 (3 : 1) = 0.24, P =0.622), consistent with a single insertion event for the LepR3 construct in the ‘NLA8’ T0 individual. Our results for the isolates ‘S005’ and ‘P042’ confirmed the AvrLm1 designations reported previously (Van de Wouw et al., 2009).
To further validate the recognition of AvrLm1 by LepR3 we transferred an AvrLm1 genomic clone driven by its native promoter to the virulent L. maculans isolate ‘3R11’ (avrLm1, avrLmS). The untransformed isolate and the transgenic ‘3R11 : AvrLm1G’ were both virulent on the ‘Topas DH16516’ and ‘Quantum’ controls, yet differed in their reaction to Rlm1 and LepR3 lines (Table 2). ‘3R11’ was virulent on the Rlm1 lines ‘Columbus 6-1’ and ‘Quinta DH24288’, the LepR3 line ‘Surpass 400’ and all three T2LepR3-transgenic homozygous lines (‘NLA8-2, ‘NLA8-3’ and ‘NLA8-9’) as expected. However, the transgenic ‘3R11 : AvrLm1G’ was avirulent on all Rlm1 or LepR3 material, including the LepR3-transgenic T2 lines, demonstrating the complementation of both R-genes by AvrLm1 (Fig. 5).
Table 2. Phenotypic reaction of AvrLm1-transgenic Leptosphaeria maculans on Brassica napus lines
The region of the B. napus chromosome A10 in which the LepR3 locus is situated had previously been shown to share a high degree of low-resolution collinearity with chromosome 5 of A. thaliana (Parkin et al., 2002, 2005). The results of our study highlight the conserved nature of this collinear region at high resolution, with perfect marker order conservation. No change in marker order was detected over this region of A. thaliana (not shown), B. rapa or B. napus (Fig. 1). This close synteny allowed the development of targeted markers that reduced the size of the LepR3 interval to 0.4 cM (Fig. 1).
In a previous study of the LmR1 gene on linkage group A07 of B. napus (Mayerhofer et al., 2005) collinearity with A. thaliana chromosome 1 (At1) was exploited to identify a number of candidate genes associated with disease resistance, though this study was complicated by an inversion that differentiated B. napus and A. thaliana in this region. By contrast, the search for candidate genes for LepR3 was facilitated by the close collinearity over the interval between markers sN8502 and sR0685 on A10 of B. napus and the corresponding region between At5g13810 and At5g11700 on At5. Although none of the A. thaliana genes were likely homologues of LepR3, the collinear nature of the relationship between A. thaliana and Brassica spp. over this region allowed the retrieval of B. rapa BAC sequences related to the target interval, one of which harbored the LepR3 homologue. While this approach to finding a candidate gene was employed successfully in this study, the recent release of the B. rapa genome (Wang et al., 2011) makes this approach mostly obsolete for finding Brassica A genome genes. However, using A. thaliana as a bridging genome may still be useful for placing Brassica B and C genome BACs within target intervals, at least until sequencing of these genomes is completed.
The LepR3 gene discovered in this study is a member of the receptor-like protein family of genes first linked to disease resistance in plants with the cloning of the Cf-9 gene for resistance to Cladosporium fulvum in tomato (Jones et al., 1994). Both Cf-9 and the well-characterized tomato RLP Ve1 (Kawchuk et al., 2001) have been previously clustered in the same ‘superclade’ as At3g05650 (Fritz-Laylin et al., 2005); the best match to LepR3 in A. thaliana (this study). A large family of 56 RLP genes is found in the A. thaliana genome (Fritz-Laylin et al., 2005). Each of the two Brassica genomes in B. napus (AACC) have evolved through the whole-genome triplication of an ancestral Brassicaceae genome closely related to A. thaliana (Lagercrantz, 1998; Nelson & Lydiate, 2006; Ziolkowski et al., 2006) thus the RLP gene family in B. napus could number close to 300 members.
The nested PCR strategy employed in this study to clone LepR3 enabled us to efficiently amplify a unique fragment without interference from any other RLP locus in the genome and without knowing the precise sequence of the target, or the added time and expense of building and screening a large-insert genomic library. This method may also prove useful for the cloning of genes from other large gene families. However, this approach has the limitation of only being applicable for use in examining candidate genes detected in the B. rapa genome. A full B. napus reference genome sequence would likely be a better resource for this approach; however, a full pathological examination of the reference B. rapa var. ‘Chiifu’ for the presence of resistance genes for all of the major B. napus pathogens would be very useful.
Investigation of the LepR3 locus in the resistant ‘Surpass 400’ and the susceptible ‘Topas DH16516’ revealed three regions harboring insertion/deletion (indel) differences between the two loci (Figs 3, S2). However, these indels do not disrupt the expression of the two alleles as we were able to detect transcripts from both loci. It is not yet known what role, if any, the truncated form of the LepR3 transcript plays in the resistance response, though we speculate that it may be the product of post-transcriptional control of transcript abundance. While alternative splicing has been described in NBS-LRR R-genes (Gassmann et al., 1999; Marathe et al., 2002; Schornack et al., 2004; Ferrier-Cana et al., 2005) we are unaware of any reports of alternative splicing or ‘exon editing’ in RLPs.
Here we provide evidence that the L. maculans effector AvrLm1 confers avirulence to both LepR3 and Rlm1. It has already been demonstrated that the addition of AvrLm1 to an avrLm1, AvrLmS’ L. maculans isolate produces a stronger avirulence reaction on ‘Surpass 400’ than the untransformed isolate (Van de Wouw et al., 2009). The recognition of a single Avr gene by two plant R-genes is not unprecedented in the L. maculans-B. napus pathosystem, since the L. maculans Avr gene AvrLm4-7 was shown to trigger a defense response in both Rlm4 and Rlm7-carrying B. napus lines (Parlange et al., 2009), although in this case both of the R-genes are located in the same genetic cluster, and may be allelic variants (Delourme et al., 2004). Only three L. maculans Avr genes have been cloned to date, yet two of them have been shown to trigger resistance responses from two separate Brassica resistance loci. This is suggestive of the broad use of these Avr genes by the pathogen and possible chromosomal rearrangement and/or horizontal transfer of R-genes among the Brassica spp.
Are Rlm1 and LepR3 the same gene? Rlm1 occupies a different chromosomal location (A07 – Delourme et al., 2004) than LepR3 (A10 – this study) thus, genetically, they should be considered separate resistance loci. Van de Wouw et al. (2009) speculated that the LepR3 resistance locus may have been produced via the translocation of Rlm1 from A07 to A10 during the original Crouch et al. (1994) resynthesis used to create ‘Surpass 400’, or that ‘Surpass 400’ contained both Rlm1 and LepR3, in addition to RlmS. During our mapping of the LepR3 locus there was no evidence to support either of these hypotheses; all markers behaved as predicted by their syntenic location on chromosome A10 of B. rapa, including the LepR3 locus itself. Had the LepR3 gene been nonsyntenic (i.e. located on A07) in B. rapa, we would not have been able to identify the candidate RLP gene on A10 of the B. rapa genome sequence. Alternatively, if both Rlm1 (A07) and LepR3 (A10) were present in ‘Surpass 400’ we would have observed a 3R : 1S segregation ratio for the BC population (two independently-assorting loci) when characterizing the plants with ‘3R5′ (AvrLm1, avrLmS). Instead, we saw the expected 1R : 1S ratio, with all of the observed resistance being explained by the A10 LepR3 locus and no distortion of associated markers to suggest it was the product of a recent translocation. It is still possible that Rlm1 and LepR3 are homologous sequences located at two independent loci within the B. napus genome, as nonsyntenic transposition of genes has been observed in Brassica genomes (Cheung et al., 2009). However, the chromosomal regions of A07 and A10 harboring the two genes do not share any homology (Parkin et al., 2005) and while recombination between homoeologous A and C genome chromosomes is frequently observed in resynthesized material (Szadkowski et al., 2010), there are no reports of nonhomoeologous translocation events occurring within the A genome of B. napus. Indeed, the Rlm1 locus was previously mapped to the same A07 location in both B. rapa and in a resynthesized B. napus (Leflon et al., 2007). The relationship between the LepR3 and Rlm1 resistance loci will be elucidated through the future fine mapping and study of the genomic region surrounding the Rlm1 gene. Meanwhile, we may be able to determine if the protein products of the redundant R-genes recognize the same or different epitopes of AVRLM1 through the study of naturally occurring or induced mutants. In this way we may be able to identify a version of AvrLm1 that confers avirulence to LepR3 and not Rlm1, similar to AvrLm7, a mutated allele of AvrLm4-7 which triggers Rlm7 but not Rlm4 (Parlange et al., 2009). However, at present, we have no evidence for such a gene.
While direct R-gene–Avr interactions have been demonstrated in a few plant–pathogen interactions (Jia et al., 2000; Deslandes et al., 2003; Dodds et al., 2006), other evidence has suggested that R-genes are not always triggered by direct interaction with pathogen effectors. The ‘guard hypothesis’ (reviewed in Jones & Dangl, 2006) suggests many R-genes are triggered indirectly by recognizing the disruption of other host cellular targets by pathogen effectors. In this scenario the R-protein is likely to be part of a multiprotein complex involved in monitoring the basal defense machinery of the cell and in triggering defense signaling by modifying the stability of the R-protein (reviewed in Belkhadir et al., 2004). Given that LepR3 codes for a protein with extracellular LRRs, it would seem likely that it would detect a fungal elicitor present outside of the cell, produced by the invading L. maculans during infection, suggesting a direct interaction. However, in studies with the well-characterized Cf RLP gene family of tomato, which is structurally similar to LepR3, no direct interaction between the RLPs and their corresponding AVR proteins has been detected. This suggests that the Cf proteins act as guards of the avirulence targets and interact with the AVR proteins in an indirect manner (Kruijt et al., 2005). For example, recognition of C. fulvum Avr2 effector by the Cf2 protein is mediated by the tomato cysteine protease Rcr3. Avr2 is a cysteine protease inhibitor that binds and inhibits Rcr3 (Rooney et al., 2005). However, unlike many fungal effectors, which are cysteine-rich, AVRLM1 only contains one cysteine residue. Cysteine-rich effectors are believed to withstand plant proteases, which are secreted into the host apoplastic space. Based on this, Gout et al. (2006) suggested that AVRLM1 may be localized to the host cytoplasm. Cloning of LepR3 and future determination of its protein interactions will help in elucidating the function and potential host target of AVRLM1.
Our results, demonstrating that AvrLm1 confers avirulence to LepR3, also offer a possible explanation as to the rapid loss of effective LepR3 resistance in B. napus material, including ‘Surpass 400′, in some parts of Australia soon after it was deployed (Sprague et al., 2006). Given that Rlm1 varieties were already in use in Australia before the release of ‘Surpass 400′ (Rouxel et al., 2003) and that the ‘breakdown’ of resistance was coincident with a large decrease in AvrLm1 allele frequency (Van de Wouw et al., 2010) it is possible that some populations of L. maculans had been driven towards a high proportion of virulent avrLm1 pathotypes because of previous exposure to Rlm1. This could have effectively enriched entire populations of the pathogen for virulence on LepR3, drastically reducing the gene's effective lifespan in these areas.
Van de Wouw et al. (2009) showed the L. maculans avirulence gene ‘AvrLmS’ triggers a second AvrLm1-independent R-gene in ‘Surpass 400’ referred to as ‘RlmS’. Our results, demonstrating that the avrLm1 isolate ‘99-53’ induces a resistance response from ‘Surpass 400’ but not Rlm1 lines or LepR3-transgenic ‘NLA8’ plants, corroborates the presence of both AvrLmS in L. maculans and RlmS in ‘Surpass 400’. This is also supported by the findings of a recent survey of western Canadian L. maculans isolates where 97.9% of the ‘avrLm1’ isolates tested still produced a resistance response on ‘Surpass 400’ (Kutcher et al., 2010). It also shows that the AvrLmS does not complement LepR3. In light of this evidence we believe the correct designation for the resistance genotype of ‘Surpass 400’ should be ‘LepR3, RlmS’. Mapping of the RlmS resistance locus and study of its interaction with AvrLmS isolates is currently in progress. In another recent report, ‘Surpass 400’ resistance was once again investigated and two resistance loci (named ‘Blm1’ and ‘Blm2’) were mapped to B. napus chromosome A10 (Long et al., 2011) using the L. maculans isolate ‘87-41’, which we have shown also carries AvrLm1 (Table 1). The Blm1 locus reported in that study corresponds to the same B. rapa BAC (KBrB080E24) described here for cloning the LepR3 gene. As the Blm1 resistance locus was mapped from the same cultivar, corresponds to the same genomic location and produces a resistance response to an AvrLm1 isolate we believe that it is LepR3.
Our results highlight the need to consider redundancy in R-gene specificities when using differential isolate sets to determine the R-gene content of Brassica lines. Rlm1 and LepR3 cannot currently be distinguished by differential phenotypic reactions; one needs to either determine the chromosomal linkage of the resistance phenotype or amplify and sequence the LepR3/lepR3 locus in order to discriminate between the presence of the two genes. We should not rely on a literal interpretation of the ‘gene-for-gene’ hypothesis, where any given avirulence protein only ever interacts with one specific R-protein.
This work is presented in remembrance of Dr S. Roger Rimmer. The authors would like to thank S. Kuzmicz, H. Lui, S. McMillan and C. Guenther for technical assistance, A. van de Wouw, B. Howlett and R. Kutcher for kindly sharing their isolates and for many helpful discussions, G. Seguin-Swartz for providing ‘Topas DH16516’, M. Curtis for providing pMDC123 and J. Utermark for helpful instruction on L. maculans transformation. This work was partly funded by Australian Research Council project LP0210571 with industry partners Council of Grain Grower Organizations Ltd and Norddeutsche Pflanzenzucht Hans-Georg Lembke KG, and the AAFC-Industry Blackleg Consortium II (partners include Bayer CropScience, Dow AgroSciences Canada, KWS SAAT, Monsanto, NSW Department of Primary Industries, Pacific Seeds, Lantmännen SW Seeds and Viterra).