Summary
The oomycete Albugo candida causes white blister rust, an important disease of Brassica crops. Distinct races of A. candida are defined by their specificity for infecting different host species.
The White Rust Resistance 4 (WRR4) locus in Col-0 accession of Arabidopsis thaliana contains three genes that encode TIR-NLR resistance proteins. The Col-0 alleles of WRR4A and WRR4B confer resistance to at least four A. candida races (2, 7 and 9 from B. juncea, B. rapa and B. oleracea, respectively, and Race 4 from Capsella bursa-pastoris). Resistance mediated by both paralogs can be overcome by Col-0-virulent isolates of Race 4.
After comparing repertoires of candidate effectors in resisted and resistance-breaking strains, we used transient co-expression in tobacco or Arabidopsis to identify effectors recognized by WRR4A and WRR4B. A library of CCG effectors from four A. candida races was screened for WRR4A- or WRR4B- dependent elicitation of hypersensitive response (HR). These CCG genes were validated for WRR-dependent HR by bombardment assays in wild type Col-0, wrr4A or wrr4B mutants.
Our analysis revealed eight WRR4A-recognized CCGs and four WRR4B-recognized CCGs. Remarkably, the N-terminal region of 100 amino acids after the secretion signal is sufficient for WRR4A recognition of these eight recognized effectors. This multiple recognition capacity potentially explains the broad-spectrum resistance to many A. candida races conferred by WRR4 paralogs.
Introduction
Co-evolution of hosts and their parasites remains a fascinating biological problem. Most plant pathogens fail to colonize most potential hosts, in part because of defense responses initiated upon pathogen perception by cell surface or intracellular immune receptors (Jones & Dangl, 2006). Successful pathogens evade host recognition and suppress defense via effectors (Toruño et al., 2016). In co-evolving systems, genetic variation in resistance and susceptibility to particular pathogen races is usually specified by loci encoding Nucleotide Binding, Leucine-rich Repeat (NLR) intracellular immune receptors (Tomborski & Krasileva, 2020). Plant NLRs are under diversifying selection compared to the rest of the genome (Monteiro & Nishimura, 2018). NLR genes are classified as TIR-NLR (TNL) with N-terminal Toll/Interleukin-1/Resistance domains, CC-NLR (CNL) with a coiled-coil (CC) domain, or RPW8-NLR (RNL) with Resistance to Powdery mildeW8 (RPW8) domains (Meyers et al. 2003; Zhang et al. 2016). Variation in NLR repertoires has been investigated by Resistance gene enrichment sequencing (RenSeq) (Jupe et al., 2013). NLR immune receptor polymorphism and diversity in 64 accessions of Arabidopsis thaliana has been defined using RenSeq (Van de Weyer et al., 2019).
NLR-mediated effector-triggered immunity (ETI) often leads to a hypersensitive cell death response (HR) upon effector recognition. Recent work using estradiol- inducible AvrRps4 activation of ETI in the absence of pattern triggered immunity (PTI), revealed mutual potentiation between PTI and ETI to confer resistance (Ngou et al., 2021). Effectors can be recognized via direct interaction with an NLR protein, described as the ‘ligand- receptor model’ (Jia et al., 2000; Dodds et al., 2006). The flax (Linum usitatissimum) L6 NLR protein directly recognizes variants of the flax rust fungus (Melampsora lini) effector AvrL567 (Ravensdale et al., 2012). Alternatively, some NLRs can detect multiple sequence-unrelated effectors indirectly. Such NLRs either “guard” host proteins that are targeted by multiple effectors (guardee model) (Dangl & Jones, 2001), or guard “decoy” proteins that have evolved to mimic the host target (van der Hoorn & Kamoun, 2008). Effector recognition by some NLRs involves a post-LRR (PL) domain. Recent structural investigations in two TNLs, RPP1 and Roq1 revealed a C- terminal jelly-roll and Ig-like domain (C-JID) in mediating effector binding to form a tetrameric resistosome upon activation (Ma et al., 2020; Martin et al., 2020).
White blister rust in a wide range of crop and wild Brassica species is caused by oomycete pathogens in the genus Albugo (Holub et al., 1995; Voglmayr & Riethmüller, 2006; Choi et al., 2007). The disease symptoms of white pustules resemble sporulation of basidiomycete rust fungi. Dispersal is via dehydrated sporangiospores (Heller & Thines, 2009). Jouet et al. (2019) verified that different phylogenetic races of A. candida exhibit distinct host specificities for infecting different crop and wild species of Brassicaceae. Important examples include Races 2, 7 and 9 from major crop species (B. juncea, B. rapa and B. oleracea, respectively) and Race 4 from the common weed Capsella bursa-pastoria. Genome comparisons revealed ancient genetic exchange and introgression amongst these A. candida races (McMullan et al., 2015). A. candida species have a remarkable capacity to suppress host resistance to other infections (Cooper et al., 2008; Belhaj et al., 2017; Prince et al., 2017) which creates an immune-compromised state in the colonized host enabling growth of non-adapted pathogens.
White Rust Resistance 4 (WRR4) from A. thaliana confers resistance to A. candida (Borhan et al., 2008). The locus contains three paralogs in A. thaliana accession Columbia (Col-0) that encode TNLs. The Col-0 allele of WRR4A confers resistance to four A. candida races in A. thaliana (Borhan et al., 2008), and also as a transgene to Ac2V in Brassica crops (Borhan et. al, 2010; Cevik et al., 2019). The Col-0 allele of WRR4B also confers resistance to Brassica-infecting A. candida races Ac2V and AcBoT (Cevik et al., 2019). Although resistance in Col-0 appears to be broad spectrum, WRR4 resistance- breaking strains (e.g., AcEx1) have been collected from white rust pustules on floral stems of wild A. thaliana or A. halleri. These are natural pathotypic variants of A. candida Race 4 (Fairhead, 2016; Jouet et al., 2019). Resistance to this Race 4 pathotype occurs in A. thaliana, such as in the accession Oystese (Oy-0) which also maps to the WRR4 locus (Fairhead, 2016; Castel et al., 2021).
A. thaliana accession Wassilewskija (Ws-2) is resistant in leaves to A. candida Race 2 and 7. The WRR4 locus in Ws-2 is disrupted compared to Col-0 by deletions or sequence variation of WRR4 paralogs. Ws-2 contains two divergent paralogs (Van de Weyer et al., 2019) and one of these (a Ws-2 allele of WRR4B) confers resistance to isolates of A. candida Race 2 (from B. juncea). Both Col-0 and Ws-2 alleles of WRR4B also confer resistance in transgenic Brassicas (Cevik et al., 2019). Conceivably, allelic variation of TNL paralogs at the WRR4 locus provides multiple genes that could control white rust in major Brassica crops. Stacking multiple WRR genes should promote durability. Thus, identifying and understanding function of A. candida effectors recognized by alleles of WRR4A and WRR4B would help choose the most effective transgene combinations to use for Albugo control in Brassica crops.
Recognized effectors from the Peronosporales (including species of Phytophthora, Pythium and downy mildews) are translocated into host cells (Kamoun, 2006; Haas et al., 2009). These effectors typically carry an N-terminal signal peptide for secretion and an RxLR motif that is implicated in host translocation (Whisson et al., 2007; Dou et al., 2008). However, in the P. infestans effector Avr3a, the RxLR motif is cleaved during secretion, and how it promotes translocation still remains unclear (Wawra et al., 2017). RxLR effectors show enhanced polymorphism and positive selection for diversification towards their C termini (Rehmany et al., 2005; Win et al., 2007).
Genome analysis of two Albugo species (Albugo laibachii and A. candida) revealed a new class of oomycete effector-like proteins that carry a “CHxC” motif (Kemen et al., 2011). Albugo lacks RxLR-encoding proteins compared to the Peronosporales. Re-sequencing of an A. candida race from B. juncea (Ac2V) using long reads revealed massive expansion of effector-like proteins with a CHxC-like motif, which was reclassified as CX2CX5G and abbreviated to CCG, resembling previously identified CHxC proteins in A. laibachii (Kemen et al., 2011; Furzer et al., 2021). CCG proteins show no homology to other oomycete secreted proteins. Their high sequence divergence resembles that of RxLR effectors (Furzer et al., 2021), and supports their investigation as candidates for being the effectors recognized by White Rust Resistance (WRR) genes.
In this study, we compared the genomes of resisted and resistance-breaking strains of A. candida to define candidate differentially-recognized CCG effectors. We screened a library of CCG secreted proteins, mainly from A. candida races 2 and 4, to identify potential effectors recognized by Col-0 alleles of WRR4A and WRR4B. Agrobacterium- mediated transient co-expression was used in a pre-screen to identify pairwise combinations of effector and R-allele that activated an HR. Twelve CCG candidates were identified including eight recognized by Col-WRR4A and four by Col-WRR4B. Positive candidates were then validated for HR recognition by a bombardment assay in Col-0 wild type and mutants (wrr4A or wrr4B) of A. thaliana. Several of these CCGs are absent or else show expression polymorphism in the Col-0 virulent isolate AcEx1. To further characterize this WRR4 recognition, we focused on WRR4A-recognized CCGs, for which the N-terminal region is sufficient for recognition. Our data reveal a novel capacity for recognition of multiple effectors of A. candida by two distinct WRR4 paralogs.
MATERIALS AND METHODS
Plant material and growth conditions
Wild type and mutant A. thaliana accessions used in this study included Col-0, Wassilewskija-2 (Ws-2), Col-0_wrr4a-6 (Borhan et al., 2008), Col-0_wrr4b (Cevik et al., 2019) and the recombinant inbred line (RIL) CW20 that was derived from a cross between Col-5 x Ws-2 and in this RIL, WRR4 locus is the only known WRR locus introgressed from Col-5 (Fairhead, 2016). Seeds were sown directly on compost and were grown at 21°C, with 10 hours of light and 14 hours of dark, at 75% humidity. For N. tabacum and N. benthamiana, plants were grown on compost at 21°C, with cycles of 16 hours of light and 8 hours of dark, at 55% humidity.
A. candida infection assay
For leaf inoculations, zoospores harvested from previous leaf infections, were suspended in water (∼105 spores/ml) and incubated on ice for 30 min for releasing of the zoospores from sporangia due to cold shock. The spore suspension was then sprayed on plants using a Humbrol spray gun (Hornby Hobbies Ltd, Sandwich, UK) with a volume equal to ∼700 μl/plant and plants were incubated at 4°C in the dark overnight for efficient zoospore germination. Infected plants were kept under 10-hour light (21°C) and 14-hour dark (16°C) cycles. Phenotypes were monitored between 7 to 10 days post inoculation (dpi) and macroscopic symptoms were readily visible during this period.
The sequential infection assay was done as described previously in McMullan et al., 2015. We developed A. candida race specific PCR primers by comparing the genome sequences for regions that were unique to the tested A. candida races AcNc2 and AcEx1. (Table S1). Primers were designed to amplify these regions from genomic DNA extracted from each isolate. Primary inoculum was sprayed onto control and test plants. In the case of AcEx1 WRR4A mediated defense suppression assays, both A. thaliana Ws-2 and CW-20 were inoculated. The inoculated plants were incubated in the dark at 4°C overnight. Same number of plants treated with water served as mock control. At 7 dpi a secondary infection with the avirulent A. candida race AcNc2 was performed on 50% of the plants while the remaining 50% were again mock inoculated with water. The co-inoculated plants were returned to the growth cabinet and incubated for a further 8 dpi. Tissue was harvested and washed in sterile water to remove surface adhering spores, and flash frozen in liquid N2. DNA was prepared using a DNeasy Plant Mini Kit (Qiagen, CA, USA) as described in manufacturer’s instructions. PCR was performed using race-specific primers and products were visualized on a 1% agarose gel.
Gene cloning and plasmid construction
Cloning of genes were carried out using Uracil-Specific Excision Reagent (USER) method (Geu-Flores et al., 2007). Genes with 5’ and 3’ regulatory sequences were cloned into LBJJ233-OD vector (containing a FAST-Red selectable marker) pre-linearized with PacI and Nt.BbvcI. For overexpression, plant genes and CCG effectors, which lack introns, were cloned into pre-linearized LBJJ234-OD (containing a FAST-Red selectable marker, CaMV 35S promoter and Ocs terminator) or pICH86988 (containing Kan selectable marker, CaMV 35S promoter and Ocs terminator). Genes were C-terminally tagged either with a His-FLAG (HF tag) or a yellow fluorescent protein (YFP) tag. Briefly, the candidate CCG effector was PCR amplified from one of the A. candida races (AcNc2, AcEm2 or Ac2V) for the high throughput screen for WRR4A- recognized CCGs and from race Ac2V for WRR4B- recognized CCGs. The genomic DNA was used as a template with KAPA HiFi Uracil+ enzyme, following the manufacturer’s protocol. To obtain mutant versions of CCG28AAs28-130 -YFP carrying mutations in the CCG motif (CCG exchanged to AAG, CAA, CAG and AAA), site directed mutagenesis was performed with QuikChange Multi Site-Directed Mutagenesis Kit (Stratagene, Santa Clara, USA) following manufacturer’s instructions. A list of primers and vectors are indicated in Table S1. WRR4A and WRR4B under 35S promoter were cloned in pICH86988 from the Col-0 genomic DNA. All the plasmids were transformed in to Escherichia coli DH10B electro-competent cells selected with appropriate antibiotics and purified using a Qiaprep Spin Miniprep Kit (Qiagen). Positive clones were transformed in Agrobacterium tumefaciens strain GV3101 and used in infiltrations for transient expression experiments.
Transient expression in N. tabacum or N. benthamiana leaves and cell death assay
A. tumefaciens strains were streaked on selective media and incubated at 28 °C for 24 hours. The streaked inoculum was transferred to liquid LB medium with appropriate antibiotic and incubated at 28 °C for 24 hours in a shaking incubator at 200 rotations/min (rpm). The resulting cultures was centrifuged at 3,000 rpm for 5 min and resuspended in infiltration buffer (10 mM MgCl2, 10 mM MES, 150 μM acetosyringone pH 5.6) at OD600 of 0.4 (2 ×108 cfu/ml). For co-expression, each bacterial suspension was adjusted to OD600 of 0.4. The abaxial surface of 4-weeks old N. tabacum or 5 weeks old N. benthamiana were infiltrated with 1 ml needleless syringe. Cell death was phenotyped two to four days after infiltration. For the WRR4B recognition assay, macroscopic cell death phenotypes were scored according to the HR index modified from (Segretin et al., 2014) ranging from 0 (no visible necrosis) to 6 (full necrosis).
Particle bombardment in A. thaliana and luciferase assay
Transient protein expression in Arabidopsis leaves was performed by biolistic gene transfer. 1.0 µm Tungsten particles (Bio-Rad) were coated with the plasmids coding for the indicated CCG genes driven under CaMV 35S promoter (Table S1). Bombardment was performed using a PDS-1000/He system (Bio-Rad) onto 4-weeks-old Arabidopsis leaves. After bombardment the leaves were incubated in small vials with the leaf petiole immersed in water, for 48 hours post bombardment (hpb). The leaves were then frozen in liquid N2 and stored at -80 °C until further processing.
For the luciferase assay a Dual Reporter Luciferase Assay system (Promega) was used. Four transiently bombarded leaf events were pooled together and crushed in lysis buffer. The extract was centrifuged at 12,000 rpm for 10 min at 4 °C. 20 µl of the lysate was then dispersed in 96 well plate in triplicates and analyzed on Varioskan Flash Instrument by injecting 100 µl of luciferase assay reagent II, which includes substrate and reaction buffer. A 10 second read time was used to measure luciferase activity for each well.
Gene expression measurement by RT-qPCR
For gene expression analysis, RNA was isolated from three biological replicates and used for reverse transcription quantitative PCR (RT-qPCR) after cDNA synthesis. Briefly, RNA was extracted using the RNeasy Plant Mini Kit (Qiagen) with the DNase treatment (Qiagen). Reverse transcription was carried out using the SuperScript IV Reverse Transcriptase (ThermoFisher). RT-qPCR was performed using CFX96 Touch Real-Time PCR (Bio-Rad). Primers for qPCR analysis of different CCGs are enlisted in Table S1. Data were analysed using the double ΔΔCT method (Livak & Schmittgen, 2001) by calculating the relative expression of candidate CCG in relation to the A. candida EF1α as a house keeping reference gene.
Protein extraction and Western Blot
Protein was extracted from Agrobacterium infiltrated N. benthamiana leaves at 72 hpi as previously described (Sarris et al., 2015). Briefly, leaves were harvested and ground in liquid N2, and extracted in GTEN buffer (10% glycerol, 100 mM Tris-HCl, pH 7.5, 1 mM EDTA, 150 mM NaCl, 5 mM 1,4-dithiothreitol (DTT), 1× Complete protease inhibitor mixture (Roche) and 0.2% (V/V) Noniodet P-40). 30 μl of the supernatant from the sample extract was used to elute the samples by boiling in loading buffer. For SDS-PAGE, samples were heated for 10 min at 95 °C for denaturation. After electrophoresis, separated proteins were transferred to PVDF (Merck) membranes for immunoblotting. Membranes were blocked for 2 hours in 5% nonfat milk, probed with horseradish peroxidase (HRP)-conjugated antibodies for overnight. Chemiluminescence detection for proteins was carried out by incubating the membrane with developing reagents (SuperSignal West Pico & West Femto), using ImageQuant LAS 4000 (Life Sciences, USA).
Statistical analysis
Statistical analysis was carried out using the GraphPad Prism 9.0 software (San Diego, USA). The statistical test employed is provided in the figure legends.
RESULTS
The WRR4ACol-0 and WRR4BCol-0/Ws-2 provide broad spectrum resistance to A. candida races
WRR4ACol-0 confers full green resistance (GR) to multiple A. candida races (Cevik et al., 2019; Borhan et al., 2008) including the Race 2 isolate Ac2V and two highly similar Race 4 isolates AcNc2 and AcEm2 (McMullan et al., 2015) (Fig. 1 and Table S2). The Arabidopsis accession Ws-2 which lacks WRR4A is fully susceptible to AcEm2 but Col-0-wrr4-6 mutant shows necrotic resistance (NR) rather than GR as observed with wild-type (WT) Col-0, due to the presence of additional WRR genes in Col-0. This suggests WRR4ACol-0 contributes to full GR to AcEm2 (Fig. 1). A Race 4 isolate AcEx1 overcomes WRR4ACol-0 mediated resistance but triggers chlorosis in infected adult Col-0 plants (Fig. 1). The Col-0wrr4-6 mutant exhibits full susceptibility to AcEx1 indicating that WRR4ACol-0 confers partial resistance to AcEx1.
The WRR4BCol-0/Ws-2 paralogs also confer resistance to Ac2V in A. thaliana and in transgenic B. juncea (Cevik et al., 2019). However, Ws-2, which lacks WRR4A, exhibits full GR phenotype in adult leaves following inoculation with either Ac2V or Ac7V indicating that WRR4BWs-2 also confers adult plant resistance to Races 2 and 7 of A. candida (Table S2).
Selection of candidate A. candida CCG effectors to test for recognition by different WRR4 paralogs
The recent PacBio based genome assembly of A. candida Race 2 isolate Ac2V (“Ac2VPB”) revealed that the CCG family of secreted proteins corresponds to ca. 10% of the A. candida secretome (Furzer et al., 2021). The genome of each A. candida race contains 60-80 CCG effector candidate genes that can vary by sequence and presence/absence polymorphism (Jouet et al., 2019; Furzer et al., 2021). This enabled identification of candidates for functional screening of CCG effectors transiently, testing for WRR gene-dependent HR. Such transient assays have been adopted for other oomycete pathogens like P. infestans, testing secreted proteins with an RxLR motif (Vleeshouwers et al., 2008; Vleeshouwers et al., 2011), and thus, identifying several recognized effector genes.
To identify CCG effector proteins that are recognized by WRR4 paralogs, we analysed CCGs predicted from Illumina-based genome assemblies of multiple A. candida races (AcNc2, AcEm2, Ac7V, AcBoT and Ac2V) (McMullan et al., 2015; Jouet et al., 2019) as well as additional CCGs predicted from Ac2VPB (Furzer et al., 2021). We identified CCGs showing presence/absence polymorphism or pseudogenisation (Jouet et al., 2019; Furzer et al., 2021). We then selected 30 candidate CCGs to screen with WRR4A prioritising those either absent or pseudogenised in WRR4A-overcoming race AcEx1. To screen with WRR4B, we selected 13 candidate CCGs that are mostly conserved in Ac2V, Ac7V and AcBoT but are absent or pseudogenised in other A. candida races (Fig. 2a,b and Fig. 3a; Table S3).
To test the selected CCG effectors for WRR4A or WRR4B recognition, we cloned CCG effectors into an expression vector with 35S promoter and transformed into Agrobacterium for infiltration of N. tabacum or N. benthamiana leaves. The transient co-delivery of CCG effector was performed either with GFP or RFP as negative control and with WRR4A or WRR4B (Cevik et al., 2019). Effectors that trigger an HR when co-expressed with the corresponding NLR were further validated by the luciferase eclipse assay modified from Allen et al., (2004) for recognition in Col-0 or in Col-0 wrr4a or wrr4b mutants (Fig. 2a).
WRR4ACol-0 confers recognition to eight different CCG effectors from A. candida
To screen candidate CCGs for their recognition by WRR4A, CCGs excluding the signal peptide were cloned from isolates of A. candida Race 2 (Ac2V) or Race 4 (AcEm2 or AcNc2). All of the effector alleles were co-infiltrated with 35S:WRR4A-expressing Agrobacterium strains (Cooper et al., 2008; Cevik et al., 2019). All of the recognized effectors and some representative non-recognized CCGs were tested by Western blot to confirm their protein expression when expressed transiently (Fig. S1). Among the 30 alleles we tested, eight CCGs (CCG28Ac2V, CCG30AcNc2, CCG33AcNc2, CCG40AcEm2, CCG67AcEm2, CCG71AcNc2, CCG79Ac2V and CCG104Ac2V) elicited HR within 36-48 hours post infiltration (hpi) when co-expressed with WRR4A but not with GFP control. Moreover, WRR4A co-expressed with GFP did not show any autoactivation (Fig. 2c).
To validate the HR observed from transient assays with the eight recognized CCGs, we carried out luciferase eclipse assays by transient delivery using particle bombardment to reveal reduced luciferase activity upon HR triggered by these recognized CCG effectors in Arabidopsis leaves in Col-0 compared to Col-0 wrr4a-6 mutant. The luciferase construct was delivered alone or with recognized CCG effector and incubated in vials with water for 48 hours post-bombardment (hpb). All the CCGs recognized by WRR4A conferred reduced luciferase activity in comparison to the luciferase only control in Col-0 but not in Col-0 wrr4a-6 mutant. The diminished luciferase activity in Col-0 indicates HR-dependent cell death in transformed leaf cells (Fig. 2d). Hence, both methods revealed eight different CCGs that are recognized by WRR4A.
The WRR4ACol-0 paralog WRR4BCol-0 recognizes four additional CCG effectors
To test recognition by WRR4B, we selected candidate CCGs from the CCG effectorome repertoire based on the resistance and susceptibility disease phenotypes (Table S2). As WRR4B confers resistance against A. candida race 2 (Ac2V), we prioritized 13 CCG candidates present mainly in crop-infecting races of A. candida and also CCGs present exclusively in Ac2V (Fig. 3a), selecting the Ac2V CCG allele for these recognition assays.
Over-expression of WRR4B in tobacco or N. benthamiana shows weak autoimmunity, even when infiltrated with the GFP control (Fig. S2 a,b). This creates a requirement for cautious interpretation of any HR phenotypes detected upon co-expression with a potentially recognized CCG effector. Hence, we used the luciferase eclipse assay for screening the set of 13 cloned CCGs for candidate WRR4B-recognized CCGs, and discovered four candidate WRR4B-recognized CCGs. We further confirmed the recognition of these four candidates (CCG45Ac2V, CCG57Ac2V, CCG61Ac2V and CCG70Ac2V) by luciferase eclipse assay by a comparison of the luciferase activity in Col-0 and Col-0-wrr4b mutant. The Col-0 wrr4b mutant showed a higher level of luciferase when co-expressed with these 4 CCGs that is diminished in Col-0 due to HR (Fig. 3b). This assay further confirms the WRR4B-specific recognition of these four CCGs from A. candida race Ac2V .To understand if any of these CCGs confer elevated HR after transient expression in N. tabacum compared to WRR4B alone, CCG45Ac2V, CCG57Ac2V, CCG61Ac2V and CCG70Ac2V were co-infiltrated with WRR4B. Only CCG45Ac2V and CCG70Ac2V showed an enhanced HR compared to the WRR4B infiltrated with GFP alone, indicating that these candidates are also recognized by WRR4B in tobacco (Fig. S2a). This HR varied between individual leaves. CCG57Ac2V and CCG61Ac2V did not show elevated HR compared to WRR4B co-infiltrated with GFP control (Fig. S2a). In order to check if the HR by CCG45Ac2V and CCG70Ac2V is also consistent in N. benthamiana, we performed co-infiltration of these CCGs with WRR4B in N. benthamiana. CCG28Ac2V co-delivered with WRR4A was used as a positive control. Consistent with N. tabacum phenotype, CCG45Ac2V and CCG70Ac2V show an elevated HR in N. benthamiana when co-expressed with WRR4B as compared to WRR4B alone (Fig. S2b, c). The CCG motif region is the most conserved part of different WRR4B-recognized CCGs (Fig. 3c). We conclude that WRR4B specifically recognizes four CCGs from A. candida that are distinct from the CCGs that are recognized by WRR4A.
To further elucidate the region of CCG that is recognized, we investigated the WRR4A-CCG interaction, as WRR4A does not show any autoimmune phenotype and also recognizes diverse CCGs across different A. candida clades (Furzer et al., 2021).
Recognition of CCGs by WRR4ACol-0 requires the N-terminal portion of the protein but does not occur via the CCG motif
To investigate the mechanism of WRR4A activation by CCGs, we tested a series of CCG effector deletions. We selected CCG28 for this analysis because it triggers the strongest WRR4A-dependent HR, with a response visible at 36 hpi. Similar to WRR4B-recognized CCGs, as the CCG motif region was the only part of the CCG protein that showed homology across different WRR4A recognized CCGs, we tested the role of the N-terminal region of CCG for recognition by WRR4A. A full-length version of CCG28 without its secretion signal triggers strong WRR4A-dependent HR. However, deletion of amino acids 28-40 in the N-terminal region after the secretion signal abolishes this recognition (Fig. 4a, b).
To define the minimal N-terminal region of CCG28 that is recognized, we made C-terminal deletions of CCG28Ac2V and tested their recognition using transient assays in N. tabacum leaves (Fig. S3). A truncation of CCG28Ac2V that includes the first 100 amino acids after the signal peptide site (CCG2828-130), including the CCG motif, is more strongly recognized by WRR4A than full length CCG28, triggering an elevated HR at 36 hpi when transiently co-expressed in N. tabacum (Fig. 4a, b; Fig. S3). This also includes a common feature with a pair of cysteines located ∼50 amino acids after the CCG motif. In contrast, a C-terminal region of CCG28 without the CCG motif, which corresponds to aa 56-543 abolishes recognition when co-expressed with WRR4A (Fig. 4a, b; Fig. S4). Hence, we conclude, that the truncated N-terminal part of CCG28 is indispensable for WRR4A recognition. Further deletion of amino acids 28-35 at the N-terminus also compromises recognition (Fig. S3 and S5). To define the shortest region of CCG28 that is recognized, we further narrowed the recognition region to 50 amino acids, corresponding to CCG2828-78. However, only YFP-tagged versions of this shortest region activate HR (Fig. S3 and S5), likely indicating that untagged versions are insufficiently stable for their interaction with WRR4A to trigger a cell death phenotype when expressed transiently.
We next tested whether WRR4A recognition of all the CCG candidates involves their N-termini. A truncation analysis was carried out for all the eight recognized CCGs by WRR4A. Upon transient expression, C-terminally truncated versions of the other seven recognized CCGs also trigger an HR phenotype with WRR4A in N. tabacum (Fig. S6a). These data suggest that the N-terminal portion of all the Avr-WRR4A is sufficient for recognizability. WRR4A recognition of CCGs requires an intact Walker A (P-loop) (Schreiber et al., 2016) because a mutation in this P-loop of WRR4A with a change from K220 >L220 abolishes the HR when WRR4A is co-infiltrated with the recognized CCGs in transient assays (Fig. S6b).
We next examined whether the previously defined CCG motif, that shows maximum homology across all recognized CCGs against WRR4A mediates effector recognizability (Fig. 4c). To this end, the truncated CCG28aa28-130, which is strongly recognized by WRR4A was used. Mutant versions of CCG28aa28-130-YFP carrying a mutation in the CCG motif (where CCG exchanged to AAG, CAA, CAG and AAA) were generated and tested for recognition in transient assays by co-infiltration with WRR4A in N. tabacum. These mutated versions were still strongly recognized by WRR4A with an HR phenotype indistinguishable from the truncated version with authentic CCG motif (Fig. 4d). These data suggest that WRR4A -CCG recognition does not occur via the CCG motif. Conceivably a structural similarity in the N-terminal portion of these recognized CCGs might be responsible for their detection by WRR4A. Consistent with this, CCG30, but not its close paralog CCG16, is recognized by WRR4A (Fig. S7).
Allelic variation and expression polymorphism of recognized WRR4A CCGs
Next, we confirmed the allelic status of the identified recognized CCG candidates across all the sequenced races of A. candida from the available Illumina assemblies of different A. candida races (McMullan et al., 2015) and the recent PacBio assembly of B. juncea-infecting race Ac2V (Furzer et al., 2021). Consistent with the resistance phenotypes in Col-0 and Col-wrr4a mutants, the recognized CCGs are present across different A. candida races. CCG33 and CCG40 are pseudogenised in the B. juncea-infecting race Ac2V (Fig. 2b). Several of the recognized alleles from these CCG candidates, show polymorphism in AcEx1 (a WRR4 resistance-breaking pathotype of Race 4). Due to an early stop codon, CCG28AcEx1 shows a protein length of only 226 amino acids as compared to the full-length alleles from other races (Fig. 5a). Another isolate of this pathotype, AcCarlisle collected on A. thaliana, shows the same early stop codon in CCG28 and virulence on Arabidopsis Col-0. The AcEx1 CCG28 allele is weakly recognized compared to full-length alleles from A. candida Races 2, 4 and 7 (Ac2V, AcEm2 and Ac7V) all of which trigger an early WRR4A-dependent HR at 36 hpi (Fig. 5b). Additionally, the WRR4A-recognized CCG71 and CCG104 are absent from AcEx1 (Fig. 2b) suggesting that absence of these effectors might contribute to evasion of WRR4A mediated resistance. On the other hand, the CCG effectors recognized by WRR4B showed a strong presence/absence polymorphism and are primarily present in the crop-infecting races which are known to be resisted by WRR4B (Fig. 3a).
To determine the extent to which these WRR4A-and WRR4B-recognized CCGs are expressed in planta, we assessed their expression profiles from the RNA-Seq data obtained over the consecutive time-points during serial infection stages of A. candida Race Ac2V (Furzer et al., 2021). WRR4A-and WRR4B-recognized CCGs in Ac2V show in planta induction during different colonization stages. These CCGs are induced specifically at different colonization time-points of 2, 4 and 6 dpi (Fig. S8a) consistent with a role in virulence.
To investigate the expression patterns of different alleles of the WRR4A-recognized CCGs, we performed a qRT-PCR expression profiling of the A. candida races Ac2V, AcEm2 and AcEx1 growing on Ws-eds1 plants (Fig. S8b). These data showed different CCG effectors increase in expression during infection. Intriguingly, CCG28 is primarily induced at 2 dpi. The Ac2V allele of CCG28 is most highly expressed followed by the AcEm2 allele. Notably, the pseudogenised allele from AcEx1 shows least expression as compared to the other alleles (Fig. 5c). The expression profiles of other recognized WRR4A-recognized CCGs follow a similar trend. The Ac2V allele is among the most highly expressed and AcEx1 shows the lowest (Fig. S8c). From these results, we conclude that the partial susceptibility of Col-0 plants we observe with AcEx1 is both due to reduced expression or due to absence of the WRR4A-recognized CCGs. In summary, these data suggest for WRR4-recognized CCGs, there exists expression polymorphism as well as presence-absence polymorphism that can enable evasion of resistance, especially when combined with the well-documented suppression of resistance upon Albugo infection.
AcEx1 has the capacity to suppress WRR4A-mediated resistance
Previous studies have shown that A. candida infection leads to strong immune suppression enabling co-infection by otherwise avirulent races, permitting sexual exchange and recombination (Cooper et al., 2008; McMullan et al., 2015). To test if AcEx1 can suppress WRR4A mediated resistance, we performed sequential inoculation experiments, on Arabidopsis accessions with and without WRR4A, monitoring A. candida races using race-specific genomic regions as a readout for race-specific PCR markers. Race-specific PCR verifies that AcNc2 can grow on Ws-2 which does not have WRR4A but not on RIL CW20 (which carries Col-5 alleles of WRR4A and WRR4B as the only WRR genes). However, pre-inoculated plants of CW20 colonized by the WRR4A resistance breaking race AcEx1 lose resistance in CW20 leaves towards AcNc2 (Fig. S9). Therefore, AcEx1 not only suppresses WRR4A mediated recognition but also enables other races to grow, that would otherwise be resisted. These data suggest that AcEx1 is able to overcome WRR4A resistance by virtue of weak recognizability due to polymorphisms in its repertoire of WRR4A recognized CCG alleles, together with strong ability for immunosuppression of WRR4A-mediated resistance.
DISCUSSION
Oomycetes in the Peronosporales (Phytophthora, Pythium and downy mildew species) are destructive pathogens. Host colonization requires effectors which carry a signal peptide and a positionally constrained RxLR motif (Win et al., 2006). A specific effector is often detected in a resistant host by a matching NLR immune receptor, often encoded by a resistance (R-) gene, that activates a defense response that thwarts pathogen success (Armstrong et al., 2005; Rehmany et al., 2005). The Irish potato famine pathogen P. infestans, RxLR effectors Avrblb2 and Avrvnt1 are recognized by Rpi-blb2 and Rpi-vnt1 (Oh et al., 2009; Pais et al., 2018). Identifying such avirulence (Avr) determinants in the pathogen is important for improving strategies to develop disease resistant crops.
The genomes of oomycetes in the Albuginales, such as A. laibachii and A. candida, are not enriched for RxLR proteins but instead contain a family of secreted proteins with a ’CHxC’ (now CCG) motif that contributes to translocation (Kemen et al., 2011; Furzer et al., 2021). These proteins show signatures of diversifying selection, with high ratios of non-synonymous to synonymous mutations, similar to RxLR effectors (Rehmany et al., 2005; Asai et al., 2018). The Ac2VPB assembly of the Brassica pathogen A. candida Race 2 has enabled refinement of CCG effector annotation (Furzer et al., 2021). CCG gene repertoires are present in all sequenced A. candida races (McMullan et al., 2015). A combined phylogeny comparison between A. laibachii and A. candida shows an expansion of CCGs from ∼30 in A. laibachii to around ∼100 in A. candida, with several A. candida-specific clades suggesting expansion and reshuffling of CCG effector repertoires to adapt to different hosts (Furzer et al., 2021). Moreover, the CCG repertoire displays elevated rates of pseudogenisation and presence/absence polymorphisms, consistent with selection for diversity while maintaining virulence functions.
In this study, we used the CCG effectoromes from different A. candida races, enabling us to screen selected CCG effectors for WRR4 recognition. This revealed eight WRR4A-recognized CCGs and four WRR4B-recognized CCGs. The allelic comparison approach helped us to select CCG candidates based on their presence/absence polymorphism in A. candida races that overcome resistance. The eight identified WRR4A-recognized CCGs when co-expressed with WRR4A show an HR at 48 hpi resulting in complete cell-death. A genetic validation of this recognition by luciferase eclipse assay after particle bombardment shows that the identified CCG effectors are specifically recognized by WRR4A. A similar approach based on allelic comparison allowed us to screen and identify four additional WRR4B-recognized CCGs. As WRR4B shows an autoimmune phenotype in transient expression experiments, the N. tabacum HR assay was more difficult to interpret than with WRR4A. The luciferase eclipse assay enabled us to identify four WRR4B-recognized CCGs; CCG45, CCG70, CCG57 and CCG61. However, only CCG45 and CCG70, which show the strongest phenotype in the luciferase eclipse assay show an enhanced HR when co-expressed transiently with WRR4B in N. tabacum or N. benthamiana. We conclude that WRR4A and WRR4B each recognize multiple and distinct CCG effectors from A. candida. Multiple recognition of effectors is known for some other NLRs that function in pairs, such as Arabidopsis RPS4/RRS1 (Sarris et al., 2015; Guo et al., 2020) and rice RGA4/RGA5 (Cesari et al., 2013). The WRR4B autoimmune phenotype hindered further analysis, and we further investigated the WRR4A-CCG interaction to define recognition requirements in more detail.
Analysis of the recognition mechanism of multiple CCGs by WRR4A revealed that the N-terminal portion of all WRR4A-recognized CCGs is sufficient for recognizability. Truncation analysis of all these CCGs shows the N terminal 12 amino acids after the secretion signal are required to trigger HR. Mutations in the CCG motif suggest that the CCG motif itself is not recognized, as mutant alleles are still recognizable. Hence, the exact role of the CCG motif and its significance in A. candida infection is still unclear. The recognized CCGs fall into different clades in the CCG phylogeny (Furzer et al., 2021). Hence, we speculate that there likely exists a structural similarity among these recognized effectors that enables their recognition. This is also suggested by the observation that CCG16, a close paralog of the WRR4A-recognized CCG30 is not recognized despite showing a higher homology to CCG30 than some other WRR4A-recognized CCGs.
This study provides evidence that WRR4A and WRR4B at the WRR4 locus in A. thaliana Col-0 have the capacity to recognize multiple effectors from A. candida likely making them highly effective for immune activation upon pathogen attack. This potentially explains the effectiveness of this resistance against many races. Interestingly, both WRR4A and WRR4B have a post-LRR C-JID domain which was recently shown to physically interact with effectors by TNLs RPP1 and Roq1 during resistosome activation (Ma et al., 2020; Martin et al., 2020). Hence, it will be intriguing to understand whether WRR4-CCG interaction is mediated by a similar mechanism. For P. infestans RxLR effectors, the C-terminal post-RxLR domain carries their effector activities (Kamoun, 2006; Kamoun, 2007). It is unknown how CCG proteins contribute to pathogen virulence and is a topic of further investigation. AcEx1 is capable of overcoming WRR4A-and WRR4B-mediated resistance, though on Col-0, but not Col-0 wrr4A, it activates chlorosis and thus is still weakly recognized. Its growth on Col-0 is due to the loss of some of the recognized A. candida effectors, and reduced expression of others.
Additionally, AcEx1 suppresses WRR4-mediated immunity against AcNc2, consistent with the hypothesis that Arabidopsis susceptibility to a specific A. candida race is determined by a balance of effector detection and effector-mediated suppression of defense. The partial susceptibility of Col-0 plants with AcEx1 is potentially explained by the reduced expression (measured by expression profiling) of otherwise detectable CCG effectors in this isolate. We suggest that strong immunosuppression of this weak recognition enables AcEx1 to colonize Col-0. Therefore, we propose that further study of WRR4A-and WRR4B-recognized CCGs might reveal insights into the colonization biology of A. candida.
Resistance to AcEx1 that maps to the WRR4 locus has been identified in other accessions of Arabidopsis (Fairhead, 2016; Castel et al., 2021). The WRR4A alleles in HR-5 and Oy-0 have an extended C-terminus, that confers AcEx1 resistance by recognizing alternative CCG effectors. The identification of WRR4A-and WRR4B-recognized CCGs will further enable understanding of the recognition mechanism at the structural level. Thus, this work contributes new insights into effector biology in obligate biotrophs, and will help inform provision of durable resistance in Brassicaceae crops by pyramiding or transgene stacking of different WRR4 paralogs which recognize diverse repertoires of CCG effectors.
Author Contributions
A.R., V.C. and J.D.G.J conceptualized and designed the research. A.R., V.C. and K.B. S.F. conducted all experiments. A.R., V.C., K.B., O.J.F. and S.F. performed the data analysis. M.H.B. and E.H gave critical intellectual input and provided material for this work. A.R., V.C. and J.D.G.J wrote the manuscript with input from all co-authors. All authors helped editing and finalizing the manuscript.
Acknowledgements
A.R. acknowledges support by EMBO LTF (ALTF-842-2015). V.C., O.J.F. and S.F. were supported by Biotechnology and Biological Sciences Research Council (BBSRC) grant BB/L011646/1. K. B. and J.D.G.J. were supported in part by ERC Advanced Investigator grant to JDGJ ‘ALBUGON’ Project ID 233376. Research in the Jones Lab is supported by the Gatsby Foundation (UK) and BBSRC. We thank Shihomi Uzuhashi for her help in initial screening of the Albugo effectors in tobacco.