Abstract
The plant immune system relies on both cell-surface and intracellular NLR (nucleotide-binding, leucine-rich repeat) receptors. NLRs respond to pathogen effectors and activate effector-triggered immunity: a cocktail of responses, often accompanied by cell death, resulting in resistance.
RPW8 encodes an unusual non-NLR Resistance (R) protein and confers broad-spectrum powdery mildew resistance. It requires genetic components also required by some NLRs, resembles the HeLo-containing protein MLKL (necroptosis executor in animals) and HET-S (cell death executor in fungi) and is targeted to the extra-haustorial membrane during powdery mildew infection by its N-terminal non-cleaved signal anchor domain. RPW8 displays extensive recent duplication events in Arabidopsis and certain alleles can induce oligomerisation-dependent activation of the NLR RPP7.
All these features enabled us to formulate hypotheses for RPW8 function: (1) RPW8 could be a cell death executor for defence against pathogens. (2) RPW8 could be a decoy for effector targets.
To test these hypotheses, we generated a quadruple knock-out mutant of the four RPW8-homologous copies in Arabidopsis Col-0, using CRISPR. The mutant still displays cell death upon activation of four well-characterised NLRs. However, it is partially impaired in powdery mildew resistance and also in bacterial resistance. Interestingly Col-0_rpw8 is delayed in flowering transition. In conclusion, RPW8 plays a broad role in immunity and plant development, beyond resistance to powdery mildew. There is no evidence that it is involved in executing ETI-associated cell death.
Introduction
Plants have co-evolved with their pathogens for millions of years, driving natural selection for effective immunity. The plant immune system relies on recognition of conserved pathogen-, microbe- or damage-associated molecular patterns (PAMPs, MAMPs and DAMPs) by pattern recognition receptors (PRRs) generally localised at the cell surface. PRR activation leads to Ca2+ influx, mitogen activated protein kinase (MAPK) activation, reactive oxygen species (ROS) production and transcriptional reprogramming, resulting in PAMP-triggered immunity (PTI) (Couto & Zipfel, 2016). Pathogens evolved effectors to colonize plants. Some effectors interfere with PTI, resulting in effector-triggered susceptibility (ETS). In turn, plants evolved Resistance- (R-) genes that recognise effectors. R-gene products (i.e. R-proteins) are mainly intracellular. Their activation results in effector-triggered immunity (ETI) which is generally accompanied by salicylic acid production and the hypersensitive response (HR, not to be confused with HR1, HR2, HR3 and HR4, which are “Homologues of RPW8” Arabidopsis genes), a form of programmed cell death at the site of infection (Jones & Dangl, 2006). Elevation of salicylic acid levels induces local defence and systemic acquired resistance (Durner et al., 1997). Localized HR is thought to stop propagation of the pathogen within the host (Morel & Dangl, 1997).
Investigation of the plant immune system revealed hundreds of R-genes in most angiosperm genomes. Most of them belong to the Nucleotide-Binding (NB), Apaf-1, R-protein and CED-4 (ARC) and Leucine-rich repeat (LRR) (NLR) family (Kourelis & Van Der Hoorn, 2018). Earlier this year, cryo-electron microscopy was employed to resolve the structure of ZAR1, a plant NLR, in its non-active and active form (Dangl & Jones, 2019). The non-active form binds ADP in its NB domain and is monomeric (Wang et al., 2019b). The active form binds ATP in its NB domain and is pentameric, forming a resistosome (Wang et al., 2019a) that imposes induced proximity on its N-terminal region. The function of the resistosome is to activate immunity, but the mode of action is not known.
Some components genetically required for NLR function have been characterised. For instance, EDS1 (Enhanced Disease Susceptibility 1) and PAD4 (Phytoalexin Dependent 4) are lipase-like proteins, conserved between many plants and required for TIR-NLR-mediated immunity (Wiermer et al., 2005). TIR-NLRs form a sub-family of plant NLRs that is defined by the presence of an N-terminal TIR (Toll-like, Interleukin-1 receptor and R-protein) domain and specific motifs in the NB-ARC domain. EDS1 is also required redundantly with ICS1 (Isochorismate Synthase 1, required for salicylic acid biosynthesis) for some CC-NLRs (Venugopal et al., 2009). CC-NLRs form a second sub-family of plant NLRs. They are defined by the presence of an N-terminal CC (coiled-coil) domain and specific motifs in the NB-ARC domain.
RPW8 (Resistance to Powdery Mildew 8) is a genetic locus identified in Arabidopsis that confers broad-spectrum resistance to powdery mildew fungi in accession Ms-0. Two homologous R-genes, RPW8.1 and RPW8.2, from the same locus contribute to resistance (Xiao et al., 2001). Transgene analysis showed that adequate expression of RPW8.1 and RPW8.2 also confers resistance to a virulent strain of the oomycete Hyaloperonospora arabidopsidis (Hpa), the cause of downy mildew disease (Wang et al., 2007). Transgenic tobacco plants expressing RPW8.1 and RPW8.2 showed enhanced resistance to powdery mildew (Xiao et al., 2007). More recently, heterologous expression of RPW8.1 in rice and RPW8.2 in grapevine has also been shown to increase resistance to Magnaporthe oryzae (Li et al., 2018) and powdery mildew (Hu et al., 2018), respectively. Both RPW8.1 and RPW8.2 (noted as RPW8 in later text, unless indicated otherwise) encode a small protein (~20 kDa) with a predicted CC domain. Intriguingly, RPW8 shares sequence homology to the N-termini of a third sub-family of plant NLRs, called RPW8-NLRs. This sub-family of NLRs often have only a few members but exists in almost all plants (Zhong & Cheng, 2016). At least in angiosperms, these RPW8-NLRs are helper NLRs required for signalling of multiple sensor NLRs (Peart et al., 2005; Bonardi et al., 2011; Dong et al., 2016; Qi et al., 2018; Adachi et al., 2019; Castel et al., 2019a; Wu et al., 2019).
Genetic analysis showed that the resistance function of RPW8 requires previously characterised immune components, including EDS1, PAD4 and the salicylic acid pathway, but does not require NDR1 (which is usually required for CC-NLRs), nor COI1 and EIN2 (which regulate jasmonic acid and ethylene signalling respectively) (Xiao et al., 2005). Because EDS1, PAD4 and salicylic acid are generally associated with TIR-NLR-mediated immunity, the above results imply a functional link between RPW8 and TIR-NLR signalling.
During their co-evolution with hosts, some filamentous pathogens, such as powdery mildew and rust fungi, and oomycetes have evolved a common invasive strategy: the formation of haustoria as their feeding structures. Recent studies have shown that the haustorium is encased by a host-derived special membrane called the extra-haustorial membrane (EHM). Interestingly, RPW8.2 is specifically targeted to the EHM during powdery mildew infection (Wang et al., 2009, 2013). The N-terminal domain of RPW8.2, along with two EHM-targeting motifs, is predicted to associate with membranes, and is required for EHM targeting and resistance function of RPW8.2. Intriguingly, the N-terminal portion (~90 amino acids) of RPW8 shares similarity with MLKL from animals and HELLP and other HeLo-domain-containing fungal proteins (Daskalov et al., 2016). MLKL can oligomerise to form pores at the membrane during necroptosis in animals (Murphy et al., 2013). HELLP has a similar function in fungi and can form prions via its C-terminus. Prionisation is associated with membrane targeting and disruption, followed by cell death. A glycine zipper motif conserved between MLKL, HELLP and RPW8 is required for HELLP membrane targeting (Daskalov et al., 2016). In parallel, RPW8 is capable of inducing HR (Xiao et al., 2001) and overexpression of RPW8 can lead to cell death (Xiao et al., 2003, 2005). However, it is not known whether RPW8’s membrane-targeting and cell death induction share the same molecular basis with that of HELLP and MLKL.
Another interesting feature of the RPW8 locus is the intraspecific gene amplification and diversity within Arabidopsis. RenSeq was conducted on 65 Arabidopsis accessions and revealed two types of sequence variation at the RPW8 locus (Barragan et al., 2019; Van de Weyer et al., 2019). Firstly, there have been a large number of recent duplication events. There are seven sub-clades of RPW8 paralogs: RPW8.1, RPW8.2, RPW8.3, HR1, HR2, HR3 and HR4. Each Arabidopsis accession contains a unique combination of these alleles with different copy numbers. For instance, Ms-0 contains five RPW8 paralogs: RPW8.1 and RPW8.2, which confer resistance to powdery mildew, and HR1, HR2 and HR3, which are not required for powdery mildew resistance (Xiao et al., 2001). Col-0, which is susceptible to powdery mildew, contains HR1, HR2, HR3, and HR4, an additional homolog most closely related to RPW8.1 (Xiao et al., 2001, 2004; Berkey et al., 2017). Interestingly, TueWa1-2 contains as many as 13 RPW8 paralogs (including three copies of RPW8.1, two of RPW8.2, four of RPW8.3, one of HR1, one of HR2, one of HR3 and one of HR4). The selective advantage of such extreme variation in duplication events is not understood. Secondly, some RPW8 genes encode proteins with intragenic C-terminal duplications. For instance, RPW8.1 from some accessions contains one or more duplications of a 21-amino acid motif (Orgil et al., 2007), and HR4 from Col-0 contains five duplications of a 14-amino acid motif (Xiao et al., 2004), whereas HR4 from ICE106 contains only two, and HR4 from Fei-0 only one such motif (Barragan et al., 2019).
A large-scale genetic analysis on intraspecific hybrid necrosis revealed that some RPW8 alleles are “incompatible” with alleles of an NLR-type R-gene at the RPP7-locus (Chae et al., 2014). Some Arabidopsis crosses display spontaneous necrosis in F1, due to the presence of two incompatible alleles, one coming from each parent. The “Dangerous Mix” (DM) loci have been mapped. Particularly, DM6 (RPP7) and DM7 (RPW8) are responsible for incompatibility in three crosses: Mrk-0 × KZ10, Lerik1-3 × Fei-0 and ICE79 × Don-0. In KZ10 and Fei-0, the RPW8 genes responsible for the phenotype are RPW8.1 and HR4 respectively (Barragan et al., 2019). RPW8.1KZ10 has three C-terminal intragenic repeats, the maximum observed among known RPW8.1 alleles. Interestingly, Arabidopsis accessions carrying an RPW8.1 allele with three C-terminal repeats all form a dangerous mix with Mrk-0. In contrast, HR4Fei-0 has only one C-terminal repeat. Arabidopsis accessions carrying an HR4 allele with one C-terminal repeat all form a dangerous mix with Lerik1-3. The number of C-terminal repeats thus correlates with incompatibility with different RPP7 alleles. However, the function of the repeat is yet to be discovered. In addition, HR4Fei-0 induces oligomerisation of RPP7bLerik1-3, resulting in HR in Nicotiana benthamiana (Li et al., 2019).
We can formulate two hypotheses about RPW8 function. (1) Based on sequence similarity, RPW8 could function as the HeLo domain-containing proteins MLKL, HET-S and HELLP from animals and fungi. This function would be a ligand-dependent oligomerization resulting in membrane targeting and disruption for programmed cell death. (2) RPW8 could also be a decoy of effector targets. NRG1 and ADR1 are low copy number and are helper NLRs (Jubic et al., 2019). Their conserved RPW8 domain could in theory provide a potent effector target to suppress immunity mediated by multiple sensor NLRs. RPW8 paralogs could encode decoys for such effectors. They could either trap the effectors to prevent their virulence function or be modified and guarded by NLRs such as RPP7. The extreme degree of recent duplication of RPW8 in Arabidopsis is consistent with this hypothesis.
To test both hypotheses, we generated a quadruple rpw8 mutant (hr1-hr2-hr3-hr4, we called Col-0_rpw8) in Arabidopsis Col-0 to test for possible altered immunity. Our results indicate a function for RPW8 beyond defence against powdery mildew pathogens.
Materials and Methods
Plant genotypes and growth conditions
Arabidopsis thaliana (Arabidopsis) accessions used in this study is Columbia-0 (Col-0). Col-0_eds1 is an eds1a-eds1b double CRISPR mutant, published as eds1-12 (Ordon et al., 2017). Seeds were sown directly on compost and plants were grown at 21°C, with 10 hours of light and 14 hours of dark, 75% humidity. For seed collection, 5-week old plants were transferred under long-day condition: 21°C, with 16 hours of light and 8 hours of dark, 75% humidity. For Nicotiana benthamiana, seeds were sown directly on compost and plants were grown at 21°C, with cycles of 16 hours of light and 8 hours of dark, 55% humidity. N. benthamiana_nrg1 is an nrg1 CRISPR mutant of N. benthamiana (Castel et al., 2019a).
Generation of an hr1-hr2-hr3-hr4 (aka rpw8) null mutant in Arabidopsis Col-0 using CRISPR
CRISPR modules were assembled using the Golden Gate cloning method. The detailed Golden Gate method and protocols can be found in (Engler et al., 2009, 2014; Weber et al., 2011). Specific details for CRISPR module assembly can be found in (Castel et al., 2019b). All the vector used can be found on Addgene (addgene.org). 12 sgRNAs targeting HR1 (TGGCGTCGTGAAGGAGTTGG[nGG], CGACGCCATCATAAGAGCCA[nGG] and GTTCATCGACTTCTTCGGTG[nGG]), HR2 (TGCTCTCCAAATCCTTCACG[nGG], GTCTCGATTCTACAATCTTG[nGG] and GGTTCTTGTCGAAGCTTATG[nGG]), HR3 (GGTTAGTGAGATTATGGCAG[nGG], GTCTTGATGCTACAATCTTT[nGG] and CGATAAGCTTAGCGAAGAAG[nGG]) and HR4 (GCTTGCTGTAATCAAAACAG[nGG], TGGAAAGTATCAGTCCGGTG[nGG] and GGAGACGCGTAAGACTTTCG[nGG]) were designed and assembled by PCR to a sgRNA backbone and 67 bp of the AtU6-26 terminator. sgRNA1 to sgRNA12 were assembled with the AtU6-26 promoter in the Golden Gate compatible level 1 pICH47761, pICH47772, pICH47781, pICH47791, pICH47732, pICH47742, pICH47751, pICH47761, pICH47772, pICH47781, pICH47791 and pICH47732 respectively. A level 1 vector pICH47811 (with expression in reverse orientation compared to the other level 1 modules) containing a human codon optimized allele of Cas9 under the control of the AtRPS5a promoter and the Pisum sativum rbcS E9 terminator (Addgene: 117505) was assembled with a FAST-Red selectable marker (Addgene: 117499) into a level M vector pAGM8031, using the end-linker pICH50892. sgRNA1 to sgRNA5 were assembled into a level M vector pAGM8067, using the end-linker pICH50872. sgRNA7 to sgRNA9 were assembled into a level M vector pAGM8043, using the end-linker pICH50914. sgRNA10 to sgRNA12 were assembled into a level M vector pAGM8043, using the dummy module pICH54033 and the end-linker pICH50881. The four level M modules were assembled into level P vector pICSL4723-P1, using the end-linker pICH79264. Level 1 vectors were cloned using BsaI enzyme and carbenicillin resistance. Level M vectors were cloned using BpiI enzyme and Spectinomycin resistance. Level P vector was cloned using BsaI enzyme and kanamycin resistance. The final vector map can be found in the Supplemental Information. It was expressed via Agrobacterium tumefaciens strain GV3101 in Arabidopsis Col-0. In the first generation after transformation, we recovered somatic mutants. The non-transgenic progeny of a somatic mutant contained a homozygous quadruple knock-out line. One mutation is a 6579 bp deletion between C173 of HR1 and G22 of HR3. The second mutation is c.28delA in HR4, causing an early stop codon at the N-terminal encoding region. The resulting line lacks functional HR1, HR2, HR3, HR4 and is T-DNA free. We called this line Col-0_rpw8.
Transcript level measurement
For gene expression analysis, RNA was isolated from three biological replicates and used for subsequent reverse transcription quantitative PCR (RT-qPCR) analysis. RNA was extracted using the RNeasy Plant Mini Kit (QIAgen) and treated with RNase-Free DNase Set (QIAGEN). Reverse transcription was carried out using the SuperScript IV Reverse Transcriptase (ThermoFisher). qPCR was performed using CFX96 TouchTM Real-Time PCR Detection System. Primers for qPCR analysis of PR1 are ATACACTCTGGTGGGCCTTACG and TACACCTCACTTTGGCACATCC. Primers for qPCR analysis of EF1α are CAGGCTGATTGTGCTGTTCTTA and GTTGTATCCGACCTTCTTCAGG. Data were analysed using the double delta Ct method (Livak & Schmittgen, 2001).
Cell death assay in Arabidopsis
Pseudomonas fluorescens engineered with a Type Three Secretion system (Pf0-1 EtHAn) (Thomas et al., 2009) carrying pBS46:AvrRps4, pBS46:AvrRps4KRVY, pBS46:AvrRpt2, pVSP61:AvrRpm1 or pVSP61:AvrPphB were grown on selective KB-medium agar plate for 48 hours at 28 °C. Bacteria were harvested from plate, re-suspended in infiltration buffer (10 mM MgCl2, pH 5.6) and concentration was adjusted to OD600 = 0.2 (~108 cfu/ml). The abaxial surface of 4-week old Arabidopsis leaves were hand-infiltrated with 1 ml needle-less syringe. Cell death was monitored 24 hours after infiltration.
Cell death assay in N. benthamiana
A. tumefaciens strains were streaked on selective media and incubated at 28 °C for 24 hours. A single colony was transferred to liquid LB medium with appropriate antibiotic and incubated at 28 °C for 24 hours in a shaking incubator (200 rotations per minute). The resulting culture was centrifuged at 3000 rotations per minute for 5 minutes and resuspended in infiltration buffer (10 mM MgCl2, 10 mM MES, pH 5.6) at OD600 = 0.4 (2×108 cfu/ml). For co-expression, each bacterial suspension was adjusted to OD600 = 0.4 in the final mix. The abaxial surface of 4-weeks old N. benthamiana were infiltrated with 1 ml needle-less syringe. Cell death was monitored three days after infiltration.
Vector used are: RRS1-RK1221Q-FLAG and RPS4-HF (Sarris et al., 2015), EDS1-V5 and SAG101-Myc (Huh et al., 2017), NRG1B-HF (Castel et al., 2019a) and HR2-mNeon (this article). HR2 was amplified from Col-0 (Fw: GGCTTAAUATGCCTCTTACCGAGATTATCG, Rv: AACCCGAUTTCAAAACGAAGCGAATTC) and mNeon c-terminal tag was amplified from pICSL50015 (Addgene: 50318) (Fw: ATCGGGTUTGGTGAGCAAGGGAGAGGAG G, Rv: GGTTTAAUTTACTTGTAAAGCTCGTCCA) using “KAPA HiFi HotStart Uracil+ ReadyMix (2X)” enzyme (KAPABIOSYSTEMS), following the manufacturer protocol. Amplicons were assembled in a USER compatible vector LBJJ234-OD (containing a FAST-Red selectable marker and a 35S / Ocs expression cassette, pre-linearized with PacI and Nt. BbvcI restriction enzymes), with the USER enzyme (NEB), following the manufacturer protocol. The final vector map can be found in the Supplemental Information.
Bacterial growth measurement
Pseudomonas syringae pv. tomato strain DC3000 carrying pVSP61:AvrRps4-HA or pVSP61 empty vector were grown on selective KB-medium agar plate for 48 hours at 28 °C. Bacteria were harvested from plate, re-suspended in infiltration buffer (10 mM MgCl2, pH 5.6) and concentration was adjusted to OD600 = 0.0005 (~2.5×105 cfu/ml). The abaxial surface of 5-week old Arabidopsis leaves were hand-infiltrated with 1 ml needle-less syringe. Plants were covered with a lid for the first 12 hours of bacterial growth. For quantification, leaf samples were harvested with a 6 mm diameter cork-borer, resulting in a ~0.283 cm2-sized leaf disc. Two leaf discs per leaf were harvested and used as single sample. For each condition, four samples were collected just after infiltration and six samples were collected 72 hours after infiltration. Samples were ground in 200 μl of infiltration buffer, serially diluted (5, 50, 500, 5000 and 50000 times) and spotted (5 to 10 μl per spot) on selective KB-medium agar plate to grow 48 hours at 28 °C. The number of colonies (cfu per drop) was monitored and the bacterial growth was expressed in cfu/cm2 of leaf tissue.
Albugo candida propagation
For propagation of Albugo candida race 2V (Rimmer et al., 2000), zoospores were suspended in water (~105 spores/ml) and incubated on ice for 30 min. The spore suspension was then sprayed on plants using a Humbrol® spray gun (~700 μl/plant) and plants were incubated at 4 °C in the dark overnight. Infected plants were kept under 10 hours light (20 °C) and 14 hours dark (16 °C) cycles. Phenotypes were monitored 12 days after spraying.
Powdery mildew propagation
For testing with powdery mildew, plants were grown under short-day (8 hr light, 16 hr dark) and 75% relative humidity at 22 °C for 7 weeks before inoculation with an adapted powdery mildew isolate Golovinomyces cichoracearum (Gc) UCSC1 or a non-adapted isolate Gc UMSG1 as previously reported (Xiao et al., 2005; Wen et al., 2011). Disease phenotypes with Gc UCSC1 were visually scored and photographed. Disease susceptibility was further assessed by counting total number of spores per mg leaf tissue. At least 6 infected leaves were weighed and combined as one leaf sample, and 4 samples were collected from 12 infected plants for disease quantification. A spore suspension of each sample was made by vortexing the leaves for 1 min in 10 ml of H2O containing 0.02 % Silwet L-77 and used for spore counting using LunaTM Automated Cell Counter (Logos biosysems). Spore counts were normalized to the fresh weight of the leaf samples. For infection tests with Gc UMSG1, which can largely penetrate the cell wall of Arabidopsis but fails to establish micro-colony capable of sporulation (Wen et al., 2011), inoculated leaves were collected at 5 dpi and subjected to trypan blue staining, and total the hyphal length was measured as previously reported (Wen et al., 2011). All infection trials were repeated three times with similar results.
Phylogenetic reconstruction of RPW8 homologs in Arabidopsis Col-0
RPW8 domain boundaries were predicted using SMART (http://smart.embl-heidelberg.de/) (Letunic & Bork, 2017). Amino acid sequences of RPW8 domains were aligned using the MUSCLE method. The evolutionary history was inferred using the Neighbor-Joining method. The evolutionary distances were computed using the Poisson correction method. All positions with less than 95% site coverage were eliminated. Evolutionary analyses were conducted using the software MEGA7.
Results
CRISPR mutagenesis enables recovery of an rpw8 quadruple mutant
In Col-0, the four RPW8 homologs HR1, HR2, HR3 and HR4 are located in an 11.8 kb cluster on Chromosome 3 (Figure 1). To generate null alleles, we designed three sgRNAs per gene, targeting their N-terminal encoding region. Cas9 activity at the sgRNA target could result in large deletion of several genes and/or indels causing early stop codons in individual genes. We expressed the 12 sgRNAs along with Cas9 from a single T-DNA in Arabidopsis Col-0. We identified a line lacking functional HR1, HR2, HR3, HR4 that is T-DNA free (see Materials and Methods for more details). We called this line Col-0_rpw8.
rpw8 is slightly autoimmune
Under normal growth conditions, Col-0_rpw8 plants appear slightly smaller than WT plants (Figure 2A). Mutations that result in activation of constitutive defence often display a dwarf phenotype (van Wersch et al., 2016). The “autoimmunity” is usually associated with elevated salicylic acid (SA) levels, resulting in elevated expression of the SA marker gene PR1. We measured PR1 expression in Col-0_rpw8 to test for autoimmunity. PR1 is significantly more highly expressed in three independent Col-0_rpw8 lines than in WT (Figure 2B). However, it is only expressed at ~0.6% of the level of EF1α. In contrast, activation of the NLR pair RRS1/RPS4 induces PR1 to ~12× the level of EF1α (Castel et al., 2019a). Autoimmune Arabidopsis mutants expressing the effector hopZ5 show PR1 expression levels of ~3 to ~260 time more than EF1α (Jayaraman et al., 2017). Thus, the autoimmunity of Col-0_rpw8 is detectable but very low.
Intriguingly, we found that after growing in short-day (8 hr light, 16 hr dark) conditions for 13 weeks, Col-0 plants started bolting, but there was no sign of bolting in Col-0_rpw8 plants (Figure 3). However, there was no noticeable difference among plants of these two genotypes when they were grown in short-day for 4-6 weeks and then shifted to long-day (16 hr light and 8 hr dark). This unexpected result implies that RPW8 homologs, or perhaps salicylic acid levels, play a role in promoting flowering under short-day conditions.
RPW8 homologs are not required for cell death mediated by four well-described NLRs
NLR activation often results in a form of cell death called the hypersensitive response (HR). The bacterial effectors AvrRpm1, AvrRpt2, AvrPphB and AvrRps4 can cause an RPM1-, RPS2-, RPS5- and RRS1/RPS4-dependent HR (Kunkel et al., 1993; Grant et al., 1995; Warren et al., 1998; Gassmann et al., 1999) respectively. We delivered these effectors into Col-0 WT and Col-0_rpw8, using Pf0-1, a non-pathogenic strain of P. fluorescens engineered with a Type III Secretion System to deliver an effector of interest into the host cell cytosol (Thomas et al., 2009). AvrRps4KRVY, an inactive form of AvrRps4, was used as a negative control. Each effector (apart from AvrRps4KRVY) can still trigger HR in Col-0_rpw8 (Figure 4). These data indicate that RPW8 homologs are not required for RPM1-, RPS2-, RPS5- or RRS1/RPS4-mediated HR.
We then tested whether RPW8 is sufficient to transduce RRS1/RPS4 signal for HR. EDS1, SAG101 and NRG1 are the three major components of TIR-NLR signalling for HR in Arabidopsis and in Nicotiana benthamiana (Lapin et al., 2019). We transiently expressed RRS1-RK1221Q (an RPS4-dependent auto-active form of RRS1-R) and RPS4 with EDS1 and/or SAG101 and/or NRG1B and/or HR2 from Arabidopsis. The minimal requirement to reconstruct RRS1/RPS4-mediated HR in N. benthamiana is EDS1/SAG101/NRG1, but HR2 is dispensable (Figure 5 and S1). This indicates that Col-0 RPW8 is neither necessary nor sufficient to transduce RRS1/RPS4 signal for HR. Parenthetically, it also indicates that N. benthamiana alleles of SAG101/EDS1/NRG1 cannot transduce the RRS1/RPS4 response, while Arabidopsis alleles can.
Col-0_rpw8 mutant supports more growth of a virulent bacterial pathogen
RRS1/RPS4-mediated HR is not RPW8-dependent (Figure 4 and Figure 5). Consistently, the growth of P. syringae pv tomato strain DC3000 carrying AvrRps4 is not affected in Col-0_rpw8 (Figure 6A and S2A-B). AvrRps4 avirulence function observed in Col-0, is lost in Col-0_eds1 but maintained in Col-0_rpw8, indicating that RPW8 is not required for RRS1/RPS4-mediated resistance. In contrast, the growth of DC3000 carrying an empty vector is significantly enhanced in Col-0_rpw8 compared to that in Col-0 (Figure 6B and S2C-D). In fact, Col-0_rpw8 is as susceptible as Col-0_eds1, indicating that loss of all RPW8 homologs compromises resistance to bacterial pathogens.
RPW8 homologs are not required for WRR4A-mediated resistance to Albugo candida
Albugo candida is an oomycete causing white rust in Brassicaceae. The race Ac2V, isolated from Brassica juncea in Canada, can grow on Arabidopsis transgressive segregants or eds1 mutants (Rimmer et al., 2000; Cevik et al., 2019). Ac2V is resisted by WRR4A, WRR4B and an unknown recessive or haplo-insufficient R-gene in Col-0 (Borhan et al., 2008; Cevik et al., 2019). We tested Col-0, Col-0_rpw8 and Col-0_eds1 with Albugo candida Ac2V and found that resistance conferred by the aforementioned R-genes requires EDS1 but not RPW8 (Figure 7).
Col-0_rpw8 is more susceptible to adapted and non-adapted powdery mildew fungi
Col-0 is susceptible to the adapted (virulent) powdery mildew isolate Gc UCSC1 but still mounts SA-dependent basal resistance against it (Xiao et al., 2005). To test if and how much the four RPW8 homologs contribute to the basal resistance, we inoculated eight weeks-old plants of Col-0 and Col-0_rpw8 with Gc UCSC1. At 10 dpi, Col-0_rpw8 plants support more fungal growth than Col-0 plants based on visual scoring in three independent experiments (one representative is shown in Figure 8A). Quantification of disease susceptibility showed that Col-0_rpw8 plants produced ~1.7× spores per mg fresh leaf tissue (Figure 8B).
In contrast, Col-0 and 24 other tested Arabidopsis accessions were completely resistant to Gc UMSG1, which can only develop very short hyphae and then is arrested shortly after spore germination (Wen et al., 2011). To test if RPW8 homologs are involved in resistance against non-adapted powdery mildew, we inoculated plants of Col-0 and Col-0_rpw8 with Gc UMSG1 and measured total hyphal length at 5 dpi. We found that Col-0_rpw8 plants supported much more extensive hyphal growth compared to Col-0 (Figure 9), indicating that RPW8 homologs in Col-0 collectively contribute to basal resistance against non- or poorly-adapted powdery mildew.
Discussion
Based on previous characterisation of RPW8 (Xiao et al., 2005; Collier et al., 2011; Barragan et al., 2019; Li et al., 2019), we propose the following two hypotheses that could explain the molecular and biochemical function of RPW8. The first hypothesis posits that RPW8 homologs and the RPW8 domain in RPW8-NLRs participate in activation of immune responses through oligomerization and membrane insertion (pore-forming), thus contributing to HR and pathogen resistance. The recent analysis revealing similarities between RPW8 and the HeLo-domain-containing proteins MLKL, HELLP and HET-S from animals and fungus (Daskalov et al., 2016) supports this hypothesis. In animals and fungi, these HeLo-domain-containing proteins can oligomerise (upon sensing of various signals) causing the HeLo-domain to form a disruptive membrane insertion structure (i.e. pore-forming), resulting in cell death (Cai et al., 2017). Based on the presence of a putative N-terminal HeLo domain, MLKL-encoding genes have recently been identified in plants (Mahdi et al., 2019). Ablation of all the three MLKL genes in Arabidopsis compromised resistance against biotrophic pathogens. These plant MLKL proteins can form tetramers and are associated with microtubules, suggesting a cell death-independent immunity mechanism (Mahdi et al., 2019). How these MLKL proteins enhance plant immunity remains unresolved. It is possible that RPW8 homologs and plant MLKL-like proteins define a HeLo-domain-containing protein family in plants and serve similar functions as their fungal and animal counterparts. If so, RPW8-NLRs such as ADR1 and NRG1 could then be renamed HeLo-NLRs. The earlier observations that RPW8.2 and other RPW8 homologs are localized to the EHM (Wang et al., 2009; Berkey et al., 2017) and over-expression of RPW8 results in massive cell death (Xiao et al., 2003) support this hypothesis. However, the precise molecular function of the putative N-terminal HeLo-domain of MLKL and RPW8 has yet to be characterized.
In this study, we showed that the HR triggered by five well-described NLRs does not require any RPW8 homologs in Arabidopsis (Figure 4). In addition, HR2 is not sufficient to reconstruct RRS1/RPS4-triggered HR in N. benthamiana (Figure 5). By contrast, loss of all four RPW8 homologs compromised resistance against P. syringae and powdery mildew (Figure 6, Figure 8 and Figure 9). These observations, together with the earlier findings that RPW8 enhances resistance against an oomycete (Hpa) and interacts with RPP7 (Wang et al., 2007; Li et al., 2019), suggest that RPW8 plays an important role in plant immunity but is dispensable for the HR. However, unlike plant MLKLs whose role in cell death is unclear (Mahdi et al., 2019), overexpression of wild-type RPW8 homologs or variants or C-terminal YFP-tagged HR3 results in cell death (Xiao et al., 2003; Wang et al., 2013; Berkey et al., 2017). Since all tested RPW8 homologs are capable of membrane targeting during infection (Wang et al., 2013; Berkey et al., 2017), RPW8 function could be to direct the associated protein complex to a specific subcellular compartment to activate local immune response, including cell death in extreme cases.
According to the second hypothesis, RPW8 proteins are decoys for RPW8-NLR-targeting effectors. However, such avirulent effectors are not known. Even if they exist, the loss of their function in Col-0_rpw8 would not been seen if other avirulent effectors get recognised by an RPW8-independent mechanism. Given that the pathogens tested in our study contain up to several hundreds of effectors, this may well be the case. Thus, our results neither validate or invalidate this RPW8 decoy hypothesis. Characterisation of an RPW8-targeted effector would enable further investigation of this question.
A detailed search for RPW8 and RPW8 domain revealed 11 predicted proteins in Col-0 (Zhong & Cheng, 2016). Four are Homologues of RPW8 (HR1, HR2, HR3 and HR4), five are RPW8-containing NLRs (ADR1, ADR1-L1, ADR1-L2, NRG1A and NRG1B), one is a non-canonical NLR (DAR5: RPW8-NB-ARC-LIM) and the last one is encoded by AT3G26470. The predicted protein encoded by AT3G26470 has 221 amino acid protein (25.3 kDa) and contains an N-terminal RPW8 domain. AT3G26470 displays a two-exons structure, typical for RPW8. We thus consider this gene a distant Homologue of RPW8 and called it HR5. However, the RPW8 domain of HR5 resembles more the RPW8 domain of ADR1-L1 than that of NRG1 or HR1, HR2, HR3 and HR4 (Figure S3). Hence, HR5 might have resulted from a recent duplication event of the ADR1-L1 RPW8 domain. Similar to the above, HR5 could be a decoy for an ADR1-L1-targeting effector, independently of any HR1, HR2, HR3 and/or HR4 function. Parsimony suggests that HR1, HR2, HR3 and HR4 share a common ancestor, which is not shared with HR5. Thus, if HR1, HR2, HR3 and HR4 play a redundant function, it is not likely to be shared with HR5.
A recent study highlighted co-evolution between EDS1, SAG101 and NRG1 within plant clades to regulate downstream pathways in TIR-NLR-mediated immunity. Particularly, these three components need to have co-evolved to be functional (Lapin et al., 2019). In this study, we tested whether HR2 is part of the minimal required components to reconstruct RRS1/RPS4-triggered HR in N. benthamiana. We found that co-expression of Arabidopsis EDS1, SAG101 and NRG1 is necessary to reconstitute RRS1/RPS4-triggered HR in N. benthamiana, but HR2 is dispensable (Fig. 5). Thus, RPW8 is neither required nor sufficient, so likely not involved, in RRS1-RPS4-mediated HR. We confirmed that EDS1, SAG101 and NRG1 need to come from the same genome to function, as previously reported (Lapin et al., 2019). However, unlike Roq1 (a TIR-NLR from N. benthamiana) that can signal via EDS1/SAG101/NRG1 from either Arabidopsis or N. benthamiana genomes (Lapin et al., 2019), RRS1-RPS4 can only fully signal via Arabidopsis EDS1/SAG101/NRG1. It indicates that the TIR-NLRs Roq1 and RPS4 are sensed differentially by EDS1, SAG101 and NRG1. Recently, TIR domains of plant NLRs have been shown to possess enzymatic activity (Horsefield et al., 2019; Wan et al., 2019). They can degrade nicotinamide adenine dinucleotide in its oxidized form (NAD+) into adenine dinucleotide ribose (ADPR), variant-cyclic ADPR (v-cADPR) and nicotinamide (Nam) (Wan et al., 2019). Our data show that N. benthamiana EDS1/SAG101/NRG1 does not sense RRS1/RPS4 activity, but Arabidopsis EDS1/SAG101/NRG1 can, even in the N. benthamiana system. This suggests a more complex signal transduction pathway between the activation of TIR-NLRs, enzymatic activity and activation of EDS1/SAG101/NRG1.
In conclusion, we generated an hr1-hr2-hr3-hr4 quadruple mutant (called rpw8 mutant) in Arabidopsis Col-0. NLR-mediated phenotypes tested remain intact in the mutant. However, Col-0_rpw8 is partially compromised in resistance against powdery mildew and P. syringae, and surprisingly in flowering under short-day growth conditions. Future research is needed to characterize the precise mechanism by which RPW8 proteins or domains contribute to immunity, in comparison to the HeLo-containing proteins in fungi and animals.
Acknowledgments
Work at The Sainsbury Laboratory (BC and JDGJ) was supported by the Gatsby Foundation (http://www.gatsby.org.uk/). Work at the University of Maryland (YW and SX) was supported by a National Science Foundation grant (IOS-1457033). The authors thanks Mark Youles (Synbio at The Sainsbury Laboratory) for help in plasmid vector development.