A downy mildew effector attenuates salicylic acid-triggered immunity in Arabidopsis by interacting with the host mediator complex

Abstract

Plants are continually exposed to pathogen attack but usually remain healthy because they can activate defences upon perception of microbes. However, pathogens have evolved to overcome plant immunity by delivering effectors into the plant cell to attenuate defence, resulting in disease. Recent studies suggest that some effectors may manipulate host transcription, but the specific mechanisms by which such effectors promote susceptibility remain unclear. We study the oomycete downy mildew pathogen of Arabidopsis, Hyaloperonospora arabidopsidis (Hpa), and show here that the nuclear-localized effector HaRxL44 interacts with Mediator subunit 19a (MED19a), resulting in the degradation of MED19a in a proteasome-dependent manner. The Mediator complex of ∼25 proteins is broadly conserved in eukaryotes and mediates the interaction between transcriptional regulators and RNA polymerase II. We found MED19a to be a positive regulator of immunity against Hpa. Expression profiling experiments reveal transcriptional changes resembling jasmonic acid/ethylene (JA/ET) signalling in the presence of HaRxL44, and also 3 d after infection with Hpa. Elevated JA/ET signalling is associated with a decrease in salicylic acid (SA)-triggered immunity (SATI) in Arabidopsis plants expressing HaRxL44 and in med19a loss-of-function mutants, whereas SATI is elevated in plants overexpressing MED19a. Using a PR1::GUS reporter, we discovered that Hpa suppresses PR1 expression specifically in cells containing haustoria, into which RxLR effectors are delivered, but not in nonhaustoriated adjacent cells, which show high PR1::GUS expression levels. Thus, HaRxL44 interferes with Mediator function by degrading MED19, shifting the balance of defence transcription from SA-responsive defence to JA/ET-signalling, and enhancing susceptibility to biotrophs by attenuating SA-dependent gene expression.

Figure 1. HaRxL44 is a nuclear-HaRxL that enhances plant susceptibility to Hpa.

(A) In silico prediction of HaRxL44 protein organization. SP, signal peptide; RFL, RxLR motif. (B) Monitoring of Hpa Waco9 sporulation at 5 d after inoculation in transgenic lines expressing HaRxL44 under the control of 35S promoter (44 lines), under the control of an “haustoriated-cell specific” promoter (dP2–44 lines), under the control of DEX inducible promoter (D44 lines). For D44 lines, plants were treated with DEX 2d after Hpa infection, in order to induce HaRxL44 expression. Expression of HaRxL44 in all the lines was monitored by Western blot (see Figure S3). Error bars represent the standard error of the mean. Asterisks represent the significance of individual unpaired t tests comparing the given column with the control. (C) Subcellular localisation of GFP-HaRxL44 4 DAI with Hpa. The green colour corresponds to the GFP signal, and the red colour corresponds to chloroplast autofluorescence. Asterisks indicate the position of the haustorium. n, nucleus.

Figure 2. MED19a is a positive regulator of nuclear immunity against Hpa.

(A) Schematic diagram of T-DNA insertions in MED19a. (B) MED19a expression in med19a-1 and med19a-2 mutants. (C) Representative images of the phenotype observed in 4-wk-old floral stem of Col-0, med19a-1, med19a-2, and med19a mutant complemented line C1. (D) Developmental phenotype of Arabidopsis transgenic lines OE-MED19a compared to Col-0. (E) Immunoblot of the Co-immunoprecipitation analysis between GFP-MED19a and MED6 and MED7. Arrows point out the interaction detected between GFP-MED19a and MED6 and MED7. (F) Subcellular localisation of GFP-MED19a in Arabidopsis plant. Scale bar, 5 µm. (G) Immunoblot of proteins extracted from two independent lines expressing GFP-MED19a. Stars indicate the expected size for GFP-MED19a. Notice the upper bands in the blot that might suggest posttranscriptional modifications. (H) Monitoring of Hpa sporulation at 5 DAI in control lines (Col-0 and GFP), med19a mutant complemented lines (C1 and C2), Mediator mutants, and MED19a OE lines. Error bars represent the standard error of the mean. Asterisks represent the significance of individual unpaired t tests comparing the given column with the control (p value<0.01).

Figure 3. HaRxL44 destabilizes MED19a in planta.

(A) Subcellular localisation of GFP-HaRxL44 (in green), RFP-MED19a (in red), and YFPc-HaRxL44 + YFPn-MED19a (BiFC, yellow) obtained by transient expression in N. benthamiana. n, nucleus. (B) Western blot analysis of protein extracted after transient expression of DEX::HaRxL44-GFP with RFP or RFP-MED19a in the presence or not of dexamethazone (DEX). Note the decrease in the level of MED19a observed in the presence of HaRxL44. (D) Co-localisation analysis between GFP-MED19a and nuclear-HaRxLs determined by transient assay in N. benthamiana. Note the lack of GFP-MED19a in the presence of RFP-HaRxL44 (arrow).

Figure 4. HaRxL44 interacts with and destabilizes MED19a in a Proteasome-dependant manner.

(A) Schematic representation of the relevant interactions obtained by Y2H between HaRxL44 and Arabidopsis cDNA library. Data extracted from Mukhtar et al. (2011) . (B) Subcellular localisation of GFP-BOI and GFP-MBR1–like determined by transient expression in N. benthamiana. (C) Immunoblotting of protein extracted from N. benthamiana leaves after transient assay, in presence or not of MG132 for 4 h. (D) Immunoblotting of protein extracted from N. benthamiana leaves after transient assay. Note the co-immunoprecipitation (Co-IP) of HA-HaRxL44 with GFP-MED19a in the presence of proteasome inhibitor.

Figure 5. Interaction between MED19a and HaRxL44 is important for HaRxL44–induced MED19a degradation via proteasome.

(A) Immunoblotting of proteins, extracted from N. benthamiana leaves after transient assay. Note the absence of Co-IP of HA-HaRxL44^M with GFP-MED19a. (B) Co-localisation analysis between GFP-MED19a and RFP-HaRxL44 or HaRxL44^M determined by transient assay in N. benthamiana. Note the lack of GFP-MED19a in the presence of RFP-HaRxL44 (arrow) but not HaRxL44^M. (C) Quantification of the number of fluorescent nucleoplasm observed in nucleus transformed with GFP-MED19a in the presence or not of RFP-HaRxL44 or RFP-HaRxL44^M. All the confocal pictures were taken with PMT 1 (494–541 nm) at Gain: 864 and PMT 2 (591–649 nm) at Gain: 844. Note the decrease in GFP-MED19a transformed cells in the presence of RFP-HaRxL44 in comparison with RFP alone or RFP-HaRxL44^M.

Figure 6. HaRxL44 expression affects JA/ET-regulated gene expressions.

Expression of the JA/ET-regulated gene reported by Jung et al. (2007) in HaRxL44 lines and 3 DAI with Hpa Waco9.

Figure 7. HaRxL44-expressing lines, med19a mutants show elevated JA/ET signalling, which is also observed after Hpa infection.

(A and B) qRT-PCR results on PDF1.2 marker gene. Data are presented as average fold induction compared with control of three biological replicates ± SD. (C) Expression pattern of PDF1.2 during a time course of Hpa Waco9 infections in Arabidopsis Col-0 extracted from expression profiling experiment. (D) Expression pattern of HaRxL44 during a time course of Hpa Waco9 infection in Arabidopsis Col-0 analysed by qRT-PCR. Data are presented as average fold induction compared with control of three biological replicates ± SD. (E) Monitoring of Pst growth in Arabidopsis Col-0. DC_Emp, Pst DC3000 strain carrying EDV vector; COR-_Emp, Pst^{COR −} strain carrying EDV vector, COR-_44, Pst^{COR −} strain carrying EDV-HaRxL44. Error bars represent the standard error of the mean (Tukey–Kramer test, p value<0.01). (F) Monitoring of Botrytis cinerea growth 5 DAI in transgenic lines expressing HaRxL44 under the control of DEX inducible promoter (D44 lines) in the presence or not of dexamethazone and under the control of 35S promoter (44 lines). Col-0, HUB1 OE, as well as hub1 KO mutants were used as controls. Error bars represent the standard error of the mean. Asterisks represent the significance of individual unpaired t tests comparing the given column with the control (p value<0.01).

Figure 8. HaRxL44 expression, MED19a mutation suppresses PR1 induction.

(A–C) qRT-PCR on SA marker genes from 5-wk-old Arabidopsis plants. (D–F) qRT-PCR on PR1 marker gene 8 h after SA treatment (200 µM) from 5-wk-old Arabidopsis plants. Data are presented as average fold induction compared with control of three biological replicates ± SD.

Figure 9. Hpa suppresses PR1 induction in infected cells.

(A) qRT-PCR on PR1 gene during a time course of infection of Hpa Waco9 in Arabidopsis Col-0. Data are presented as average fold induction compared with control of three biological replicates ± SD. (B) GUS staining of pro(PR1)::GUS in Arabidopsis leaves 6 DAI Hpa Waco9. Red arrows indicate Hpa hyphae's print surrounded by GUS stained cells. Note that no GUS was detected in Hpa-haustoriated mesophyll cell (black stars), while GUS staining was restricted to nonhaustoriated mesophyll cells (red stars). (C) qRT-PCR on PR1 gene 6 DAI Hpa Waco9 in Arabidopsis Col-0, med19a-1, and med19a-2 KO mutants. (D) Western blot on proteins extracted from med19a mutant complemented with GFP-MED19a after Hpa infection in comparison with mock treatment, using GFP antibody.