A dominant-negative avirulence effector of the barley powdery mildew fungus provides mechanistic insight into barley MLA immune receptor activation

Abstract

Nucleotide-binding leucine-rich repeat receptors (NLRs) recognize pathogen effectors to mediate plant disease resistance often involving host cell death. Effectors escape NLR recognition through polymorphisms, allowing the pathogen to proliferate on previously resistant host plants. The powdery mildew effector AVRA13-1 is recognized by the barley NLR MLA13 and activates host cell death. We demonstrate here that a virulent form of AVRA13, called AVRA13-V2, escapes MLA13 recognition by substituting a serine for a leucine residue at the C-terminus. Counterintuitively, this substitution in AVRA13-V2 resulted in an enhanced MLA13 association and prevented the detection of AVRA13-1 by MLA13. Therefore, AVRA13-V2 is a dominant-negative form of AVRA13 and has probably contributed to the breakdown of Mla13 resistance. Despite this dominant-negative activity, AVRA13-V2 failed to suppress host cell death mediated by the MLA13 autoactive MHD variant. Neither AVRA13-1 nor AVRA13-V2 interacted with the MLA13 autoactive variant, implying that the binding moiety in MLA13 that mediates association with AVRA13-1 is altered after receptor activation. We also show that mutations in the MLA13 coiled-coil domain, which were thought to impair Ca2+ channel activity and NLR function, instead resulted in MLA13 autoactive cell death. Our results constitute an important step to define intermediate receptor conformations during NLR activation.

Keywords: Blumeria graminis; AVR; MLA; Mildew Locus A; NLR; barley; cell death; fungal effector; powdery mildew; resistance.

Fig. 1.

The C-terminus of AVR_A13 effectors controls interaction with and activation of MLA13. (A) Amino acid alignment of AVR_A13 variants analysed for interaction with MLA13 and activation of MLA13-mediated cell death. Signal peptide (SP) residues are underlined; amino acids in blue and pink highlight the amino acid variation between AVR_A13-V2 and AVR_A13-1, respectively. (B and C) Nicotiana benthamiana leaves were transformed transiently with 35S:Mla13-4Myc (pGWB517) with one of the AVR_a13 variants lacking SPs cloned between the 35S promoter and a C-terminal mYFP sequence or empty vector (EV). (B) Cell death was determined 3 d post-transformation, and figures shown are representatives of at least nine independent leaves from at least three independent plants. (C) Protein stability of the AVR_A13 variants fused to mYFP corresponding to constructs of (B). Leaf tissue was harvested 2 d post-infiltration. Total protein was extracted, separated by gel electrophoresis, and probed by anti-GFP. (D and E) Yeast cells were co-transformed with Mla13 fused N-terminally to the LexA-binding domain (BD) sequence and AVR_a13 variants lacking SPs fused N-terminally to the B42 activation domain (AD) and 1×HA tag sequence as indicated. Growth of transformants was determined on selective growth medium containing raffinose and galactose as carbon sources but lacking uracil, histidine, and tryptophan (–UHW), and interaction of proteins was determined by leucine reporter activity reflected by growth of yeast on selective medium containing raffinose and galactose as carbon sources but lacking uracil, histidine, tryptophan, and leucine (–UHWL). Figures shown are representatives of at least three experiments, and pictures were taken 6–8 d after drop-out. (E) Protein levels of BD-MLA13 and AD-AVR_A variants corresponding to yeast of (D). Yeast transformants were grown in raffinose- and galactose-containing selective medium lacking uracil, tryptophan, and histidine to OD₆₀₀=1. Then, cells were harvested, total protein extracted, separated by gel electrophoresis, and western blots were probed with anti-LexA or anti-HA as indicated. CBB: Coomassie brilliant blue.

Fig. 2.

AVR_A13-V2 can act as dominant-negative effector on MLA13. Nicotiana benthamiana leaves were co-transformed transiently with cDNAs of Mla1 or Mla7 or MLA13 (pGWB vectors) with AVR_a1, AVR_a7-2, AVR_a13-1, or empty vector (EV) as indicated and either AVR_a13-V1, AVR_A13-V2, or EV fused to epitope tags as indicated. All constructs were expressed from the 35S promoter. (A and B) Cell death was determined 3–4 d post-transformation and (B) scored from 0 to 3 based on the cell death scale indicated. All values obtained in at least three independent experiments are indicated by dots; error bars=SE. Differences between samples were assessed by non-parametric Kruskal–Wallis and subsequent Dunn’s tests for each MLA variant. Calculated P-values were as follows: Mla1, P=0.824; Mla7, P=0.551; and Mla13, P=1.00E-06. Samples marked by identical letters in the plots do not differ significantly (P<0.05) in the Tukey test for the corresponding MLA. (C) Protein levels corresponding to samples of (B). Leaf tissue was harvested 2 d post-infiltration. Total protein was extracted and recovered by GFP-Trap (AVR_a1 and AVR_a7-2) separated by gel electrophoresis, and probed by anti-HA (MLAs), anti-Myc (AVR_A13-V2-4×Myc), or anti-GFP (AVR_A1–mYFP, AVR_A7-2–mYFP, and AVR_A13-1–mYFP) as indicated. CBB: Coomassie brilliant blue. (D) Barley protoplasts were transfected with pUBQ:luciferase (4.5 µg) and genes encoding Mla1, Mla7, or Mla13 and either an EV (reference sample) or AVR_a1, AVR_a7-2, or AVR_a13-1 lacking their respective signal peptides (SPs), respectively. Additionally, an EV or AVR_a13-V1 or AVR_a13-V2 lacking their respective SPs was co-expressed. The piPKb002 vector was used for all Mla and AVR_a constructs and, for each transfection, 9 µg of Mla-containing vector and 4.5 µg of each AVR_a-containing vector or EV were transfected. Luciferase activity was measured at 16 h post-transfection, and relative luciferase activity determined by setting the reference samples (Mla+EV) to 1. Differences between samples were assessed by non-parametric Kruskal–Wallis and subsequent Dunn tests for each MLA variant. Calculated P-values were as follows: Mla1: P=0.412; Mla7, P=0.683; and Mla13, P=1.9E-04. Samples marked by identical letters in the plots do not differ significantly (P<0.05) in the Dunn test for the corresponding MLA. n.s=not significant.

Fig. 3.

Amino acid exchanges in the nucleotide-binding site of MLA13 compromise AVR_A13 effector binding. (A, B) Yeast cells were co-transformed with Mla13 wild type (wt) or mutant variants Mla13^D502V (MHD) or Mla13 K207R (P-loop) fused N-terminally to the LexA-binding domain (BD) sequence and AVR_a13 variants lacking SPs fused N-terminally to the B42 activation domain (AD) and 1×HA tag sequence as indicated. (A) Growth of transformants was determined on selective growth medium containing raffinose and galactose as carbon sources but lacking uracil, histidine, and tryptophan (–UHW), and interaction of proteins was determined by leucine reporter activity reflected by growth of yeast on selective medium containing raffinose and galactose as carbon sources, but lacking uracil, histidine, tryptophan, and leucine (–UHWL). Figures shown are representatives of at least three experiments, and pictures were taken 6–8 d after drop-out. (B) Protein levels of BD-MLA13 variants and AD-AVR_A variants corresponding to yeast of (A). Yeast transformants were grown in raffinose- and galactose-containing selective medium lacking uracil, tryptophan, and histidine to OD₆₀₀=1. Cells were harvested, total protein extracted, separated by gel electrophoresis, and western blots were probed with anti-LexA or anti-HA as indicated. (C and D) Nicotiana benthamiana leaves were co-transformed transiently with cDNAs of AVR_a13-V1, AVR_a13-V2, or empty vector (EV) together with constructs encoding either MLA13 or MLA13^D502V (pAM-PAT vector) as indicated and under the control of the 35S promoter sequence at a 2:1 ratio. (C) Cell death was determined 2 d (MLA13 MHD) to 5 d (MLA13) post-transformation and scored from 0 to 3 based on the cell death scale indicated. All values obtained in at least three independent experiments are indicated by dots;| error bars=SD. Differences between samples were assessed by non-parametric Kruskal–Wallis and subsequent Dunn’s tests for each MLA variant. Calculated P-values were as follows: MLA13, P=5E-05; MLA13 MHD, P=0.078. Samples marked by identical letters in the plots did not differ significantly (P<0.05) in the Dunn test for the corresponding MLA. (D) Protein levels corresponding to samples of (C). Leaf tissue was harvested 36 h post-infiltration. Total protein was extracted, separated by gel electrophoresis, and probed by anti-Myc (MLAs) or anti-GFP (AVR_A13-V2) western blotting as indicated. CBB: Coomassie brilliant blue.

Fig. 4.

Amino acid exchanges in the coiled-coil (CC) domain de-regulate MLA13 autoinhibition. (A) Amino acid changes in MLA13 mutant variants. The D2A_E17A and the F99E variants encode changes in the MLA13 CC domain, which spans from amino acid 1 to 160. The K207R, D284A, D502V, and H501G variants encode changes in the nucleotide-binding site (NB, amino acids 161– 549). The S902F_F935I variant affects the leucine-rich repeats (LRRs, amino acids 550–942) which are followed by a short C-terminal amino acid sequence. (B and C) Nicotiana benthamiana leaves were transformed transiently with cDNAs of one of the Mla13 variants as indicated (pGWB517 vector) either with or without AVR_a13-1 lacking SPs and fused C-terminally to an mYFP sequence. All constructs are under the control of the 35S promotor. (B) Cell death was determined 3 d post-transformation; n≥9. (C) Protein stability of the MLA variants fused to 4×Myc corresponding to constructs of (B). Leaf tissue was harvested 2 d post-infiltration. Total protein was extracted, separated by gel electrophoresis, and probed by anti-Myc western blotting as indicated. (D and E) Yeast cells were co-transformed with Mla13 variants fused N-terminally to the LexA-binding domain (BD) sequence and AVR_a13-V2 lacking SPs fused N-terminally to the B42 activation domain (AD) and 1×HA tag sequence as indicated. Growth of transformants was determined on selective growth medium containing raffinose and galactose as carbon sources but lacking uracil, histidine, and tryptophan (–UHW), and interaction of proteins was determined by leucine reporter activity reflected by growth of yeast on selective medium containing raffinose and galactose as carbon sources but lacking uracil, histidine, tryptophan, and leucine (–UHWL). Figures shown are representatives of at least three experiments, and pictures were taken 6–8 d after drop-out. (E) Protein levels of BD-MLA13 variants and AD-AVR_A13-V2 corresponding to yeast of (D). Yeast transformants were grown in raffinose- and galactose-containing selective medium lacking uracil, tryptophan, and histidine to OD₆₀₀=1. Then, cells were harvested, total protein extracted, separated by gel electrophoresis, and western blots were probed with anti-LexA or anti-HA as indicated. CBB: Coomassie brilliant blue. (F) N. benthamiana leaves were co-transformed transiently with cDNAs of AVR_a13-V1, AVR_a13-V2, or empty vector (EV) together with constructs encoding the MLA13 variant as indicated and under the control of the 35S promoter sequence at a 2:1 ratio. Cell death was determined based on the cell death scale indicated. All values obtained in at least two independent experiments are indicated by dots, error bars=SD. Differences between samples were assessed by non-parametric Kruskal–Wallis and subsequent Dunn’s tests for each MLA variant. Calculated P-values were as follows: MLA13, P=9.38E-07; MLA13^D2A_E17A, P=0.77. n.s.=no significant difference.

Fig. 5.

Calcium channel activity is required for Mla13-mediated cell death in barley. (A) Barley protoplasts of lines CI 16155 (cultivar Manchuria Mla13) and CI2330 (Manchuria) were transfected with pUBQ:luciferase (6 µg) and piPKb002 containing AVR_a13-1 cDNA without signal peptide (5 µg) or a piPKb002 empty vector control (5 µg) and recovered in the presence of LaCl₃ at the concentrations indicated. Luciferase activity was determined 16 h post-transfection/addition of LaCl₃ as a proxy for cell death and normalized against the respective EV sample. Error bars=SE. Differences between samples were assessed using non-parametric Kruskal–Wallis and subsequent Dunn’s post-hoc tests. P=6.179e-10. Samples marked by identical letters in the plot did not differ significantly (P<0.05) in Dunn’s test. (B) Protoplasts derived from cultivar Manchuria CI2330 leaves transfected with pZmUBQ:AVR_a13-1-mYFP were harvested 16 h post-transfection/LaCl₃ treatment. Total protein was extracted, separated by gel electrophoresis, and western blots were probed with anti-GFP. CBB: Coomassie brilliant blue.