Phosphatidylinositol monophosphate-binding interface in the oomycete RXLR effector AVR3a is required for its stability in host cells to modulate plant immunity

Abstract

The oomycete pathogen Phytophthora infestans causes potato late blight, one of the most economically damaging plant diseases worldwide. P. infestans produces AVR3a, an essential modular virulence effector with an N-terminal RXLR domain that is required for host-cell entry. In host cells, AVR3a stabilizes and inhibits the function of the E3 ubiquitin ligase CMPG1, a key factor in host immune responses including cell death triggered by the pathogen-derived elicitor protein INF1 elicitin. To elucidate the molecular basis of AVR3a effector function, we determined the structure of Phytophthora capsici AVR3a4, a close homolog of P. infestans AVR3a. Our structural and functional analyses reveal that the effector domain of AVR3a contains a conserved, positively charged patch and that this region, rather than the RXLR domain, is required for binding to phosphatidylinositol monophosphates (PIPs) in vitro. Mutations affecting PIP binding do not abolish AVR3a recognition by the resistance protein R3a but reduce its ability to suppress INF1-triggered cell death in planta. Similarly, stabilization of CMPG1 in planta is diminished by these mutations. The steady-state levels of non-PIP-binding mutant proteins in planta are reduced greatly, although these proteins are stable in vitro. Furthermore, overexpression of a phosphatidylinositol phosphate 5-kinase results in reduction of AVR3a levels in planta. Our results suggest that the PIP-binding ability of the AVR3a effector domain is essential for its accumulation inside host cells to suppress CMPG1-dependent immunity.

Fig. 1.

Structural analysis of AVR3a4 and AVR3a. (A) Multiple sequence alignment of the effector domain of Phytophthora AVR3a and homologs. AVR3a4 and AVR3a11 are from P. capsici, Avr1b is from P. sojae, and AVR3a is from P. infestans. The helical regions and corresponding amino acid positions for AVR3a4 are shown above the alignment and for AVR3a are shown below the alignment. (B and C) Ribbon diagrams of the structures of (B) AVR3a4 and (C) AVR3a. (D) The important residues for R3a recognition (K80, green; I103, cyan) and INF1-induced cell death (Y147, purple) are mapped on the ribbon diagram (Left) and the surface structure (Right) of AVR3a. (E and F) Surface charge distribution of (E) AVR3a4 and (F) AVR3a. Electrostatic potential was calculated with PyMol (DeLano Scientific). Positively and negatively charged surfaces are shown in blue and red, respectively.

Fig. 2.

The positively charged patch of the effector domain of AVR3a is required for binding to PIPs. (A) Protein lipid overlay assay of AVR3a and AVR3a4. One hundred picomoles of various lipids were spotted onto nitrocellulose membranes and incubated overnight with E. coli-expressed AVR3a-GST or AVR3a4-GST proteins at 4 °C. After rigorous washing, the bound proteins were detected using anti–GST-HRP antibodies. PA, phosphatidic acid; PC, phosphatidylcholine; PE, phosphatidylethanolamine; PI, phosphatidylinositol; PI3P, PI-3-phosphate; PI4P, PI-4-phosphate; PI5P, PI-5-phosphate; PI3,4P2, PI-3,4-biphosphate; PI3,5P2, PI-3,5-biphosphate; PI4,5P2, PI-4,5-biphosphate; PI3,4,5P3, PI-3,4,5-triphosphate; PS, phosphatidylserine; S1P, sphingosine-1-phosphate. (B) PIP binding of the RXLR mutant (RLLR/AAAA), the N-terminal region including the RXLR domain (23–59), and the effector domain (60–147) of AVR3a. Serial dilutions (200, 100, 50, 25, 12.5, and 6.25 pmol) of PI3P, PI4P, and PI5P were spotted onto nitrocellulose membranes. The protein lipid overlay assay was performed as in Fig. 2A. (C) Positions of positively charged amino acids Arg81, Lys85, Lys86, and Lys89 on the AVR3a structure. (D) PIP binding of AVR3a variants with mutations in the positively charged surface patch. The protein lipid overlay assay was performed as in Fig. 2A. EM, AVR3a^EM; WT, wild-type AVR3a^KI.

Fig. 3.

A positively charged patch in the effector domain, but not the RXLR domain, of Avr1b is required for PIP binding. (A) Positions of positively charged amino acids Lys75, Lys79, Lys80, and Lys83 in the Avr1b structure. (B) PIP binding of the RXLR mutants Avr1b^RFLR/AAAA and Avr1b^RFLR/QFLR and AVR1b with mutations at the positively charged patch, Avr1b^K79E, and Avr1b^{K79E, K83E}. The lipid overlay assay was performed as in Fig. 2A.

Fig. 4.

PIP binding is not required for R3a-dependent recognition but is required for the suppression of INF1-induced cell death. (A) Percentage of R3a cell death induced by expression of AVR3a derivatives WT (AVR3a^KI), AVR3a^R81E, AVR3a^K85E, AVR3a^K86E, AVR3a^K89E, AVR3a^EM, and AVR3a^Y147del by A. tumefaciens infiltration. Cell death in N. benthamiana leaves was scored 5 d after infiltration. (B) Samples of infiltration sites coexpressing AVR3a derivatives with R3a. (C) Percentage of sites with INF1-induced cell death upon coexpression of INF1 with the AVR3 derivatives WT (AVR3a^KI), AVR3a^R81E, AVR3a^K85E, AVR3a^K86E, AVR3a^K89E, AVR3a^EM, and AVR3a^Y147del. Cell death was scored 5 d after infiltration with INF1. Error bars in A and C indicate SD of the means (n = 3). (D) Symptoms observed at infiltration sites coexpressing AVR3a derivatives with INF1.

Fig. 5.

PIP binding is required for the accumulation of CMPG1 and AVR3a in planta. (A) Immunoblots probed with anti-myc and anti-FLAG antibodies following coexpression of 4×-myc-CMPG1 and FLAG-AVR3a derivatives, respectively, in N. benthamiana. Protein loading is shown by Coomassie blue (CBB) staining. (B) Immunoblots probed with anti-FLAG antibody following expression of FLAG-AVR3a mutants in N. benthamiana (in vivo) and purified His-tagged AVR3a mutant proteins after incubation overnight at room temperature (in vitro). (C) Circular dichroism spectra of the purified His-tagged AVR3a proteins (WT, black; K85E, gray). (D) Interaction of AVR3a derivatives with CMPG1 in the yeast two-hybrid system. The expression of lacZ was monitored to check the interaction. (E) PIP binding of the Lys85 derivatives WT (AVR3a^KI), AVR3a^K85E, AVR3a^K85A, and AVR3a^K85G. The protein lipid overlay assay was performed as in Fig. 2A. (F) Immunoblots probed with anti-FLAG antibody following expression of FLAG-tagged Lys85 derivatives in N. benthamiana. (G) Percentage of sites with INF1-induced cell death upon coexpression of INF1 with Lys85 derivatives. Error bars indicate SD of the means (n = 3).

Fig. 6.

PIP5K1 OX suppresses the accumulation of AVR3a in planta. (A) Immunoblots probed with anti-FLAG antibody, anti–H⁺-ATPase antibody, and anti-UDPase antibody, following expression of FLAG-AVR3a in N. benthamiana. The protein was extracted with or without Nonidet P-40 (NP-40). Protein loading is shown by Coomassie blue (CBB) staining. (B) Immunoblots probed with anti-FLAG antibody, anti–H⁺-ATPase antibody, and anti-UDPase antibody, following expression of FLAG-AVR3a in N. benthamiana leaves overexpressing PIP5K1. (C) Percentage of sites with INF1-induced cell death upon coexpression of INF1 with AVR3a and PIP5K1. Cell death was scored 5 d after infiltration with INF1. Error bars indicate SD of the means (n = 3). (D) Symptoms observed at infiltration sites coexpressing INF1 with AVR3a and PIP5K1. A. tumefaciens strains carrying pGWB12 empty vector (EV1) or pEAQ-HT empty vector (EV2) were infiltrated as controls for AVR3a and PIP5K1, respectively.