Effector-dependent activation and oligomerization of plant NRC class helper NLRs by sensor NLR immune receptors Rpi-amr3 and Rpi-amr1

Abstract

Plant pathogens compromise crop yields. Plants have evolved robust innate immunity that depends in part on intracellular Nucleotide-binding, Leucine rich-Repeat (NLR) immune receptors that activate defense responses upon detection of pathogen-derived effectors. Most "sensor" NLRs that detect effectors require the activity of "helper" NLRs, but how helper NLRs support sensor NLR function is poorly understood. Many Solanaceae NLRs require NRC (NLR-Required for Cell death) class of helper NLRs. We show here that Rpi-amr3, a sensor NLR from Solanum americanum, detects AVRamr3 from the potato late blight pathogen, Phytophthora infestans, and activates oligomerization of helper NLRs NRC2 and NRC4 into high-molecular-weight resistosomes. In contrast, recognition of P. infestans effector AVRamr1 by another sensor NLR Rpi-amr1 induces formation of only the NRC2 resistosome. The activated NRC2 oligomer becomes enriched in membrane fractions. ATP-binding motifs of both Rpi-amr3 and NRC2 are required for NRC2 resistosome formation, but not for the interaction of Rpi-amr3 with its cognate effector. NRC2 resistosome can be activated by Rpi-amr3 upon detection of AVRamr3 homologs from other Phytophthora species. Mechanistic understanding of NRC resistosome formation will underpin engineering crops with durable disease resistance.

Keywords: Rpi-amr; NLR activation; NRC; blue native-PAGE; plant immunity.

Figure 1. Rpi‐amr3 and AVRamr3 form protein complexes unaltered by NRC2^EEE‐Myc

1. Schematic model of NRC2‐dependent resistance by sensor NLRs Rpi‐amr3, Rpi‐amr1 and cognate effectors AVRamr3 and AVRamr1, respectively. Each domain is labeled and represented with a different color. CC, coiled‐coil; NB‐ARC, nucleotide binding domain shared by APAF‐1, R genes, CED‐4; LRR, Leucine‐rich repeat; SP, signal peptide; RXLR, conserved motif found in Phytophthora effectors; WY domains, domains in RXLR effectors with conserved Trp (W) and Tyr (Y) residues; NLS, nuclear localization signal.

2. Blue native‐PAGE loading of protein extracts from nrc2/3/4 knockout N. benthamiana plants after immunoprecipitation with anti‐Flag antibody. Co‐migration of Rpi‐amr3‐HF and AVRamr3‐V5 are indicated (*, red). Same samples were loaded twice on one blue native‐PAGE gel, transferred onto one membrane, and then the membrane was cut into two and immunoblotted separately. GUS‐V5, β‐glucuronidase fused with V5 tag.

3. NRC2^EEE‐Myc does not alter association between Rpi‐amr3 and AVRamr3. Samples of (B) were SDS‐boiled and loaded on SDS–PAGE. Input samples were taken prior to immunoprecipitation to show expression of all proteins.

Data information: Ponceau S staining of rubisco large subunit serve as loading control for panel (C). Molecular weight markers (in kilodaltons, kDa) are shown on the right. Experiments were done with at least three biological replicates with similar results. Representative images are shown. Source data are available online for this figure.

Figure EV1. AVRamr1 is an RXLR effector with WY domains interacting with Rpi‐amr1

1. Predicted protein structure of AVRamr1 indicate 3 WY domains. Protein structure model was predicted using AlphaFold and visualized using ChimeraX software. Confidence level of b factors are indicated in colors.

2. Third WY domain (WY3, blue) of AVRamr1 shows conserved structure of WY domains with four α‐helices as well as the Trp (220W) and Tyr (252Y) residues. The WY3 domain and the WY residues overlap with previously identified WY domain structure of AVR3a11 (PDB ID: 3ZR8) (green).

3. Predicted lDDT plot for AVRamr1 structure prediction.

4. Predicted aligned error plot for AVRamr1 structure prediction.

5. Rpi‐amr1 recognizes AVRamr1 and induces HR. Wild‐type N. benthamiana plants were transiently infiltrated, and leaf samples were imaged at 5 dpi for HR.

6. Rpi‐amr1 interacts with AVRamr1 in planta. Constructs with truncations of luciferase (Nluc or Cluc) were transiently expressed in nrc2/3/4 KO N. benthamiana plants and imaged at 3 dpi.

Data information: Experiments were performed with at least three biological replicates with similar results. Source data are available online for this figure.

Figure EV2. NRC2 activation leads to degradation of Rpi‐amr3 and AVRamr3

1. Protein extracts from nrc2/3/4 knockout N. benthamiana plants expressing AVRamr3‐V5 and/or NRC2^EEE‐Myc were loaded on blue native‐PAGE. Aliquot of the protein extracts that were SDS‐boiled serve as control.

2. NRC2‐Myc co‐expression leads to degradation of Rpi‐amr3 and AVRamr3. Immunoprecipitation with anti‐Flag antibody of protein extracts in nrc2/3/4 knockout N. benthamiana plants. Aliquot of samples were SDS‐boiled and loaded on SDS–PAGE.

3. Immunoprecipitated and eluted samples from (B) were loaded on blue native‐PAGE. Rpi‐amr3 and AVRamr3 complex is indicated (*, red).

Data information: Ponceau S staining serves as loading control for panels (A and B). Molecular markers are shown on the right. Similar results were observed with at least three biological replicates. Source data are available online for this figure.

Figure EV3. Small proportion of Rpi‐amr1 or Rpi‐amr3 can self‐associate

1. A

   Rpi‐amr1 and Rpi‐amr3 can self‐associate weakly in vivo regardless of effector co‐expression. Protein samples were immunoprecipitated with anti‐FLAG antibodies and were blotted for HA‐tagged Rpi‐amr (anti‐HA), and AVRamr (anti‐V5). Both low and high exposures of the blots are shown. Solid line in membrane blotted with anti‐HA indicate gaps between the samples.

2. B, C

   Rpi‐amr1 and Rpi‐amr3 are mostly monomers in vivo. Immunoprecipitated samples of (A) were loaded on blue native‐PAGE. High exposure versions of (B) are shown in (C).

3. D

   Rpi‐amr1 and Rpi‐amr3 can self‐associate weakly in vivo regardless of effector co‐expression and form slower‐migrating protein complexes. Protein samples were immunoprecipitated with anti‐FLAG antibodies, loaded onto blue native‐PAGE, and were blotted for HA‐tagged Rpi‐amr (anti‐HA).

4. E

   AVRamr1 and AVRamr3 form a protein complex with Rpi‐amr1 and Rpi‐amr3, respectively. Protein extracts from N. benthamiana nrc2/3/4 knockout plants were immunoprecipitated with anti‐Flag antibodies and loaded on blue native‐PAGE.

Data information: Ponceau S staining serve as loading control in panel (A). Molecular weight markers are shown on the right. Experiments were done with at least three biological replicates with similar results. Source data are available online for this figure.

Figure 2. NRC2^EEE oligomerizes upon effector detection by Rpi‐amr3 or Rpi‐amr1

1. NRC2^EEE‐Myc is oligomerized upon effector‐dependent activation of Rpi‐amr3. Protein lysates from Fig 1B and C were loaded on blue native‐PAGE. SDS‐boiled protein lysate samples serve as control for actual size of NRC2^EEE‐Myc. Oligomerized NRC2^EEE‐Myc is indicated (*, red).

2. NRC2^EEE‐Myc oligomerizes upon effector‐dependent activation of Rpi‐amr1. Protein lysates from nrc2/3/4 knockout N. benthamiana plants were loaded on blue native‐PAGE. Oligomerized NRC2^EEE‐Myc is indicated (*, red).

3. Samples from (B) were SDS‐boiled and loaded on SDS–PAGE. Protein accumulation of Rpi‐amr1‐Flag, NRC2^EEE‐Myc, AVRamr1‐V5 and AVRamr3‐V5 are shown.

4. Semi‐log plots of NRC2^EEE‐Myc proteins loaded on blue native‐PAGE gels. At least three biological replicates were used for analysis of each sample and data points are plotted on the boxplot (blue, inactive; red, active). Wilcoxon test was conducted in a pairwise manner, and statistical significance is indicated (*, P < 0.05; **, P < 0.01, ***, P < 0.001).

Data information: Ponceau S staining serve as loading control for panels (A and C). Molecular weight markers are shown on the right. Experiments were done with at least three biological replicates with similar results. Standard curves of protein migration derived from semi‐log plots of molecular weight markers on individual gels can be found in Dataset EV2. Source data are available online for this figure.

Figure 3. Rpi‐amr3 and Rpi‐amr1 are not present in the oligomerized NRC2 protein complex

1. Experimental design for 2D‐PAGE (blue native‐PAGE/SDS–PAGE). Agro‐infiltrated nrc2/3/4 knockout N. benthamiana plants were collected at 3 dpi for protein extraction. Protein extracts were loaded on blue native‐PAGE (1D) to separate high molecular weight (HMW) protein complexes (hypothesized as a pentamer) from low‐molecular‐weight (LMW) protein complexes. Subsequently, blue native‐PAGE gels were loaded on SDS–PAGE (2D) for separation of protein complexes into individual proteins.

2. NRC2^EEE and Rpi‐amr3 migration in the absence of effector. N. benthamiana nrc2/3/4 knockout plants were transiently infiltrated with Rpi‐amr3‐HF, NRC2^EEE‐Myc and GUS‐V5 followed by 2D‐PAGE. NRC2^EEE‐Myc proteins (*, red) and Rpi‐amr3‐HF proteins (*, blue) is indicated.

3. NRC2^EEE and Rpi‐amr3 migration in the presence of effector. N. benthamiana nrc2/3/4 knockout plants were transiently infiltrated with Rpi‐amr3‐HF, NRC2^EEE‐Myc and AVRamr3‐V5 followed by 2D‐PAGE. NRC2^EEE‐Myc proteins (*, red) and Rpi‐amr3‐HF protein (*, blue) is indicated.

4. Re‐localization of NRC2^EEE‐Myc upon effector‐dependent activation of Rpi‐amr1. Lysates of nrc2/3/4 knockout plants transiently expressing Rpi‐amr1‐Flag, NRC2^EEE‐Myc with AVRamr3‐V5 or AVRamr1‐V5 were fractionated into total (T), soluble (S), and pellet (P) fractions.

5. Re‐localization of NRC2^EEE‐Myc upon effector‐dependent activation of Rpi‐amr3. Lysates of nrc2/3/4 knockout plants transiently expressing Rpi‐amr3‐HF, NRC2^EEE‐Myc with AVRamr3‐V5 or AVRamr1‐V5 were fractionated into total (T), soluble (S), and pellet (P) fractions.

Data information: Protein lysates boiled in SDS were loaded on the same gel as control of protein size (SDS input) for panels (B and C). Molecular weight markers for blue native‐PAGE are labeled on top for panels (B and C), and SDS–PAGE markers are labeled on the right. Ponceau S staining of rubisco large subunit serves as control for panels (B and C). MPK6, which localizes mainly to cytosol, and H⁺‐ATPase, which localize mainly to membrane fractions serve as fractionation control in panels (D and E). Experiments were repeated with three biological replicates with similar results. Source data are available online for this figure.

Figure 4. P‐loop of Rpi‐amr3 and NRC2^EEE is required for NRC2^EEE oligomerization

1. P‐loop of Rpi‐amr3 is required for AVRamr3‐dependent HR in N. benthamiana. Representative leaf phenotype of HR (hypersensitive response) in wild‐type N. benthamiana is shown. The number of leaves tested and occurrences of HR are indicated in parentheses.

2. P‐loop of NRC2 is required for AVRamr3‐dependent HR in N. benthamiana nrc2/3/4 knockout plants. Representative leaf phenotype of HR is shown. The number of leaves tested and occurrences of HR are indicated in parentheses.

3. NRC2^EEE‐Myc requires functional P‐loop of both Rpi‐amr3 and NRC2 for oligomerization. Protein lysates from N. benthamiana nrc2/3/4 knockout plants expressing wild‐type or P‐loop mutants of Rpi‐amr3 or NRC2^EEE were loaded for blue native‐PAGE. Membranes were immunoblotted with anti‐Myc.

Data information: At least three biological replicates with multiple technical replicates were tested for panels (A and B). SDS‐boiled samples of the protein lysates were loaded onto SDS–PAGE for control in panel (C). Ponceau S staining serves as loading control. Molecular weight markers are shown on the right. Experiments were done with at least three biological replicates with similar results for panel (C). Source data are available online for this figure.

Figure 5. NRC4^AAA‐Myc is not oligomerized by activation of Rpi‐amr1

1. Cartoon depicting genetic requirement of Rpi‐amr1 and Rpi‐amr3‐dependent activation of NRCs. Arrows are shown to illustrate which NRC helper NLR supports each sensor NLR.

2. Protein lysates expressing NRC4^AAA‐Myc with Rpi‐amr3‐HF or Rpi‐amr1‐Flag in the presence and absence of cognate effector were loaded on blue native‐PAGE. Proteins were transiently expressed in N. benthamiana nrc2/3/4 knockout plants. Non‐specific bands are indicated with *.

Data information: At least three biological replicates with technical replicates were tested. SDS‐boiled samples of the protein lysates were loaded onto SDS–PAGE for control. Ponceau S staining serves as loading control. Molecular weight markers are shown on the right. Experiments were done with more than three biological replicates with similar results for panel (B). Source data are available online for this figure.

Figure EV4. Rpi‐amr3 P‐loop mutant interacts with and forms stable complex with AVRamr3

1. P‐loop of Rpi‐amr3 is dispensable for association with AVRamr3. Protein extracts from N. benthamiana nrc2/3/4 knockout plants were immunoprecipitated with anti‐Flag antibody and blue native‐PAGE was performed. Membranes were immunoblotted with anti‐Flag.

2. AVRamr3 associates with both wild‐type Rpi‐amr3 and Rpi‐amr3^P‐loop. Protein extracts from N. benthamiana nrc2/3/4 knockout plants were immunoprecipitated with anti‐Flag antibody and blue native‐PAGE was performed. Membranes were immunoblotted with anti‐V5.

Data information: Ponceau S staining serves as loading control for panels (A and B). Molecular markers are shown on the right. Similar results were observed in at least three biological replicates. Source data are available online for this figure.

Figure 6. Recognized AVRamr3 alleles from different Phytophthora species can trigger oligomerization of NRC2^EEE

1. Cartoon depicting different alleles of AVRamr3 from different Phytophthora species, P. infestans, P. capsici, and P. parasitica. Response of the corresponding alleles by Rpi‐amr3, and thus occurrence of HR is indicated as + (recognition) or − (no recognition).

2. NRC2^EEE‐Myc oligomerizes in response‐dependent manner. Protein lysates from N. benthamiana nrc2/3/4 knockout transiently expressing NRC2^EEE‐Myc, Rpi‐amr3‐HF and AVRamr3 alleles from P. infestans, P. capsici, and P. parasitica were loaded on blue native‐PAGE and blotted for NRC2^EEE (anti‐Myc).

3. Recognition is correlated with interaction and protein complex formation of AVRamr3 with Rpi‐amr3. Protein extracts from Fig 5B were immunoprecipitated with anti‐Flag antibody, separated on blue native‐PAGE, and AVRamr3‐GFP proteins of P. infestans, P. capsici, and P. parasitica were visualized.

Data information: SDS‐boiled input and IP eluates were loaded onto SDS–PAGE as control and blotted for Rpi‐amr3 (anti‐Flag), and AVRamr3 alleles (anti‐GFP). Ponceau S staining serves as loading control. Molecular markers are indicated on the right. Experiments were done with more than three biological replicates with similar results. Source data are available online for this figure.

Figure EV5. Non‐recognized alleles of AVRamr3 do not trigger NRC2 oligomerization and do not interact with Rpi‐amr3

1. Schematic depiction of AVRamr3 from Phytophthora infestans, an unrecognized, truncated version of AVRamr3 (AVRamr3 T9), and AVRamr3 from P. capsici. Response of the corresponding alleles by Rpi‐amr3, and thus occurrence of HR is indicated as + (recognition) or − (no recognition).

2. NRC2^EEE‐Myc oligomerizes in recognition‐dependent manner. Protein extracts from N. benthamiana nrc2/3/4 knockout were loaded on blue native‐PAGE, and blotted for Myc.

3. Recognition is correlated with interaction and protein complex formation of AVRamr3 with Rpi‐amr3. Protein extracts from (B) were immunoprecipitated with anti‐Flag antibody and were loaded on blue native‐PAGE and blotted for V5.

4. Recognition is correlated with interaction and protein complex formation of Rpi‐amr3 with AVRamr3. Protein extracts from (B) were immunoprecipitated with anti‐Flag antibody and were loaded on blue native‐PAGE and blotted for Flag.

Data information: SDS‐boiled input and IP eluates were loaded onto SDS–PAGE as control. Ponceau S staining serves as loading control in panels (B and C). Molecular markers are indicated on the right. Experiments were done with at least three biological replicates with similar results. Ponceau S loading for (D) was same as with (C). Source data are available online for this figure.

Figure 7. Model for activation of helper NLR NRCs upon recognition of AVRamr by sensor NLRs

In the pre‐activated state, the sensor NLRs, such as Rpi‐amr1 and Rpi‐amr3, reside in the cells mainly as monomers. Conceivably, additional proteins, such as chaperones, could be bound to these monomeric NLRs. Interaction with AVRamr converts Rpi‐amr into an activated form (indicated by an *) that can interact with and activate NRC2 into an activated form. These conversion of NRC2 by activated Rpi‐amr may lead to oligomerization, or alternatively, additional NRC2 protomers may be subsequently oligomerized by Rpi‐amr‐activated NRC2 protomer.