Editor's evaluation: Effector target-guided engineering of an integrated domain expands the disease resistance profile of a rice NLR immune receptor

Abstract

Full text Figures and data Side by side Abstract Editor's evaluation Introduction Results Discussion Materials and methods Data availability References Decision letter Author response Article and author information Metrics Abstract A subset of plant intracellular NLR immune receptors detect effector proteins, secreted by phytopathogens to promote infection, through unconventional integrated domains which resemble the effector’s host targets. Direct binding of effectors to these integrated domains activates plant defenses. The rice NLR receptor Pik-1 binds the Magnaporthe oryzae effector AVR-Pik through an integrated heavy metal-associated (HMA) domain. However, the stealthy alleles AVR-PikC and AVR-PikF avoid interaction with Pik-HMA and evade host defenses. Here, we exploited knowledge of the biochemical interactions between AVR-Pik and its host target, OsHIPP19, to engineer novel Pik-1 variants that respond to AVR-PikC/F. First, we exchanged the HMA domain of Pikp-1 for OsHIPP19-HMA, demonstrating that effector targets can be incorporated into NLR receptors to provide novel recognition profiles. Second, we used the structure of OsHIPP19-HMA to guide the mutagenesis of Pikp-HMA to expand its recognition profile. We demonstrate that the extended recognition profiles of engineered Pikp-1 variants correlate with effector binding in planta and in vitro, and with the gain of new contacts across the effector/HMA interface. Crucially, transgenic rice producing the engineered Pikp-1 variants was resistant to blast fungus isolates carrying AVR-PikC or AVR-PikF. These results demonstrate that effector target-guided engineering of NLR receptors can provide new-to-nature disease resistance in crops. Editor's evaluation Engineering NLR proteins to improve disease resistance in crop plants is a major goal of the field. This study applies knowledge from structural and evolutionary studies of the rice NLR protein Pik-1 and cognate effector protein AVR-Pik to engineering of new disease resistance genes. The authors nicely demonstrate that it is indeed possible to engineer resistance proteins with broad recognition specificity for the rice blast fungus. The work is of interest to colleagues in synthetic biology, protein engineering and plant-pathogen interactions. https://doi.org/10.7554/eLife.81123.sa0 Decision letter Reviews on Sciety eLife's review process Introduction Intracellular nucleotide-binding and leucine-rich repeat (NLR) domain-containing immune receptors are essential components of the plant innate immune system (Jones and Dangl, 2006; Jones et al., 2016). These receptors detect effector proteins which are delivered into host cells by invading pathogens and pests to promote virulence. NLR receptors have a modular domain architecture typically consisting of an N-terminal coiled-coil (CC or CCR) or Toll/Interleukin-1 receptor (TIR) domain, a central nucleotide-binding (NB-ARC) domain, and a C-terminal leucine-rich repeat (LRR) domain. In addition, some NLRs contain non-canonical domains which are integrated into the protein architecture either at the N- or C-termini or between canonical domains (Sarris et al., 2016; Cesari et al., 2014; Kroj et al., 2016; Bailey et al., 2018). These integrated domains (IDs) resemble effector virulence targets and either directly bind or are modified by effector proteins to activate NLR-mediated immune signaling (Cesari et al., 2014; Oikawa et al., 2020; Maidment et al., 2021; Maqbool et al., 2015; Sarris et al., 2015; Mukhi et al., 2021). Many of the characterized resistance (R) genes used to confer disease resistance in crop breeding programs encode NLR proteins (Kourelis and van der Hoorn, 2018). However, NLR-mediated resistance can be overcome through silencing or deletion of effectors in pathogen genomes, the gain of new effectors or effector functions, or mutation to evade NLR activation (Raffaele et al., 2010; Yoshida et al., 2016). Engineering NLRs to detect currently unrecognized effector proteins would provide new opportunities to control plant pathogens. Early attempts to engineer NLRs focused on random mutagenesis followed by gain-of-function screening, with some success in both expanding recognition profiles to new effector variants and increasing the sensitivity of the receptor (Giannakopoulou et al., 2015; Segretin et al., 2014; Farnham and Baulcombe, 2006; Harris et al., 2013). More recently, modification of the effector target PBS1, which is guarded by the NLR protein RPS5 led to successful engineering of novel recognition by this system (Kim et al., 2016; Pottinger et al., 2020; Pottinger and Innes, 2020; Helm et al., 2019). RPS5 is activated by cleavage of the Arabidopsis thaliana protein kinase PBS1 by the Pseudomonas syringae effector AvrPphB. By varying the PBS1 cleavage site, the RPS5/PBS1 system has been engineered to recognize proteases from different pathogens (Pottinger et al., 2020; Helm et al., 2019). Using a different strategy, a protein domain targeted for degradation by the phytoplasma effector SAP05 was fused to the C-terminus of the TIR-NLR RRS1-R. RRS1-R represses the immune cell death-triggering activity of a second TIR-NLR, RPS4. While transient co-expression of the engineered RRS1-R, RPS4, and the phytoplasma effector SAP05 led to cell death in N. tabacum, transgenic A. thaliana plants were not resistant to phytoplasma carrying the SAP05 effector (Wang et al., 2021). Integrated domains can facilitate the recognition of structure- and sequence-diverse effectors which target similar host proteins. This is exemplified by the TIR-NLR pair RRS1 and RPS4, which mediate recognition of the structurally distinct effectors PopP2 from Ralstonia solanacearum and AvrRps4 from Pseudomonas syringae pv. pisi (Sarris et al., 2015; Mukhi et al., 2021; Deslandes et al., 2003; Saucet et al., 2015; Huh et al., 2017; Le Roux et al., 2015). Furthermore, the potential to replace naturally occurring integrated domains with nanobodies of defined specificity to confer disease resistance has recently been demonstrated (Kourelis et al., 2023). Modification of existing integrated domains, or the incorporation of entirely new protein domains into an NLR structure could deliver new recognition specificities and extend the toolbox of resistance genes available to combat crop pathogens and pests. The paired rice CC-NLR proteins Pik-1 and Pik-2 cooperatively activate plant defense in response to the blast pathogen effector AVR-Pik (Ashikawa et al., 2008; Kanzaki et al., 2012; Zdrzałek et al., 2020). The sensor NLR Pik-1 contains an integrated HMA domain between the CC and NB-ARC domains (Maqbool et al., 2015, Figure 1—figure supplement 1a). Direct binding of AVR-Pik to the HMA domain is required to activate Pik-mediated immunity (Maqbool et al., 2015; De la Concepcion et al., 2018; De la Concepcion et al., 2021). Multiple Pik alleles have been described in different rice cultivars, with most amino acid polymorphisms located within the integrated HMA domain of Pik-1. Five Pik alleles (Pikp, Pikm, Pikh, Piks, and Pik*) have been functionally characterized for their response to blast isolates carrying different AVR-Pik variants (Maqbool et al., 2015; Kanzaki et al., 2012; De la Concepcion et al., 2018; De la Concepcion et al., 2021, Figure 1—figure supplement 1c). To date, six AVR-Pik variants (A-F) have been described, which differ in five amino acid positions at the HMA-binding interface (Kanzaki et al., 2012; Yoshida et al., 2009; Longya et al., 2019, Figure 1—figure supplement 1b). These polymorphisms influence the binding of the effector to the integrated HMA domain of Pik-1 (Maqbool et al., 2015; De la Concepcion et al., 2018; Longya et al., 2019). Interestingly, the Asp67 and Lys78 polymorphisms of AVR-PikC and AVR-PikF, respectively, disrupt interactions between the effector and all tested integrated Pik-HMA domains (Maidment et al., 2021; De la Concepcion et al., 2021; Longya et al., 2019). To date, none of the characterized Pik alleles can confer disease resistance to blast isolates carrying AVR-PikC or AVR-PikF (Maqbool et al., 2015; Kanzaki et al., 2012; De la Concepcion et al., 2018; De la Concepcion et al., 2021). The molecular basis of interaction between AVR-Pik effectors and the integrated HMA domains of Pikp-1, Pikm-1, and Pikh-1 has been well explored (Maqbool et al., 2015; De la Concepcion et al., 2018; De la Concepcion et al., 2021). Pikp-1 is only able to recognize the AVR-PikD variant, however, the introduction of two amino acid changes (Asn261Lys and Lys262Glu) extends recognition to AVR-PikE and AVR-PikA, phenocopying the recognition profile of Pikm-1 and Pikh-1 (De la Concepcion et al., 2019). The NLR pair RGA5 and RGA4 detect the blast pathogen effectors AVR-Pia and AVR1-CO39, with activation requiring binding of the effector to an integrated HMA domain at the C-terminus of RGA5 (Okuyama et al., 2011; Cesari et al., 2013; Ortiz et al., 2017). Crystal structures of the RGA5-HMA/AVR1-CO39 and Pik-HMA/AVR-Pik complexes were used to engineer the RGA5-HMA domain to bind AVR-PikD in addition to its cognate effectors AVR-Pia and AVR1-CO39 and deliver cell death in Nicotiana benthamiana, but not disease resistance in transgenic rice (Cesari et al., 2022). More recently, RGA5 has been engineered to bind the non-cognate effector AVR-Pib. Transgenic rice carrying the engineered RGA5 variant was resistant to AVR-Pib-expressing M. oryzae strains, with resistance comparable to that displayed by the (untransformed) rice cultivar K14 which carries the Pib CC-NLR resistance gene (Liu et al., 2021). These studies demonstrate the potential for engineering integrated domains to alter the recognition profile of the NLR protein, however, engineering new-to-nature effector recognition is yet to be reported. The AVR-Pik effector targets members of the rice heavy metal-associated isoprenylated plant protein (HIPP) and heavy metal-associated plant protein (HPP) families through direct interaction with their HMA domain, supporting the hypothesis that NLR integrated domains are likely derived from host proteins (Oikawa et al., 2020; Maidment et al., 2021). In a previous study, we showed that all AVR-Pik effector variants bind to the HMA domain of OsHIPP19 with high affinity and elucidated the structural basis of this interaction by determining the crystal structure of a OsHIPP19-HMA/AVR-PikF complex (Maidment et al., 2021). This shows that effector variants that are not bound by Pik-HMA domains, and escape immune recognition, retain a tight binding for HMA domains of their putative host targets. Here, we leverage our understanding of the interaction between OsHIPP19 and AVR-Pik to engineer the integrated HMA domain of Pik-1 to expand recognition to the stealthy AVR-PikC and AVR-PikF variants, enabling new-to-nature disease resistance profiles in an NLR. We use two parallel strategies to engineer recognition. First, we demonstrate that exchanging the HMA domain of Pikp-1 for that of OsHIPP19 (including additional amino acid substitutions to prevent autoactivity), gives a chimeric Pik-1 which binds AVR-Pik effectors and triggers AVR-PikC- and AVR-PikF-dependent cell death in N. benthamiana. Second, guided by the structure of the OsHIPP19-HMA/AVR-PikF complex, we use targeted mutagenesis of Pikp-1 to give a second engineered Pik-1 receptor capable of binding to AVR-PikC and AVR-PikF and triggering cell death in N. benthamiana. Finally, we show that transgenic rice expressing either of these engineered Pik-1 proteins is resistant to blast pathogen strains carrying AVR-PikC or AVR-PikF, while rice expressing wild-type Pikp is susceptible. This work highlights how a biochemical and structural understanding of the interaction between a pathogen effector and its host target can guide the rational engineering of NLR proteins with the novel, and new-to-nature, disease resistance profiles. Results A Pikp-1OsHIPP19 chimera extends binding and response to previously unrecognized AVR-Pik variants Previously, we reported that all AVR-Pik variants, including AVR-PikC and AVR-PikF, bind to the HMA domain of OsHIPP19 with high affinity (Maidment et al., 2021). The HMA domains of Pikp-1 and OsHIPP19 share 51% amino acid identity and are structurally similar; the RMSD (as calculated in Coot using secondary structure matching) between Pikp-HMA (PDB 6G10) and OsHIPP19-HMA (PDB 7B1I) is 0.97 Å across 71 amino acids. We hypothesized that exchanging the HMA domain of Pikp-1 for the HMA domain of OsHIPP19 would result in an NLR capable of binding and responding to AVR-PikC and AVR-PikF. For this exchange, amino acids 188–263 (inclusive) of Pikp-1 were replaced with amino acids 2–77 of OsHIPP19 to give the chimeric NLR protein Pikp-1OsHIPP19 (Figure 1—figure supplement 2a). To test whether Pikp-1OsHIPP19 could associate with AVR-Pik effector variants in planta, we performed co-immunoprecipitation experiments in N. benthamiana. Each of the myc-AVR-Pik variants (an N-terminal myc tag was used for effectors in all experiments in N. benthamiana) was transiently co-expressed with Pikp-1OsHIPP19-HF (C-terminal 6xHis/3xFLAG tag, used for all Pikp-1 constructs expressed in N. benthamiana) by agroinfiltration. Pikp-2 was not included to prevent the onset of cell death, which reduces protein levels in the plant cell extract, and previous work has shown that AVR-Pik can associate with Pik-1 in the absence of Pik-2 (De la Concepcion et al., 2018). Following immunoprecipitation with anti-FLAG beads to enrich for Pikp-1OsHIPP19, western blot analysis showed that all AVR-Pik variants, except AVR-PikF, co-precipitated with Pikp-1OsHIPP19 (Figure 1—figure supplement 3). As a control, and consistent with previous studies, AVR-PikD associated with Pikp-1, while AVR-PikC did not. We then transiently co-expressed epitope-tagged Pik-1, Pikp-2, and AVR-Pik in N. benthamiana using cell death as a proxy for immune activation (Maqbool et al., 2015; De la Concepcion et al., 2018; De la Concepcion et al., 2021; Longya et al., 2019). We found that when co-expressed with Pikp-2-HA (C-terminal hemagglutinin tag, used for all Pikp-2 constructs expressed in N. benthamiana), Pikp-1OsHIPP19 is autoactive and triggers spontaneous cell death in the absence of the effector (Figure 1a and b, Figure 1—figure supplement 4a). This autoactivity requires an intact P-loop and MHD motif in Pikp-2, as cell death is abolished when Pikp-1OsHIPP19 is transiently co-expressed with either Pikp-2K217R or Pikp-2D559V (Figure 1—figure supplement 5a, Figure 1—figure supplement 5b, Figure 1—figure supplement 6). Cell death was reduced, but not abolished, when Pikp-1OsHIPP19 with a Lys296Arg mutation in the P-loop motif (Pikp-1OsHIPP19_K296R) was transiently co-expressed with Pikp-2 (Figure 1—figure supplement 5a, Figure 1—figure supplement 5b, Figure 1—figure supplement 6). Western blot analysis indicated that all fusion proteins were produced (Figure 1—figure supplement 5c, Figure 1—figure supplement 7a). Figure 1 with 11 supplements see all Download asset Open asset The Pikp-1OsHIPP19-mbl7 chimera expands binding and response to previously unrecognized AVR-Pik effector variants. (A) Representative leaf image showing the Pikp-1OsHIPP19 chimera is autoactive in N. benthamiana in a Pikp-2 dependent manner. Nucleotide-binding and leucine-rich repeat (NLR)-mediated responses appear as autofluorescence imaged under UV light. Pikp-mediated response to AVR-PikD (positive control, top left), Pikp-1OsHIPP19/Pikp-2 without effector shows autoactivity (top right), Pikp-1OsHIPP19/Pikp-2 response remains in the presence of AVR-PikD (middle right). Other leaf positions represent relevant negative controls. (B) Pikp-mediated response scoring represented as dot plots to summarize 30 repeats of the experiment shown in (A) across three independent experiments (Materials and methods, Figure 1—figure supplement 4a). Fluorescence intensity is scored as previously described (Maqbool et al., 2015; De la Concepcion et al., 2018). (C) The Pikp-1OsHIPP19-mbl7 chimera does not display autoactive cell death in N. benthamiana (bottom right), as seen for Pikp-1OsHIPP19 (middle left), but retains response to AVR-PikD and expands Pikp-mediated response to AVR-PikC (middle right). (D) Pikp-mediated response scoring represented as dot plots to summarize 60 repeats of the experiment shown in (C) across three independent experiments (Materials and methods, Figure 1—figure supplement 4b). (E) As for (C), but showing the expanded Pikp-mediated response to AVR-PikF. (F) As for (D) but for 60 repeats of the experiment in (E) across three independent experiments (Materials and methods, Figure 1—figure supplement 4c). (G) Western blots following co-immunoprecipitation reveal that the Pikp-1OsHIPP19-mbl7 chimera associates with all AVR-Pik effector variants in N. benthamiana. Plant cell lysates were probed for the expression of Pikp-1/Pikp-1OsHIPP19-mbl7 and AVR-Pik effector variants using anti-FLAG and anti-Myc antiserum, respectively. Total protein extracts were visualized by Ponceau Staining. Figure 1—source data 1 Unedited and uncropped blot for panel G, IP Pik-1 α-FLAG, Pik-1 α-FLAG, with relevant bands labeled. https://cdn.elifesciences.org/articles/81123/elife-81123-fig1-data1-v1.tiff Download elife-81123-fig1-data1-v1.tiff Figure 1—source data 2 Unedited and uncropped blot for panel G, IP Pik-1 α-FLAG, Pik-1 α-FLAG. https://cdn.elifesciences.org/articles/81123/elife-81123-fig1-data2-v1.tiff Download elife-81123-fig1-data2-v1.tiff Figure 1—source data 3 Unedited and uncropped blot for panel G, IP Pik-1 α-FLAG, AVRPik α-myc, with relevant bands labeled. https://cdn.elifesciences.org/articles/81123/elife-81123-fig1-data3-v1.tiff Download elife-81123-fig1-data3-v1.tiff Figure 1—source data 4 Unedited and uncropped blot for panel G, IP Pik-1 α-FLAG, AVRPik α-myc. https://cdn.elifesciences.org/articles/81123/elife-81123-fig1-data4-v1.tiff Download elife-81123-fig1-data4-v1.tiff Figure 1—source data 5 Unedited and uncropped blot for panel G, input, Pik-1 α-FLAG, with relevant bands labeled. https://cdn.elifesciences.org/articles/81123/elife-81123-fig1-data5-v1.tiff Download elife-81123-fig1-data5-v1.tiff Figure 1—source data 6 Unedited and uncropped blot for panel G, input, Pik-1 α-FLAG. https://cdn.elifesciences.org/articles/81123/elife-81123-fig1-data6-v1.tiff Download elife-81123-fig1-data6-v1.tiff Figure 1—source data 7 Unedited and uncropped blot for panel G, input, AVRPik α-myc, with relevant bands labeled. https://cdn.elifesciences.org/articles/81123/elife-81123-fig1-data7-v1.tiff Download elife-81123-fig1-data7-v1.tiff Figure 1—source data 8 Unedited and uncropped blot for panel G, input, AVRPik α-myc. https://cdn.elifesciences.org/articles/81123/elife-81123-fig1-data8-v1.tiff Download elife-81123-fig1-data8-v1.tiff Figure 1—source data 9 Unedited and uncropped blot for panel G, Ponceau stain, with relevant bands labeled. https://cdn.elifesciences.org/articles/81123/elife-81123-fig1-data9-v1.tiff Download elife-81123-fig1-data9-v1.tiff Figure 1—source data 10 Unedited and uncropped blot for panel G, Ponceau stain. https://cdn.elifesciences.org/articles/81123/elife-81123-fig1-data10-v1.tiff Download elife-81123-fig1-data10-v1.tiff Figure 1—source data 11 Cell death scores used for dot plots in panel B. https://cdn.elifesciences.org/articles/81123/elife-81123-fig1-data11-v1.zip Download elife-81123-fig1-data11-v1.zip Figure 1—source data 12 Cell death scores used for dot plots in panel D. https://cdn.elifesciences.org/articles/81123/elife-81123-fig1-data12-v1.zip Download elife-81123-fig1-data12-v1.zip Figure 1—source data 13 Cell death scores used for dot plots in panel F. https://cdn.elifesciences.org/articles/81123/elife-81123-fig1-data13-v1.zip Download elife-81123-fig1-data13-v1.zip A previous study showed that autoactivity following HMA domain exchange could be abolished by reverting the degenerate metal-binding motif of the HMA domain (‘MxCxxC’) to the corresponding amino acids in Pikp-1 (Białas et al., 2021). Based on this observation, we exchanged seven amino acids (encompassing the entire MxCxxC motif) in the β1-α1 loop of the Pikp-1OsHIPP19 chimera for the corresponding amino acids in Pikp-1 (Figure 1—figure supplement 2b, Figure 1—figure supplement 8). The resulting chimera, Pikp-1OsHIPP19-mbl7 hereafter (mbl7 refers to seven amino acids in the ‘metal-binding loop’), was not autoactive, and did not trigger spontaneous cell death in the absence of the effector. Crucially, Pikp-1OsHIPP19-mbl7 retained the ability to trigger cell death in N. benthamiana when co-expressed with Pikp-2 and AVR-PikD. Further, Pikp-1OsHIPP19-mbl7 also triggered cell death in N. benthamiana when co-expressed with Pikp-2 and AVR-PikC or AVR-PikF (Figure 1c–f, Figure 1—figure supplement 4b–c, Figure 1—figure supplement 9). Western blot analysis showed that all proteins for these cell death assays were produced in leaf tissue (Figure 1—figure supplement 7b, Figure 1—figure supplement 7c). We also showed that Pikp-1OsHIPP19-mbl7 triggered cell death in N. benthamiana when co-expressed with Pikp-2 and AVR-PikA, AVR-PikE (Figure 1—figure supplement 10, Figure 1—figure supplement 11). Finally, we confirmed that Pikp-1OsHIPP19-mbl7 retains binding to AVR-Pik variants by co-immunoprecipitation in N. benthamiana, including AVR-PikF (Figure 1g). Structure-guided mutagenesis of Pikp-1 extends response to previously unrecognized AVR-Pik variants Alongside the HMA-domain exchange strategy, we also used a structure-guided approach to target point mutations in Pikp-HMA that could extend the effector recognition profile of Pikp without triggering autoimmunity. The interaction surfaces between integrated Pik-HMA domains and AVR-Pik effectors are well-characterized, with crystal structures revealing three predominant interfaces (termed 1–3) between the proteins (Maqbool et al., 2015; De la Concepcion et al., 2018; De la Concepcion et al., 2021; De la Concepcion et al., 2019). These interfaces are also observed in the structure of the HMA domain of OsHIPP19 in complex with AVR-PikF (PDB accession code 7B1I Maidment et al., 2021). The Asp67 and Lys78 polymorphisms that distinguish AVR-PikC and AVR-PikF from AVR-PikE and AVR-PikA, respectively, are located at interface 2. In the crystal structure of Pikh-HMA/AVR-PikC, the side chain of AVR-PikCAsp67 extends towards a loop in the HMA domain containing Pikh-HMAAsp224. This loop is shifted away from the effector, likely due to steric clash and/or repulsion by the two Asp sidechains, and intermolecular hydrogen bonds between Pikh-HMAAsp224 and AVR-PikCArg64 are disrupted. We hypothesized that compensatory mutations at interface 2 could mitigate against the disruption caused by AVR-PikCAsp67. Therefore, we introduced Asp224Ala and Asp224Lys mutations in the Pikp-1NK-KE background (De la Concepcion et al., 2019) and tested these constructs in cell death assays in N. benthamiana. Neither mutation extended the Pikp-1NK-KE-mediated cell death response to AVR-PikC, and both mutations reduced the extent of the response to AVR-PikD (Figure 2—figure supplement 1, Figure 2—figure supplement 2). Pikh-1 and Pikp-1NK-KE differ from Pikp-1 by one and two amino acids, respectively, at interface 3. These amino acid differences are sufficient to extend binding and cell death response to AVR-PikE and AVR-PikA, even though the residues that distinguish these variants from AVR-PikD are located at interface 2. We, therefore, predicted that we could engineer a modified Pik-1 that interacts with AVR-PikC/AVR-PikF by mutating other interfaces in the HMA domain to compensate for disruption at the site of the polymorphic residue. The crystal structure of the OsHIPP19-HMA/AVR-PikF complex revealed additional hydrogen bond interactions at interface 3 relative to integrated HMAs in the complex with AVR-Pik variants (Maidment et al., 2021). The side chain of OsHIPP19Glu72 was particularly striking. The corresponding residue in all described Pik-1 HMA domains is serine, and while the hydroxyl group of the serine side chain only forms an intramolecular hydrogen bond within the HMA domain, the bulkier OsHIPP19Glu72 side chain extends across the interface and forms a direct hydrogen bond with the effector (Figure 2a). We, therefore, introduced a Ser258Glu mutation in the Pikp-1NK-KE background to give the triple mutant Pikp-1SNK-EKE and tested the ability of this protein to respond to AVR-Pik variants in N. benthamiana cell death assays. Firstly, we confirmed that Pikp-1SNK-EKE was not autoactive, evidenced by a lack of cell death following co-expression with Pikp-2 only (in the absence of effectors). Next, we established that Pikp-1SNK-EKE remained functional and caused cell death on co-expression with Pikp-2 and AVR-PikD. Crucially, transient expression of Pikp-1SNK-EKE, but not Pikp-1NK-KE, with Pikp-2 and either AVR-PikC or AVR-PikF, triggered cell death suggestive of new effector response specificities (Figure 2b–e, Figure 2—figure supplement 3, Figure 2—figure supplement 4). Western blot analysis indicated that all fusion proteins were produced (Figure 2—figure supplement 5). Pikp-1NK-KE has previously been shown to cause cell death when co-expressed with Pikp-2 and AVR-PikE or AVR-PikA, and consistent with these findings, Pikp-1SNK-EKE also triggered cell death when co-expressed with Pikp-2 and either of these effector variants (Figure 2—figure supplement 6, Figure 2—figure supplement 7). Figure 2 with 9 supplements see all Download asset Open asset Structure-guided mutagenesis of Pikp-1 expands response to previously unrecognized AVR-Pik effector variants. (A) Comparison of the crystal structures of AVR-Pik effector variants in complex with Pik-HMA domains (PDB entries 6G10, 6FU9, and 7A8X) and AVR-PikF in complex OsHIPP19 (PDB entry 7B1I) suggests the addition of an S258E mutation to the NK-KE mutations described previously (De la Concepcion et al., 2019) could introduce new contacts across the protein:protein interface. Protein structures are represented as ribbons with relevant side chains displayed as cylinders. Dashed lines indicate hydrogen bonds. Relevant water molecules are represented as red spheres. (B) The PikpSNK-EKE mutant gains response to AVR-PikC (right, middle) where no response is observed for PikpNK-KE (left, middle). Further, the PikpSNK-EKE mutant is not autoactive (right, bottom) and retains response to AVR-PikD (right, top). All infiltration spots contain Pikp-2. (C) Pikp-mediated response scoring represented as dot plots to summarize 60 repeats of the experiment shown in (B) across three independent experiments (Materials and methods, Figure 2—figure supplement 3a). (D) and (E) as described for (B) and (C) but with AVR-PikF and 57 repeats across three independent experiments (Materials and methods, Figure 2—figure supplement 3b). Figure 2—source data 1 Cell death scores used for dot plots in panel C. https://cdn.elifesciences.org/articles/81123/elife-81123-fig2-data1-v1.zip Download elife-81123-fig2-data1-v1.zip Figure 2—source data 2 Cell death scores used for dot plots in panel E. https://cdn.elifesciences.org/articles/81123/elife-81123-fig2-data2-v1.zip Download elife-81123-fig2-data2-v1.zip We tested whether the Ser258Glu mutation alone was sufficient to extend the cell death response to AVR-PikC or AVR-PikF using the cell death assay. When Pikp-1S258E was co-infiltrated with Pikp-2 and either AVR-PikC or AVR-PikF no cell death was observed (Figure 2—figure supplement 8, Figure 2—figure supplement 9), demonstrating that the triple mutation is necessary for response to these effectors. The Ser258Glu mutation extends the binding of Pikp-HMANK-KE to AVR-PikC and AVR-PikF in vitro The extent of the Pik/AVR-Pik-dependent cell death response in N. benthamiana largely correlates with binding affinity in vitro and in planta (Maqbool et al., 2015; De la Concepcion et al., 2018; De la Concepcion et al., 2021; De la Concepcion et al., 2019). To test whether the Pikp-1SNK-EKE response to AVR-PikC or AVR-PikF in N. benthamiana correlates with increased binding to the modified HMA domain, we first used surface plasmon resonance (SPR) with purified proteins. Pik-HMA domains and AVR-Pik effectors were purified from E. coli cultures using established protocols for the production of these proteins (Maidment et al., 2021; Maqbool et al., 2015; De la Concepcion et al., 2018; De la Concepcion et al., 2021; De la Concepcion et al., 2019). For SPR, AVR-Pik effector variants were immobilized on a Ni2+-NTA sensor chip via a C-terminal 6xHis tag. Pikp-HMA, PikpNK-KE-HMA or PikpSNK-EKE-HMA has flowed over the surface of the chip at three different concentrations (4 nM, 40 nM, and 100 nM). The binding (Robs, measured in response units, RU) was recorded and expressed as a percentage of the maximum theoretical responses (%Rmax), assuming a 2:1 HMA:effector interaction model (Maqbool et al., 2015). Consistent with previous studies, Pikp-HMA did not bind AVR-PikC (nor AVR-PikF), and weak binding was observed for PikpNK-KE-HMA to AVR-PikC (and also to AVR-PikF). By contrast, PikpSNK-EKE-HMA bound to both AVR-PikC and AVR-PikF with higher apparent affinity (larger %Rmax) than PikpNK-KE-HMA (Figure 3a–b). This result was consistent across the three concentrations investigated, though binding (and %Rmax) was low for all three HMA domains at 4 nM (Figure 3—figure supplement 1). Figure 3 with 4 supplements see all Download asset Open asset The SNK-EKE triple mutation extends Pikp-1 binding to AVR-PikC and AVR-PikF in vitro and in planta by facilitating new contacts across the pr